From: Darrick J. Wong <djwong@xxxxxxxxxx>
In the kernel, commit 8ebbf262d4684 ("xfs: don't block in busy flushing
when freeing extents") changed the allocator behavior such that AGFL
fixing can return -EAGAIN in response to detection of a deadlock with
the transaction busy extent list. If this happens, we're supposed to
requeue the EFI so that we can roll the transaction and try the item
again.
If a requeue happens, we should not free the xefi pointer in
xfs_extent_free_finish_item or else the retry will walk off a dangling
pointer. There is no extent busy list in userspace so this should
never happen, but let's fix the logic bomb anyway.
We should have ported kernel commit 0853b5de42b47 ("xfs: allow extent
free intents to be retried") to userspace, but neither Carlos nor I
noticed this fine detail. :(
Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx>
---
libxfs/defer_item.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/libxfs/defer_item.c b/libxfs/defer_item.c
index 3f519252046..8731d1834be 100644
--- a/libxfs/defer_item.c
+++ b/libxfs/defer_item.c
@@ -115,6 +115,13 @@ xfs_extent_free_finish_item(
error = xfs_free_extent(tp, xefi->xefi_pag, agbno,
xefi->xefi_blockcount, &oinfo, XFS_AG_RESV_NONE);
+ /*
+ * Don't free the XEFI if we need a new transaction to complete
+ * processing of it.
+ */
+ if (error == -EAGAIN)
+ return error;
+
xfs_extent_free_put_group(xefi);
kmem_cache_free(xfs_extfree_item_cache, xefi);
return error;
