We periodically check the available rt blocks when filling up zones
and start GC if needed, but we may run completely out in between
filling zones, so start GC(unless already running) if we can't reserve
writable space.

This should only happen as a corner case in setups with very few
backing zones.

Fixes: 080d01c41d44 ("xfs: implement zoned garbage collection")
Signed-off-by: Hans Holmberg <hans.holmberg@wdc.com>
---

This issue was found in a yet-to-be-upstreamed xfstest.

 fs/xfs/xfs_zone_space_resv.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/fs/xfs/xfs_zone_space_resv.c b/fs/xfs/xfs_zone_space_resv.c
index 4bf1b18aa7a7..4433a060b7ff 100644
--- a/fs/xfs/xfs_zone_space_resv.c
+++ b/fs/xfs/xfs_zone_space_resv.c
@@ -159,6 +159,16 @@ xfs_zoned_reserve_available(
 		if (error != -ENOSPC)
 			break;
 
+		/*
+		 * Make sure to start GC if it is not running already. As we
+		 * check the rtavailable count when filling up zones, GC is
+		 * normally already running at this point, but in some setups
+		 * with very few zones we may completely run out of non-
+		 * reserved blocks in between filling zones.
+		 */
+		if (!xfs_is_zonegc_running(mp))
+			wake_up_process(zi->zi_gc_thread);
+
 		/*
 		 * If there is no reclaimable group left and we aren't still
 		 * processing a pending GC request give up as we're fully out
-- 
2.34.1

On Mon, 10 Mar 2025 13:39:59 +0000, Hans Holmberg wrote:
> We periodically check the available rt blocks when filling up zones
> and start GC if needed, but we may run completely out in between
> filling zones, so start GC(unless already running) if we can't reserve
> writable space.
> 
> This should only happen as a corner case in setups with very few
> backing zones.
> 
> [...]

Applied to for-next, thanks!

[1/1] xfs: trigger zone GC when out of available rt blocks
      commit: b7bc85480b03765a7993262f2c333628c36fbc45

Best regards,
-- 
Carlos Maiolino <cem@kernel.org>

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>