The driver could possibly sleep while in atomic context resulting
in the following call trace while CONFIG_DEBUG_ATOMIC_SLEEP=y is
set:
[ 675.116953] BUG: spinlock bad magic on CPU#2, bash/2481
[ 675.116966] lock: 0xffff8d6052a88f50, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
[ 675.116978] CPU: 2 PID: 2481 Comm: bash Tainted: G S 6.6.0-rc1-next-20230911 #1
[ 675.116986] Hardware name: HPE ProLiant DL360 Gen10/ProLiant DL360 Gen10, BIOS U32 01/23/2021
[ 675.116991] Call Trace:
[ 675.116997] <TASK>
[ 675.117002] dump_stack_lvl+0x36/0x50
[ 675.117014] do_raw_spin_lock+0x79/0xc0
[ 675.117032] pds_vfio_reset+0x1d/0x60 [pds_vfio_pci]
[ 675.117049] pci_reset_function+0x4b/0x70
[ 675.117061] reset_store+0x5b/0xa0
[ 675.117074] kernfs_fop_write_iter+0x137/0x1d0
[ 675.117087] vfs_write+0x2de/0x410
[ 675.117101] ksys_write+0x5d/0xd0
[ 675.117111] do_syscall_64+0x3b/0x90
[ 675.117122] entry_SYSCALL_64_after_hwframe+0x6e/0xd8
[ 675.117135] RIP: 0033:0x7f9ebbd1fa28
[ 675.117141] Code: 89 02 48 c7 c0 ff ff ff ff eb b3 0f 1f 80 00 00 00 00 f3 0f 1e fa 48 8d 05 15 4d 2a 00 8b 00 85 c0 75 17 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 58 c3 0f 1f 80 00 00 00 00 41 54 49 89 d4 55
[ 675.117148] RSP: 002b:00007ffdff410728 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
[ 675.117156] RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 00007f9ebbd1fa28
[ 675.117161] RDX: 0000000000000002 RSI: 000055ffc5fdf7c0 RDI: 0000000000000001
[ 675.117166] RBP: 000055ffc5fdf7c0 R08: 000000000000000a R09: 00007f9ebbd7fae0
[ 675.117170] R10: 000000000000000a R11: 0000000000000246 R12: 00007f9ebbfc06e0
[ 675.117174] R13: 0000000000000002 R14: 00007f9ebbfbb860 R15: 0000000000000002
[ 675.117180] </TASK>
This can happen if pds_vfio_put_restore_file() and/or
pds_vfio_put_save_file() grab the mutex_lock(&lm_file->lock)
while the spin_lock(&pds_vfio->reset_lock) is held, which can
happen during while calling pds_vfio_state_mutex_unlock().
Fix this by releasing the spin_unlock(&pds_vfio->reset_lock) before
calling pds_vfio_put_restore_file() and pds_vfio_put_save_file() and
re-acquiring spin_lock(&pds_vfio->reset_lock) after the previously
mentioned functions are called to protect setting the subsequent
state/deferred reset settings.
The only possible concerns are other threads that may call
pds_vfio_put_restore_file() and/or pds_vfio_put_save_file(). However,
those paths are already protected by the state mutex_lock().
Reported-by: Dan Carpenter <[email protected]>
Closes: https://lore.kernel.org/kvm/[email protected]/
Reviewed-by: Shannon Nelson <[email protected]>
Signed-off-by: Brett Creeley <[email protected]>
---
drivers/vfio/pci/pds/vfio_dev.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/vfio/pci/pds/vfio_dev.c b/drivers/vfio/pci/pds/vfio_dev.c
index b46174f5eb09..0807d163f83e 100644
--- a/drivers/vfio/pci/pds/vfio_dev.c
+++ b/drivers/vfio/pci/pds/vfio_dev.c
@@ -33,8 +33,10 @@ void pds_vfio_state_mutex_unlock(struct pds_vfio_pci_device *pds_vfio)
if (pds_vfio->deferred_reset) {
pds_vfio->deferred_reset = false;
if (pds_vfio->state == VFIO_DEVICE_STATE_ERROR) {
+ spin_unlock(&pds_vfio->reset_lock);
pds_vfio_put_restore_file(pds_vfio);
pds_vfio_put_save_file(pds_vfio);
+ spin_lock(&pds_vfio->reset_lock);
pds_vfio_dirty_disable(pds_vfio, false);
}
pds_vfio->state = pds_vfio->deferred_reset_state;
--
2.17.1