Re: [PATCH] xfs: don't let racing writers consume the free zones reserved for GC

[Christoph Hellwig <hch@lst.de>]

On Sun, Sep 27, 2026 at 02:48:41PM +0200, Hans Holmberg wrote:
> xfs_try_open_zone() checks the free zone count to avoid grabbing zones
> required for GC forward progress, but does so without decreasing the
> counter, letting multiple user writers through to call xfs_open_zone()
> and collectivly gobble up all free zones, stalling gc and eventually
> leading to user threads getting stuck waiting for free space.
> 
> Fold the check and the accounting into a single atomic claim in
> xfs_open_zone() instead.  The free zone count is decremented before the
> zone is looked up and given back if none could be grabbed, so the reserve
> can no longer be raced away.  User allocations only claim a zone if that
> leaves the reserve behind, while GC just needs one free zone and thus can
> still use the zones that are kept for it.

Ouch.   Yes, this looks good.  One nit:

> @@ -460,11 +483,12 @@ xfs_open_zone(
>  		if (atomic_inc_not_zero(&xg->xg_active_ref))
>  			goto found;
>  	xas_unlock(&xas);
> +
> +	atomic_inc(&zi->zi_nr_free_zones);
>  	return NULL;

This NULL return case now is a bug as the earlier atomic decrement
of zi_nr_free_zones should protect against it ever happening.

So this should have a comment and some kind of warning (WARN_ON maybe,
the usual XFS_IS_CORRUPT doesn't really fit for a pure in-core issue).