Re: [PATCH 0/2] selftests: ublk: fix the fault_inject delay

[Jens Axboe]

On Thu, 08 Oct 2026 22:45:38 +0800, Qiliang Yuan wrote:
> The delay set by --delay_us on the fault_inject target is lost in two
> ways. Its timespec is read after it has gone out of scope, which makes
> generic_06 fail, and any other completion on the ring ends it early when
> more than one I/O is in flight. Patch 1 keeps the timespec alive until
> the SQE is submitted, and patch 2 makes the delay a pure timeout.
> 
> With both patches, a 1 s delay holds every I/O for 1.000 s at queue
> depth 4.
> 
> [...]

Applied, thanks!

[1/2] selftests: ublk: don't keep the fault_inject timespec on the stack
      commit: 955cba87d9ed334fbe0b669b35b4abcddd7f9929
[2/2] selftests: ublk: make the fault_inject delay a pure timeout
      commit: 69ae59173a5668d015ab2733c8628ed542f00ccc

Best regards,
-- 
Jens Axboe