> Jens Axboe <axboe@xxxxxxxxx> 于2021年9月13日周一 上午10:50写道：
>
> On 9/12/21 8:26 PM, Hao Sun wrote:
> > Hi
> >
> > Healer found a C reproducer for this crash ("INFO: task hung in
> > io_ring_exit_work").
> >
> > HEAD commit: 4b93c544e90e-thunderbolt: test: split up test cases
>
> Does this reproduce on 5.15-rc1? We had a few hang cases fixed for io-wq
> since that commit 6 days ago.
Just tried it. No, it did not crash the kernel on 5.15-rc1.
Following log was printed repeatedly:
[ 647.478557][ T6807] loop0: p1 p2 < > p3 p4
[ 647.478922][ T6807] loop0: p1 size 11290111 extends beyond EOD, truncated
[ 647.481111][ T6807] loop0: p3 size 1914664839 extends beyond EOD, truncated
[ 647.482512][ T6807] loop0: p4 size 3389030400 extends beyond EOD, truncated
It seems that this crash was fixed. Sorry for the dup report.
>
> --
> Jens Axboe
>
