Re: [PATCH] xfs: drain inodegc before quota teardown on mount failure

[Christoph Hellwig <hch@infradead.org>]

On Thu, May 28, 2026 at 09:28:02PM +0800, Cen Zhang wrote:
> In my local attempts, a mount-failure errortag after
> xfs_qm_mount_quotas() was not enough by itself, because ordinary test
> setup did not reliably leave an inodegc item that would enter the
> dqattach/eofblocks path during the mount-failure unwind.
> 
> The draft I have locally uses two DEBUG-only errortags:
> 
>   - mount_fail_after_qm_mount_quotas
> 
>     This injects -EIO immediately after xfs_qm_mount_quotas().
> 
>   - force_inodegc_eofblocks
> 
>     This makes a regular inode released after quotacheck scanning go
>     through inodegc/eofblocks cleanup, and delays that inodegc work so
>     the mount-failure unwind can reach quota teardown first.

This sounds pretty useful, but I actually have a different
question:

>   Workqueue: xfs-inodegc/vdc xfs_inodegc_worker
>   xfs_qm_dqget_cache_lookup()
>   xfs_qm_dqget_inode()
>   xfs_qm_dqattach_locked()
>   xfs_qm_dqattach()
>   xfs_free_eofblocks()
>   xfs_inactive()
>   xfs_inodegc_worker()

What kind of inode do we actually get here?  I guess your test
case reproduces an unclean shutdown and this processes inodes
that otherwise would have been processed on unmount?

I'm kinda tempted to just make xfs_can_free_eofblocks return false
during early mount (or rather during mount failure), as trying
to fix all the corner cases here seems like an uphill battle (
also see the sashiko review)