2019-05-28 10:10:12 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-12-07 13:06:11 -05:00
|
|
|
|
2006-01-11 12:17:46 -08:00
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2012-07-26 21:08:32 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2017-02-02 17:54:15 +01:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
2013-09-11 14:26:10 -07:00
|
|
|
|
2011-12-07 13:06:11 -05:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2011-12-07 13:06:11 -05:00
|
|
|
|
2013-04-12 01:50:06 +01:00
|
|
|
|
2013-05-01 17:51:54 -07:00
|
|
|
|
2018-10-30 15:09:49 -07:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
2017-02-04 01:20:53 +01:00
|
|
|
|
2018-11-01 23:07:23 +00:00
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
2019-06-01 18:09:44 -04:00
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2017-02-04 01:20:53 +01:00
|
|
|
|
2005-11-07 17:19:07 -05:00
|
|
|
|
2007-07-15 23:41:25 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
2022-01-21 22:13:27 -08:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
2023-10-11 19:55:00 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-02-06 01:37:57 -08:00
|
|
|
|
2014-02-27 14:40:10 -05:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2023-10-25 16:01:59 +02:00
|
|
|
|
2024-07-19 13:41:48 +02:00
|
|
|
|
2024-07-04 10:58:34 +02:00
|
|
|
|
2023-10-25 16:01:59 +02:00
|
|
|
|
2023-10-11 19:55:00 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-16 21:34:53 -04:00
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2006-01-16 22:14:23 -08:00
|
|
|
|
2023-10-11 19:55:00 +03:00
|
|
|
|
2007-10-29 14:17:23 -06:00
|
|
|
|
2006-01-16 22:14:23 -08:00
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2005-11-07 17:16:09 -05:00
|
|
|
|
|
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:01:59 +02:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
|
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
|
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
2009-06-24 03:12:00 -04:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2018-06-11 12:31:36 -04:00
|
|
|
|
2011-11-25 00:50:41 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-10-31 21:40:21 -07:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
2020-10-31 21:40:21 -07:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
|
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
|
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
2008-07-21 18:06:36 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
memcg: enable accounting for mnt_cache entries
Patch series "memcg accounting from OpenVZ", v7.
OpenVZ uses memory accounting 20+ years since v2.2.x linux kernels.
Initially we used our own accounting subsystem, then partially committed
it to upstream, and a few years ago switched to cgroups v1. Now we're
rebasing again, revising our old patches and trying to push them upstream.
We try to protect the host system from any misuse of kernel memory
allocation triggered by untrusted users inside the containers.
Patch-set is addressed mostly to cgroups maintainers and cgroups@ mailing
list, though I would be very grateful for any comments from maintainersi
of affected subsystems or other people added in cc:
Compared to the upstream, we additionally account the following kernel objects:
- network devices and its Tx/Rx queues
- ipv4/v6 addresses and routing-related objects
- inet_bind_bucket cache objects
- VLAN group arrays
- ipv6/sit: ip_tunnel_prl
- scm_fp_list objects used by SCM_RIGHTS messages of Unix sockets
- nsproxy and namespace objects itself
- IPC objects: semaphores, message queues and share memory segments
- mounts
- pollfd and select bits arrays
- signals and posix timers
- file lock
- fasync_struct used by the file lease code and driver's fasync queues
- tty objects
- per-mm LDT
We have an incorrect/incomplete/obsoleted accounting for few other kernel
objects: sk_filter, af_packets, netlink and xt_counters for iptables.
They require rework and probably will be dropped at all.
Also we're going to add an accounting for nft, however it is not ready
yet.
We have not tested performance on upstream, however, our performance team
compares our current RHEL7-based production kernel and reports that they
are at least not worse as the according original RHEL7 kernel.
This patch (of 10):
The kernel allocates ~400 bytes of 'struct mount' for any new mount.
Creating a new mount namespace clones most of the parent mounts, and this
can be repeated many times. Additionally, each mount allocates up to
PATH_MAX=4096 bytes for mnt->mnt_devname.
It makes sense to account for these allocations to restrict the host's
memory consumption from inside the memcg-limited container.
Link: https://lkml.kernel.org/r/045db11f-4a45-7c9b-2664-5b32c2b44943@virtuozzo.com
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Yutian Yang <nglaive@gmail.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Andrei Vagin <avagin@gmail.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dmitry Safonov <0x7f454c46@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jiri Slaby <jirislaby@kernel.org>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Serge Hallyn <serge@hallyn.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Zefan Li <lizefan.x@bytedance.com>
Cc: Borislav Petkov <bp@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-02 14:55:10 -07:00
|
|
|
|
|
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
2008-07-21 18:06:36 +08:00
|
|
|
|
2008-03-26 22:11:34 +01:00
|
|
|
|
|
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
|
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
|
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-22 19:37:01 -07:00
|
|
|
|
2016-10-24 16:16:13 -05:00
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
2024-12-15 21:17:05 +01:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
2008-07-21 18:06:36 +08:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
2015-02-13 14:36:41 -08:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2008-07-21 18:06:36 +08:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
2008-07-21 18:06:36 +08:00
|
|
|
|
2011-11-25 02:35:16 -05:00
|
|
|
|
2008-07-21 18:06:36 +08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
2023-06-20 13:28:32 +02:00
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
2023-06-20 13:28:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-01-07 17:50:10 +11:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-01-07 17:50:10 +11:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-11-25 13:07:11 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2023-06-20 13:28:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
2011-01-07 17:50:10 +11:00
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-09-08 16:29:00 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2020-09-22 09:44:18 -07:00
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
2020-09-22 09:44:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
2017-09-05 12:53:12 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-09-22 09:44:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-09-05 12:53:12 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-07-18 15:44:43 +02:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
2018-07-18 15:44:43 +02:00
|
|
|
|
2017-09-05 12:53:12 +02:00
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:55 +10:00
|
|
|
|
|
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2023-09-08 16:29:00 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
2020-09-22 09:44:18 -07:00
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2012-06-12 16:20:35 +02:00
|
|
|
|
|
|
|
|
|
2017-09-05 12:53:12 +02:00
|
|
|
|
|
|
|
|
|
2023-09-08 16:28:59 +03:00
|
|
|
|
2017-09-05 12:53:12 +02:00
|
|
|
|
|
|
|
|
|
2011-12-09 08:06:57 -05:00
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:07 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:51 +01:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-01-07 17:50:10 +11:00
|
|
|
|
2021-01-21 14:19:51 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:07 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:51 +01:00
|
|
|
|
|
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2021-01-21 14:19:51 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
[PATCH] r/o bind mounts: track numbers of writers to mounts
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-02-15 14:37:59 -08:00
|
|
|
|
2008-02-15 14:37:30 -08:00
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:33 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:10 +01:00
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:33 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-20 13:28:32 +02:00
|
|
|
|
|
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-21 12:11:31 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
2015-02-13 14:36:41 -08:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 22:53:09 -05:00
|
|
|
|
2009-04-26 20:25:54 +10:00
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2014-01-24 12:17:54 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2015-05-08 11:43:53 -04:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-05-08 11:43:53 -04:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2015-05-08 11:43:53 -04:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
2025-04-28 23:56:14 -04:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2015-05-08 11:43:53 -04:00
|
|
|
|
2018-08-09 17:51:32 -04:00
|
|
|
|
2025-04-27 15:41:51 -04:00
|
|
|
|
2018-08-09 17:51:32 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-05-08 11:43:53 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-07-05 12:22:46 -04:00
|
|
|
|
2015-05-08 11:43:53 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:40 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-10-01 16:11:26 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2013-10-01 16:11:26 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2013-10-01 16:11:26 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:20:17 -05:00
|
|
|
|
2012-06-25 12:55:28 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:20:17 -05:00
|
|
|
|
2016-11-20 19:45:28 -05:00
|
|
|
|
2005-11-07 17:20:17 -05:00
|
|
|
|
2011-11-24 18:22:03 -05:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:20:17 -05:00
|
|
|
|
|
|
|
|
|
2013-10-04 19:15:13 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2013-10-04 19:15:13 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2013-10-04 19:15:13 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-03-04 18:12:45 +02:00
|
|
|
|
2013-10-04 19:15:13 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-02-24 17:32:34 -08:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-02-24 17:32:34 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2014-02-24 17:32:34 -08:00
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2014-02-24 17:32:34 -08:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2018-10-03 10:18:33 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2013-09-05 14:39:11 +02:00
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-22 19:37:01 -07:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2005-11-07 17:15:49 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2005-11-07 17:15:49 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-29 12:58:42 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2019-06-29 12:58:42 -04:00
|
|
|
|
2011-11-24 22:19:58 -05:00
|
|
|
|
2011-11-24 22:25:07 -05:00
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2013-09-22 19:37:01 -07:00
|
|
|
|
2019-06-29 12:58:42 -04:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2019-06-29 12:58:42 -04:00
|
|
|
|
2014-12-29 13:03:41 -06:00
|
|
|
|
|
|
|
|
|
2015-01-15 22:58:33 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-29 12:58:42 -04:00
|
|
|
|
2015-01-15 22:58:33 -06:00
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
2011-11-24 21:28:22 -05:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2011-11-25 03:19:09 -05:00
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2011-11-25 03:19:09 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2013-09-22 19:37:01 -07:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
|
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
2014-08-10 03:44:55 -04:00
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-08-10 03:44:55 -04:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-12-15 21:17:05 +01:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2011-11-24 22:19:58 -05:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2011-11-24 22:19:58 -05:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2011-11-25 02:19:55 -05:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-01-14 22:30:21 -05:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-24 22:19:58 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-24 20:38:33 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-04 03:19:03 -05:00
|
|
|
|
2018-11-04 06:48:34 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-02-01 06:06:16 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2017-02-01 06:06:16 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:24:27 -05:00
|
|
|
|
2005-11-07 17:17:22 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-24 21:24:27 -05:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2012-07-31 13:13:04 -07:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2018-04-19 22:03:08 -04:00
|
|
|
|
2024-12-15 21:17:05 +01:00
|
|
|
|
2013-03-29 21:04:39 -07:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2012-07-31 13:13:04 -07:00
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 19:33:09 -05:00
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 20:55:08 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
2014-01-24 12:17:54 +00:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
|
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
|
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-08-07 17:44:45 +09:00
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
2017-08-07 17:44:45 +09:00
|
|
|
|
|
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 00:33:11 -05:00
|
|
|
|
2005-11-07 17:13:39 -05:00
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2020-10-31 21:40:21 -07:00
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
fix mntput/mntput race
mntput_no_expire() does the calculation of total refcount under mount_lock;
unfortunately, the decrement (as well as all increments) are done outside
of it, leading to false positives in the "are we dropping the last reference"
test. Consider the following situation:
* mnt is a lazy-umounted mount, kept alive by two opened files. One
of those files gets closed. Total refcount of mnt is 2. On CPU 42
mntput(mnt) (called from __fput()) drops one reference, decrementing component
* After it has looked at component #0, the process on CPU 0 does
mntget(), incrementing component #0, gets preempted and gets to run again -
on CPU 69. There it does mntput(), which drops the reference (component #69)
and proceeds to spin on mount_lock.
* On CPU 42 our first mntput() finishes counting. It observes the
decrement of component #69, but not the increment of component #0. As the
result, the total it gets is not 1 as it should've been - it's 0. At which
point we decide that vfsmount needs to be killed and proceed to free it and
shut the filesystem down. However, there's still another opened file
on that filesystem, with reference to (now freed) vfsmount, etc. and we are
screwed.
It's not a wide race, but it can be reproduced with artificial slowdown of
the mnt_get_count() loop, and it should be easier to hit on SMP KVM setups.
Fix consists of moving the refcount decrement under mount_lock; the tricky
part is that we want (and can) keep the fast case (i.e. mount that still
has non-NULL ->mnt_ns) entirely out of mount_lock. All places that zero
mnt->mnt_ns are dropping some reference to mnt and they call synchronize_rcu()
before that mntput(). IOW, if mntput() observes (under rcu_read_lock())
a non-NULL ->mnt_ns, it is guaranteed that there is another reference yet to
be dropped.
Reported-by: Jann Horn <jannh@google.com>
Tested-by: Jann Horn <jannh@google.com>
Fixes: 48a066e72d97 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-08-09 17:21:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2011-01-14 22:30:21 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2018-08-09 17:51:32 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fix mntput/mntput race
mntput_no_expire() does the calculation of total refcount under mount_lock;
unfortunately, the decrement (as well as all increments) are done outside
of it, leading to false positives in the "are we dropping the last reference"
test. Consider the following situation:
* mnt is a lazy-umounted mount, kept alive by two opened files. One
of those files gets closed. Total refcount of mnt is 2. On CPU 42
mntput(mnt) (called from __fput()) drops one reference, decrementing component
* After it has looked at component #0, the process on CPU 0 does
mntget(), incrementing component #0, gets preempted and gets to run again -
on CPU 69. There it does mntput(), which drops the reference (component #69)
and proceeds to spin on mount_lock.
* On CPU 42 our first mntput() finishes counting. It observes the
decrement of component #69, but not the increment of component #0. As the
result, the total it gets is not 1 as it should've been - it's 0. At which
point we decide that vfsmount needs to be killed and proceed to free it and
shut the filesystem down. However, there's still another opened file
on that filesystem, with reference to (now freed) vfsmount, etc. and we are
screwed.
It's not a wide race, but it can be reproduced with artificial slowdown of
the mnt_get_count() loop, and it should be easier to hit on SMP KVM setups.
Fix consists of moving the refcount decrement under mount_lock; the tricky
part is that we want (and can) keep the fast case (i.e. mount that still
has non-NULL ->mnt_ns) entirely out of mount_lock. All places that zero
mnt->mnt_ns are dropping some reference to mnt and they call synchronize_rcu()
before that mntput(). IOW, if mntput() observes (under rcu_read_lock())
a non-NULL ->mnt_ns, it is guaranteed that there is another reference yet to
be dropped.
Reported-by: Jann Horn <jannh@google.com>
Tested-by: Jann Horn <jannh@google.com>
Fixes: 48a066e72d97 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-08-09 17:21:17 -04:00
|
|
|
|
2020-10-31 21:40:21 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-08 13:32:02 +09:30
|
|
|
|
2011-11-21 12:11:30 +01:00
|
|
|
|
2014-12-23 21:37:03 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
2014-12-23 21:37:03 -06:00
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2013-09-28 12:41:25 -04:00
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
task_work: cleanup notification modes
A previous commit changed the notification mode from true/false to an
int, allowing notify-no, notify-yes, or signal-notify. This was
backwards compatible in the sense that any existing true/false user
would translate to either 0 (on notification sent) or 1, the latter
which mapped to TWA_RESUME. TWA_SIGNAL was assigned a value of 2.
Clean this up properly, and define a proper enum for the notification
mode. Now we have:
- TWA_NONE. This is 0, same as before the original change, meaning no
notification requested.
- TWA_RESUME. This is 1, same as before the original change, meaning
that we use TIF_NOTIFY_RESUME.
- TWA_SIGNAL. This uses TIF_SIGPENDING/JOBCTL_TASK_WORK for the
notification.
Clean up all the callers, switching their 0/1/false/true to using the
appropriate TWA_* mode for notifications.
Fixes: e91b48162332 ("task_work: teach task_work_add() to do signal_wake_up()")
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-10-16 09:02:26 -06:00
|
|
|
|
2014-08-08 13:08:20 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:13:39 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 00:57:42 -05:00
|
|
|
|
2023-10-04 13:19:16 +02:00
|
|
|
|
2011-11-25 00:57:42 -05:00
|
|
|
|
2023-10-04 13:19:16 +02:00
|
|
|
|
2011-11-25 00:57:42 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-01-27 13:46:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-03-17 19:52:25 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-24 08:03:41 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 19:45:28 -05:00
|
|
|
|
2005-11-07 17:13:39 -05:00
|
|
|
|
2014-08-07 09:12:31 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:13:39 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-07 16:55:37 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2011-12-06 12:21:54 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
fs: use seq_open_private() for proc_mounts
A patchset to remove support for passing pre-allocated struct seq_file to
seq_open(). Such feature is undocumented and prone to error.
In particular, if seq_release() is used in release handler, it will
kfree() a pointer which was not allocated by seq_open().
So this patchset drops support for pre-allocated struct seq_file: it's
only of use in proc_namespace.c and can be easily replaced by using
seq_open_private()/seq_release_private().
Additionally, it documents the use of file->private_data to hold pointer
to struct seq_file by seq_open().
This patch (of 3):
Since patch described below, from v2.6.15-rc1, seq_open() could use a
struct seq_file already allocated by the caller if the pointer to the
structure is stored in file->private_data before calling the function.
Commit 1abe77b0fc4b485927f1f798ae81a752677e1d05
Author: Al Viro <viro@zeniv.linux.org.uk>
Date: Mon Nov 7 17:15:34 2005 -0500
[PATCH] allow callers of seq_open do allocation themselves
Allow caller of seq_open() to kmalloc() seq_file + whatever else they
want and set ->private_data to it. seq_open() will then abstain from
doing allocation itself.
Such behavior is only used by mounts_open_common().
In order to drop support for such uncommon feature, proc_mounts is
converted to use seq_open_private(), which take care of allocating the
proc_mounts structure, making it available through ->private in struct
seq_file.
Conversely, proc_mounts is converted to use seq_release_private(), in
order to release the private structure allocated by seq_open_private().
Then, ->private is used directly instead of proc_mounts() macro to access
to the proc_mounts structure.
Link: http://lkml.kernel.org/r/cover.1433193673.git.ydroneaud@opteya.com
Signed-off-by: Yann Droneaud <ydroneaud@opteya.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-30 14:57:30 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2005-11-07 17:17:51 -05:00
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2007-07-15 23:39:55 -07:00
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:17:51 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-12-06 12:21:54 -05:00
|
|
|
|
2008-03-27 13:06:25 +01:00
|
|
|
|
fs: use seq_open_private() for proc_mounts
A patchset to remove support for passing pre-allocated struct seq_file to
seq_open(). Such feature is undocumented and prone to error.
In particular, if seq_release() is used in release handler, it will
kfree() a pointer which was not allocated by seq_open().
So this patchset drops support for pre-allocated struct seq_file: it's
only of use in proc_namespace.c and can be easily replaced by using
seq_open_private()/seq_release_private().
Additionally, it documents the use of file->private_data to hold pointer
to struct seq_file by seq_open().
This patch (of 3):
Since patch described below, from v2.6.15-rc1, seq_open() could use a
struct seq_file already allocated by the caller if the pointer to the
structure is stored in file->private_data before calling the function.
Commit 1abe77b0fc4b485927f1f798ae81a752677e1d05
Author: Al Viro <viro@zeniv.linux.org.uk>
Date: Mon Nov 7 17:15:34 2005 -0500
[PATCH] allow callers of seq_open do allocation themselves
Allow caller of seq_open() to kmalloc() seq_file + whatever else they
want and set ->private_data to it. seq_open() will then abstain from
doing allocation itself.
Such behavior is only used by mounts_open_common().
In order to drop support for such uncommon feature, proc_mounts is
converted to use seq_open_private(), which take care of allocating the
proc_mounts structure, making it available through ->private in struct
seq_file.
Conversely, proc_mounts is converted to use seq_release_private(), in
order to release the private structure allocated by seq_open_private().
Then, ->private is used directly instead of proc_mounts() macro to access
to the proc_mounts structure.
Link: http://lkml.kernel.org/r/cover.1433193673.git.ydroneaud@opteya.com
Signed-off-by: Yann Droneaud <ydroneaud@opteya.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-30 14:57:30 -07:00
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
2011-12-06 12:21:54 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2008-03-27 13:06:24 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-12-06 12:21:54 -05:00
|
|
|
|
2006-03-20 13:44:12 -05:00
|
|
|
|
2020-05-14 16:44:24 +02:00
|
|
|
|
2008-03-27 13:06:24 +01:00
|
|
|
|
2006-03-20 13:44:12 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2021-03-17 19:52:25 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2005-11-07 17:17:22 -05:00
|
|
|
|
|
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2006-03-27 01:14:51 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-03-27 01:14:51 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-03-27 01:14:51 -08:00
|
|
|
|
2010-01-16 12:56:08 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2006-03-27 01:14:51 -08:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2010-01-16 12:56:08 -05:00
|
|
|
|
2005-11-07 17:20:17 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2015-04-02 20:33:53 -05:00
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2015-04-02 20:33:53 -05:00
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 10:39:08 -04:00
|
|
|
|
|
|
|
|
|
2015-04-02 20:33:53 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-30 10:33:18 +11:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:17:04 -05:00
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2013-03-16 14:35:16 -04:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2013-03-16 14:35:16 -04:00
|
|
|
|
|
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-04-01 18:30:06 -05:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2015-07-17 14:15:30 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
|
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2005-11-07 17:17:04 -05:00
|
|
|
|
2014-12-18 13:10:48 -06:00
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2005-11-07 17:17:04 -05:00
|
|
|
|
2015-01-03 05:39:35 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-12-18 13:10:48 -06:00
|
|
|
|
2014-12-22 18:30:08 -06:00
|
|
|
|
|
|
|
|
|
2024-12-15 21:17:05 +01:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-12-22 18:30:08 -06:00
|
|
|
|
2014-12-18 13:10:48 -06:00
|
|
|
|
2014-12-22 19:12:07 -06:00
|
|
|
|
2014-12-18 13:10:48 -06:00
|
|
|
|
|
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2005-11-07 17:17:04 -05:00
|
|
|
|
2024-08-06 11:47:10 +08:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2011-01-15 20:08:44 -05:00
|
|
|
|
2005-11-07 17:20:17 -05:00
|
|
|
|
2014-12-18 13:10:48 -06:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
2014-12-23 21:37:03 -06:00
|
|
|
|
2014-12-18 13:10:48 -06:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2011-11-25 02:19:55 -05:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2015-01-10 19:01:08 -05:00
|
|
|
|
2015-07-17 14:15:30 -05:00
|
|
|
|
2011-11-24 21:47:05 -05:00
|
|
|
|
2014-08-30 18:32:05 -04:00
|
|
|
|
2014-12-23 21:37:03 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-03-21 23:59:49 -04:00
|
|
|
|
2011-11-24 20:43:10 -05:00
|
|
|
|
2019-07-24 12:45:46 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 14:39:34 -04:00
|
|
|
|
2008-03-22 00:46:23 -04:00
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-24 22:37:54 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 00:57:42 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-04-24 07:21:56 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-09-18 17:58:08 -05:00
|
|
|
|
2014-10-08 12:32:47 -07:00
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-10-22 10:21:38 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2024-12-15 21:17:05 +01:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2025-04-28 23:56:14 -04:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-12-15 21:17:05 +01:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-10-22 10:21:38 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2013-03-16 14:35:16 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-10-03 01:31:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2013-10-03 01:31:18 -07:00
|
|
|
|
2019-06-29 12:06:51 -04:00
|
|
|
|
2013-10-03 01:31:18 -07:00
|
|
|
|
|
|
|
|
|
2016-04-15 14:24:41 -07:00
|
|
|
|
2013-10-03 01:31:18 -07:00
|
|
|
|
|
|
|
|
|
2014-12-23 21:37:03 -06:00
|
|
|
|
2015-07-17 14:54:27 -05:00
|
|
|
|
2019-07-04 16:57:51 -04:00
|
|
|
|
2014-12-23 21:37:03 -06:00
|
|
|
|
2015-04-01 18:30:06 -05:00
|
|
|
|
2013-10-03 01:31:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2013-10-03 01:31:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-04 17:25:09 +01:00
|
|
|
|
2013-02-22 22:45:42 -05:00
|
|
|
|
|
|
|
|
|
2022-03-01 00:05:29 -05:00
|
|
|
|
2013-02-22 22:45:42 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-08-19 14:56:38 -04:00
|
|
|
|
2015-11-16 09:49:34 -05:00
|
|
|
|
2021-08-19 14:56:38 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-11-16 09:49:34 -05:00
|
|
|
|
|
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
2025-03-18 12:29:21 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-02-22 22:45:42 -05:00
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
2018-10-22 10:21:38 -05:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
2025-03-18 12:29:21 -04:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-04 15:25:34 -05:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2008-02-14 19:34:31 -08:00
|
|
|
|
2020-07-21 10:54:34 +02:00
|
|
|
|
2011-11-25 00:33:11 -05:00
|
|
|
|
2020-08-06 16:07:10 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2020-07-23 08:23:08 +02:00
|
|
|
|
2020-07-21 10:54:34 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-04 15:25:34 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-07-21 10:54:34 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-03-11 11:34:40 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:16:09 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2009-01-14 14:14:12 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-03-11 11:34:40 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2024-12-11 13:11:17 +01:00
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
2024-12-11 13:11:17 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
|
|
|
|
|
nsproxy: attach to namespaces via pidfds
For quite a while we have been thinking about using pidfds to attach to
namespaces. This patchset has existed for about a year already but we've
wanted to wait to see how the general api would be received and adopted.
Now that more and more programs in userspace have started using pidfds
for process management it's time to send this one out.
This patch makes it possible to use pidfds to attach to the namespaces
of another process, i.e. they can be passed as the first argument to the
setns() syscall. When only a single namespace type is specified the
semantics are equivalent to passing an nsfd. That means
setns(nsfd, CLONE_NEWNET) equals setns(pidfd, CLONE_NEWNET). However,
when a pidfd is passed, multiple namespace flags can be specified in the
second setns() argument and setns() will attach the caller to all the
specified namespaces all at once or to none of them. Specifying 0 is not
valid together with a pidfd.
Here are just two obvious examples:
setns(pidfd, CLONE_NEWPID | CLONE_NEWNS | CLONE_NEWNET);
setns(pidfd, CLONE_NEWUSER);
Allowing to also attach subsets of namespaces supports various use-cases
where callers setns to a subset of namespaces to retain privilege, perform
an action and then re-attach another subset of namespaces.
If the need arises, as Eric suggested, we can extend this patchset to
assume even more context than just attaching all namespaces. His suggestion
specifically was about assuming the process' root directory when
setns(pidfd, 0) or setns(pidfd, SETNS_PIDFD) is specified. For now, just
keep it flexible in terms of supporting subsets of namespaces but let's
wait until we have users asking for even more context to be assumed. At
that point we can add an extension.
The obvious example where this is useful is a standard container
manager interacting with a running container: pushing and pulling files
or directories, injecting mounts, attaching/execing any kind of process,
managing network devices all these operations require attaching to all
or at least multiple namespaces at the same time. Given that nowadays
most containers are spawned with all namespaces enabled we're currently
looking at at least 14 syscalls, 7 to open the /proc/<pid>/ns/<ns>
nsfds, another 7 to actually perform the namespace switch. With time
namespaces we're looking at about 16 syscalls.
(We could amortize the first 7 or 8 syscalls for opening the nsfds by
stashing them in each container's monitor process but that would mean
we need to send around those file descriptors through unix sockets
everytime we want to interact with the container or keep on-disk
state. Even in scenarios where a caller wants to join a particular
namespace in a particular order callers still profit from batching
other namespaces. That mostly applies to the user namespace but
all container runtimes I found join the user namespace first no matter
if it privileges or deprivileges the container similar to how unshare
behaves.)
With pidfds this becomes a single syscall no matter how many namespaces
are supposed to be attached to.
A decently designed, large-scale container manager usually isn't the
parent of any of the containers it spawns so the containers don't die
when it crashes or needs to update or reinitialize. This means that
for the manager to interact with containers through pids is inherently
racy especially on systems where the maximum pid number is not
significicantly bumped. This is even more problematic since we often spawn
and manage thousands or ten-thousands of containers. Interacting with a
container through a pid thus can become risky quite quickly. Especially
since we allow for an administrator to enable advanced features such as
syscall interception where we're performing syscalls in lieu of the
container. In all of those cases we use pidfds if they are available and
we pass them around as stable references. Using them to setns() to the
target process' namespaces is as reliable as using nsfds. Either the
target process is already dead and we get ESRCH or we manage to attach
to its namespaces but we can't accidently attach to another process'
namespaces. So pidfds lend themselves to be used with this api.
The other main advantage is that with this change the pidfd becomes the
only relevant token for most container interactions and it's the only
token we need to create and send around.
Apart from significiantly reducing the number of syscalls from double
digit to single digit which is a decent reason post-spectre/meltdown
this also allows to switch to a set of namespaces atomically, i.e.
either attaching to all the specified namespaces succeeds or we fail. If
we fail we haven't changed a single namespace. There are currently three
namespaces that can fail (other than for ENOMEM which really is not
very interesting since we then have other problems anyway) for
non-trivial reasons, user, mount, and pid namespaces. We can fail to
attach to a pid namespace if it is not our current active pid namespace
or a descendant of it. We can fail to attach to a user namespace because
we are multi-threaded or because our current mount namespace shares
filesystem state with other tasks, or because we're trying to setns()
to the same user namespace, i.e. the target task has the same user
namespace as we do. We can fail to attach to a mount namespace because
it shares filesystem state with other tasks or because we fail to lookup
the new root for the new mount namespace. In most non-pathological
scenarios these issues can be somewhat mitigated. But there are cases where
we're half-attached to some namespace and failing to attach to another one.
I've talked about some of these problem during the hallway track (something
only the pre-COVID-19 generation will remember) of Plumbers in Los Angeles
in 2018(?). Even if all these issues could be avoided with super careful
userspace coding it would be nicer to have this done in-kernel. Pidfds seem
to lend themselves nicely for this.
The other neat thing about this is that setns() becomes an actual
counterpart to the namespace bits of unshare().
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Serge Hallyn <serge@hallyn.com>
Cc: Jann Horn <jannh@google.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Link: https://lore.kernel.org/r/20200505140432.181565-3-christian.brauner@ubuntu.com
2020-05-05 16:04:31 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-07-19 13:41:52 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-11-01 03:13:17 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
2005-11-07 17:17:22 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2018-10-25 09:04:18 -05:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
2018-10-25 09:04:18 -05:00
|
|
|
|
|
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
2018-10-25 09:04:18 -05:00
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
|
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
|
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2024-06-07 01:39:12 +08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
2016-11-20 19:45:28 -05:00
|
|
|
|
2007-06-07 12:20:32 -04:00
|
|
|
|
2011-11-24 20:55:08 -05:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2015-01-07 14:28:26 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 14:49:45 -04:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2013-08-14 12:44:39 +03:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2007-06-07 12:20:32 -04:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
2007-06-07 12:20:32 -04:00
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2018-10-25 12:05:11 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2013-03-16 14:42:19 -04:00
|
|
|
|
2007-06-07 12:20:32 -04:00
|
|
|
|
|
|
|
|
|
2025-06-01 14:23:52 -04:00
|
|
|
|
2021-08-09 10:19:47 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 14:23:52 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
|
|
|
|
|
2021-03-17 19:52:25 -07:00
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
2021-03-17 19:52:25 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 19:45:28 -05:00
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-08-09 10:19:47 +02:00
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
2021-08-09 10:19:47 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 20:11:06 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 14:23:52 -04:00
|
|
|
|
2021-08-09 10:19:47 +02:00
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
|
|
|
|
|
2021-08-09 10:19:47 +02:00
|
|
|
|
|
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-06-04 10:48:19 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
2021-08-09 10:19:47 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-10-24 00:14:36 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-01-30 22:51:25 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 02:19:55 -05:00
|
|
|
|
2010-01-30 22:51:25 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 02:19:55 -05:00
|
|
|
|
|
|
|
|
|
2010-01-30 22:51:25 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-01-30 13:15:45 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-25 01:05:37 -05:00
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-25 01:05:37 -05:00
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-13 22:42:30 -05:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2022-02-13 22:42:30 -05:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2022-02-13 22:42:30 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-13 22:42:30 -05:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
2022-02-13 22:42:30 -05:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:01 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:01 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:20:03 -05:00
|
|
|
|
2005-11-07 17:21:01 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2011-11-24 19:59:16 -05:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2019-01-30 13:15:45 -05:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 01:05:37 -05:00
|
|
|
|
2011-11-24 19:59:16 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
2014-03-21 10:14:08 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-24 20:43:10 -05:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2005-11-07 17:20:03 -05:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
2005-11-07 17:20:03 -05:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2014-03-20 20:34:43 -04:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2019-01-30 13:15:45 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-17 23:22:14 +02:00
|
|
|
|
2025-06-22 18:03:29 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
smarter propagate_mnt()
The current mainline has copies propagated to *all* nodes, then
tears down the copies we made for nodes that do not contain
counterparts of the desired mountpoint. That sets the right
propagation graph for the copies (at teardown time we move
the slaves of removed node to a surviving peer or directly
to master), but we end up paying a fairly steep price in
useless allocations. It's fairly easy to create a situation
where N calls of mount(2) create exactly N bindings, with
O(N^2) vfsmounts allocated and freed in process.
Fortunately, it is possible to avoid those allocations/freeings.
The trick is to create copies in the right order and find which
one would've eventually become a master with the current algorithm.
It turns out to be possible in O(nodes getting propagation) time
and with no extra allocations at all.
One part is that we need to make sure that eventual master will be
created before its slaves, so we need to walk the propagation
tree in a different order - by peer groups. And iterate through
the peers before dealing with the next group.
Another thing is finding the (earlier) copy that will be a master
of one we are about to create; to do that we are (temporary) marking
the masters of mountpoints we are attaching the copies to.
Either we are in a peer of the last mountpoint we'd dealt with,
or we have the following situation: we are attaching to mountpoint M,
the last copy S_0 had been attached to M_0 and there are sequences
S_0...S_n, M_0...M_n such that S_{i+1} is a master of S_{i},
S_{i} mounted on M{i} and we need to create a slave of the first S_{k}
such that M is getting propagation from M_{k}. It means that the master
of M_{k} will be among the sequence of masters of M. On the
other hand, the nearest marked node in that sequence will either
be the master of M_{k} or the master of M_{k-1} (the latter -
in the case if M_{k-1} is a slave of something M gets propagation
from, but in a wrong peer group).
So we go through the sequence of masters of M until we find
a marked one (P). Let N be the one before it. Then we go through
the sequence of masters of S_0 until we find one (say, S) mounted
on a node D that has P as master and check if D is a peer of N.
If it is, S will be the master of new copy, if not - the master of S
will be.
That's it for the hard part; the rest is fairly simple. Iterator
is in next_group(), handling of one prospective mountpoint is
propagate_one().
It seems to survive all tests and gives a noticably better performance
than the current mainline for setups that are seriously using shared
subtrees.
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-02-27 09:35:45 -05:00
|
|
|
|
|
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
smarter propagate_mnt()
The current mainline has copies propagated to *all* nodes, then
tears down the copies we made for nodes that do not contain
counterparts of the desired mountpoint. That sets the right
propagation graph for the copies (at teardown time we move
the slaves of removed node to a surviving peer or directly
to master), but we end up paying a fairly steep price in
useless allocations. It's fairly easy to create a situation
where N calls of mount(2) create exactly N bindings, with
O(N^2) vfsmounts allocated and freed in process.
Fortunately, it is possible to avoid those allocations/freeings.
The trick is to create copies in the right order and find which
one would've eventually become a master with the current algorithm.
It turns out to be possible in O(nodes getting propagation) time
and with no extra allocations at all.
One part is that we need to make sure that eventual master will be
created before its slaves, so we need to walk the propagation
tree in a different order - by peer groups. And iterate through
the peers before dealing with the next group.
Another thing is finding the (earlier) copy that will be a master
of one we are about to create; to do that we are (temporary) marking
the masters of mountpoints we are attaching the copies to.
Either we are in a peer of the last mountpoint we'd dealt with,
or we have the following situation: we are attaching to mountpoint M,
the last copy S_0 had been attached to M_0 and there are sequences
S_0...S_n, M_0...M_n such that S_{i+1} is a master of S_{i},
S_{i} mounted on M{i} and we need to create a slave of the first S_{k}
such that M is getting propagation from M_{k}. It means that the master
of M_{k} will be among the sequence of masters of M. On the
other hand, the nearest marked node in that sequence will either
be the master of M_{k} or the master of M_{k-1} (the latter -
in the case if M_{k-1} is a slave of something M gets propagation
from, but in a wrong peer group).
So we go through the sequence of masters of M until we find
a marked one (P). Let N be the one before it. Then we go through
the sequence of masters of S_0 until we find one (say, S) mounted
on a node D that has P as master and check if D is a peer of N.
If it is, S will be the master of new copy, if not - the master of S
will be.
That's it for the hard part; the rest is fairly simple. Iterator
is in next_group(), handling of one prospective mountpoint is
propagate_one().
It seems to survive all tests and gives a noticably better performance
than the current mainline for setups that are seriously using shared
subtrees.
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-02-27 09:35:45 -05:00
|
|
|
|
|
|
|
|
|
2014-03-21 10:14:08 -04:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2016-09-28 00:27:17 -05:00
|
|
|
|
2017-01-20 18:28:35 +13:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2005-11-07 17:19:50 -05:00
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2023-05-03 13:18:41 +02:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
2023-05-03 13:18:41 +02:00
|
|
|
|
|
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:41 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2016-01-22 15:40:57 -05:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2023-05-03 13:18:41 +02:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:41 +02:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
2023-05-03 13:18:41 +02:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
[ Upstream commit 0d039eac6e5950f9d1ecc9e410c2fd1feaeab3b6 ]
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac392815628 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-04-23 02:30:34 +01:00
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2013-03-16 14:49:45 -04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
VFS: (Scripted) Convert S_ISLNK/DIR/REG(dentry->d_inode) to d_is_*(dentry)
Convert the following where appropriate:
(1) S_ISLNK(dentry->d_inode) to d_is_symlink(dentry).
(2) S_ISREG(dentry->d_inode) to d_is_reg(dentry).
(3) S_ISDIR(dentry->d_inode) to d_is_dir(dentry). This is actually more
complicated than it appears as some calls should be converted to
d_can_lookup() instead. The difference is whether the directory in
question is a real dir with a ->lookup op or whether it's a fake dir with
a ->d_automount op.
In some circumstances, we can subsume checks for dentry->d_inode not being
NULL into this, provided we the code isn't in a filesystem that expects
d_inode to be NULL if the dirent really *is* negative (ie. if we're going to
use d_inode() rather than d_backing_inode() to get the inode pointer).
Note that the dentry type field may be set to something other than
DCACHE_MISS_TYPE when d_inode is NULL in the case of unionmount, where the VFS
manages the fall-through from a negative dentry to a lower layer. In such a
case, the dentry type of the negative union dentry is set to the same as the
type of the lower dentry.
However, if you know d_inode is not NULL at the call site, then you can use
the d_is_xxx() functions even in a filesystem.
There is one further complication: a 0,0 chardev dentry may be labelled
DCACHE_WHITEOUT_TYPE rather than DCACHE_SPECIAL_TYPE. Strictly, this was
intended for special directory entry types that don't have attached inodes.
The following perl+coccinelle script was used:
use strict;
my @callers;
open($fd, 'git grep -l \'S_IS[A-Z].*->d_inode\' |') ||
die "Can't grep for S_ISDIR and co. callers";
@callers = <$fd>;
close($fd);
unless (@callers) {
print "No matches\n";
exit(0);
}
my @cocci = (
'@@',
'expression E;',
'@@',
'',
'- S_ISLNK(E->d_inode->i_mode)',
'+ d_is_symlink(E)',
'',
'@@',
'expression E;',
'@@',
'',
'- S_ISDIR(E->d_inode->i_mode)',
'+ d_is_dir(E)',
'',
'@@',
'expression E;',
'@@',
'',
'- S_ISREG(E->d_inode->i_mode)',
'+ d_is_reg(E)' );
my $coccifile = "tmp.sp.cocci";
open($fd, ">$coccifile") || die $coccifile;
print($fd "$_\n") || die $coccifile foreach (@cocci);
close($fd);
foreach my $file (@callers) {
chomp $file;
print "Processing ", $file, "\n";
system("spatch", "--sp-file", $coccifile, $file, "--in-place", "--no-show-diff") == 0 ||
die "spatch failed";
}
[AV: overlayfs parts skipped]
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-01-29 12:02:35 +00:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2025-08-14 01:44:31 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-08-26 11:07:22 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2010-08-26 11:07:22 -07:00
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2010-08-26 11:07:22 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:19:07 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2005-11-07 17:19:07 -05:00
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2011-11-24 19:54:23 -05:00
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2010-08-26 11:07:22 -07:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2005-11-07 17:19:07 -05:00
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2005-11-07 17:19:07 -05:00
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2010-08-26 11:07:22 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2025-08-14 01:44:31 -04:00
|
|
|
|
|
|
|
|
|
2025-06-04 12:27:08 -04:00
|
|
|
|
2025-08-14 01:44:31 -04:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-24 20:43:10 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2008-03-27 13:06:23 +01:00
|
|
|
|
2005-11-07 17:19:07 -05:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 14:23:52 -04:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-10-11 11:42:01 -04:00
|
|
|
|
2008-02-08 04:22:12 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2008-08-02 00:51:11 -04:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2013-02-22 22:49:10 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-09-26 20:36:09 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
2017-07-04 17:25:09 +01:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
2008-02-14 19:34:32 -08:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
take the targets of /proc/*/ns/* symlinks to separate fs
New pseudo-filesystem: nsfs. Targets of /proc/*/ns/* live there now.
It's not mountable (not even registered, so it's not in /proc/filesystems,
etc.). Files on it *are* bindable - we explicitly permit that in do_loopback().
This stuff lives in fs/nsfs.c now; proc_ns_fget() moved there as well.
get_proc_ns() is a macro now (it's simply returning ->i_private; would
have been an inline, if not for header ordering headache).
proc_ns_inode() is an ex-parrot. The interface used in procfs is
ns_get_path(path, task, ops) and ns_get_name(buf, size, task, ops).
Dentries and inodes are never hashed; a non-counting reference to dentry
is stashed in ns_common (removed by ->d_prune()) and reused by ns_get_path()
if present. See ns_get_path()/ns_prune_dentry/nsfs_evict() for details
of that mechanism.
As the result, proc_ns_follow_link() has stopped poking in nd->path.mnt;
it does nd_jump_link() on a consistent <vfsmount,dentry> pair it gets
from ns_get_path().
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-11-01 10:57:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
|
|
|
|
|
2013-04-09 17:33:29 +04:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2005-11-07 17:15:04 -05:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2005-11-07 17:15:04 -05:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-11-07 17:16:29 -05:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2005-11-07 17:15:04 -05:00
|
|
|
|
2008-08-02 00:51:11 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-10-15 11:35:02 +01:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-02-15 14:38:00 -08:00
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
2008-02-15 14:38:00 -08:00
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-02-15 14:38:00 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:48 +01:00
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-03-22 14:39:22 -07:00
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
2024-07-30 10:58:13 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
2023-03-14 17:09:06 +02:00
|
|
|
|
2019-10-16 19:48:14 -07:00
|
|
|
|
|
|
|
|
|
2023-03-14 17:09:06 +02:00
|
|
|
|
|
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
2022-03-22 14:39:22 -07:00
|
|
|
|
2024-07-30 10:58:13 +02:00
|
|
|
|
|
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:50 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:48 +01:00
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:48 +01:00
|
|
|
|
2021-01-21 14:19:50 +01:00
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
2008-02-15 14:38:00 -08:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2008-08-02 00:51:11 -04:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
2014-07-28 17:10:56 -07:00
|
|
|
|
mnt: Correct permission checks in do_remount
While invesgiating the issue where in "mount --bind -oremount,ro ..."
would result in later "mount --bind -oremount,rw" succeeding even if
the mount started off locked I realized that there are several
additional mount flags that should be locked and are not.
In particular MNT_NOSUID, MNT_NODEV, MNT_NOEXEC, and the atime
flags in addition to MNT_READONLY should all be locked. These
flags are all per superblock, can all be changed with MS_BIND,
and should not be changable if set by a more privileged user.
The following additions to the current logic are added in this patch.
- nosuid may not be clearable by a less privileged user.
- nodev may not be clearable by a less privielged user.
- noexec may not be clearable by a less privileged user.
- atime flags may not be changeable by a less privileged user.
The logic with atime is that always setting atime on access is a
global policy and backup software and auditing software could break if
atime bits are not updated (when they are configured to be updated),
and serious performance degradation could result (DOS attack) if atime
updates happen when they have been explicitly disabled. Therefore an
unprivileged user should not be able to mess with the atime bits set
by a more privileged user.
The additional restrictions are implemented with the addition of
MNT_LOCK_NOSUID, MNT_LOCK_NODEV, MNT_LOCK_NOEXEC, and MNT_LOCK_ATIME
mnt flags.
Taken together these changes and the fixes for MNT_LOCK_READONLY
should make it safe for an unprivileged user to create a user
namespace and to call "mount --bind -o remount,... ..." without
the danger of mount flags being changed maliciously.
Cc: stable@vger.kernel.org
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
2014-07-28 17:26:07 -07:00
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-03 16:09:14 -05:00
|
|
|
|
2023-11-22 12:17:37 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-07-14 14:45:41 +02:00
|
|
|
|
2023-11-22 12:17:37 -05:00
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:48 +01:00
|
|
|
|
|
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
2021-01-21 14:19:48 +01:00
|
|
|
|
|
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
|
|
|
|
|
2008-09-26 19:01:20 -07:00
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 09:28:36 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 20:01:19 -05:00
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-25 01:05:37 -05:00
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-08-14 01:44:31 -04:00
|
|
|
|
|
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
2025-08-14 01:44:31 -04:00
|
|
|
|
|
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 14:23:52 -04:00
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-03 17:57:27 -04:00
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 14:02:26 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-01 14:02:26 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2025-06-01 14:02:26 -04:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2025-06-01 14:02:26 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-21 00:15:55 +01:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
2011-11-24 21:47:05 -05:00
|
|
|
|
2011-11-24 19:59:16 -05:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2013-02-22 22:49:10 -05:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
Add a dentry op to allow processes to be held during pathwalk transit
Add a dentry op (d_manage) to permit a filesystem to hold a process and make it
sleep when it tries to transit away from one of that filesystem's directories
during a pathwalk. The operation is keyed off a new dentry flag
(DCACHE_MANAGE_TRANSIT).
The filesystem is allowed to be selective about which processes it holds and
which it permits to continue on or prohibits from transiting from each flagged
directory. This will allow autofs to hold up client processes whilst letting
its userspace daemon through to maintain the directory or the stuff behind it
or mounted upon it.
The ->d_manage() dentry operation:
int (*d_manage)(struct path *path, bool mounting_here);
takes a pointer to the directory about to be transited away from and a flag
indicating whether the transit is undertaken by do_add_mount() or
do_move_mount() skipping through a pile of filesystems mounted on a mountpoint.
It should return 0 if successful and to let the process continue on its way;
-EISDIR to prohibit the caller from skipping to overmounted filesystems or
automounting, and to use this directory; or some other error code to return to
the user.
->d_manage() is called with namespace_sem writelocked if mounting_here is true
and no other locks held, so it may sleep. However, if mounting_here is true,
it may not initiate or wait for a mount or unmount upon the parameter
directory, even if the act is actually performed by userspace.
Within fs/namei.c, follow_managed() is extended to check with d_manage() first
on each managed directory, before transiting away from it or attempting to
automount upon it.
follow_down() is renamed follow_down_one() and should only be used where the
filesystem deliberately intends to avoid management steps (e.g. autofs).
A new follow_down() is added that incorporates the loop done by all other
callers of follow_down() (do_add/move_mount(), autofs and NFSD; whilst AFS, NFS
and CIFS do use it, their use is removed by converting them to use
d_automount()). The new follow_down() calls d_manage() as appropriate. It
also takes an extra parameter to indicate if it is being called from mount code
(with namespace_sem writelocked) which it passes to d_manage(). follow_down()
ignores automount points so that it can be used to mount on them.
__follow_mount_rcu() is made to abort rcu-walk mode if it hits a directory with
DCACHE_MANAGE_TRANSIT set on the basis that we're probably going to have to
sleep. It would be possible to enter d_manage() in rcu-walk mode too, and have
that determine whether to abort or not itself. That would allow the autofs
daemon to continue on in rcu-walk mode.
Note that DCACHE_MANAGE_TRANSIT on a directory should be cleared when it isn't
required as every tranist from that directory will cause d_manage() to be
invoked. It can always be set again when necessary.
==========================
WHAT THIS MEANS FOR AUTOFS
==========================
Autofs currently uses the lookup() inode op and the d_revalidate() dentry op to
trigger the automounting of indirect mounts, and both of these can be called
with i_mutex held.
autofs knows that the i_mutex will be held by the caller in lookup(), and so
can drop it before invoking the daemon - but this isn't so for d_revalidate(),
since the lock is only held on _some_ of the code paths that call it. This
means that autofs can't risk dropping i_mutex from its d_revalidate() function
before it calls the daemon.
The bug could manifest itself as, for example, a process that's trying to
validate an automount dentry that gets made to wait because that dentry is
expired and needs cleaning up:
mkdir S ffffffff8014e05a 0 32580 24956
Call Trace:
[<ffffffff885371fd>] :autofs4:autofs4_wait+0x674/0x897
[<ffffffff80127f7d>] avc_has_perm+0x46/0x58
[<ffffffff8009fdcf>] autoremove_wake_function+0x0/0x2e
[<ffffffff88537be6>] :autofs4:autofs4_expire_wait+0x41/0x6b
[<ffffffff88535cfc>] :autofs4:autofs4_revalidate+0x91/0x149
[<ffffffff80036d96>] __lookup_hash+0xa0/0x12f
[<ffffffff80057a2f>] lookup_create+0x46/0x80
[<ffffffff800e6e31>] sys_mkdirat+0x56/0xe4
versus the automount daemon which wants to remove that dentry, but can't
because the normal process is holding the i_mutex lock:
automount D ffffffff8014e05a 0 32581 1 32561
Call Trace:
[<ffffffff80063c3f>] __mutex_lock_slowpath+0x60/0x9b
[<ffffffff8000ccf1>] do_path_lookup+0x2ca/0x2f1
[<ffffffff80063c89>] .text.lock.mutex+0xf/0x14
[<ffffffff800e6d55>] do_rmdir+0x77/0xde
[<ffffffff8005d229>] tracesys+0x71/0xe0
[<ffffffff8005d28d>] tracesys+0xd5/0xe0
which means that the system is deadlocked.
This patch allows autofs to hold up normal processes whilst the daemon goes
ahead and does things to the dentry tree behind the automouter point without
risking a deadlock as almost no locks are held in d_manage() and none in
d_automount().
Signed-off-by: David Howells <dhowells@redhat.com>
Was-Acked-by: Ian Kent <raven@themaw.net>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-01-14 18:45:26 +00:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2019-06-29 13:27:44 -07:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
2019-06-29 13:27:44 -07:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:20:03 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-11-07 17:21:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 01:05:37 -05:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
2011-11-25 01:05:37 -05:00
|
|
|
|
2011-11-24 21:47:05 -05:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2008-02-14 19:34:32 -08:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:31 +00:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2008-08-02 00:51:11 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2021-06-18 20:27:57 -04:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
smarter propagate_mnt()
The current mainline has copies propagated to *all* nodes, then
tears down the copies we made for nodes that do not contain
counterparts of the desired mountpoint. That sets the right
propagation graph for the copies (at teardown time we move
the slaves of removed node to a surviving peer or directly
to master), but we end up paying a fairly steep price in
useless allocations. It's fairly easy to create a situation
where N calls of mount(2) create exactly N bindings, with
O(N^2) vfsmounts allocated and freed in process.
Fortunately, it is possible to avoid those allocations/freeings.
The trick is to create copies in the right order and find which
one would've eventually become a master with the current algorithm.
It turns out to be possible in O(nodes getting propagation) time
and with no extra allocations at all.
One part is that we need to make sure that eventual master will be
created before its slaves, so we need to walk the propagation
tree in a different order - by peer groups. And iterate through
the peers before dealing with the next group.
Another thing is finding the (earlier) copy that will be a master
of one we are about to create; to do that we are (temporary) marking
the masters of mountpoints we are attaching the copies to.
Either we are in a peer of the last mountpoint we'd dealt with,
or we have the following situation: we are attaching to mountpoint M,
the last copy S_0 had been attached to M_0 and there are sequences
S_0...S_n, M_0...M_n such that S_{i+1} is a master of S_{i},
S_{i} mounted on M{i} and we need to create a slave of the first S_{k}
such that M is getting propagation from M_{k}. It means that the master
of M_{k} will be among the sequence of masters of M. On the
other hand, the nearest marked node in that sequence will either
be the master of M_{k} or the master of M_{k-1} (the latter -
in the case if M_{k-1} is a slave of something M gets propagation
from, but in a wrong peer group).
So we go through the sequence of masters of M until we find
a marked one (P). Let N be the one before it. Then we go through
the sequence of masters of S_0 until we find one (say, S) mounted
on a node D that has P as master and check if D is a peer of N.
If it is, S will be the master of new copy, if not - the master of S
will be.
That's it for the hard part; the rest is fairly simple. Iterator
is in next_group(), handling of one prospective mountpoint is
propagate_one().
It seems to survive all tests and gives a noticably better performance
than the current mainline for setups that are seriously using shared
subtrees.
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-02-27 09:35:45 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2012-09-21 08:19:02 -04:00
|
|
|
|
|
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2012-09-21 08:19:02 -04:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2012-09-21 08:19:02 -04:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
VFS: (Scripted) Convert S_ISLNK/DIR/REG(dentry->d_inode) to d_is_*(dentry)
Convert the following where appropriate:
(1) S_ISLNK(dentry->d_inode) to d_is_symlink(dentry).
(2) S_ISREG(dentry->d_inode) to d_is_reg(dentry).
(3) S_ISDIR(dentry->d_inode) to d_is_dir(dentry). This is actually more
complicated than it appears as some calls should be converted to
d_can_lookup() instead. The difference is whether the directory in
question is a real dir with a ->lookup op or whether it's a fake dir with
a ->d_automount op.
In some circumstances, we can subsume checks for dentry->d_inode not being
NULL into this, provided we the code isn't in a filesystem that expects
d_inode to be NULL if the dirent really *is* negative (ie. if we're going to
use d_inode() rather than d_backing_inode() to get the inode pointer).
Note that the dentry type field may be set to something other than
DCACHE_MISS_TYPE when d_inode is NULL in the case of unionmount, where the VFS
manages the fall-through from a negative dentry to a lower layer. In such a
case, the dentry type of the negative union dentry is set to the same as the
type of the lower dentry.
However, if you know d_inode is not NULL at the call site, then you can use
the d_is_xxx() functions even in a filesystem.
There is one further complication: a 0,0 chardev dentry may be labelled
DCACHE_WHITEOUT_TYPE rather than DCACHE_SPECIAL_TYPE. Strictly, this was
intended for special directory entry types that don't have attached inodes.
The following perl+coccinelle script was used:
use strict;
my @callers;
open($fd, 'git grep -l \'S_IS[A-Z].*->d_inode\' |') ||
die "Can't grep for S_ISDIR and co. callers";
@callers = <$fd>;
close($fd);
unless (@callers) {
print "No matches\n";
exit(0);
}
my @cocci = (
'@@',
'expression E;',
'@@',
'',
'- S_ISLNK(E->d_inode->i_mode)',
'+ d_is_symlink(E)',
'',
'@@',
'expression E;',
'@@',
'',
'- S_ISDIR(E->d_inode->i_mode)',
'+ d_is_dir(E)',
'',
'@@',
'expression E;',
'@@',
'',
'- S_ISREG(E->d_inode->i_mode)',
'+ d_is_reg(E)' );
my $coccifile = "tmp.sp.cocci";
open($fd, ">$coccifile") || die $coccifile;
print($fd "$_\n") || die $coccifile foreach (@cocci);
close($fd);
foreach my $file (@callers) {
chomp $file;
print "Processing ", $file, "\n";
system("spatch", "--sp-file", $coccifile, $file, "--in-place", "--no-show-diff") == 0 ||
die "spatch failed";
}
[AV: overlayfs parts skipped]
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-01-29 12:02:35 +00:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2011-11-25 00:30:56 -05:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2011-01-17 01:47:59 -05:00
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-12-20 15:04:50 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-04-15 14:17:12 -07:00
|
|
|
|
|
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-10-16 19:48:14 -07:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
|
|
|
|
|
2015-05-08 23:22:29 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2012-10-11 11:42:01 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2012-07-26 21:42:03 -07:00
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2012-07-26 21:42:03 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2012-07-26 21:42:03 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-06-09 16:06:06 -05:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
2012-07-26 21:42:03 -07:00
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-22 12:17:37 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
|
|
|
|
|
2019-05-13 12:57:22 -04:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:18:51 -05:00
|
|
|
|
|
|
|
|
|
2011-01-17 01:41:58 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2021-06-18 20:27:57 -04:00
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
2020-01-11 11:27:46 -05:00
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
2020-01-11 11:27:46 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
|
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
2011-01-17 01:47:59 -05:00
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
|
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
|
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
2020-01-11 10:14:09 -05:00
|
|
|
|
|
|
|
|
|
fix automount/automount race properly
Protection against automount/automount races (two threads hitting the same
referral point at the same time) is based upon do_add_mount() prevention of
identical overmounts - trying to overmount the root of mounted tree with
the same tree fails with -EBUSY. It's unreliable (the other thread might've
mounted something on top of the automount it has triggered) *and* causes
no end of headache for follow_automount() and its caller, since
finish_automount() behaves like do_new_mount() - if the mountpoint to be is
overmounted, it mounts on top what's overmounting it. It's not only wrong
(we want to go into what's overmounting the automount point and quietly
discard what we planned to mount there), it introduces the possibility of
original parent mount getting dropped. That's what 8aef18845266 (VFS: Fix
vfsmount overput on simultaneous automount) deals with, but it can't do
anything about the reliability of conflict detection - if something had
been overmounted the other thread's automount (e.g. that other thread
having stepped into automount in mount(2)), we don't get that -EBUSY and
the result is
referral point under automounted NFS under explicit overmount
under another copy of automounted NFS
What we need is finish_automount() *NOT* digging into overmounts - if it
finds one, it should just quietly discard the thing it was asked to mount.
And don't bother with actually crossing into the results of finish_automount() -
the same loop that calls follow_automount() will do that just fine on the
next iteration.
IOW, instead of calling lock_mount() have finish_automount() do it manually,
_without_ the "move into overmount and retry" part. And leave crossing into
the results to the caller of follow_automount(), which simplifies it a lot.
Moral: if you end up with a lot of glue working around the calling conventions
of something, perhaps these calling conventions are simply wrong...
Fixes: 8aef18845266 (VFS: Fix vfsmount overput on simultaneous automount)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2020-01-11 10:44:29 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-01-17 01:47:59 -05:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
2011-01-17 01:47:59 -05:00
|
|
|
|
|
|
|
|
|
2011-01-17 01:35:23 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-01-14 19:10:03 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2011-01-14 19:10:03 +00:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2011-01-14 19:10:03 +00:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2011-01-14 19:10:03 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:07:43 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2011-11-25 00:57:42 -05:00
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2008-03-22 00:21:53 -04:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2008-03-22 00:21:53 -04:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2013-03-16 14:42:19 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:15:14 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
2011-11-24 21:15:14 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
2011-11-24 21:15:14 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 21:35:16 -05:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 23:24:33 -05:00
|
|
|
|
2011-11-24 22:19:58 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: brlock vfsmount_lock
fs: brlock vfsmount_lock
Use a brlock for the vfsmount lock. It must be taken for write whenever
modifying the mount hash or associated fields, and may be taken for read when
performing mount hash lookups.
A new lock is added for the mnt-id allocator, so it doesn't need to take
the heavy vfsmount write-lock.
The number of atomics should remain the same for fastpath rlock cases, though
code would be slightly slower due to per-cpu access. Scalability is not not be
much improved in common cases yet, due to other locks (ie. dcache_lock) getting
in the way. However path lookups crossing mountpoints should be one case where
scalability is improved (currently requiring the global lock).
The slowpath is slower due to use of brlock. On a 64 core, 64 socket, 32 node
Altix system (high latency to remote nodes), a simple umount microbenchmark
(mount --bind mnt mnt2 ; umount mnt2 loop 1000 times), before this patch it
took 6.8s, afterwards took 7.1s, about 5% slower.
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2010-08-18 04:37:39 +10:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
2013-03-16 14:39:34 -04:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
2011-11-24 21:07:43 -05:00
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
2008-03-22 00:46:23 -04:00
|
|
|
|
2008-03-22 00:21:53 -04:00
|
|
|
|
2011-11-24 21:07:43 -05:00
|
|
|
|
2011-11-25 00:22:05 -05:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2014-12-24 07:20:01 -06:00
|
|
|
|
2008-03-22 00:21:53 -04:00
|
|
|
|
|
|
|
|
|
2006-06-09 09:34:17 -04:00
|
|
|
|
|
|
|
|
|
2020-09-17 10:22:34 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
2020-07-01 17:46:06 +01:00
|
|
|
|
2005-11-07 17:16:09 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2020-07-01 17:46:06 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2020-07-01 17:46:06 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2020-07-01 17:46:06 +01:00
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2020-09-17 10:22:34 +02:00
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
2019-01-22 12:21:52 +05:30
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-07-21 11:12:08 +02:00
|
|
|
|
2012-10-11 11:42:01 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
2020-06-04 09:41:08 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-06-04 09:41:08 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-08-19 14:56:38 -04:00
|
|
|
|
|
|
|
|
|
2009-10-04 21:49:49 +09:00
|
|
|
|
2009-04-19 18:40:43 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-03-26 17:53:14 +00:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-01-09 20:52:17 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-03-26 17:49:56 +00:00
|
|
|
|
|
|
|
|
|
2018-04-20 13:35:02 +01:00
|
|
|
|
2008-02-15 14:38:00 -08:00
|
|
|
|
2020-08-27 11:09:46 -06:00
|
|
|
|
|
|
|
|
|
2006-01-09 20:52:17 -08:00
|
|
|
|
2014-07-28 17:36:04 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-06-04 09:41:08 +02:00
|
|
|
|
2014-07-28 17:36:04 -07:00
|
|
|
|
|
|
|
|
|
2017-07-17 08:45:35 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-10-08 00:28:21 -04:00
|
|
|
|
2017-10-11 07:01:31 +02:00
|
|
|
|
2017-10-08 00:28:21 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2018-11-01 23:07:25 +00:00
|
|
|
|
2020-06-04 09:41:08 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-08-02 00:51:11 -04:00
|
|
|
|
2020-06-04 09:41:08 +02:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-07-26 21:08:32 -07:00
|
|
|
|
|
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
2012-07-26 21:08:32 -07:00
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
|
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
2011-06-15 10:21:48 -07:00
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
|
|
|
|
|
2016-09-22 13:08:36 -05:00
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
2021-09-02 14:55:27 -07:00
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
|
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-06-15 10:21:48 -07:00
|
|
|
|
2014-11-01 02:32:53 -04:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
2020-08-03 13:16:42 +03:00
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
2012-07-26 21:08:32 -07:00
|
|
|
|
2016-08-08 14:37:37 -05:00
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-06-20 20:42:34 +02:00
|
|
|
|
2013-09-28 20:47:57 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2008-05-10 20:44:54 -04:00
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2013-09-28 20:47:57 -04:00
|
|
|
|
2011-11-24 20:55:08 -05:00
|
|
|
|
2012-07-31 13:13:04 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-09-28 20:47:57 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-03-16 15:12:40 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
2013-09-28 20:47:57 -04:00
|
|
|
|
2019-01-30 13:15:45 -05:00
|
|
|
|
2012-07-31 13:13:04 -07:00
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2013-03-16 14:49:45 -04:00
|
|
|
|
2024-10-16 19:49:48 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-06-25 12:55:18 +01:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2019-01-30 13:15:45 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-12-06 13:32:36 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
2011-11-24 20:55:08 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2013-09-28 20:47:57 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-09-28 20:47:57 -04:00
|
|
|
|
|
|
|
|
|
2011-11-24 18:57:30 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-11-25 03:06:56 -05:00
|
|
|
|
|
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
|
|
|
|
|
2022-11-24 22:55:57 -05:00
|
|
|
|
2013-03-30 01:35:18 -07:00
|
|
|
|
2022-11-24 22:55:57 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
2013-03-16 14:49:45 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2011-01-14 22:30:21 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-01-14 22:30:21 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-02-07 12:59:00 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
2011-11-22 12:31:21 -05:00
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-22 12:31:21 -05:00
|
|
|
|
|
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
|
|
|
|
|
2011-11-22 12:31:21 -05:00
|
|
|
|
2011-11-16 21:43:59 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
init: use do_mount() instead of ksys_mount()
In prepare_namespace(), do_mount() can be used instead of ksys_mount()
as the first and third argument are const strings in the kernel, the
second and fourth argument are passed through anyway, and the fifth
argument is NULL.
In do_mount_root(), ksys_mount() is called with the first and third
argument being already kernelspace strings, which do not need to be
copied over from userspace to kernelspace (again). The second and
fourth arguments are passed through to do_mount() anyway. The fifth
argument, while already residing in kernelspace, needs to be put into
a page of its own. Then, do_mount() can be used instead of
ksys_mount().
Once this is done, there are no in-kernel users to ksys_mount() left,
which can therefore be removed.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2018-10-23 22:41:09 +02:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2014-08-28 11:26:03 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2014-08-28 11:26:03 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
|
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2015-12-14 18:44:44 -05:00
|
|
|
|
fs: fix overflow in sys_mount() for in-kernel calls
sys_mount() reads/copies a whole page for its "type" parameter. When
do_mount_root() passes a kernel address that points to an object which is
smaller than a whole page, copy_mount_options() will happily go past this
memory object, possibly dereferencing "wild" pointers that could be in any
state (hence the kmemcheck warning, which shows that parts of the next
page are not even allocated).
(The likelihood of something going wrong here is pretty low -- first of
all this only applies to kernel calls to sys_mount(), which are mostly
found in the boot code. Secondly, I guess if the page was not mapped,
exact_copy_from_user() _would_ in fact handle it correctly because of its
access_ok(), etc. checks.)
But it is much nicer to avoid the dubious reads altogether, by stopping as
soon as we find a NUL byte. Is there a good reason why we can't do
something like this, using the already existing strndup_from_user()?
[akpm@linux-foundation.org: make copy_mount_string() static]
[AV: fix compat mount breakage, which involves undoing akpm's change above]
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: al <al@dizzy.pdmi.ras.ru>
2009-09-18 13:05:45 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2021-06-01 11:33:59 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:52 +01:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:52 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-06-01 11:33:59 +02:00
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:52 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:52 +01:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:52 +01:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-05-31 14:12:01 -04:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-05-31 14:12:01 -04:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
2024-05-31 14:12:01 -04:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-08-19 14:56:38 -04:00
|
|
|
|
|
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2019-06-12 11:43:13 -07:00
|
|
|
|
2018-11-01 23:36:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-10-15 11:35:02 +01:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-07-15 13:07:13 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
|
|
|
|
|
2018-11-05 17:40:30 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
2011-11-24 22:00:28 -05:00
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
|
|
|
|
|
2011-11-24 22:00:28 -05:00
|
|
|
|
2011-11-24 22:25:07 -05:00
|
|
|
|
2011-11-24 22:19:58 -05:00
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
2011-11-24 22:00:28 -05:00
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
|
|
|
|
|
2016-11-14 22:14:35 +01:00
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
2015-11-17 14:40:10 +08:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2011-11-24 22:00:28 -05:00
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2011-11-23 19:34:49 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-01-08 01:03:18 -08:00
|
|
|
|
2020-04-14 18:48:37 +02:00
|
|
|
|
2006-01-08 01:03:18 -08:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-01-14 14:14:16 +01:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2013-02-22 22:45:42 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2019-07-14 16:42:44 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-07-14 16:42:44 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-07-22 09:59:21 -04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2010-08-10 11:41:36 +02:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-24 19:41:16 -05:00
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2013-03-29 21:04:39 -07:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2009-05-04 03:32:03 +04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2011-11-24 21:47:05 -05:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2011-11-24 21:47:05 -05:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2008-02-14 19:34:32 -08:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2014-10-08 10:42:27 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2013-03-29 21:04:39 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-02-14 19:34:32 -08:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2008-02-14 19:34:32 -08:00
|
|
|
|
fs: allow to mount beneath top mount
Various distributions are adding or are in the process of adding support
for system extensions and in the future configuration extensions through
various tools. A more detailed explanation on system and configuration
extensions can be found on the manpage which is listed below at [1].
System extension images may – dynamically at runtime — extend the /usr/
and /opt/ directory hierarchies with additional files. This is
particularly useful on immutable system images where a /usr/ and/or
/opt/ hierarchy residing on a read-only file system shall be extended
temporarily at runtime without making any persistent modifications.
When one or more system extension images are activated, their /usr/ and
/opt/ hierarchies are combined via overlayfs with the same hierarchies
of the host OS, and the host /usr/ and /opt/ overmounted with it
("merging"). When they are deactivated, the mount point is disassembled
— again revealing the unmodified original host version of the hierarchy
("unmerging"). Merging thus makes the extension's resources suddenly
appear below the /usr/ and /opt/ hierarchies as if they were included in
the base OS image itself. Unmerging makes them disappear again, leaving
in place only the files that were shipped with the base OS image itself.
System configuration images are similar but operate on directories
containing system or service configuration.
On nearly all modern distributions mount propagation plays a crucial
role and the rootfs of the OS is a shared mount in a peer group (usually
with peer group id 1):
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:1 29 1
On such systems all services and containers run in a separate mount
namespace and are pivot_root()ed into their rootfs. A separate mount
namespace is almost always used as it is the minimal isolation mechanism
services have. But usually they are even much more isolated up to the
point where they almost become indistinguishable from containers.
Mount propagation again plays a crucial role here. The rootfs of all
these services is a slave mount to the peer group of the host rootfs.
This is done so the service will receive mount propagation events from
the host when certain files or directories are updated.
In addition, the rootfs of each service, container, and sandbox is also
a shared mount in its separate peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/ / ext4 shared:24 master:1 71 47
For people not too familiar with mount propagation, the master:1 means
that this is a slave mount to peer group 1. Which as one can see is the
host rootfs as indicated by shared:1 above. The shared:24 indicates that
the service rootfs is a shared mount in a separate peer group with peer
group id 24.
A service may run other services. Such nested services will also have a
rootfs mount that is a slave to the peer group of the outer service
rootfs mount.
For containers things are just slighly different. A container's rootfs
isn't a slave to the service's or host rootfs' peer group. The rootfs
mount of a container is simply a shared mount in its own peer group:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/home/ubuntu/debian-tree / ext4 shared:99 61 60
So whereas services are isolated OS components a container is treated
like a separate world and mount propagation into it is restricted to a
single well known mount that is a slave to the peer group of the shared
mount /run on the host:
TARGET SOURCE FSTYPE PROPAGATION MNT_ID PARENT_ID
/propagate/debian-tree /run/host/incoming tmpfs master:5 71 68
Here, the master:5 indicates that this mount is a slave to the peer
group with peer group id 5. This allows to propagate mounts into the
container and served as a workaround for not being able to insert mounts
into mount namespaces directly. But the new mount api does support
inserting mounts directly. For the interested reader the blogpost in [2]
might be worth reading where I explain the old and the new approach to
inserting mounts into mount namespaces.
Containers of course, can themselves be run as services. They often run
full systems themselves which means they again run services and
containers with the exact same propagation settings explained above.
The whole system is designed so that it can be easily updated, including
all services in various fine-grained ways without having to enter every
single service's mount namespace which would be prohibitively expensive.
The mount propagation layout has been carefully chosen so it is possible
to propagate updates for system extensions and configurations from the
host into all services.
The simplest model to update the whole system is to mount on top of
/usr, /opt, or /etc on the host. The new mount on /usr, /opt, or /etc
will then propagate into every service. This works cleanly the first
time. However, when the system is updated multiple times it becomes
necessary to unmount the first update on /opt, /usr, /etc and then
propagate the new update. But this means, there's an interval where the
old base system is accessible. This has to be avoided to protect against
downgrade attacks.
The vfs already exposes a mechanism to userspace whereby mounts can be
mounted beneath an existing mount. Such mounts are internally referred
to as "tucked". The patch series exposes the ability to mount beneath a
top mount through the new MOVE_MOUNT_BENEATH flag for the move_mount()
system call. This allows userspace to seamlessly upgrade mounts. After
this series the only thing that will have changed is that mounting
beneath an existing mount can be done explicitly instead of just
implicitly.
Today, there are two scenarios where a mount can be mounted beneath an
existing mount instead of on top of it:
(1) When a service or container is started in a new mount namespace and
pivot_root()s into its new rootfs. The way this is done is by
mounting the new rootfs beneath the old rootfs:
fd_newroot = open("/var/lib/machines/fedora", ...);
fd_oldroot = open("/", ...);
fchdir(fd_newroot);
pivot_root(".", ".");
After the pivot_root(".", ".") call the new rootfs is mounted
beneath the old rootfs which can then be unmounted to reveal the
underlying mount:
fchdir(fd_oldroot);
umount2(".", MNT_DETACH);
Since pivot_root() moves the caller into a new rootfs no mounts must
be propagated out of the new rootfs as a consequence of the
pivot_root() call. Thus, the mounts cannot be shared.
(2) When a mount is propagated to a mount that already has another mount
mounted on the same dentry.
The easiest example for this is to create a new mount namespace. The
following commands will create a mount namespace where the rootfs
mount / will be a slave to the peer group of the host rootfs /
mount's peer group. IOW, it will receive propagation from the host:
mount --make-shared /
unshare --mount --propagation=slave
Now a new mount on the /mnt dentry in that mount namespace is
created. (As it can be confusing it should be spelled out that the
tmpfs mount on the /mnt dentry that was just created doesn't
propagate back to the host because the rootfs mount / of the mount
namespace isn't a peer of the host rootfs.):
mount -t tmpfs tmpfs /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt tmpfs tmpfs
Now another terminal in the host mount namespace can observe that
the mount indeed hasn't propagated back to into the host mount
namespace. A new mount can now be created on top of the /mnt dentry
with the rootfs mount / as its parent:
mount --bind /opt /mnt
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 shared:1
The mount namespace that was created earlier can now observe that
the bind mount created on the host has propagated into it:
TARGET SOURCE FSTYPE PROPAGATION
└─/mnt /dev/sda2[/opt] ext4 master:1
└─/mnt tmpfs tmpfs
But instead of having been mounted on top of the tmpfs mount at the
/mnt dentry the /opt mount has been mounted on top of the rootfs
mount at the /mnt dentry. And the tmpfs mount has been remounted on
top of the propagated /opt mount at the /opt dentry. So in other
words, the propagated mount has been mounted beneath the preexisting
mount in that mount namespace.
Mount namespaces make this easy to illustrate but it's also easy to
mount beneath an existing mount in the same mount namespace
(The following example assumes a shared rootfs mount / with peer
group id 1):
mount --bind /opt /opt
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/opt] ext4 188 29 shared:1
If another mount is mounted on top of the /opt mount at the /opt
dentry:
mount --bind /tmp /opt
The following clunky mount tree will result:
TARGET SOURCE FSTYPE MNT_ID PARENT_ID PROPAGATION
└─/opt /dev/sda2[/tmp] ext4 405 29 shared:1
└─/opt /dev/sda2[/opt] ext4 188 405 shared:1
└─/opt /dev/sda2[/tmp] ext4 404 188 shared:1
The /tmp mount is mounted beneath the /opt mount and another copy is
mounted on top of the /opt mount. This happens because the rootfs /
and the /opt mount are shared mounts in the same peer group.
When the new /tmp mount is supposed to be mounted at the /opt dentry
then the /tmp mount first propagates to the root mount at the /opt
dentry. But there already is the /opt mount mounted at the /opt
dentry. So the old /opt mount at the /opt dentry will be mounted on
top of the new /tmp mount at the /tmp dentry, i.e. @opt->mnt_parent
is @tmp and @opt->mnt_mountpoint is /tmp (Note that @opt->mnt_root
is /opt which is what shows up as /opt under SOURCE). So again, a
mount will be mounted beneath a preexisting mount.
(Fwiw, a few iterations of mount --bind /opt /opt in a loop on a
shared rootfs is a good example of what could be referred to as
mount explosion.)
The main point is that such mounts allows userspace to umount a top
mount and reveal an underlying mount. So for example, umounting the
tmpfs mount on /mnt that was created in example (1) using mount
namespaces reveals the /opt mount which was mounted beneath it.
In (2) where a mount was mounted beneath the top mount in the same mount
namespace unmounting the top mount would unmount both the top mount and
the mount beneath. In the process the original mount would be remounted
on top of the rootfs mount / at the /opt dentry again.
This again, is a result of mount propagation only this time it's umount
propagation. However, this can be avoided by simply making the parent
mount / of the @opt mount a private or slave mount. Then the top mount
and the original mount can be unmounted to reveal the mount beneath.
These two examples are fairly arcane and are merely added to make it
clear how mount propagation has effects on current and future features.
More common use-cases will just be things like:
mount -t btrfs /dev/sdA /mnt
mount -t xfs /dev/sdB --beneath /mnt
umount /mnt
after which we'll have updated from a btrfs filesystem to a xfs
filesystem without ever revealing the underlying mountpoint.
The crux is that the proposed mechanism already exists and that it is so
powerful as to cover cases where mounts are supposed to be updated with
new versions. Crucially, it offers an important flexibility. Namely that
updates to a system may either be forced or can be delayed and the
umount of the top mount be left to a service if it is a cooperative one.
This adds a new flag to move_mount() that allows to explicitly move a
beneath the top mount adhering to the following semantics:
* Mounts cannot be mounted beneath the rootfs. This restriction
encompasses the rootfs but also chroots via chroot() and pivot_root().
To mount a mount beneath the rootfs or a chroot, pivot_root() can be
used as illustrated above.
* The source mount must be a private mount to force the kernel to
allocate a new, unused peer group id. This isn't a required
restriction but a voluntary one. It avoids repeating a semantical
quirk that already exists today. If bind mounts which already have a
peer group id are inserted into mount trees that have the same peer
group id this can cause a lot of mount propagation events to be
generated (For example, consider running mount --bind /opt /opt in a
loop where the parent mount is a shared mount.).
* Avoid getting rid of the top mount in the kernel. Cooperative services
need to be able to unmount the top mount themselves.
This also avoids a good deal of additional complexity. The umount
would have to be propagated which would be another rather expensive
operation. So namespace_lock() and lock_mount_hash() would potentially
have to be held for a long time for both a mount and umount
propagation. That should be avoided.
* The path to mount beneath must be mounted and attached.
* The top mount and its parent must be in the caller's mount namespace
and the caller must be able to mount in that mount namespace.
* The caller must be able to unmount the top mount to prove that they
could reveal the underlying mount.
* The propagation tree is calculated based on the destination mount's
parent mount and the destination mount's mountpoint on the parent
mount. Of course, if the parent of the destination mount and the
destination mount are shared mounts in the same peer group and the
mountpoint of the new mount to be mounted is a subdir of their
->mnt_root then both will receive a mount of /opt. That's probably
easier to understand with an example. Assuming a standard shared
rootfs /:
mount --bind /opt /opt
mount --bind /tmp /opt
will cause the same mount tree as:
mount --bind /opt /opt
mount --beneath /tmp /opt
because both / and /opt are shared mounts/peers in the same peer
group and the /opt dentry is a subdirectory of both the parent's and
the child's ->mnt_root. If a mount tree like that is created it almost
always is an accident or abuse of mount propagation. Realistically
what most people probably mean in this scenarios is:
mount --bind /opt /opt
mount --make-private /opt
mount --make-shared /opt
This forces the allocation of a new separate peer group for the /opt
mount. Aferwards a mount --bind or mount --beneath actually makes
sense as the / and /opt mount belong to different peer groups. Before
that it's likely just confusion about what the user wanted to achieve.
* Refuse MOVE_MOUNT_BENEATH if:
(1) the @mnt_from has been overmounted in between path resolution and
acquiring @namespace_sem when locking @mnt_to. This avoids the
proliferation of shadow mounts.
(2) if @to_mnt is moved to a different mountpoint while acquiring
@namespace_sem to lock @to_mnt.
(3) if @to_mnt is unmounted while acquiring @namespace_sem to lock
@to_mnt.
(4) if the parent of the target mount propagates to the target mount
at the same mountpoint.
This would mean mounting @mnt_from on @mnt_to->mnt_parent and then
propagating a copy @c of @mnt_from onto @mnt_to. This defeats the
whole purpose of mounting @mnt_from beneath @mnt_to.
(5) if the parent mount @mnt_to->mnt_parent propagates to @mnt_from at
the same mountpoint.
If @mnt_to->mnt_parent propagates to @mnt_from this would mean
propagating a copy @c of @mnt_from on top of @mnt_from. Afterwards
@mnt_from would be mounted on top of @mnt_to->mnt_parent and
@mnt_to would be unmounted from @mnt->mnt_parent and remounted on
@mnt_from. But since @c is already mounted on @mnt_from, @mnt_to
would ultimately be remounted on top of @c. Afterwards, @mnt_from
would be covered by a copy @c of @mnt_from and @c would be covered
by @mnt_from itself. This defeats the whole purpose of mounting
@mnt_from beneath @mnt_to.
Cases (1) to (3) are required as they deal with races that would cause
bugs or unexpected behavior for users. Cases (4) and (5) refuse
semantical quirks that would not be a bug but would cause weird mount
trees to be created. While they can already be created via other means
(mount --bind /opt /opt x n) there's no reason to repeat past mistakes
in new features.
Link: https://man7.org/linux/man-pages/man8/systemd-sysext.8.html [1]
Link: https://brauner.io/2023/02/28/mounting-into-mount-namespaces.html [2]
Link: https://github.com/flatcar/sysext-bakery
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1
Link: https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_2
Link: https://github.com/systemd/systemd/pull/26013
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Message-Id: <20230202-fs-move-mount-replace-v4-4-98f3d80d7eaa@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-05-03 13:18:42 +02:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2014-10-08 10:42:57 -07:00
|
|
|
|
|
|
|
|
|
2017-01-03 14:18:43 +13:00
|
|
|
|
2013-09-29 11:24:49 -04:00
|
|
|
|
2008-07-22 09:59:21 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2019-06-30 19:18:53 -04:00
|
|
|
|
|
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2008-03-22 18:00:39 -04:00
|
|
|
|
2011-03-18 08:55:38 -04:00
|
|
|
|
2008-07-22 09:59:21 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2008-07-22 09:59:21 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-22 13:44:37 +01:00
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-12-03 12:16:58 +01:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-09-03 17:16:12 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:08 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-05-10 11:58:40 +02:00
|
|
|
|
|
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
2022-02-03 14:14:08 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-04-20 15:19:25 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-04-20 15:19:25 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-03 14:14:09 +01:00
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2023-05-03 13:18:39 +02:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-02-06 11:22:09 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
2024-02-06 11:22:09 +01:00
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-28 23:04:20 -05:00
|
|
|
|
|
|
|
|
|
2022-02-03 14:14:11 +01:00
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-03-30 09:13:16 +02:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-06-06 00:30:50 -04:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-06-06 00:30:50 -04:00
|
|
|
|
2024-05-31 14:12:01 -04:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
2024-05-31 14:12:01 -04:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-05-31 14:12:01 -04:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-12-03 12:17:07 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
2023-01-13 12:49:33 +01:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-08-16 11:47:52 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-06-06 00:30:50 -04:00
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-10-26 12:51:27 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-12-30 20:23:09 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-01-21 14:19:54 +01:00
|
|
|
|
fs: add mount_setattr()
This implements the missing mount_setattr() syscall. While the new mount
api allows to change the properties of a superblock there is currently
no way to change the properties of a mount or a mount tree using file
descriptors which the new mount api is based on. In addition the old
mount api has the restriction that mount options cannot be applied
recursively. This hasn't changed since changing mount options on a
per-mount basis was implemented in [1] and has been a frequent request
not just for convenience but also for security reasons. The legacy
mount syscall is unable to accommodate this behavior without introducing
a whole new set of flags because MS_REC | MS_REMOUNT | MS_BIND |
MS_RDONLY | MS_NOEXEC | [...] only apply the mount option to the topmost
mount. Changing MS_REC to apply to the whole mount tree would mean
introducing a significant uapi change and would likely cause significant
regressions.
The new mount_setattr() syscall allows to recursively clear and set
mount options in one shot. Multiple calls to change mount options
requesting the same changes are idempotent:
int mount_setattr(int dfd, const char *path, unsigned flags,
struct mount_attr *uattr, size_t usize);
Flags to modify path resolution behavior are specified in the @flags
argument. Currently, AT_EMPTY_PATH, AT_RECURSIVE, AT_SYMLINK_NOFOLLOW,
and AT_NO_AUTOMOUNT are supported. If useful, additional lookup flags to
restrict path resolution as introduced with openat2() might be supported
in the future.
The mount_setattr() syscall can be expected to grow over time and is
designed with extensibility in mind. It follows the extensible syscall
pattern we have used with other syscalls such as openat2(), clone3(),
sched_{set,get}attr(), and others.
The set of mount options is passed in the uapi struct mount_attr which
currently has the following layout:
struct mount_attr {
__u64 attr_set;
__u64 attr_clr;
__u64 propagation;
__u64 userns_fd;
};
The @attr_set and @attr_clr members are used to clear and set mount
options. This way a user can e.g. request that a set of flags is to be
raised such as turning mounts readonly by raising MOUNT_ATTR_RDONLY in
@attr_set while at the same time requesting that another set of flags is
to be lowered such as removing noexec from a mount tree by specifying
MOUNT_ATTR_NOEXEC in @attr_clr.
Note, since the MOUNT_ATTR_<atime> values are an enum starting from 0,
not a bitmap, users wanting to transition to a different atime setting
cannot simply specify the atime setting in @attr_set, but must also
specify MOUNT_ATTR__ATIME in the @attr_clr field. So we ensure that
MOUNT_ATTR__ATIME can't be partially set in @attr_clr and that @attr_set
can't have any atime bits set if MOUNT_ATTR__ATIME isn't set in
@attr_clr.
The @propagation field lets callers specify the propagation type of a
mount tree. Propagation is a single property that has four different
settings and as such is not really a flag argument but an enum.
Specifically, it would be unclear what setting and clearing propagation
settings in combination would amount to. The legacy mount() syscall thus
forbids the combination of multiple propagation settings too. The goal
is to keep the semantics of mount propagation somewhat simple as they
are overly complex as it is.
The @userns_fd field lets user specify a user namespace whose idmapping
becomes the idmapping of the mount. This is implemented and explained in
detail in the next patch.
[1]: commit 2e4b7fcd9260 ("[PATCH] r/o bind mounts: honor mount writer counts at remount")
Link: https://lore.kernel.org/r/20210121131959.646623-35-christian.brauner@ubuntu.com
Cc: David Howells <dhowells@redhat.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-api@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 14:19:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:01 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:49 -04:00
|
|
|
|
2024-06-24 11:49:47 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-29 16:12:53 +01:00
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
|
|
|
|
|
2025-01-29 16:12:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-29 16:12:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-30 13:15:00 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-30 13:15:00 +01:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-30 13:15:00 +01:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-30 13:15:00 +01:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
2025-01-30 13:15:00 +01:00
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-01-30 13:15:00 +01:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: find rootfs mount of the mount namespace
The method we used was predicated on the assumption that the mount
immediately following the root mount of the mount namespace would be the
rootfs mount of the namespace. That's not always the case though. For
example:
ID PARENT ID
408 412 0:60 /containers/overlay-containers/bc391117192b32071b22ef2083ebe7735d5c390f87a5779e02faf79ba0746ceb/userdata/hosts /etc/hosts rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=954664k,nr_inodes=238666,mode=700,uid=1000,gid=1000,inode64
409 414 0:61 / /dev/shm rw,nosuid,nodev,noexec,relatime - tmpfs shm rw,size=64000k,uid=1000,gid=1000,inode64
410 412 0:60 /containers/overlay-containers/bc391117192b32071b22ef2083ebe7735d5c390f87a5779e02faf79ba0746ceb/userdata/.containerenv /run/.containerenv rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=954664k,nr_inodes=238666,mode=700,uid=1000,gid=1000,inode64
411 412 0:60 /containers/overlay-containers/bc391117192b32071b22ef2083ebe7735d5c390f87a5779e02faf79ba0746ceb/userdata/hostname /etc/hostname rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=954664k,nr_inodes=238666,mode=700,uid=1000,gid=1000,inode64
412 363 0:65 / / rw,relatime - overlay overlay rw,lowerdir=/home/user1/.local/share/containers/storage/overlay/l/JS65SUCGTPCP2EEBHLRP4UCFI5:/home/user1/.local/share/containers/storage/overlay/l/DLW22KVDWUNI4242D6SDJ5GKCL [...]
413 412 0:68 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
414 412 0:69 / /dev rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,uid=1000,gid=1000,inode64
415 412 0:70 / /sys ro,nosuid,nodev,noexec,relatime - sysfs sysfs rw
416 414 0:71 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
417 414 0:67 / /dev/mqueue rw,nosuid,nodev,noexec,relatime - mqueue mqueue rw
418 415 0:27 / /sys/fs/cgroup ro,nosuid,nodev,noexec,relatime - cgroup2 cgroup2 rw,nsdelegate,memory_recursiveprot
419 414 0:6 /null /dev/null rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
420 414 0:6 /zero /dev/zero rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
422 414 0:6 /full /dev/full rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
423 414 0:6 /tty /dev/tty rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
430 414 0:6 /random /dev/random rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
431 414 0:6 /urandom /dev/urandom rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
433 413 0:72 / /proc/acpi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
440 413 0:6 /null /proc/kcore ro,nosuid - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
441 413 0:6 /null /proc/keys ro,nosuid - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
442 413 0:6 /null /proc/timer_list ro,nosuid - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
443 413 0:73 / /proc/scsi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
444 415 0:74 / /sys/firmware ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
445 415 0:75 / /sys/dev/block ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
446 413 0:68 /bus /proc/bus ro,nosuid,nodev,noexec,relatime - proc proc rw
447 413 0:68 /fs /proc/fs ro,nosuid,nodev,noexec,relatime - proc proc rw
448 413 0:68 /irq /proc/irq ro,nosuid,nodev,noexec,relatime - proc proc rw
449 413 0:68 /sys /proc/sys ro,nosuid,nodev,noexec,relatime - proc proc rw
450 413 0:68 /sysrq-trigger /proc/sysrq-trigger ro,nosuid,nodev,noexec,relatime - proc proc rw
364 414 0:71 /0 /dev/console rw,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
In this mount table the root mount of the mount namespace is the mount
with id 363 (It isn't visible because it's literally just what the
rootfs mount is mounted upon and usually it's just a copy of the real
rootfs).
The rootfs mount that's mounted on the root mount of the mount namespace
is the mount with id 412. But the mount namespace contains mounts that
were created before the rootfs mount and thus have earlier mount ids. So
the first call to listmnt_next() would return the mount with the mount
id 408 and not the rootfs mount.
So we need to find the actual rootfs mount mounted on the root mount of
the mount namespace. This logic is also present in mntns_install() where
vfs_path_lookup() is used. We can't use this though as we're holding the
namespace semaphore. We could look at the children of the root mount of
the mount namespace directly but that also seems a bit out of place
while we have the rbtree. So let's just iterate through the rbtree
starting from the root mount of the mount namespace and find the mount
whose parent is the root mount of the mount namespace. That mount will
usually appear very early in the rbtree and afaik there can only be one.
IOW, it would be very strange if we ended up with a root mount of a
mount namespace that has shadow mounts.
Fixes: 0a3deb11858a ("fs: Allow listmount() in foreign mount namespace") # mainline only
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-05 13:08:54 +02:00
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fs: find rootfs mount of the mount namespace
The method we used was predicated on the assumption that the mount
immediately following the root mount of the mount namespace would be the
rootfs mount of the namespace. That's not always the case though. For
example:
ID PARENT ID
408 412 0:60 /containers/overlay-containers/bc391117192b32071b22ef2083ebe7735d5c390f87a5779e02faf79ba0746ceb/userdata/hosts /etc/hosts rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=954664k,nr_inodes=238666,mode=700,uid=1000,gid=1000,inode64
409 414 0:61 / /dev/shm rw,nosuid,nodev,noexec,relatime - tmpfs shm rw,size=64000k,uid=1000,gid=1000,inode64
410 412 0:60 /containers/overlay-containers/bc391117192b32071b22ef2083ebe7735d5c390f87a5779e02faf79ba0746ceb/userdata/.containerenv /run/.containerenv rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=954664k,nr_inodes=238666,mode=700,uid=1000,gid=1000,inode64
411 412 0:60 /containers/overlay-containers/bc391117192b32071b22ef2083ebe7735d5c390f87a5779e02faf79ba0746ceb/userdata/hostname /etc/hostname rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=954664k,nr_inodes=238666,mode=700,uid=1000,gid=1000,inode64
412 363 0:65 / / rw,relatime - overlay overlay rw,lowerdir=/home/user1/.local/share/containers/storage/overlay/l/JS65SUCGTPCP2EEBHLRP4UCFI5:/home/user1/.local/share/containers/storage/overlay/l/DLW22KVDWUNI4242D6SDJ5GKCL [...]
413 412 0:68 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
414 412 0:69 / /dev rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,uid=1000,gid=1000,inode64
415 412 0:70 / /sys ro,nosuid,nodev,noexec,relatime - sysfs sysfs rw
416 414 0:71 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
417 414 0:67 / /dev/mqueue rw,nosuid,nodev,noexec,relatime - mqueue mqueue rw
418 415 0:27 / /sys/fs/cgroup ro,nosuid,nodev,noexec,relatime - cgroup2 cgroup2 rw,nsdelegate,memory_recursiveprot
419 414 0:6 /null /dev/null rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
420 414 0:6 /zero /dev/zero rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
422 414 0:6 /full /dev/full rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
423 414 0:6 /tty /dev/tty rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
430 414 0:6 /random /dev/random rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
431 414 0:6 /urandom /dev/urandom rw,nosuid,noexec - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
433 413 0:72 / /proc/acpi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
440 413 0:6 /null /proc/kcore ro,nosuid - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
441 413 0:6 /null /proc/keys ro,nosuid - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
442 413 0:6 /null /proc/timer_list ro,nosuid - devtmpfs devtmpfs rw,size=4096k,nr_inodes=1179282,mode=755,inode64
443 413 0:73 / /proc/scsi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
444 415 0:74 / /sys/firmware ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
445 415 0:75 / /sys/dev/block ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
446 413 0:68 /bus /proc/bus ro,nosuid,nodev,noexec,relatime - proc proc rw
447 413 0:68 /fs /proc/fs ro,nosuid,nodev,noexec,relatime - proc proc rw
448 413 0:68 /irq /proc/irq ro,nosuid,nodev,noexec,relatime - proc proc rw
449 413 0:68 /sys /proc/sys ro,nosuid,nodev,noexec,relatime - proc proc rw
450 413 0:68 /sysrq-trigger /proc/sysrq-trigger ro,nosuid,nodev,noexec,relatime - proc proc rw
364 414 0:71 /0 /dev/console rw,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
In this mount table the root mount of the mount namespace is the mount
with id 363 (It isn't visible because it's literally just what the
rootfs mount is mounted upon and usually it's just a copy of the real
rootfs).
The rootfs mount that's mounted on the root mount of the mount namespace
is the mount with id 412. But the mount namespace contains mounts that
were created before the rootfs mount and thus have earlier mount ids. So
the first call to listmnt_next() would return the mount with the mount
id 408 and not the rootfs mount.
So we need to find the actual rootfs mount mounted on the root mount of
the mount namespace. This logic is also present in mntns_install() where
vfs_path_lookup() is used. We can't use this though as we're holding the
namespace semaphore. We could look at the children of the root mount of
the mount namespace directly but that also seems a bit out of place
while we have the rbtree. So let's just iterate through the rbtree
starting from the root mount of the mount namespace and find the mount
whose parent is the root mount of the mount namespace. That mount will
usually appear very early in the rbtree and afaik there can only be one.
IOW, it would be very strange if we ended up with a root mount of a
mount namespace that has shadow mounts.
Fixes: 0a3deb11858a ("fs: Allow listmount() in foreign mount namespace") # mainline only
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-05 13:08:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
2023-11-19 19:03:33 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
2024-06-24 15:40:52 -04:00
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-29 12:27:15 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
2023-11-29 12:27:15 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-07-04 10:58:34 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-29 12:27:15 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-09-23 09:35:36 -07:00
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
|
|
|
|
|
2024-09-23 09:35:36 -07:00
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
|
|
|
|
|
2024-09-23 09:35:36 -07:00
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:49 -04:00
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-29 12:27:15 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
2024-06-24 11:49:49 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
2024-06-25 14:33:45 +02:00
|
|
|
|
|
|
|
|
|
2023-11-19 20:58:49 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:02 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-07 16:55:36 +02:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-07 16:55:36 +02:00
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
2024-06-24 11:49:44 -04:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-07 16:55:37 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-07 16:55:37 +02:00
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-07-04 10:58:35 +02:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
2024-06-07 16:55:37 +02:00
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
2024-01-12 09:09:14 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-29 12:27:15 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
2024-07-04 10:58:35 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-07-19 13:41:49 +02:00
|
|
|
|
2024-06-24 11:49:48 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
2024-07-04 10:58:35 +02:00
|
|
|
|
2024-06-07 16:55:37 +02:00
|
|
|
|
2024-07-04 17:00:19 +02:00
|
|
|
|
|
|
|
|
|
2024-06-10 14:49:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:03 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
2008-02-14 19:34:39 -08:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2019-05-30 17:48:35 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
fs: scale mntget/mntput
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:11 +11:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
2009-06-23 17:29:49 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
|
|
|
|
|
2006-12-08 02:37:56 -08:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-12-06 13:32:36 -05:00
|
|
|
|
|
|
|
|
|
umount: Do not allow unmounting rootfs.
Andrew Vagin <avagin@parallels.com> writes:
> #define _GNU_SOURCE
> #include <sys/types.h>
> #include <sys/stat.h>
> #include <fcntl.h>
> #include <sched.h>
> #include <unistd.h>
> #include <sys/mount.h>
>
> int main(int argc, char **argv)
> {
> int fd;
>
> fd = open("/proc/self/ns/mnt", O_RDONLY);
> if (fd < 0)
> return 1;
> while (1) {
> if (umount2("/", MNT_DETACH) ||
> setns(fd, CLONE_NEWNS))
> break;
> }
>
> return 0;
> }
>
> root@ubuntu:/home/avagin# gcc -Wall nsenter.c -o nsenter
> root@ubuntu:/home/avagin# strace ./nsenter
> execve("./nsenter", ["./nsenter"], [/* 22 vars */]) = 0
> ...
> open("/proc/self/ns/mnt", O_RDONLY) = 3
> umount("/", MNT_DETACH) = 0
> setns(3, 131072) = 0
> umount("/", MNT_DETACH
>
causes:
> [ 260.548301] ------------[ cut here ]------------
> [ 260.550941] kernel BUG at /build/buildd/linux-3.13.0/fs/pnode.c:372!
> [ 260.552068] invalid opcode: 0000 [#1] SMP
> [ 260.552068] Modules linked in: xt_CHECKSUM iptable_mangle xt_tcpudp xt_addrtype xt_conntrack ipt_MASQUERADE iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack bridge stp llc dm_thin_pool dm_persistent_data dm_bufio dm_bio_prison iptable_filter ip_tables x_tables crct10dif_pclmul crc32_pclmul ghash_clmulni_intel binfmt_misc nfsd auth_rpcgss nfs_acl aesni_intel nfs lockd aes_x86_64 sunrpc fscache lrw gf128mul glue_helper ablk_helper cryptd serio_raw ppdev parport_pc lp parport btrfs xor raid6_pq libcrc32c psmouse floppy
> [ 260.552068] CPU: 0 PID: 1723 Comm: nsenter Not tainted 3.13.0-30-generic #55-Ubuntu
> [ 260.552068] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
> [ 260.552068] task: ffff8800376097f0 ti: ffff880074824000 task.ti: ffff880074824000
> [ 260.552068] RIP: 0010:[<ffffffff811e9483>] [<ffffffff811e9483>] propagate_umount+0x123/0x130
> [ 260.552068] RSP: 0018:ffff880074825e98 EFLAGS: 00010246
> [ 260.552068] RAX: ffff88007c741140 RBX: 0000000000000002 RCX: ffff88007c741190
> [ 260.552068] RDX: ffff88007c741190 RSI: ffff880074825ec0 RDI: ffff880074825ec0
> [ 260.552068] RBP: ffff880074825eb0 R08: 00000000000172e0 R09: ffff88007fc172e0
> [ 260.552068] R10: ffffffff811cc642 R11: ffffea0001d59000 R12: ffff88007c741140
> [ 260.552068] R13: ffff88007c741140 R14: ffff88007c741140 R15: 0000000000000000
> [ 260.552068] FS: 00007fd5c7e41740(0000) GS:ffff88007fc00000(0000) knlGS:0000000000000000
> [ 260.552068] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 260.552068] CR2: 00007fd5c7968050 CR3: 0000000070124000 CR4: 00000000000406f0
> [ 260.552068] Stack:
> [ 260.552068] 0000000000000002 0000000000000002 ffff88007c631000 ffff880074825ed8
> [ 260.552068] ffffffff811dcfac ffff88007c741140 0000000000000002 ffff88007c741160
> [ 260.552068] ffff880074825f38 ffffffff811dd12b ffffffff811cc642 0000000075640000
> [ 260.552068] Call Trace:
> [ 260.552068] [<ffffffff811dcfac>] umount_tree+0x20c/0x260
> [ 260.552068] [<ffffffff811dd12b>] do_umount+0x12b/0x300
> [ 260.552068] [<ffffffff811cc642>] ? final_putname+0x22/0x50
> [ 260.552068] [<ffffffff811cc849>] ? putname+0x29/0x40
> [ 260.552068] [<ffffffff811dd88c>] SyS_umount+0xdc/0x100
> [ 260.552068] [<ffffffff8172aeff>] tracesys+0xe1/0xe6
> [ 260.552068] Code: 89 50 08 48 8b 50 08 48 89 02 49 89 45 08 e9 72 ff ff ff 0f 1f 44 00 00 4c 89 e6 4c 89 e7 e8 f5 f6 ff ff 48 89 c3 e9 39 ff ff ff <0f> 0b 66 2e 0f 1f 84 00 00 00 00 00 90 66 66 66 66 90 55 b8 01
> [ 260.552068] RIP [<ffffffff811e9483>] propagate_umount+0x123/0x130
> [ 260.552068] RSP <ffff880074825e98>
> [ 260.611451] ---[ end trace 11c33d85f1d4c652 ]--
Which in practice is totally uninteresting. Only the global root user can
do it, and it is just a stupid thing to do.
However that is no excuse to allow a silly way to oops the kernel.
We can avoid this silly problem by setting MNT_LOCKED on the rootfs
mount point and thus avoid needing any special cases in the unmount
code.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
2014-10-07 12:19:53 -07:00
|
|
|
|
2008-02-14 19:34:39 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-24 11:49:46 -04:00
|
|
|
|
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2007-10-16 23:26:30 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-09-29 01:58:57 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-11-23 12:14:10 -05:00
|
|
|
|
memcg: enable accounting for mnt_cache entries
Patch series "memcg accounting from OpenVZ", v7.
OpenVZ uses memory accounting 20+ years since v2.2.x linux kernels.
Initially we used our own accounting subsystem, then partially committed
it to upstream, and a few years ago switched to cgroups v1. Now we're
rebasing again, revising our old patches and trying to push them upstream.
We try to protect the host system from any misuse of kernel memory
allocation triggered by untrusted users inside the containers.
Patch-set is addressed mostly to cgroups maintainers and cgroups@ mailing
list, though I would be very grateful for any comments from maintainersi
of affected subsystems or other people added in cc:
Compared to the upstream, we additionally account the following kernel objects:
- network devices and its Tx/Rx queues
- ipv4/v6 addresses and routing-related objects
- inet_bind_bucket cache objects
- VLAN group arrays
- ipv6/sit: ip_tunnel_prl
- scm_fp_list objects used by SCM_RIGHTS messages of Unix sockets
- nsproxy and namespace objects itself
- IPC objects: semaphores, message queues and share memory segments
- mounts
- pollfd and select bits arrays
- signals and posix timers
- file lock
- fasync_struct used by the file lease code and driver's fasync queues
- tty objects
- per-mm LDT
We have an incorrect/incomplete/obsoleted accounting for few other kernel
objects: sk_filter, af_packets, netlink and xt_counters for iptables.
They require rework and probably will be dropped at all.
Also we're going to add an accounting for nft, however it is not ready
yet.
We have not tested performance on upstream, however, our performance team
compares our current RHEL7-based production kernel and reports that they
are at least not worse as the according original RHEL7 kernel.
This patch (of 10):
The kernel allocates ~400 bytes of 'struct mount' for any new mount.
Creating a new mount namespace clones most of the parent mounts, and this
can be repeated many times. Additionally, each mount allocates up to
PATH_MAX=4096 bytes for mnt->mnt_devname.
It makes sense to account for these allocations to restrict the host's
memory consumption from inside the memcg-limited container.
Link: https://lkml.kernel.org/r/045db11f-4a45-7c9b-2664-5b32c2b44943@virtuozzo.com
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Yutian Yang <nglaive@gmail.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Andrei Vagin <avagin@gmail.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dmitry Safonov <0x7f454c46@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jiri Slaby <jirislaby@kernel.org>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Serge Hallyn <serge@hallyn.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Zefan Li <lizefan.x@bytedance.com>
Cc: Borislav Petkov <bp@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-02 14:55:10 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
2014-03-20 21:10:51 -04:00
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
2017-07-06 15:39:11 -07:00
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-07-06 15:39:11 -07:00
|
|
|
|
2014-02-28 13:46:44 -05:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2013-03-15 10:53:28 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
2013-11-28 14:54:43 -05:00
|
|
|
|
|
|
|
|
|
2006-09-29 01:58:57 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2008-04-30 00:55:09 -07:00
|
|
|
|
2007-10-29 14:17:23 -06:00
|
|
|
|
|
|
|
|
|
2008-04-30 00:55:09 -07:00
|
|
|
|
2019-06-01 18:09:44 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2020-08-03 13:16:42 +03:00
|
|
|
|
2009-06-22 15:09:13 -04:00
|
|
|
|
2013-09-16 21:19:20 -04:00
|
|
|
|
2012-07-26 21:08:32 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-01 23:07:26 +00:00
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2011-07-19 09:32:38 -07:00
|
|
|
|
2018-11-01 23:07:26 +00:00
|
|
|
|
2011-07-19 09:32:38 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-06-09 00:59:08 -04:00
|
|
|
|
2011-07-19 09:32:38 -07:00
|
|
|
|
|
|
|
|
|
2011-03-17 22:08:28 -04:00
|
|
|
|
2018-11-01 23:07:26 +00:00
|
|
|
|
2011-07-19 09:32:38 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-01-27 13:46:51 -05:00
|
|
|
|
|
|
|
|
|
2013-09-29 22:06:07 -04:00
|
|
|
|
2011-07-19 09:32:38 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fix apparmor dereferencing potentially freed dentry, sanitize __d_path() API
__d_path() API is asking for trouble and in case of apparmor d_namespace_path()
getting just that. The root cause is that when __d_path() misses the root
it had been told to look for, it stores the location of the most remote ancestor
in *root. Without grabbing references. Sure, at the moment of call it had
been pinned down by what we have in *path. And if we raced with umount -l, we
could have very well stopped at vfsmount/dentry that got freed as soon as
prepend_path() dropped vfsmount_lock.
It is safe to compare these pointers with pre-existing (and known to be still
alive) vfsmount and dentry, as long as all we are asking is "is it the same
address?". Dereferencing is not safe and apparmor ended up stepping into
that. d_namespace_path() really wants to examine the place where we stopped,
even if it's not connected to our namespace. As the result, it looked
at ->d_sb->s_magic of a dentry that might've been already freed by that point.
All other callers had been careful enough to avoid that, but it's really
a bad interface - it invites that kind of trouble.
The fix is fairly straightforward, even though it's bigger than I'd like:
* prepend_path() root argument becomes const.
* __d_path() is never called with NULL/NULL root. It was a kludge
to start with. Instead, we have an explicit function - d_absolute_root().
Same as __d_path(), except that it doesn't get root passed and stops where
it stops. apparmor and tomoyo are using it.
* __d_path() returns NULL on path outside of root. The main
caller is show_mountinfo() and that's precisely what we pass root for - to
skip those outside chroot jail. Those who don't want that can (and do)
use d_path().
* __d_path() root argument becomes const. Everyone agrees, I hope.
* apparmor does *NOT* try to use __d_path() or any of its variants
when it sees that path->mnt is an internal vfsmount. In that case it's
definitely not mounted anywhere and dentry_path() is exactly what we want
there. Handling of sysctl()-triggered weirdness is moved to that place.
* if apparmor is asked to do pathname relative to chroot jail
and __d_path() tells it we it's not in that jail, the sucker just calls
d_absolute_path() instead. That's the other remaining caller of __d_path(),
BTW.
* seq_path_root() does _NOT_ return -ENAMETOOLONG (it's stupid anyway -
the normal seq_file logics will take care of growing the buffer and redoing
the call of ->show() just fine). However, if it gets path not reachable
from root, it returns SEQ_SKIP. The only caller adjusted (i.e. stopped
ignoring the return value as it used to do).
Reviewed-by: John Johansen <john.johansen@canonical.com>
ACKed-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
2011-12-05 08:43:34 -05:00
|
|
|
|
2020-06-04 10:48:19 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-01-27 13:46:51 -05:00
|
|
|
|
2020-06-04 10:48:19 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fix apparmor dereferencing potentially freed dentry, sanitize __d_path() API
__d_path() API is asking for trouble and in case of apparmor d_namespace_path()
getting just that. The root cause is that when __d_path() misses the root
it had been told to look for, it stores the location of the most remote ancestor
in *root. Without grabbing references. Sure, at the moment of call it had
been pinned down by what we have in *path. And if we raced with umount -l, we
could have very well stopped at vfsmount/dentry that got freed as soon as
prepend_path() dropped vfsmount_lock.
It is safe to compare these pointers with pre-existing (and known to be still
alive) vfsmount and dentry, as long as all we are asking is "is it the same
address?". Dereferencing is not safe and apparmor ended up stepping into
that. d_namespace_path() really wants to examine the place where we stopped,
even if it's not connected to our namespace. As the result, it looked
at ->d_sb->s_magic of a dentry that might've been already freed by that point.
All other callers had been careful enough to avoid that, but it's really
a bad interface - it invites that kind of trouble.
The fix is fairly straightforward, even though it's bigger than I'd like:
* prepend_path() root argument becomes const.
* __d_path() is never called with NULL/NULL root. It was a kludge
to start with. Instead, we have an explicit function - d_absolute_root().
Same as __d_path(), except that it doesn't get root passed and stops where
it stops. apparmor and tomoyo are using it.
* __d_path() returns NULL on path outside of root. The main
caller is show_mountinfo() and that's precisely what we pass root for - to
skip those outside chroot jail. Those who don't want that can (and do)
use d_path().
* __d_path() root argument becomes const. Everyone agrees, I hope.
* apparmor does *NOT* try to use __d_path() or any of its variants
when it sees that path->mnt is an internal vfsmount. In that case it's
definitely not mounted anywhere and dentry_path() is exactly what we want
there. Handling of sysctl()-triggered weirdness is moved to that place.
* if apparmor is asked to do pathname relative to chroot jail
and __d_path() tells it we it's not in that jail, the sucker just calls
d_absolute_path() instead. That's the other remaining caller of __d_path(),
BTW.
* seq_path_root() does _NOT_ return -ENAMETOOLONG (it's stupid anyway -
the normal seq_file logics will take care of growing the buffer and redoing
the call of ->show() just fine). However, if it gets path not reachable
from root, it returns SEQ_SKIP. The only caller adjusted (i.e. stopped
ignoring the return value as it used to do).
Reviewed-by: John Johansen <john.johansen@canonical.com>
ACKed-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
2011-12-05 08:43:34 -05:00
|
|
|
|
|
|
|
|
|
2011-11-25 00:46:35 -05:00
|
|
|
|
fix apparmor dereferencing potentially freed dentry, sanitize __d_path() API
__d_path() API is asking for trouble and in case of apparmor d_namespace_path()
getting just that. The root cause is that when __d_path() misses the root
it had been told to look for, it stores the location of the most remote ancestor
in *root. Without grabbing references. Sure, at the moment of call it had
been pinned down by what we have in *path. And if we raced with umount -l, we
could have very well stopped at vfsmount/dentry that got freed as soon as
prepend_path() dropped vfsmount_lock.
It is safe to compare these pointers with pre-existing (and known to be still
alive) vfsmount and dentry, as long as all we are asking is "is it the same
address?". Dereferencing is not safe and apparmor ended up stepping into
that. d_namespace_path() really wants to examine the place where we stopped,
even if it's not connected to our namespace. As the result, it looked
at ->d_sb->s_magic of a dentry that might've been already freed by that point.
All other callers had been careful enough to avoid that, but it's really
a bad interface - it invites that kind of trouble.
The fix is fairly straightforward, even though it's bigger than I'd like:
* prepend_path() root argument becomes const.
* __d_path() is never called with NULL/NULL root. It was a kludge
to start with. Instead, we have an explicit function - d_absolute_root().
Same as __d_path(), except that it doesn't get root passed and stops where
it stops. apparmor and tomoyo are using it.
* __d_path() returns NULL on path outside of root. The main
caller is show_mountinfo() and that's precisely what we pass root for - to
skip those outside chroot jail. Those who don't want that can (and do)
use d_path().
* __d_path() root argument becomes const. Everyone agrees, I hope.
* apparmor does *NOT* try to use __d_path() or any of its variants
when it sees that path->mnt is an internal vfsmount. In that case it's
definitely not mounted anywhere and dentry_path() is exactly what we want
there. Handling of sysctl()-triggered weirdness is moved to that place.
* if apparmor is asked to do pathname relative to chroot jail
and __d_path() tells it we it's not in that jail, the sucker just calls
d_absolute_path() instead. That's the other remaining caller of __d_path(),
BTW.
* seq_path_root() does _NOT_ return -ENAMETOOLONG (it's stupid anyway -
the normal seq_file logics will take care of growing the buffer and redoing
the call of ->show() just fine). However, if it gets path not reachable
from root, it returns SEQ_SKIP. The only caller adjusted (i.e. stopped
ignoring the return value as it used to do).
Reviewed-by: John Johansen <john.johansen@canonical.com>
ACKed-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
2011-12-05 08:43:34 -05:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2013-03-15 01:45:51 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
|
|
|
|
|
2016-06-09 16:06:06 -05:00
|
|
|
|
2013-03-24 14:28:27 -07:00
|
|
|
|
2015-05-08 23:49:47 -05:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
2013-03-24 14:28:27 -07:00
|
|
|
|
2013-09-16 21:37:36 -04:00
|
|
|
|
2023-10-25 16:02:00 +02:00
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
2015-06-04 09:43:11 -05:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
|
|
|
|
|
2015-05-08 16:36:50 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-06-15 06:59:49 -05:00
|
|
|
|
2015-06-04 09:43:11 -05:00
|
|
|
|
|
|
|
|
|
2016-06-10 12:21:40 -05:00
|
|
|
|
2017-07-17 08:45:34 +01:00
|
|
|
|
2016-06-10 12:21:40 -05:00
|
|
|
|
|
|
|
|
|
2015-05-08 23:49:47 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-06-04 09:43:11 -05:00
|
|
|
|
2015-05-08 23:49:47 -05:00
|
|
|
|
|
|
|
|
|
2015-06-04 09:43:11 -05:00
|
|
|
|
|
|
|
|
|
2015-05-08 23:49:47 -05:00
|
|
|
|
|
|
|
|
|
2015-01-07 08:10:09 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-01-07 08:10:09 -06:00
|
|
|
|
2016-05-27 14:50:05 -05:00
|
|
|
|
2015-01-07 08:10:09 -06:00
|
|
|
|
2024-08-06 11:47:10 +08:00
|
|
|
|
2015-05-13 20:51:09 -05:00
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
2013-03-24 14:28:27 -07:00
|
|
|
|
2015-05-08 23:49:47 -05:00
|
|
|
|
2015-06-04 09:43:11 -05:00
|
|
|
|
|
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-03-24 14:28:27 -07:00
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
2013-09-16 21:37:36 -04:00
|
|
|
|
2013-03-30 19:57:41 -07:00
|
|
|
|
2013-03-24 14:28:27 -07:00
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
2016-06-09 16:06:06 -05:00
|
|
|
|
2016-06-15 06:59:49 -05:00
|
|
|
|
2016-06-09 16:06:06 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
2016-06-09 16:06:06 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-06-15 06:59:49 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-04 07:43:08 -05:00
|
|
|
|
2016-06-09 16:06:06 -05:00
|
|
|
|
|
|
|
|
|
fs: Treat foreign mounts as nosuid
If a process gets access to a mount from a different user
namespace, that process should not be able to take advantage of
setuid files or selinux entrypoints from that filesystem. Prevent
this by treating mounts from other mount namespaces and those not
owned by current_user_ns() or an ancestor as nosuid.
This will make it safer to allow more complex filesystems to be
mounted in non-root user namespaces.
This does not remove the need for MNT_LOCK_NOSUID. The setuid,
setgid, and file capability bits can no longer be abused if code in
a user namespace were to clear nosuid on an untrusted filesystem,
but this patch, by itself, is insufficient to protect the system
from abuse of files that, when execed, would increase MAC privilege.
As a more concrete explanation, any task that can manipulate a
vfsmount associated with a given user namespace already has
capabilities in that namespace and all of its descendents. If they
can cause a malicious setuid, setgid, or file-caps executable to
appear in that mount, then that executable will only allow them to
elevate privileges in exactly the set of namespaces in which they
are already privileges.
On the other hand, if they can cause a malicious executable to
appear with a dangerous MAC label, running it could change the
caller's security context in a way that should not have been
possible, even inside the namespace in which the task is confined.
As a hardening measure, this would have made CVE-2014-5207 much
more difficult to exploit.
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Acked-by: James Morris <james.l.morris@oracle.com>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2016-06-23 16:41:05 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-11-01 00:37:32 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2014-11-01 00:00:23 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
2014-02-03 19:13:49 -08:00
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2014-11-01 00:00:23 -04:00
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2014-02-03 19:13:49 -08:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-11-01 00:37:32 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2014-11-01 00:00:23 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
2020-05-05 16:04:30 +02:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2020-05-05 16:04:30 +02:00
|
|
|
|
|
|
|
|
|
2017-04-15 17:31:22 -04:00
|
|
|
|
2020-05-05 16:04:30 +02:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2017-04-15 17:31:22 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2012-07-26 21:42:03 -07:00
|
|
|
|
2020-05-05 16:04:30 +02:00
|
|
|
|
|
|
|
|
|
2012-09-13 16:38:03 +08:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2019-01-30 13:30:21 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-04-15 17:31:22 -04:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-04-15 17:31:22 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2017-06-08 17:32:29 -07:00
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-06 00:47:13 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-06 00:47:13 -07:00
|
|
|
|
2010-03-07 18:49:36 -08:00
|
|
|
|
2022-01-21 22:13:27 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
