On Monday, October 12, 2015 at 07:29:43 AM, Thomas Chou wrote:
Convert dma_alloc_coherent to use memalign.
Signed-off-by: Thomas Chou thomas@wytron.com.tw
v2 use memalign. v3 check memalign() return for out of memory. v4 use malloc_cache_aligned().
arch/nios2/include/asm/dma-mapping.h | 24 +++++++++++------------- 1 file changed, 11 insertions(+), 13 deletions(-)
diff --git a/arch/nios2/include/asm/dma-mapping.h b/arch/nios2/include/asm/dma-mapping.h index 1350e3b..bfaf662 100644 --- a/arch/nios2/include/asm/dma-mapping.h +++ b/arch/nios2/include/asm/dma-mapping.h @@ -1,23 +1,21 @@ #ifndef __ASM_NIOS2_DMA_MAPPING_H #define __ASM_NIOS2_DMA_MAPPING_H
-/* dma_alloc_coherent() return cache-line aligned allocation which is mapped +#include <memalign.h> +#include <asm/io.h>
+/*
- dma_alloc_coherent() return cache-line aligned allocation which is
mapped * to uncached io region.
- IO_REGION_BASE should be defined in board config header file
*/
- 0x80000000 for nommu, 0xe0000000 for mmu
static inline void *dma_alloc_coherent(size_t len, unsigned long *handle) {
- void *addr = malloc(len + CONFIG_SYS_DCACHELINE_SIZE);
- if (!addr)
return 0;- flush_dcache((unsigned long)addr, len + CONFIG_SYS_DCACHELINE_SIZE);
- *handle = ((unsigned long)addr +
(CONFIG_SYS_DCACHELINE_SIZE - 1)) &~(CONFIG_SYS_DCACHELINE_SIZE - 1) & ~(IO_REGION_BASE);- return (void *)(*handle | IO_REGION_BASE);
- *handle = (unsigned long)malloc_cache_aligned(len);
- if (!*handle)
return NULL;- flush_dcache_range(*handle, *handle + len);
Wouldn't invalidate_dcache_range() be enough here ? You don't care about the data in the newly allocated area at this point I guess -- either you fill them in and then flush, for DMA from CPU to device OR you receive data from device to CPU and then you invalidate this buffer again.
- return ioremap(*handle, len);
}
#endif /* __ASM_NIOS2_DMA_MAPPING_H */
Best regards, Marek Vasut