| Message ID | 1339176514-13137-1-git-send-email-marex@denx.de |
|---|---|
| State | Rejected |
| Delegated to: | Tom Rini |
| Headers | show |
Dear Marek Vasut, In message <1339176514-13137-1-git-send-email-marex@denx.de> you wrote: > > -#define kzalloc(size, flags) calloc(size, 1) > +#define kzalloc(size, flags) calloc(1, size) Does this make any practical difference? I mean, are you aware of any problem that gets fixed by this patch, or even any kind of performance degradation? "common/dlmalloc.src" says: calloc(size_t unit, size_t quantity); Returns a pointer to quantity * unit bytes, with all locations set to zero. Multiplication being commutative, I see zero effect in this patch? Best regards, Wolfgang Denk
Dear Wolfgang Denk, > Dear Marek Vasut, > > In message <1339176514-13137-1-git-send-email-marex@denx.de> you wrote: > > -#define kzalloc(size, flags) calloc(size, 1) > > +#define kzalloc(size, flags) calloc(1, size) > > Does this make any practical difference? > > I mean, are you aware of any problem that gets fixed by this patch, or > even any kind of performance degradation? No, I just blindly found it out when I was enraged and was hacking on filesystem code. > "common/dlmalloc.src" says: > > calloc(size_t unit, size_t quantity); > Returns a pointer to quantity * unit bytes, with all locations > set to zero. > > Multiplication being commutative, I see zero effect in this patch? Weeeeell, not in every algebraic system. It has zero effect, it's only about correctness [1]. Will applying this break anything? [1] http://pubs.opengroup.org/onlinepubs/009695399/functions/calloc.html > Best regards, > > Wolfgang Denk Best regards, Marek Vasut
Dear Marek Vasut, In message <201206090148.40933.marex@denx.de> you wrote: > > > Multiplication being commutative, I see zero effect in this patch? > > Weeeeell, not in every algebraic system. It has zero effect, it's only about > correctness [1]. Will applying this break anything? In which way is "a * b" more (or less?) correct than "b * a" ? What is the difference between allocating a contiguous sequence of 4096 single bytes, or allocating a single block of size 4096? Best regards, Wolfgang Denk
Dear Wolfgang Denk, > Dear Marek Vasut, > > In message <201206090148.40933.marex@denx.de> you wrote: > > > Multiplication being commutative, I see zero effect in this patch? > > > > Weeeeell, not in every algebraic system. It has zero effect, it's only > > about correctness [1]. Will applying this break anything? > > In which way is "a * b" more (or less?) correct than "b * a" ? In our natural integer system, it's all perfectly OK. > What is the difference between allocating a contiguous sequence of > 4096 single bytes, or allocating a single block of size 4096? None. > Best regards, > > Wolfgang Denk Best regards, Marek Vasut
diff --git a/include/linux/compat.h b/include/linux/compat.h index 593b07f..9a9d7b0 100644 --- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -18,7 +18,7 @@ #define KERN_DEBUG #define kmalloc(size, flags) malloc(size) -#define kzalloc(size, flags) calloc(size, 1) +#define kzalloc(size, flags) calloc(1, size) #define vmalloc(size) malloc(size) #define kfree(ptr) free(ptr) #define vfree(ptr) free(ptr)
Quote from the manpage: [...] void *calloc(size_t nmemb, size_t size); [...] DESCRIPTION [...] The calloc() function allocates memory for an array of nmemb elements of size bytes each and returns a pointer to the allocated memory. The memory is set to zero. If nmemb or size is 0, then calloc() returns either NULL, or a unique pointer value that can later be successfully passed to free(). Signed-off-by: Marek Vasut <marex@denx.de> Cc: Wolfgang Denk <wd@denx.de> --- include/linux/compat.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)