ext4: use pre-allocated es in __es_remove_extent()
authorBaokun Li <libaokun1@huawei.com>
Mon, 24 Apr 2023 03:38:39 +0000 (11:38 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 3 Dec 2023 06:32:10 +0000 (07:32 +0100)
commitf1c236936674251649f826bb39a8b1e42332a3e0
tree92333b49d6d803cb1eb3c548e74904b552909ccb
parentce581f8631a4a772556f113656160e3befe06da5
ext4: use pre-allocated es in __es_remove_extent()

[ Upstream commit bda3efaf774fb687c2b7a555aaec3006b14a8857 ]

When splitting extent, if the second extent can not be dropped, we return
-ENOMEM and use GFP_NOFAIL to preallocate an extent_status outside of
i_es_lock and pass it to __es_remove_extent() to be used as the second
extent. This ensures that __es_remove_extent() is executed successfully,
thus ensuring consistency in the extent status tree. If the second extent
is not undroppable, we simply drop it and return 0. Then retry is no longer
necessary, remove it.

Now, __es_remove_extent() will always remove what it should, maybe more.

Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230424033846.4732-6-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Stable-dep-of: 8e387c89e96b ("ext4: make sure allocate pending entry not fail")
Signed-off-by: Sasha Levin <sashal@kernel.org>
fs/ext4/extents_status.c