Hi Yunlei,
On Tue, Mar 02, 2021 at 07:38:50PM +0800, Yunlei He wrote:
> If file enable verity failed, should truncate anything wrote
> past i_size, including cache pages. Move the truncation to
> the end of function, in case of f2fs set xattr failed.
>
> Fixes: 95ae251fe828 ("f2fs: add fs-verity support")
> Cc: <stable@xxxxxxxxxxxxxxx> # v5.4+
> Signed-off-by: Yunlei He <heyunlei@xxxxxxxxxxx>
> ---
> fs/f2fs/verity.c | 11 +++++++----
> 1 file changed, 7 insertions(+), 4 deletions(-)
>
> diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c
> index 054ec852b5ea..610f2a9b4928 100644
> --- a/fs/f2fs/verity.c
> +++ b/fs/f2fs/verity.c
> @@ -169,10 +169,6 @@ static int f2fs_end_enable_verity(struct file *filp, const void *desc,
> err = filemap_write_and_wait(inode->i_mapping);
> }
>
> - /* If we failed, truncate anything we wrote past i_size. */
> - if (desc == NULL || err)
> - f2fs_truncate(inode);
> -
> clear_inode_flag(inode, FI_VERITY_IN_PROGRESS);
>
> if (desc != NULL && !err) {
> @@ -185,6 +181,13 @@ static int f2fs_end_enable_verity(struct file *filp, const void *desc,
> f2fs_mark_inode_dirty_sync(inode, true);
> }
> }
> +
> + /* If we failed, truncate anything we wrote past i_size. */
> + if (desc == NULL || err) {
> + truncate_inode_pages(inode->i_mapping, inode->i_size);
> + f2fs_truncate(inode);
> + }
> +
> return err;
This is better, but we really should properly separate the success path from the
error paths in this function; otherwise it's too hard to understand. Also, the
same bugs need to be fixed in ext4 too, and the commit message could be better.
I went ahead and sent out a new patchset which addresses all this
(https://lkml.kernel.org/linux-f2fs-devel/20210302200420.137977-1-ebiggers@xxxxxxxxxx/T/#u);
can you take a look at that instead? Thanks!
- Eric
