UPSTREAM: f2fs: should not truncate blocks during roll-forward recovery
If the file preallocated blocks and fsync'ed, we should not truncate them during
roll-forward recovery which will recover i_size correctly back.
Bug: 223740163
Fixes: d4dd19ec1ea0 ("f2fs: do not expose unwritten blocks to user by DIO")
Cc: <stable@vger.kernel.org> # 5.17+
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
(cherry picked from commit 4d8ec91208196e0e19195f1e7d6be9de5873f242)
Change-Id: I5e974a564667115455b53a18f31902a875d86dee
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index ad5fde3..5fffc22 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -550,7 +550,8 @@
}
f2fs_set_inode_flags(inode);
- if (file_should_truncate(inode)) {
+ if (file_should_truncate(inode) &&
+ !is_sbi_flag_set(sbi, SBI_POR_DOING)) {
ret = f2fs_truncate(inode);
if (ret)
goto bad_inode;