Move malloc_cache_aligned() to malloc.h
PATCH 1/1 Move malloc_cache_aligned() to malloc.h
At present malloc.h is included everywhere since it recently was added to common.h in this commit:
4519668 mtd/nand/ubi: assortment of alignment fixes
This seems wasteful and unnecessary. We have been trying to trim down common.h and put separate functions into separate header files and that change goes in the opposite direction.
Move malloc_cache_aligned() to malloc.h so that this can be avoided.
---
fs/ubifs/super.c | 2 ++ include/common.h | 9 --------- include/malloc.h | 12 ++++++++++++ lib/zlib/zutil.c | 4 +++- 4 files changed, 17 insertions(+), 10 deletions(-)
fs/ubifs/super.c +2 -0
@@ -28,6 +28,8 @@#include <linux/writeback.h> #else +#include <common.h> +#include <malloc.h> #include <linux/compat.h> #include <linux/stat.h> #include <linux/err.h>
include/common.h +0 -9
@@ -1060,15 +1060,6 @@ int cpu_release(int nr, int argc, char * const argv[]);#define DEFINE_CACHE_ALIGN_BUFFER (type, name, size) \ DEFINE_ALIGN_BUFFER(type, name, size, ARCH_DMA_MINALIGN) -#ifndef __ASSEMBLY__ -#include <malloc.h> - -static inline void *malloc_cache_aligned (size_t size) -{ -return memalign (ARCH_DMA_MINALIGN ,ALIGN (size,ARCH_DMA_MINALIGN )); -} -#endif -/* * check_member() - Check the offset of a structure member *
include/malloc.h +12 -0
@@ -953,6 +953,18 @@ extern ulong mem_malloc_start;extern ulong mem_malloc_end;extern ulong mem_malloc_brk; +/* +* ARCH_DMA_MINALIGN is defined by including common.h which should be the +* first include in a U-Boot C file. Protect this declaration for imported +* files which don't follow this convention, to avoid compile errors. +*/ +#ifdef ARCH_DMA_MINALIGN +static inline void *malloc_cache_aligned (size_t size) +{ +return memalign (ARCH_DMA_MINALIGN ,ALIGN (size,ARCH_DMA_MINALIGN )); +} +#endif +void mem_malloc_init (ulong start,ulong size);#ifdef __cplusplus
lib/zlib/zutil.c +3 -1
@@ -43,7 +43,9 @@ void z_error (m)*/ #ifndef MY_ZCALLOC /* Any system without a special alloc function */ -#ifndef __UBOOT__ +#ifdef __UBOOT__ +#include <malloc.h> +#else #ifndef STDC extern voidp malloc OF((uInt size)); extern voidp calloc OF((uInt items, uInt size));
Signature · 1 line
2.5.0.457.gab17608
PATCH 1/1 Move malloc_cache_aligned() to malloc.h
Quoted 4 lines · Simon Glass
At present
malloc.his included everywhere since it recently was added tocommon.hin this commit:
Quoted 8 lines
4519668 mtd/nand/ubi: assortment of alignment fixes
This seems wasteful and unnecessary. We have been trying to trim down
common.hand put separate functions into separate header files and that
change goes in the opposite direction.Move
malloc_cache_aligned()tomalloc.hso that this can be avoided.
Hm, but we've been there before:
PATCH 1/1 Move malloc_cache_aligned() to malloc.h
Quoted 12 lines · Marcel Ziswiler
Quoted 11 lines · Simon Glass
At present
malloc.his included everywhere since it recently was added tocommon.hin this commit:4519668 mtd/nand/ubi: assortment of alignment fixes
This seems wasteful and unnecessary. We have been trying to trim down
common.hand put separate functions into separate header files and that
change goes in the opposite direction.
Quoted 5 lines
Quoted 1 line
Move
malloc_cache_aligned()tomalloc.hso that this can be avoided.Hm, but we've been there before:
Maybe all the cache alignment stuff could go in its own header?
-Scott
PATCH 1/1 Move malloc_cache_aligned() to malloc.h
Hi Scott,
Quoted 13 lines · Scott Wood
Quoted 12 lines · Marcel Ziswiler
Quoted 11 lines · Simon Glass
At present
malloc.his included everywhere since it recently was added tocommon.hin this commit:4519668 mtd/nand/ubi: assortment of alignment fixes
This seems wasteful and unnecessary. We have been trying to trim down
common.hand put separate functions into separate header files and that
change goes in the opposite direction.
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
Move
malloc_cache_aligned()tomalloc.hso that this can be avoided.Hm, but we've been there before:
Maybe all the cache alignment stuff could go in its own header?
Yes I like that idea, it seems like the best solution.
Regards,
Simon