[Arjan van de Ven - Sat, Sep 22, 2007 at 12:46:59PM -0700]
| On Sat, 22 Sep 2007 12:33:55 +0400
| Cyrill Gorcunov <gorcunov@gmail.com> wrote:
|
| > Hi list,
| >
| > could anyone tell me why there is no official memzero function (or
| > macros) in the kernel.
|
| it doesn't add value.... memset with a constant 0 is just as fast
| (since the compiler knows it's 0) than any wrapper around it, and the
| syntax around it is otherwise the same.
|
It seems I expressed wrong. I'm worried about code duplication. Look
simple grep for memzero tells us that in particular:
...
-- arch/x86_64/boot/compressed/misc.c:110:#define memzero(s, n) memset ((s), 0, (n))
-- init/do_mounts_rd.c:279:#define memzero(s, n) memset ((s), 0, (n))
-- init/initramfs.c:377:#define memzero(s, n) memset ((s), 0, (n))
-- lib/inflate.c:331: memzero(stk->c, sizeof(stk->c));
...
So instead of several 'define' that are the _same_ maybe better just use
_single common_ define? That's all I wanna ask. (Btw, it seems ARM has
a special case for memzero ;)
|
| > As I see a lot of kernel parts calls for it
| > (defying own macros as alias to memset). Maybe there is a special
| > reason not to do so? Actually my suggestion is to define _one_
| > general macros for this.
|
| my suggestion is to nuke all the macros and just use memset().
|
Quite clear, thanks. So if that is OK - I'm shutting up ;)
Cyrill
-
