Re: [PATCH v2] xfs: start gc on zonegc_low_space attribute updates

[Dave Chinner <dgc@kernel.org>]

On Wed, Mar 25, 2026 at 01:43:12PM +0100, Hans Holmberg wrote:
> Start gc if the agressiveness of zone garbage collection is changed
> by the user (if the file system is not read only).
> 
> Without this change, the new setting will not be taken into account
> until the gc thread is woken up by e.g. a write.
> 
> Cc: <stable@vger.kernel.org> # v6.15
> Fixes: 845abeb1f06a8a ("xfs: add tunable threshold parameter for triggering zone GC")
> Signed-off-by: Hans Holmberg <hans.holmberg@wdc.com>
> ---
> 
> v2:
> - Added a new helper to wake up the gc thread in stead of unparking it,
>   which is required to make this work properly.
> - Added protection against races with unmounts as sysfs gets torn down
>   after the zone info struct is freed. This also avoids unneded
>   wakeups during remount.

Isn't that a deadlock vector?

i.e. unmount takes s_umount, concurrently userspace writes a new
value to sysfs file. sysfs file write blocks on s_umount, unmount
blocks holding s_umount waiting for sysfs file reference count to go
to zero to destroy it?

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