From: Jan Kara <jack@xxxxxxx>
commit b2bbb92f7042e8075fb036bf97043339576330c3 upstream.
Commit 81414b4dd48 ("ext4: remove redundant sb checksum
recomputation") removed checksum recalculation after updating
superblock free space / inode counters in ext4_fill_super() based on
the fact that we will recalculate the checksum on superblock
writeout.
That is correct assumption but until the writeout happens (which can
take a long time) the checksum is incorrect in the buffer cache and if
programs such as tune2fs or resize2fs is called shortly after a file
system is mounted can fail. So return back the checksum recalculation
and add a comment explaining why.
Fixes: 81414b4dd48f ("ext4: remove redundant sb checksum recomputation")
Cc: stable@xxxxxxxxxx
Reported-by: Boyang Xue <bxue@xxxxxxxxxx>
Signed-off-by: Jan Kara <jack@xxxxxxx>
Signed-off-by: Theodore Ts'o <tytso@xxxxxxx>
Link: https://lore.kernel.org/r/20210812124737.21981-1-jack@xxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
fs/ext4/super.c | 8 ++++++++
1 file changed, 8 insertions(+)
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -5032,6 +5032,14 @@ no_journal:
err = percpu_counter_init(&sbi->s_freeinodes_counter, freei,
GFP_KERNEL);
}
+ /*
+ * Update the checksum after updating free space/inode
+ * counters. Otherwise the superblock can have an incorrect
+ * checksum in the buffer cache until it is written out and
+ * e2fsprogs programs trying to open a file system immediately
+ * after it is mounted can fail.
+ */
+ ext4_superblock_csum_set(sb);
if (!err)
err = percpu_counter_init(&sbi->s_dirs_counter,
ext4_count_dirs(sb), GFP_KERNEL);
