PT-2026-43917 · Linux · Linux
Published
2026-05-27
·
Updated
2026-05-27
·
CVE-2026-46050
None
No severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
In the Linux kernel, the following vulnerability has been resolved:
md/raid10: fix deadlock with check operation and nowait requests
When an array check is running it will raise the barrier at which point
normal requests will become blocked and increment the nr pending value to
signal there is work pending inside of wait barrier(). NOWAIT requests
do not block and so will return immediately with an error, and additionally
do not increment nr pending in wait barrier(). Upstream change commit
43806c3d5b9b ("raid10: cleanup memleak at raid10 make request") added a
call to raid end bio io() to fix a memory leak when NOWAIT requests hit
this condition. raid end bio io() eventually calls allow barrier() and
it will unconditionally do an atomic dec and test(&conf->nr pending) even
though the corresponding increment on nr pending didn't happen in the
NOWAIT case.
This can be easily seen by starting a check operation while an application
is doing nowait IO on the same array. This results in a deadlocked state
due to nr pending value underflowing and so the md resync thread gets stuck
waiting for nr pending to == 0.
Output of r10conf state of the array when we hit this condition:
crash> struct r10conf
barrier = 1,
nr pending = {
counter = -41
},
nr waiting = 15,
nr queued = 0,
Example of md sync thread stuck waiting on raise barrier() and other
requests stuck in wait barrier():
md1 resync
[<0>] raise barrier+0xce/0x1c0
[<0>] raid10 sync request+0x1ca/0x1ed0
[<0>] md do sync+0x779/0x1110
[<0>] md thread+0x90/0x160
[<0>] kthread+0xbe/0xf0
[<0>] ret from fork+0x34/0x50
[<0>] ret from fork asm+0x1a/0x30
kworker/u1040:2+flush-253:4
[<0>] wait barrier+0x1de/0x220
[<0>] regular request wait+0x30/0x180
[<0>] raid10 make request+0x261/0x1000
[<0>] md handle request+0x13b/0x230
[<0>] submit bio+0x107/0x1f0
[<0>] submit bio noacct nocheck+0x16f/0x390
[<0>] ext4 io submit+0x24/0x40
[<0>] ext4 do writepages+0x254/0xc80
[<0>] ext4 writepages+0x84/0x120
[<0>] do writepages+0x7a/0x260
[<0>] writeback single inode+0x3d/0x300
[<0>] writeback sb inodes+0x1dd/0x470
[<0>] writeback inodes wb+0x4c/0xe0
[<0>] wb writeback+0x18b/0x2d0
[<0>] wb workfn+0x2a1/0x400
[<0>] process one work+0x149/0x330
[<0>] worker thread+0x2d2/0x410
[<0>] kthread+0xbe/0xf0
[<0>] ret from fork+0x34/0x50
[<0>] ret from fork asm+0x1a/0x30
Found an issue in the description? Have something to add? Feel free to write us 👾
Related Identifiers
Affected Products
Linux