During the change to new btrfs extent-oriented qgroup implement, due to
it doesn't use the old __qgroup_excl_accounting() for exclusive extent,
it didn't free the reserved bytes.
The bug will cause limit function go crazy as the reserved space is
never freed, increasing limit will have no effect and still cause
EQOUT.
The fix is easy, just free reserved bytes for newly created exclusive
extent as what it does before.
Reported-by: Tsutomu Itoh <t-i...@jp.fujitsu.com>
Signed-off-by: Yang Dongsheng <yangds.f...@cn.fujitsu.com>
Signed-off-by: Qu Wenruo <quwen...@cn.fujitsu.com>
---
To Chris:
Would you please consider merging this patch in the late v4.2 merge
window?
As it's a huge regression and the fix is small enough and just does
what it did before in __qgroup_excl_accounting() function.
The corresponding test case will follow soon.
And sorry for the regression I introduced.
Thanks,
Qu
---
fs/btrfs/qgroup.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index d5f1f03..1667567 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -1618,6 +1618,11 @@ static int qgroup_update_counters(struct btrfs_fs_info
*fs_info,
/* Exclusive -> exclusive, nothing changed */
}
}
+
+ /* For exclusive extent, free its reserved bytes too */
+ if (nr_old_roots == 0 && nr_new_roots == 1 &&
+ cur_new_count == nr_new_roots)
+ qg->reserved -= num_bytes;
if (dirty)
qgroup_dirty(fs_info, qg);
}
--
2.5.0
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
