The new auditing standard for the subsystem will be to only use
__aligned_64 in uapi headers to try and prevent 32/64 compat bugs
from existing in the future.
Changing all existing usage will help ensure new developers copy the
right idea.
The before/after of this patch was tested using pahole on 32 and 64
bit compiles to confirm it has no change in the structure layout, so
this patch is a NOP.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
With 32 bit compilation several of the fields become misaligned here.
Fixing this is an ABI break for 32 bit rxe and it is in well used
portions of the rxe ABI.
To handle this we bump the ABI version, as expected. However the user
space driver doesn't handle it properly today, so all existing user
space continues to work.
Updated userspace will start to require the necessary kernel version.
We don't expect there to be any 32 bit users of rxe. Most likely cases,
such as ARM 32 already generally don't work because rxe does not handle
the CPU cache properly on its shared with userspace pages.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
rss_caps in struct mlx4_uverbs_ex_query_device_resp is misaligned on
32 bit compared to 64 bit, add explicit padding.
The rss caps were introduced recently and are very rarely used in user
space, mainly for DPDK.
We don't expect there to be a real 32 bit user, so this change is done
without compat considerations.
Fixes: 09d208b258 ("IB/mlx4: Add report for RSS capabilities by vendor channel")
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
struct qedr_alloc_ucontext_resp is a different length in 32 and 64
bit compiles due to implicit compiler padding.
The structs alloc_pd_uresp, create_cq_uresp and create_qp_uresp are
not padded by the compiler, but in user space the compiler pads them
due to the way the core and driver structs are concatenated. Make
this padding explicit and consistent for future sanity.
The kernel driver can already handle the user buffer being smaller
than required and copies correctly, so no compat or ABI break happens
from introducing the explicit padding.
Acked-by: Michal Kalderon <michal.kalderon@cavium.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
The rdma_ucm_event_resp is a different length on 32 and 64 bit compiles.
The kernel requires it to be the expected length or longer so 32 bit
builds running on a 64 bit kernel will not work.
Retain full compat by having all kernels accept a struct with or without
the trailing reserved field.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
To help automatic detection we want pahole to report the same struct
layouts for 32 and 64 bit compiles. These cases are all implicit
padding added at the end of embedded structs as part of a union.
The added reserved fields have no impact on the ABI.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Generally we do a preload before doing idr allocation. This also help
improve the allocation success rate in memory pressure.
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Even after the previous patch to drop lo_ctl_mutex while calling
vfs_getattr(), there are other cases where we can end up sleeping for a
long time while holding lo_ctl_mutex. Let's avoid the uninterruptible
sleep from the ioctls.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We hit an issue where a loop device on NFS was stuck in
loop_get_status() doing vfs_getattr() after the NFS server died, which
caused a pile-up of uninterruptible processes waiting on lo_ctl_mutex.
There's no reason to hold this lock while we wait on the filesystem;
let's drop it so that other processes can do their thing. We need to
grab a reference on lo_backing_file while we use it, and we can get rid
of the check on lo_device, which has been unnecessary since commit
a34c0ae9ebd6 ("[PATCH] loop: remove the bio remapping capability") in
the linux-history tree.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ensure that device exists prior to accessing its properties.
Reported-by: <syzbot+71655d44855ac3e76366@syzkaller.appspotmail.com>
Fixes: 7521663857 ("RDMA/cma: Export rdma cm interface to userspace")
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
In an effort to pick up channels that are in a funky state we
optimistically tried to open all channels that we found, with the
addition that we failed if the other side did not handshake the opening.
But as we're starting the modem a second time all channels are found -
in a "funky" state - and we try to open them. But the modem firmware
requires the IPCRTR to be up in order to initialize. So any channels we
try to open before that will fail and will not be opened again.
This takes care of the regression, at the cost of reintroducing the
previous behavior of handling of channels with "funky" states.
Reverts commit c12fc4519f ("rpmsg: smd: Create device for all channels")
Reported-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Edward Cree says:
====================
sfc: rework locking around filter management
The use of a spinlock to protect filter state combined with the need for a
sleeping operation (MCDI) to apply that state to the NIC (on EF10) led to
unfixable race conditions, around the handling of filter restoration after
an MC reboot.
So, this patch series removes the requirement to be able to modify the SW
filter table from atomic context, by using a workqueue to request
asynchronous filter operations (which are needed for ARFS). Then, the
filter table locks are changed to mutexes, replacing the dance of spinlocks
and 'busy' flags. Also, a mutex is added to protect the RSS context state,
since otherwise a similar race is possible around restoring that after an
MC reboot. While we're at it, fix a couple of other related bugs.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The FLOW_RSS flag was causing us to insert UDP filters when TCP was wanted.
Fixes: 42356d9a13 ("sfc: support RSS spreading of ethtool ntuple filters")
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Otherwise races are possible between ethtool ops and
efx_ef10_rx_restore_rss_contexts().
Also, don't try to perform the restore on every reset, only after an MC
reboot, otherwise we'll leak RSS contexts on the NIC.
Fixes: 42356d9a13 ("sfc: support RSS spreading of ethtool ntuple filters")
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If some other operation gets the MCDI lock ahead of us and performs an MC
reboot, then our attempt to insert the filter will fail with EINVAL,
because the destination VI (spec->dmaq_id, MC_CMD_FILTER_OP_IN_RX_QUEUE) does
not exist. But the caller's request (which might e.g. be an ethtool ntuple
request from userland) isn't invalid, it just got unlucky; so return EAGAIN.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With this change, the spinlock efx->filter_lock is no longer used and is
thus removed.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
efx->filter_lock remains in place for use on farch, but EF10 now ignores it.
EFX_EF10_FILTER_FLAG_BUSY is no longer needed, hence it is removed.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of having an efx->type->filter_rfs_insert() method, just use
workitems with a worker function that calls efx->type->filter_insert().
The only user of this is efx_filter_rfs(), which now queues a call to
efx_filter_rfs_work().
Similarly, efx_filter_rfs_expire() is now a worker function called on a
new channel->filter_work work_struct, so the method
efx->type->filter_rfs_expire_one() is no longer called in atomic context.
We also add a new mutex efx->rps_mutex to protect the RPS state (efx->
rps_expire_channel, efx->rps_expire_index, and channel->rps_flow_id) so
that the taking of efx->filter_lock can be moved to
efx->type->filter_rfs_expire_one().
Thus, all filter table functions are now called in a sleepable context,
allowing them to use sleeping locks in a future patch.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Kirill Tkhai says:
====================
Make pernet_operations always read locked
All the pernet_operations are converted, and the last one
is in this patchset (nfsd_net_ops acked by J. Bruce Fields).
So, it's the time to kill pernet_operations::async field,
and make setup_net() and cleanup_net() always require
the rwsem only read locked.
All further pernet_operations have to be developed to fit
this rule. Some of previous patches added a comment to
struct pernet_operations about that.
Also, this patchset renames net_sem to pernet_ops_rwsem
to make the target area of the rwsem is more clear visible,
and adds more comments.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This adds comments to different places to improve
readability.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net_sem is some undefined area name, so it will be better
to make the area more defined.
Rename it to pernet_ops_rwsem for better readability and
better intelligibility.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Synchronous pernet_operations are not allowed anymore.
All are asynchronous. So, drop the structure member.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All pernet_operations are reviewed and converted, hooray!
Reflect this in core code: setup_net() and cleanup_net()
will take down_read() always.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These pernet_operations look similar to rpcsec_gss_net_ops,
they just create and destroy another caches. So, they also
can be async.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use relaxed I/O on the hot path. This achieves significant performance
improvements. On a 10G link, this makes a basic iperf TCP test go from
an average of 4.5 Gbits/sec to about 9.40 Gbits/sec.
Signed-off-by: Yan Markman <ymarkman@marvell.com>
[Maxime: Commit message, cosmetic changes]
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit c5ad119fb6
("net: sched: pfifo_fast use skb_array") driver is exposed
to an issue where it is hitting NULL skbs while handling TX
completions. Driver uses mmiowb() to flush the writes to the
doorbell bar which is a write-combined bar, however on x86
mmiowb() does not flush the write combined buffer.
This patch fixes this problem by replacing mmiowb() with wmb()
after the write combined doorbell write so that writes are
flushed and synchronized from more than one processor.
V1->V2:
-------
This patch was marked as "superseded" in patchwork.
(Not really sure for what reason).Resending it as v2.
Signed-off-by: Ariel Elior <ariel.elior@cavium.com>
Signed-off-by: Manish Chopra <manish.chopra@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We tried to remove vq poll from wait queue, but do not check whether
or not it was in a list before. This will lead double free. Fixing
this by switching to use vhost_poll_stop() which zeros poll->wqh after
removing poll from waitqueue to make sure it won't be freed twice.
Cc: Darren Kenny <darren.kenny@oracle.com>
Reported-by: syzbot+c0272972b01b872e604a@syzkaller.appspotmail.com
Fixes: 2b8b328b61 ("vhost_net: handle polling errors when setting backend")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As found by the ubsan checker, the value of the 'index' variable can be
out of range for the bc[] array:
UBSAN: Undefined behaviour in arch/parisc/kernel/drivers.c:655:21
index 6 is out of range for type 'char [6]'
Backtrace:
[<104fa850>] __ubsan_handle_out_of_bounds+0x68/0x80
[<1019d83c>] check_parent+0xc0/0x170
[<1019d91c>] descend_children+0x30/0x6c
[<1059e164>] device_for_each_child+0x60/0x98
[<1019cd54>] parse_tree_node+0x40/0x54
[<1019d86c>] check_parent+0xf0/0x170
[<1019d91c>] descend_children+0x30/0x6c
[<1059e164>] device_for_each_child+0x60/0x98
[<1019d938>] descend_children+0x4c/0x6c
[<1059e164>] device_for_each_child+0x60/0x98
[<1019cd54>] parse_tree_node+0x40/0x54
[<1019cffc>] hwpath_to_device+0xa4/0xc4
Signed-off-by: Helge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org
Qemu now supports emulating PA-RISC machines. For that a forked version
of SeaBIOS available at https://github.com/hdeller/seabios-hppa is used
which requires some information about the emulated machine.
This patch adds code to generate a header file with the necessary
information for SeaBIOS. The information is extracted from the firmware
the current kernel is running on.
Tested on a B160L workstation.
Signed-off-by: Helge Deller <deller@gmx.de>
Carlo Pisani noticed that his C3600 workstation behaved unstable during heavy
I/O on the PCI bus with a VIA VT6421 IDE/SATA PCI card.
To avoid such instability, this patch switches the LBA PCI bus from Hard Fail
mode into Soft Fail mode. In this mode the bus will return -1UL for timed out
MMIO transactions, which is exactly how the x86 (and most other architectures)
PCI busses behave.
This patch is based on a proposal by Grant Grundler and Kyle McMartin 10
years ago:
https://www.spinics.net/lists/linux-parisc/msg01027.html
Cc: Carlo Pisani <carlojpisani@gmail.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Reviewed-by: Grant Grundler <grantgrundler@gmail.com>
Signed-off-by: Helge Deller <deller@gmx.de>
Make sure that the HPMC (High Priority Machine Check) handler is 16-byte
aligned and that it's length in the IVT is a multiple of 16 bytes.
Otherwise PDC may decide not to call the HPMC crash handler.
Signed-off-by: Helge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org
Smatch warns that is_tiny can be used uninitialized:
arch/parisc/math-emu/fcnvff.c:297 dbl_to_sgl_fcnvff()
error: uninitialized symbol 'is_tiny'.
This code is very old so that suggests the bug doesn't have a huge
affect in real life. But I've read the code and it seems like a
reasonable warning. Either way it should be harmless to initialize it
to false and silence the static checker warning.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Helge Deller <deller@gmx.de>
On parisc we want to be as much as possible compatible to the major
architectures like x86. Those architectures have MAP_TYPE defined as 0x0f which
covers MAP_SHARED and MAP_PRIVATE and leaves two more bits unused.
In contrast, on parisc we have MAP_TYPE defined to 0x03 which covers MAP_SHARED
and MAP_PRIVATE only. But we don't have the 2 bits free as x86.
Usually that's not a problem, but during the discussions for pmem+dax support
the idea came up to use the two remaining bits of MAP_TYPE (on x86 and others)
for the new MAP_DIRECT and MAP_SYNC flags. One requirement is, that an old
kernel should correctly handle MAP_DIRECT and MAP_SYNC and fail on those if
set. This only works if MAP_TYPE has 4 bits.
Even though the pmem+dax people now choosed another solution via
MAP_SHARED_VALIDATE, let's still proceed to be more compatible to x86 by adding
two more bits for future usage.
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: John David Anglin <dave.anglin@bell.net>
There is a change in how command line parsing is done in this version.
Excludes and includes are now ordered with the file list. Since
the spec file puts the file list before the exclude list it means newer
tar ignores the excludes and packs all the build output into the
kernel-devel RPM resulting in a huge package.
Simple argument re-ordering fixes the problem.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Since d5d332d3f7, a couple of links in scripts/dtc/include-prefixes
are additionally required in order to build device trees with the header
package.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Reviewed-by: Riku Voipio <riku.voipio@linaro.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Add nodes of the AVE ethernet controller for PXs3 and the boards.
This SoC has two controllers.
Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Add pinctrl groups of ethernet, such as "ether1_rgmii" and "ether1_rmii".
These are used for second ethernet instance.
Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>