2014-12-22 19:05:19

by Fabian Frédérick

Subject: [PATCH 1/1 linux-next] ocfs2: remove unnecessary sizeof(char)

sizeof(char) is always 1.

Suggested-by: Joe Perches <[email protected]>
Signed-off-by: Fabian Frederick <[email protected]>
---
fs/ocfs2/journal.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index e238bbc..750597c 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -129,8 +129,8 @@ int ocfs2_compute_replay_slots(struct ocfs2_super *osb)
if (osb->replay_map)
return 0;

- replay_map = kzalloc(sizeof(struct ocfs2_replay_map) +
- (osb->max_slots * sizeof(char)), GFP_KERNEL);
+ replay_map = kzalloc(sizeof(struct ocfs2_replay_map) + osb->max_slots,
+ GFP_KERNEL);

if (!replay_map) {
mlog_errno(-ENOMEM);
--
1.9.1

2014-12-22 21:35:05

by Andrew Morton

Subject: Re: [Ocfs2-devel] [PATCH 1/1 linux-next] ocfs2: remove unnecessary sizeof(char)

On Mon, 22 Dec 2014 20:05:09 +0100 Fabian Frederick <[email protected]> wrote:

> sizeof(char) is always 1.
>
> ...
>
> --- a/fs/ocfs2/journal.c
> +++ b/fs/ocfs2/journal.c
> @@ -129,8 +129,8 @@ int ocfs2_compute_replay_slots(struct ocfs2_super *osb)
> if (osb->replay_map)
> return 0;
>
> - replay_map = kzalloc(sizeof(struct ocfs2_replay_map) +
> - (osb->max_slots * sizeof(char)), GFP_KERNEL);
> + replay_map = kzalloc(sizeof(struct ocfs2_replay_map) + osb->max_slots,
> + GFP_KERNEL);
>
> if (!replay_map) {
> mlog_errno(-ENOMEM);

I dunno. The code at present isn't particularly idiomatic, but it has
some documentation value and says "I know what I'm doing".

It would be better if it was

kzalloc(sizeof(struct ocfs2_replay_map) *
sizeof(struct ocfs2_replay_map.rm_replay_slots[0]), ...);

And it would be better if C permitted that ;)

kzalloc(sizeof(struct ocfs2_replay_map) *
sizeof((struct ocfs2_replay_map *)0)->rm_replay_slots[0]), ...);

yuk.

2014-12-23 11:14:50

by Rasmus Villemoes

Subject: Re: [Ocfs2-devel] [PATCH 1/1 linux-next] ocfs2: remove unnecessary sizeof(char)

On Mon, Dec 22 2014, Andrew Morton <[email protected]> wrote:

> On Mon, 22 Dec 2014 20:05:09 +0100 Fabian Frederick <[email protected]> wrote:
>
>>
>> - replay_map = kzalloc(sizeof(struct ocfs2_replay_map) +
>> - (osb->max_slots * sizeof(char)), GFP_KERNEL);
>> + replay_map = kzalloc(sizeof(struct ocfs2_replay_map) + osb->max_slots,
>> + GFP_KERNEL);
>>
>
> I dunno. The code at present isn't particularly idiomatic, but it has
> some documentation value and says "I know what I'm doing".
>
> It would be better if it was
>
> kzalloc(sizeof(struct ocfs2_replay_map) *
> sizeof(struct ocfs2_replay_map.rm_replay_slots[0]), ...);
>
> And it would be better if C permitted that ;)
>
> kzalloc(sizeof(struct ocfs2_replay_map) *
> sizeof((struct ocfs2_replay_map *)0)->rm_replay_slots[0]), ...);
>
> yuk.

Well, the yuckiness could be hidden behind FIELD_SIZEOF. CodingStyle
explicitly mentions that macro. No opinion on whether it's worth it in
this case.

Rasmus