Re: [bug report] fstests generic/774 hang again

[Dave Chinner <dgc@kernel.org>]

On Thu, Sep 17, 2026 at 09:28:15AM +0100, John Garry wrote:
> On 9/16/26 23:23, Dave Chinner wrote:
> > On Wed, Sep 16, 2026 at 11:23:16AM +0100, John Garry wrote:
> > > On 15/09/2026 16:41, John Garry wrote:
> > > > On 9/15/26 15:50, Darrick J. Wong wrote:
> > > > > > thanks
> > > > > > 
> > > > > > About the kernel code, I am wondering if using IOMAP_DIO_FORCE_WAIT for
> > > > > > CoW-based atomics could help avoid this issue as we seem to be
> > > > > > bogged down
> > > > > > in lock contention.
> > > > > Which lock is being contended, anyway?  It looks like the ILOCK?
> > > > 
> > > > Yeah, I think so Re. ilock. I had some perf data illustrating this from
> > > > last year, which I can't seem to find, so I will re-generate it.
> > > 
> > > Here is perf call graph snippet when running fio with 72x threads issuing
> > > 4KB atomic writes on 1MB file:
> > > 
> > > --50.46%--xfs_file_dio_write_atomic
> > > |
> > > |--37.19%--iomap_dio_rw
> > > |          |
> > > |      --37.14%--__iomap_dio_rw
> > > |           |
> > > |       --36.50%--iomap_iter
> > > |            |
> > > |             --36.49%--xfs_atomic_write_cow_iomap_next
> > > |                  |
> > > |                  |--27.34%--xfs_ilock
> > > |                  |          |
> > > |                  |      --27.34%--down_write
> > > |                  |           |
> > > |                  |       --27.28%--rwsem_down_write_slowpath
> > > |                  |            |
> > > |                  |            |--24.65%--osq_lock
> > > |                  |            |
> > > |                  |             --2.35%--rwsem_spin_on_owner
> > > |                  |
> > > |                  |--7.62%--xfs_trans_alloc_inode
> > > |                  |          |
> > > |                  |      --7.59%--xfs_ilock
> > 
> > Where's the other half of the lock contention?
> 
> I was expecting xfs_reflink_end_atomic_cow() -> xfs_ilock() to be a point of
> contention as well, but it does not show up. I suppose that it because all
> those threads at completion stage are sleeping.
> 
> However, there is this which I omitted:
> 
> --13.21%--xfs_file_write_checks
>    |
>       --13.18%--kiocb_modified
>          |
>             --13.17%--file_update_time_flags
>                |
>                   --13.16%--xfs_vn_update_time
>                      |
>                         --12.99%--xfs_ilock
>                            |
>                               --12.98%--down_write

Fair enough, I'd expect that from the all the threads blocked on the
lock there in the hung task output.

....

> JFYI, I have another log for a similar hang which we see one thread deadlock
> for log space, below.

Looks very similar. So, same questiosn about log size, etc.

-Dave.
-- 
Dave Chinner
dgc@kernel.org