Re: blktests failures with v6.19 kernel

[Nilay Shroff <nilay@linux.ibm.com>]

Hi Chaitanya,

On 2/13/26 1:27 PM, Shinichiro Kawasaki wrote:
> Hi all,
> 
> I ran the latest blktests (git hash: b5b699341102) with the v6.19 kernel. I
> observed 6 failures listed below. Comparing with the previous report with the
> v6.19-rc1 kernel [1], two failures were resolved (nvme/033 and srp) and three
> failures are newly observed (nvme/061, zbd/009 and zbd/013). Recently, kmemleak
> support was introduced to blktests. Two out of the three new failures were
> detected by kmemleak. Your actions to fix the failures will be appreciated as
> always.
> 
> [1] https://lore.kernel.org/linux-block/a078671f-10b3-47e7-acbb-4251c8744523@wdc.com/
> 
> 
> List of failures
> ================
> #1: nvme/005,063 (tcp transport)
> #2: nvme/058 (fc transport)
> #3: nvme/061 (rdma transport, siw driver)(new)(kmemleak)
> #4: nbd/002
> #5: zbd/009 (new)(kmemleak)
> #6: zbd/013 (new)
> 
> 
> Failure description
> ===================
> 
> #1: nvme/005,063 (tcp transport)
> 
>     The test case nvme/005 and 063 fail for tcp transport due to the lockdep
>     WARN related to the three locks q->q_usage_counter, q->elevator_lock and
>     set->srcu. Refer to the nvme/063 failure report for v6.16-rc1 kernel [2].
> 
>     [2] https://lore.kernel.org/linux-block/4fdm37so3o4xricdgfosgmohn63aa7wj3ua4e5vpihoamwg3ui@fq42f5q5t5ic/

For the lockdep failure reported above for nvme/063, it seems we already had
solution but it appears that it's not yet upstreamed, check this:
https://lore.kernel.org/all/20251125061142.18094-1-ckulkarnilinux@gmail.com/

Can you please update and resend the above patch per the last feedback? I think
this should fix the lockdep reported under nvme/063.

Thanks,
--Nilay