On Apr 1, 2024, at 8:48 PM, Li zeming <zeming@xxxxxxxxxxxx> wrote:
>
> ablocks is assigned first, so it does not need to initialize the
> assignment.
While it is true that "ablocks" is currently set before use,
this is happening a long way away from the variable declaration
and also "ablocks" is used after the "cleanup:" label error case:
cleanup:
if (bh) {
if (buffer_locked(bh))
unlock_buffer(bh);
brelse(bh);
}
if (err) {
/* free all allocated blocks in error case */
for (i = 0; i < depth; i++) {
if (!ablocks[i])
continue;
ext4_free_blocks(handle, inode, NULL, ablocks[i], 1,
EXT4_FREE_BLOCKS_METADATA);
}
}
kfree(ablocks);
So there is definitely a risk that a code change in the future
would introduce hard-to-debug problems, crashes, or even just
spurious static code analysis warnings. My recommendation would
be to keep this 1-cycle local variable initialization in place
rather than spend hours or days trying to debug and fix a crash
here in the future.
Cheers, Andreas
>
> Signed-off-by: Li zeming <zeming@xxxxxxxxxxxx>
> ---
> fs/ext4/extents.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index 4ab96f01a6f31..caace8c3fd3c1 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -1061,7 +1061,7 @@ static int ext4_ext_split(handle_t *handle, struct inode *inode,
> int i = at, k, m, a;
> ext4_fsblk_t newblock, oldblock;
> __le32 border;
> - ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
> + ext4_fsblk_t *ablocks; /* array of allocated blocks */
> gfp_t gfp_flags = GFP_NOFS;
> int err = 0;
> size_t ext_size = 0;
> --
> 2.18.2
>
Cheers, Andreas
Attachment:
signature.asc
Description: Message signed with OpenPGP
