Re: [PATCH] btrfs: fix outstanding extents calculation

[Filipe Manana <fdmanana@kernel.org>]

On Mon, Mar 28, 2022 at 01:14:02PM +0000, Johannes Thumshirn wrote:
> On 28/03/2022 15:09, Filipe Manana wrote:
> > On Mon, Mar 28, 2022 at 09:32:05PM +0900, Naohiro Aota wrote:
> >> Running generic/406 causes the following WARNING in btrfs_destroy_inode()
> >> which tells there is outstanding extents left.
> > 
> > I can't trigger the warning with generic/406.
> > Any special setup or config to trigger it?
> > 
> > The change looks fine to me, however I'm curious why this isn't triggered
> > with generic/406, nor anyone else reported it before, since the test is
> > fully deterministic.
> > 
> 
> I am able to trigger the WARN() with a different test (which is for a different,
> not yet solved problem) on my zoned setup. With this patch, the WARN() is gone.

I have no doubts about the fix being correct.
I'm just puzzled why I can't trigger it with generic/406, given that it's a very
deterministic test.

If there's any special config or setup (mount options, zoned fs, etc), I would
like to have it explicitly mentioned in the changelog.