On 2022/3/5 12:00, Fengnan Chang wrote:
When compressed file has blocks, start atomic write will be failed, but
You mean f2fs_ioc_start_atomic_write will succeed, but compressed flag
will be remained in inode?
still set FI_ATOMIC_FILE flag, if write partial cluster and commit
atomic write will cause data corruption.
Could you please explain more about why data will be corrupted?
Thanks,
Fixes: 4c8ff7095bef (f2fs: support data compression)
Fixes: 7eab7a696827 (f2fs: compress: remove unneeded read when rewrite whole
cluster)
Signed-off-by: Fengnan Chang <[email protected]>
---
fs/f2fs/data.c | 4 +---
fs/f2fs/file.c | 3 ++-
2 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 6b5f389ba998..5cbee4ed0982 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -3358,8 +3358,7 @@ static int f2fs_write_begin(struct file *file, struct
address_space *mapping,
int ret;
*fsdata = NULL;
-
- if (len == PAGE_SIZE)
+ if (len == PAGE_SIZE && !(f2fs_is_atomic_file(inode)))
goto repeat;
ret = f2fs_prepare_compress_overwrite(inode, pagep,
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index cfdc41f87f5d..26a648ef9e1f 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -2009,7 +2009,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp)
inode_lock(inode);
- f2fs_disable_compressed_file(inode);
+ if (!f2fs_disable_compressed_file(inode))
+ return -EINVAL;
if (f2fs_is_atomic_file(inode)) {
if (is_inode_flag_set(inode, FI_ATOMIC_REVOKE_REQUEST))
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
