Commit graph

1,425 commits

Author SHA1 Message Date
Linus Torvalds
91b436fc92 20 smb client fixes
-----BEGIN PGP SIGNATURE-----
 
 iQGzBAABCgAdFiEE6fsu8pdIjtWE/DpLiiy9cAdyT1EFAmjpMA0ACgkQiiy9cAdy
 T1Emcgv/WWk7+nRyHQwnv4fLkIMXt+OwJUhVrHvxoMUjbi4L2635wyOWQJ2pYAdt
 oy3J6ozk/TkvSI7KzUGAo+5vVXjAkX3fKWWA8R2jrVcmSOgXGjVSm4NSQ2bO8B35
 7OXyW8ke1V5swzRGUhmn0Ppv6As8Hb77puZD7NzcYLLKhRTeuL9DuPyy2kNYYnIN
 WL6Cs3wApRAwNw7d6bI/0IQZPG2z0F0ShnLEytlgCxrifpMLSwj+8rfu8oy1Ks0S
 9NzeZtku9RIct+zkgDUS0Msg6+w6greynaLy2UzN6Ki4pgOp48sbPZ3TRVT0iNPC
 amtgqazY3+1Pbs6cdGcSv2VeJCJ9RvF9dzqmQFEIsPysOhVCjDvyQ3oQru2a7lsQ
 ImPqwmMGSW7t1jtBJsKDxVKcANgDM0w6GdPGaBPlN3Zpw+jVO1UgA+7fWKTW+30L
 JiKjBQYYXcAieFJsHUkcpWDTi6asr0Xcmx4Sy2nnJ9msDmdsevlLQzY7hBSFySjn
 /tvSQiQf
 =/qIG
 -----END PGP SIGNATURE-----

Merge tag 'v6.18-rc-part2-smb-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull more smb client updates from Steve French:

 - fix i_size in fallocate

 - two truncate fixes

 - utime fix

 - minor cleanups

 - SMB1 fixes

 - improve error check in read

 - improve perf of copy file_range (copy_chunk)

* tag 'v6.18-rc-part2-smb-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: update internal version number
  cifs: Add comments for DeletePending assignments in open functions
  cifs: Add fallback code path for cifs_mkdir_setinfo()
  cifs: Allow fallback code in smb_set_file_info() also for directories
  cifs: Query EA $LXMOD in cifs_query_path_info() for WSL reparse points
  smb: client: remove cfids_invalidation_worker
  smb: client: remove redudant assignment in cifs_strict_fsync()
  smb: client: fix race with fallocate(2) and AIO+DIO
  smb: client: fix missing timestamp updates after utime(2)
  smb: client: fix missing timestamp updates after ftruncate(2)
  smb: client: fix missing timestamp updates with O_TRUNC
  cifs: Fix copy_to_iter return value check
  smb: client: batch SRV_COPYCHUNK entries to cut round trips
  smb: client: Omit an if branch in smb2_find_smb_tcon()
  smb: client: Return directly after a failed genlmsg_new() in cifs_swn_send_register_message()
  smb: client: Use common code in cifs_do_create()
  smb: client: Improve unlocking of a mutex in cifs_get_swn_reg()
  smb: client: Return a status code only as a constant in cifs_spnego_key_instantiate()
  smb: client: Use common code in cifs_lookup()
  smb: client: Reduce the scopes for a few variables in two functions
2025-10-10 11:23:57 -07:00
Steve French
b30c32c784 cifs: update internal version number
to 2.57

Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-10 11:10:01 -05:00
Pali Rohár
fa9fe87150 cifs: Add comments for DeletePending assignments in open functions
On more places is set DeletePending member to 0. Add comments why is 0 the
correct value. Paths in DELETE_PENDING state cannot be opened by new calls.
So if the newly issued open for that path succeed then it means that the
path cannot be in DELETE_PENDING state.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 23:01:24 -05:00
Pali Rohár
92210ccd87 cifs: Add fallback code path for cifs_mkdir_setinfo()
Use SMBSetInformation() as a fallback function (when CIFSSMBSetPathInfo()
fails) which can set attribudes on the directory, including changing
read-only attribute.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 23:01:24 -05:00
Pali Rohár
88cae132dc cifs: Allow fallback code in smb_set_file_info() also for directories
On NT systems, it is possible to do SMB open call also for directories.
Open argument CREATE_NOT_DIR disallows opening directories. So in fallback
code path in smb_set_file_info() remove CREATE_NOT_DIR restriction to allow
it also for directories.

Similar fallback is implemented also in CIFSSMBSetPathInfoFB() function and
this function already allows to call operation for directories.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 23:01:24 -05:00
Pali Rohár
057ac50638 cifs: Query EA $LXMOD in cifs_query_path_info() for WSL reparse points
EA $LXMOD is required for WSL non-symlink reparse points.

Fixes: ef86ab131d ("cifs: Fix querying of WSL CHR and BLK reparse points over SMB1")
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 23:01:24 -05:00
Enzo Matsumiya
7ae6152b78 smb: client: remove cfids_invalidation_worker
We can do the same cleanup on laundromat.

On invalidate_all_cached_dirs(), run laundromat worker with 0 timeout
and flush it for immediate + sync cleanup.

Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 11:18:09 -05:00
Paulo Alcantara
be3898a395 smb: client: remove redudant assignment in cifs_strict_fsync()
Remove redudant assignment of @rc as it will be overwritten by the
following cifs_file_flush() call.

Reported-by: Steve French <stfrench@microsoft.com>
Addresses-Coverity: 1665925
Fixes: 210627b0aca9 ("smb: client: fix missing timestamp updates with O_TRUNC")
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 11:16:25 -05:00
Paulo Alcantara
dba9f997c9 smb: client: fix race with fallocate(2) and AIO+DIO
AIO+DIO may extend the file size, hence we need to make sure ->i_size
is stable across the entire fallocate(2) operation, otherwise it would
become a truncate and then inode size reduced back down when it
finishes.

Fix this by calling netfs_wait_for_outstanding_io() right after
acquiring ->i_rwsem exclusively in cifs_fallocate() and then guarantee
a stable ->i_size across fallocate(2).

Also call netfs_wait_for_outstanding_io() after truncating pagecache
to avoid any potential races with writeback.

Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Reviewed-by: David Howells <dhowells@redhat.com>
Fixes: 210627b0aca9 ("smb: client: fix missing timestamp updates with O_TRUNC")
Cc: Frank Sorenson <sorenson@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:42:14 -05:00
Paulo Alcantara
b95cd1bdf5 smb: client: fix missing timestamp updates after utime(2)
Don't reuse open handle when changing timestamps to prevent the server
from disabling automatic timestamp updates as per MS-FSA 2.1.4.17.

---8<---
import os
import time

filename = '/mnt/foo'

def print_stat(prefix):
    st = os.stat(filename)
    print(prefix, ': ', time.ctime(st.st_atime), time.ctime(st.st_ctime))

fd = os.open(filename, os.O_CREAT|os.O_TRUNC|os.O_WRONLY, 0o644)
print_stat('old')
os.utime(fd, None)
time.sleep(2)
os.write(fd, b'foo')
os.close(fd)
time.sleep(2)
print_stat('new')
---8<---

Before patch:

$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old :  Fri Oct  3 14:01:21 2025 Fri Oct  3 14:01:21 2025
new :  Fri Oct  3 14:01:21 2025 Fri Oct  3 14:01:21 2025

After patch:

$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old :  Fri Oct  3 17:03:34 2025 Fri Oct  3 17:03:34 2025
new :  Fri Oct  3 17:03:36 2025 Fri Oct  3 17:03:36 2025

Fixes: b6f2a0f89d ("cifs: for compound requests, use open handle if possible")
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: Frank Sorenson <sorenson@redhat.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:42:14 -05:00
Paulo Alcantara
57ce9f7793 smb: client: fix missing timestamp updates after ftruncate(2)
Mask off ATTR_MTIME|ATTR_CTIME bits on ATTR_SIZE (e.g. ftruncate(2))
to prevent the client from sending set info calls and then disabling
automatic timestamp updates on server side as per MS-FSA 2.1.4.17.

---8<---
import os
import time

filename = '/mnt/foo'

def print_stat(prefix):
    st = os.stat(filename)
    print(prefix, ': ', time.ctime(st.st_atime), time.ctime(st.st_ctime))

fd = os.open(filename, os.O_CREAT|os.O_TRUNC|os.O_WRONLY, 0o644)
print_stat('old')
os.ftruncate(fd, 10)
time.sleep(2)
os.write(fd, b'foo')
os.close(fd)
time.sleep(2)
print_stat('new')
---8<---

Before patch:

$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old :  Fri Oct  3 13:47:03 2025 Fri Oct  3 13:47:03 2025
new :  Fri Oct  3 13:47:00 2025 Fri Oct  3 13:47:03 2025

After patch:

$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old :  Fri Oct  3 13:48:39 2025 Fri Oct  3 13:48:39 2025
new :  Fri Oct  3 13:48:41 2025 Fri Oct  3 13:48:41 2025

Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: Frank Sorenson <sorenson@redhat.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:42:14 -05:00
Paulo Alcantara
110fee6b9b smb: client: fix missing timestamp updates with O_TRUNC
Don't call ->set_file_info() on open handle to prevent the server from
stopping [cm]time updates automatically as per MS-FSA 2.1.4.17.

Fix this by checking for ATTR_OPEN bit earlier in cifs_setattr() to
prevent ->set_file_info() from being called when opening a file with
O_TRUNC.  Do the truncation in ->open() instead.

This also saves two roundtrips when opening a file with O_TRUNC and
there are currently no open handles to be reused.

Before patch:

$ mount.cifs //srv/share /mnt -o ...
$ cd /mnt
$ exec 3>foo; stat -c 'old: %z %y' foo; sleep 2; echo test >&3; exec 3>&-; sleep 2; stat -c 'new: %z %y' foo
old: 2025-10-03 13:26:23.151030500 -0300 2025-10-03 13:26:23.151030500 -0300
new: 2025-10-03 13:26:23.151030500 -0300 2025-10-03 13:26:23.151030500 -0300

After patch:

$ mount.cifs //srv/share /mnt -o ...
$ cd /mnt
$ exec 3>foo; stat -c 'old: %z %y' foo; sleep 2; echo test >&3; exec 3>&-; sleep 2; stat -c 'new: %z %y' foo
$ exec 3>foo; stat -c 'old: %z %y' foo; sleep 2; echo test >&3; exec 3>&-; sleep 2; stat -c 'new: %z %y' foo
old: 2025-10-03 13:28:13.911933800 -0300 2025-10-03 13:28:13.911933800 -0300
new: 2025-10-03 13:28:26.647492700 -0300 2025-10-03 13:28:26.647492700 -0300

Reported-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:42:14 -05:00
Fushuai Wang
0cc380d0e1 cifs: Fix copy_to_iter return value check
The return value of copy_to_iter() function will never be negative,
it is the number of bytes copied, or zero if nothing was copied.
Update the check to treat 0 as an error, and return -1 in that case.

Fixes: d08089f649 ("cifs: Change the I/O paths to use an iterator rather than a page list")
Acked-by: Tom Talpey <tom@talpey.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Signed-off-by: Fushuai Wang <wangfushuai@baidu.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:42:14 -05:00
Henrique Carvalho
68d2e2ca1c smb: client: batch SRV_COPYCHUNK entries to cut round trips
smb2_copychunk_range() used to send a single SRV_COPYCHUNK per
SRV_COPYCHUNK_COPY IOCTL.

Implement variable Chunks[] array in struct copychunk_ioctl and fill it
with struct copychunk (MS-SMB2 2.2.31.1.1), bounded by server-advertised
limits.

This reduces the number of IOCTL requests for large copies.

While we are at it, rename a couple variables to follow the terminology
used in the specification.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:42:14 -05:00
Markus Elfring
1643cd51ba smb: client: Omit an if branch in smb2_find_smb_tcon()
Statements from an if branch and the end of this function implementation
were equivalent.
Thus delete duplicate source code.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-09 10:40:43 -05:00
Markus Elfring
4468490251 smb: client: Return directly after a failed genlmsg_new() in cifs_swn_send_register_message()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* Return directly after a call of the function “genlmsg_new” failed
  at the beginning.

* Delete the label “fail” which became unnecessary
  with this refactoring.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-07 14:28:16 -05:00
Markus Elfring
ce47f74985 smb: client: Use common code in cifs_do_create()
Use a label once more so that a bit of common code can be better reused
at the end of this function implementation.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Enzo Matsumiya <ematsumiya@suse.de>
Reviewed-by: David Howells <dhowells@redhat.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-07 13:31:27 -05:00
Markus Elfring
e2080b70c5 smb: client: Improve unlocking of a mutex in cifs_get_swn_reg()
Use two additional labels so that another bit of common code can be better
reused at the end of this function implementation.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-07 11:12:23 -05:00
Markus Elfring
0a98b40b8f smb: client: Return a status code only as a constant in cifs_spnego_key_instantiate()
* Return a status code without storing it in an intermediate variable.

* Delete the local variable “ret” and the label “error”
  which became unnecessary with this refactoring.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-07 11:12:19 -05:00
Markus Elfring
61da08ecb5 smb: client: Use common code in cifs_lookup()
Use three additional labels so that another bit of common code can be
better reused at the end of this function implementation.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-06 16:43:44 -05:00
Markus Elfring
e7933f5b01 smb: client: Reduce the scopes for a few variables in two functions
* cifs_lookup():
  Move the definition for the local variable “cfid” into an else branch
  so that the corresponding setting will only be performed if a NULL inode
  was detected during lookup by this function.

* cifs_d_revalidate():
  Move the definition for the local variables “inode” and “rc” into
  an if branch so that the corresponding setting will only be performed
  after a d_really_is_positive() call.

  Move the definition for the local variable “cfid” into an else branch
  so that the corresponding setting will only be performed if further data
  processing will be needed for an open_cached_dir_by_dentry() call.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-06 16:43:39 -05:00
Linus Torvalds
50647a1176 file->f_path constification
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYKAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCaN3daAAKCRBZ7Krx/gZQ
 6zNWAP9kD6rOJRNqDgea4pibDPa47Tps/WM5tsDv3dsLliY29gEA6sveOWZ3guAj
 4oY3ts/NtHLWXvhI7Vd/1mr2aTKEZQk=
 =YNK+
 -----END PGP SIGNATURE-----

Merge tag 'pull-f_path' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull file->f_path constification from Al Viro:
 "Only one thing was modifying ->f_path of an opened file - acct(2).

  Massaging that away and constifying a bunch of struct path * arguments
  in functions that might be given &file->f_path ends up with the
  situation where we can turn ->f_path into an anon union of const
  struct path f_path and struct path __f_path, the latter modified only
  in a few places in fs/{file_table,open,namei}.c, all for struct file
  instances that are yet to be opened"

* tag 'pull-f_path' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (23 commits)
  Have cc(1) catch attempts to modify ->f_path
  kernel/acct.c: saner struct file treatment
  configfs:get_target() - release path as soon as we grab configfs_item reference
  apparmor/af_unix: constify struct path * arguments
  ovl_is_real_file: constify realpath argument
  ovl_sync_file(): constify path argument
  ovl_lower_dir(): constify path argument
  ovl_get_verity_digest(): constify path argument
  ovl_validate_verity(): constify {meta,data}path arguments
  ovl_ensure_verity_loaded(): constify datapath argument
  ksmbd_vfs_set_init_posix_acl(): constify path argument
  ksmbd_vfs_inherit_posix_acl(): constify path argument
  ksmbd_vfs_kern_path_unlock(): constify path argument
  ksmbd_vfs_path_lookup_locked(): root_share_path can be const struct path *
  check_export(): constify path argument
  export_operations->open(): constify path argument
  rqst_exp_get_by_name(): constify path argument
  nfs: constify path argument of __vfs_getattr()
  bpf...d_path(): constify path argument
  done_path_create(): constify path argument
  ...
2025-10-03 16:32:36 -07:00
Linus Torvalds
63e62baaa7 16 smb3 client fixes
-----BEGIN PGP SIGNATURE-----
 
 iQGzBAABCgAdFiEE6fsu8pdIjtWE/DpLiiy9cAdyT1EFAmjf7TUACgkQiiy9cAdy
 T1EeoQv/cCbhCv5S+wM9mtqtJt3ea+0E4+fXc012198jgphgd39s5cV07flVj4vA
 P5gC0ZGQmTsCOhcmYzxOefUpyq5N+ckIY8PoQKEXyHf3e/0GqczkdDokkhB8eOkh
 aC/arAcQvYXzcR+RwLdkRa0W/kwoXJ6iQ4+44KS7U6AUzWOR3H21UAJ2tg96xNLk
 Qp4dKOG3i1RcBMMX2iuFKbF693XQPhoefs+/kPzWe2R99IIfJrFT7rLPhuwV+iJ7
 TIHgXk+UEUiDizLsSi9SPs59X0oOrQ2BlVI6mR0sUso94yQanItla9TTVTVTy2Tz
 tB7FUbF5x7a0Yaj3pVai5MXLRaKDEP38KWoM5mUsDyxDLblE5cHLtfxZzPwK8VO0
 oTVRDQu9DigOmfyiwJ9tu3O5/ZnAr0A5pJEYHHfue84xuBRb1mUPyzX+ML2yym7u
 8K+CUOKXS3M8x4Ep+88MukSX3PNEhJ1ohleyQghJHQVqrcA4tp289X+GJRuG9Q1b
 NiII7Tk0
 =4Nt1
 -----END PGP SIGNATURE-----

Merge tag '6.18-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client updates from Steve French:

 - Fix oops in crypt message

 - Remove duplicate arc4 code

 - Fix potential io_uring reconnect

 - Two important directory leases fixes and three perf improvements

 - Three minor cleanups

 - Four debug improvements (e.g. for showing more information on leases,
   and one for adding more helpful information on reconnect)

* tag '6.18-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: client: force multichannel=off when max_channels=1
  smb client: fix bug with newly created file in cached dir
  smb: client: short-circuit negative lookups when parent dir is fully cached
  smb: client: short-circuit in open_cached_dir_by_dentry() if !dentry
  smb: client: remove pointless cfid->has_lease check
  smb: client: transport: minor indentation style fix
  smb: client: transport: avoid reconnects triggered by pending task work
  smb: client: remove unused fid_lock
  smb: client: update cfid->last_access_time in open_cached_dir_by_dentry()
  smb: client: ensure open_cached_dir_by_dentry() only returns valid cfid
  smb: client: account smb directory cache usage and per-tcon totals
  smb: client: add drop_dir_cache module parameter to invalidate cached dirents
  smb: client: show lease state as R/H/W (or NONE) in open_files
  smb: client: fix crypto buffers in non-linear memory
  smb: Use arc4 library instead of duplicate arc4 code
  smb: client: add tcon information to smb2_reconnect() debug messages
2025-10-03 14:13:23 -07:00
Linus Torvalds
b3fee71e66 nine ksmbd server fixes
-----BEGIN PGP SIGNATURE-----
 
 iQGzBAABCgAdFiEE6fsu8pdIjtWE/DpLiiy9cAdyT1EFAmjdnKIACgkQiiy9cAdy
 T1GSdQv/ZOTEtqd1g5CKih9gD8QczGDndh1e650fqU7WIPhgoZ9c+Qo50KsXwo7B
 pnLp0Rjy8OymfRcm2DZVDiEkiR1ocRTNiPOwwMK7ET3+F8WS0elxu4utF5/MWgkR
 QA0CJ1LlulAFeSkp2KLiXzaRUabGrEyIz/+WCl7Rz/DnLRNB2f5+45T4TNozbH0a
 W2YaQ/M1udrLJ2g+B+VYH8D5GLGrcnFOJROJ21Wt6mYSbRseUkRawr89WD+lQ4wL
 U64vOda4AHR1CIgGnaXPU3but1chdvoiUzgd3p7/Qt9GJNGOL1iH5m0sG/tf53FX
 370NhFPeF3zKeje+j0nVkvr1xAcLWQCdIMhN3MRleuepNeNtmC18+TojHg4flf1Z
 axuWDjIBZGtGUByGewigW2D6Q46HI8DnjQzsAGTRVrJ2itIJsCiAPvlhM1mw3gCp
 5mtClLxuHouUH47C86YVAKN77poXVjx4gc6XCTqt+NqHX0ZJgA8y9TkpxQLR1GvJ
 Vgn05FsQ
 =rwfk
 -----END PGP SIGNATURE-----

Merge tag 'v6.18rc1-part1-ksmbd-server-fixes' of git://git.samba.org/ksmbd

Pull smb server fixes from Steve French:

 - Fix potential UAFs and corruptions in rpc open and close

 - Fix copy_file_range when ranges overlap

 - Improve session, share, connection lookup performance

 - Fix potential hash collisions in share and session lists

 - Debugging improvement - making per-connection threads easier to
   identify

 - Improve socket creation

 - Fix return code mapping for posix query fs info

 - Add support for limiting the maximum number of connections per IP
   address, extending the existing connection limiting mechanism to
   enforce per-IP connection limits alongside the global connection
   limit

* tag 'v6.18rc1-part1-ksmbd-server-fixes' of git://git.samba.org/ksmbd:
  ksmbd: increase session and share hash table bits
  ksmbd: replace connection list with hash table
  ksmbd: add an error print when maximum IP connections limit is reached
  ksmbd: add max ip connections parameter
  ksmbd: fix error code overwriting in smb2_get_info_filesystem()
  ksmbd: copy overlapped range within the same file
  ksmbd: use sock_create_kern interface to create kernel socket
  ksmbd: make ksmbd thread names distinct by client IP
  ksmbd: Fix race condition in RPC handle list access
2025-10-03 14:09:59 -07:00
Linus Torvalds
829745b75a finish_no_open calling conventions change
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYKAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCaNh82wAKCRBZ7Krx/gZQ
 67qmAP9l34sMvmagYDve+C+BtndI5FO2N8vHCB0VuWpgDvM7XgD/Zy4Iuup0fePh
 paGQgpwjYQ7/ghc4KjOUAw4EpfJ+SAE=
 =cjj2
 -----END PGP SIGNATURE-----

Merge tag 'pull-finish_no_open' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull finish_no_open updates from Al Viro:
 "finish_no_open calling conventions change to simplify callers"

* tag 'pull-finish_no_open' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  slightly simplify nfs_atomic_open()
  simplify gfs2_atomic_open()
  simplify fuse_atomic_open()
  simplify nfs_atomic_open_v23()
  simplify vboxsf_dir_atomic_open()
  simplify cifs_atomic_open()
  9p: simplify v9fs_vfs_atomic_open_dotl()
  9p: simplify v9fs_vfs_atomic_open()
  allow finish_no_open(file, ERR_PTR(-E...))
2025-10-03 10:59:31 -07:00
Linus Torvalds
51e9889ab1 vfs_parse_fs_string() stuff
change on vfs_parse_fs_string() calling conventions - get rid of
 the length argument (almost all callers pass strlen() of the
 string argument there), add vfs_parse_fs_qstr() for the cases
 that do want separate length
 
 Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYKAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCaNhllQAKCRBZ7Krx/gZQ
 6wyiAP9TmyFLBWKC/sDNtRAiGRybEtcwvVZj1agpx0kZIWshUwD7BVg4AfDs+vN3
 RoYYg9DR1SP5kZF7h2Ve1T39XDq6ZQU=
 =YZFG
 -----END PGP SIGNATURE-----

Merge tag 'pull-fs_context' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull fs_context updates from Al Viro:
 "Change vfs_parse_fs_string() calling conventions

  Get rid of the length argument (almost all callers pass strlen() of
  the string argument there), add vfs_parse_fs_qstr() for the cases that
  do want separate length"

* tag 'pull-fs_context' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  do_nfs4_mount(): switch to vfs_parse_fs_string()
  change the calling conventions for vfs_parse_fs_string()
2025-10-03 10:51:44 -07:00
Rajasi Mandal
37e263e68c cifs: client: force multichannel=off when max_channels=1
Previously, specifying both multichannel and max_channels=1 as mount
options would leave multichannel enabled, even though it is not
meaningful when only one channel is allowed. This led to confusion and
inconsistent behavior, as the client would advertise multichannel
capability but never establish secondary channels.

Fix this by forcing multichannel to false whenever max_channels=1,
ensuring the mount configuration is consistent and matches user intent.
This prevents the client from advertising or attempting multichannel
support when it is not possible.

Signed-off-by: Rajasi Mandal <rajasimandal@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:42:15 -05:00
Bharath SM
aa12118dbc smb client: fix bug with newly created file in cached dir
Test generic/637 spotted a problem with create of a new file in a
cached directory (by the same client) could cause cases where the
new file does not show up properly in ls on that client until the
lease times out.

Fixes: 037e1bae58 ("smb: client: use ParentLeaseKey in cifs_do_create")
Cc: stable@vger.kernel.org
Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:42:15 -05:00
Henrique Carvalho
316025335a smb: client: short-circuit negative lookups when parent dir is fully cached
When the parent directory has a valid and complete cached enumeration we
can assume that negative dentries are not present in the directory, thus
we can return without issuing a request.

This reduces traffic for common ENOENT when the directory entries are
cached.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:42:15 -05:00
Henrique Carvalho
55580ad027 smb: client: short-circuit in open_cached_dir_by_dentry() if !dentry
When dentry is NULL, the current code acquires the spinlock and traverses
the entire list, but the condition (dentry && cfid->dentry == dentry)
ensures no match will ever be found.

Return -ENOENT early in this case, avoiding unnecessary lock acquisition
and list traversal.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:42:15 -05:00
Henrique Carvalho
2f6a4af028 smb: client: remove pointless cfid->has_lease check
open_cached_dir() will only return a valid cfid, which has both
has_lease = true and time != 0.

Remove the pointless check of cfid->has_lease right after
open_cached_dir() returns no error.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:42:04 -05:00
Fiona Ebner
6c7fd18423 smb: client: transport: minor indentation style fix
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:25:25 -05:00
Fiona Ebner
00be6f26a2 smb: client: transport: avoid reconnects triggered by pending task work
When io_uring is used in the same task as CIFS, there might be
unnecessary reconnects, causing issues in user-space applications
like QEMU with a log like:

> CIFS: VFS: \\10.10.100.81 Error -512 sending data on socket to server

Certain io_uring completions might be added to task_work with
notify_method being TWA_SIGNAL and thus TIF_NOTIFY_SIGNAL is set for
the task.

In __smb_send_rqst(), signals are masked before calling
smb_send_kvec(), but the masking does not apply to TIF_NOTIFY_SIGNAL.

If sk_stream_wait_memory() is reached via sock_sendmsg() while
TIF_NOTIFY_SIGNAL is set, signal_pending(current) will evaluate to
true there, and -EINTR will be propagated all the way from
sk_stream_wait_memory() to sock_sendmsg() in smb_send_kvec().
Afterwards, __smb_send_rqst() will see that not everything was written
and reconnect.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:25:21 -05:00
Henrique Carvalho
17ef15fa80 smb: client: remove unused fid_lock
The fid_lock in struct cached_fid does not currently provide any real
synchronization. Previously, it had the intention to prevent a double
release of the dentry, but every change to cfid->dentry is already
protected either by cfid_list_lock (while the entry is in the list) or
happens after the cfid has been removed (so no other thread should find
it).

Since there is no scenario in which fid_lock prevents any race, it is
vestigial and can be removed along with its associated
spin_lock()/spin_unlock() calls.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:05:19 -05:00
Henrique Carvalho
5676398315 smb: client: update cfid->last_access_time in open_cached_dir_by_dentry()
open_cached_dir_by_dentry() was missing an update of
cfid->last_access_time to jiffies, similar to what open_cached_dir()
has.

Add it to the function.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 22:01:03 -05:00
Steve French
a365f2c049 smb: client: ensure open_cached_dir_by_dentry() only returns valid cfid
open_cached_dir_by_dentry() was exposing an invalid cached directory to
callers. The validity check outside the function was exclusively based
on cfid->time.

Add validity check before returning success and introduce
is_valid_cached_dir() helper for consistent checks across the code.

Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviwed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 21:49:59 -05:00
Bharath SM
63eb8bd6c8 smb: client: account smb directory cache usage and per-tcon totals
Add lightweight accounting for directory lease cache usage
to aid debugging and limiting cache size in future. Track
per-directory entry/byte counts and maintain per-tcon
aggregates. Also expose the totals in /proc/fs/cifs/open_dirs.

Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 21:49:53 -05:00
Bharath SM
dde6667fa3 smb: client: add drop_dir_cache module parameter to invalidate cached dirents
Add write-only /sys/module/cifs/parameters/drop_dir_cache. Writing a
non-zero value iterates all tcons and calls invalidate_all_cached_dirs()
to drop cached directory entries. This is useful to force a dirent cache
drop across mounts for debugging and testing purpose.

Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 20:36:23 -05:00
Bharath SM
ac3ad9845b smb: client: show lease state as R/H/W (or NONE) in open_files
Print the lease/oplock caching state for each open file as a
compact string of letters: R (read), H (handle), W (write).

Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-10-01 12:41:46 -05:00
Namjae Jeon
e28c5bc456 ksmbd: increase session and share hash table bits
Increases the number of bits for the hash table from 3 to 12.
The thousands of sessions and shares can be connected.
So the current 3-bit size can lead to frequent hash collisions.

Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:55 -05:00
Namjae Jeon
0bcc831be5 ksmbd: replace connection list with hash table
Replace connection list with hash table to improve lookup performance.

Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:55 -05:00
Namjae Jeon
6747885781 ksmbd: add an error print when maximum IP connections limit is reached
This change introduces an error print using pr_info_ratelimited()
to prevent excessive logging. This message will inform the user that
the limit for maximum IP connections has been hit and what that
current count is, which can be useful for debugging and monitoring
connection limits.

Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Namjae Jeon
d8b6dc9256 ksmbd: add max ip connections parameter
This parameter set the maximum number of connections per ip address.
The default is 8.

Cc: stable@vger.kernel.org
Fixes: c0d41112f1 ("ksmbd: extend the connection limiting mechanism to support IPv6")
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Matvey Kovalev
88daf2f448 ksmbd: fix error code overwriting in smb2_get_info_filesystem()
If client doesn't negotiate with SMB3.1.1 POSIX Extensions,
then proper error code won't be returned due to overwriting.

Return error immediately.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: e2f34481b2 ("cifsd: add server-side procedures for SMB3")
Cc: stable@vger.kernel.org
Signed-off-by: Matvey Kovalev <matvey.kovalev@ispras.ru>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Namjae Jeon
c20988c217 ksmbd: copy overlapped range within the same file
cifs.ko request to copy overlapped range within the same file.
ksmbd is using vfs_copy_file_range for this, vfs_copy_file_range() does not
allow overlapped copying within the same file.
This patch use do_splice_direct() if offset and length are overlapped.

Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Namjae Jeon
3677ca67b9 ksmbd: use sock_create_kern interface to create kernel socket
we should use sock_create_kern() if the socket resides in kernel space.

Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Namjae Jeon
5da92a251e ksmbd: make ksmbd thread names distinct by client IP
This patch makes ksmbd thread names distinct by client IP address.

100943 ?        S      0:00 [ksmbd:::ffff:10.177.110.57]
 or
101752 ?        S      0:00 [ksmbd:10.177.110.57]

Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Yunseong Kim
305853cce3 ksmbd: Fix race condition in RPC handle list access
The 'sess->rpc_handle_list' XArray manages RPC handles within a ksmbd
session. Access to this list is intended to be protected by
'sess->rpc_lock' (an rw_semaphore). However, the locking implementation was
flawed, leading to potential race conditions.

In ksmbd_session_rpc_open(), the code incorrectly acquired only a read lock
before calling xa_store() and xa_erase(). Since these operations modify
the XArray structure, a write lock is required to ensure exclusive access
and prevent data corruption from concurrent modifications.

Furthermore, ksmbd_session_rpc_method() accessed the list using xa_load()
without holding any lock at all. This could lead to reading inconsistent
data or a potential use-after-free if an entry is concurrently removed and
the pointer is dereferenced.

Fix these issues by:
1. Using down_write() and up_write() in ksmbd_session_rpc_open()
   to ensure exclusive access during XArray modification, and ensuring
   the lock is correctly released on error paths.
2. Adding down_read() and up_read() in ksmbd_session_rpc_method()
   to safely protect the lookup.

Fixes: a1f46c99d9 ("ksmbd: fix use-after-free in ksmbd_session_rpc_open")
Fixes: b685757c7b ("ksmbd: Implements sess->rpc_handle_list as xarray")
Cc: stable@vger.kernel.org
Signed-off-by: Yunseong Kim <ysk@kzalloc.com>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-30 21:37:54 -05:00
Enzo Matsumiya
998a67b954 smb: client: fix crypto buffers in non-linear memory
The crypto API, through the scatterlist API, expects input buffers to be
in linear memory.  We handle this with the cifs_sg_set_buf() helper
that converts vmalloc'd memory to their corresponding pages.

However, when we allocate our aead_request buffer (@creq in
smb2ops.c::crypt_message()), we do so with kvzalloc(), which possibly
puts aead_request->__ctx in vmalloc area.

AEAD algorithm then uses ->__ctx for its private/internal data and
operations, and uses sg_set_buf() for such data on a few places.

This works fine as long as @creq falls into kmalloc zone (small
requests) or vmalloc'd memory is still within linear range.

Tasks' stacks are vmalloc'd by default (CONFIG_VMAP_STACK=y), so too
many tasks will increment the base stacks' addresses to a point where
virt_addr_valid(buf) will fail (BUG() in sg_set_buf()) when that
happens.

In practice: too many parallel reads and writes on an encrypted mount
will trigger this bug.

To fix this, always alloc @creq with kmalloc() instead.
Also drop the @sensitive_size variable/arguments since
kfree_sensitive() doesn't need it.

Backtrace:

[  945.272081] ------------[ cut here ]------------
[  945.272774] kernel BUG at include/linux/scatterlist.h:209!
[  945.273520] Oops: invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC NOPTI
[  945.274412] CPU: 7 UID: 0 PID: 56 Comm: kworker/u33:0 Kdump: loaded Not tainted 6.15.0-lku-11779-g8e9d6efccdd7-dirty #1 PREEMPT(voluntary)
[  945.275736] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-2-gc13ff2cd-prebuilt.qemu.org 04/01/2014
[  945.276877] Workqueue: writeback wb_workfn (flush-cifs-2)
[  945.277457] RIP: 0010:crypto_gcm_init_common+0x1f9/0x220
[  945.278018] Code: b0 00 00 00 48 83 c4 08 5b 5d 41 5c 41 5d 41 5e 41 5f c3 cc cc cc cc 48 c7 c0 00 00 00 80 48 2b 05 5c 58 e5 00 e9 58 ff ff ff <0f> 0b 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b 48 c7 04 24 01 00 00 00 48 8b
[  945.279992] RSP: 0018:ffffc90000a27360 EFLAGS: 00010246
[  945.280578] RAX: 0000000000000000 RBX: ffffc90001d85060 RCX: 0000000000000030
[  945.281376] RDX: 0000000000080000 RSI: 0000000000000000 RDI: ffffc90081d85070
[  945.282145] RBP: ffffc90001d85010 R08: ffffc90001d85000 R09: 0000000000000000
[  945.282898] R10: ffffc90001d85090 R11: 0000000000001000 R12: ffffc90001d85070
[  945.283656] R13: ffff888113522948 R14: ffffc90001d85060 R15: ffffc90001d85010
[  945.284407] FS:  0000000000000000(0000) GS:ffff8882e66cf000(0000) knlGS:0000000000000000
[  945.285262] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  945.285884] CR2: 00007fa7ffdd31f4 CR3: 000000010540d000 CR4: 0000000000350ef0
[  945.286683] Call Trace:
[  945.286952]  <TASK>
[  945.287184]  ? crypt_message+0x33f/0xad0 [cifs]
[  945.287719]  crypto_gcm_encrypt+0x36/0xe0
[  945.288152]  crypt_message+0x54a/0xad0 [cifs]
[  945.288724]  smb3_init_transform_rq+0x277/0x300 [cifs]
[  945.289300]  smb_send_rqst+0xa3/0x160 [cifs]
[  945.289944]  cifs_call_async+0x178/0x340 [cifs]
[  945.290514]  ? __pfx_smb2_writev_callback+0x10/0x10 [cifs]
[  945.291177]  smb2_async_writev+0x3e3/0x670 [cifs]
[  945.291759]  ? find_held_lock+0x32/0x90
[  945.292212]  ? netfs_advance_write+0xf2/0x310
[  945.292723]  netfs_advance_write+0xf2/0x310
[  945.293210]  netfs_write_folio+0x346/0xcc0
[  945.293689]  ? __pfx__raw_spin_unlock_irq+0x10/0x10
[  945.294250]  netfs_writepages+0x117/0x460
[  945.294724]  do_writepages+0xbe/0x170
[  945.295152]  ? find_held_lock+0x32/0x90
[  945.295600]  ? kvm_sched_clock_read+0x11/0x20
[  945.296103]  __writeback_single_inode+0x56/0x4b0
[  945.296643]  writeback_sb_inodes+0x229/0x550
[  945.297140]  __writeback_inodes_wb+0x4c/0xe0
[  945.297642]  wb_writeback+0x2f1/0x3f0
[  945.298069]  wb_workfn+0x300/0x490
[  945.298472]  process_one_work+0x1fe/0x590
[  945.298949]  worker_thread+0x1ce/0x3c0
[  945.299397]  ? __pfx_worker_thread+0x10/0x10
[  945.299900]  kthread+0x119/0x210
[  945.300285]  ? __pfx_kthread+0x10/0x10
[  945.300729]  ret_from_fork+0x119/0x1b0
[  945.301163]  ? __pfx_kthread+0x10/0x10
[  945.301601]  ret_from_fork_asm+0x1a/0x30
[  945.302055]  </TASK>

Fixes: d08089f649 ("cifs: Change the I/O paths to use an iterator rather than a page list")
Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-29 21:21:23 -05:00
Eric Biggers
8e979aab34 smb: Use arc4 library instead of duplicate arc4 code
fs/smb/common/cifs_arc4.c has an implementation of ARC4, but a copy of
this same code is also present in lib/crypto/arc4.c to serve the other
users of this legacy algorithm in the kernel.  Remove the duplicate
implementation in fs/smb/, which seems to have been added because of a
misunderstanding, and just use the lib/crypto/ one.

Signed-off-by: Eric Biggers <ebiggers@kernel.org>
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
2025-09-29 21:21:23 -05:00