[PATCH 1/2] drm/ttm: remove unused placement flags

zhoucm1 david1.zhou at amd.com
Fri Sep 9 06:41:16 UTC 2016



On 2016年09月08日 21:41, Christian König wrote:
> From: Christian König <christian.koenig at amd.com>
>
> Either never used or not used in quite a while.
No, I remember Flora's Direct GMA is using them like GDS use PRIV0-2. 
And you cannot make sure there isn't no one using them in other closed 
projects, right?
If you removed now, that obviously will break her implementation and 
brings  her many troubles.


Regards,
David Zhou
>
> Signed-off-by: Christian König <christian.koenig at amd.com>
> ---
>   drivers/gpu/drm/ttm/ttm_bo.c    |  2 +-
>   include/drm/ttm/ttm_placement.h | 19 -------------------
>   2 files changed, 1 insertion(+), 20 deletions(-)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
> index bc37f02..4d2e8f2 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
> @@ -59,7 +59,7 @@ static inline int ttm_mem_type_from_place(const struct ttm_place *place,
>   {
>   	int i;
>   
> -	for (i = 0; i <= TTM_PL_PRIV5; i++)
> +	for (i = 0; i <= TTM_PL_PRIV2; i++)
>   		if (place->flags & (1 << i)) {
>   			*mem_type = i;
>   			return 0;
> diff --git a/include/drm/ttm/ttm_placement.h b/include/drm/ttm/ttm_placement.h
> index 8ed44f9..20219d9 100644
> --- a/include/drm/ttm/ttm_placement.h
> +++ b/include/drm/ttm/ttm_placement.h
> @@ -40,10 +40,6 @@
>   #define TTM_PL_PRIV0            3
>   #define TTM_PL_PRIV1            4
>   #define TTM_PL_PRIV2            5
> -#define TTM_PL_PRIV3            6
> -#define TTM_PL_PRIV4            7
> -#define TTM_PL_PRIV5            8
> -#define TTM_PL_SWAPPED          15
>   
>   #define TTM_PL_FLAG_SYSTEM      (1 << TTM_PL_SYSTEM)
>   #define TTM_PL_FLAG_TT          (1 << TTM_PL_TT)
> @@ -51,10 +47,6 @@
>   #define TTM_PL_FLAG_PRIV0       (1 << TTM_PL_PRIV0)
>   #define TTM_PL_FLAG_PRIV1       (1 << TTM_PL_PRIV1)
>   #define TTM_PL_FLAG_PRIV2       (1 << TTM_PL_PRIV2)
> -#define TTM_PL_FLAG_PRIV3       (1 << TTM_PL_PRIV3)
> -#define TTM_PL_FLAG_PRIV4       (1 << TTM_PL_PRIV4)
> -#define TTM_PL_FLAG_PRIV5       (1 << TTM_PL_PRIV5)
> -#define TTM_PL_FLAG_SWAPPED     (1 << TTM_PL_SWAPPED)
>   #define TTM_PL_MASK_MEM         0x0000FFFF
>   
>   /*
> @@ -72,7 +64,6 @@
>   #define TTM_PL_FLAG_CACHED      (1 << 16)
>   #define TTM_PL_FLAG_UNCACHED    (1 << 17)
>   #define TTM_PL_FLAG_WC          (1 << 18)
> -#define TTM_PL_FLAG_SHARED      (1 << 20)
>   #define TTM_PL_FLAG_NO_EVICT    (1 << 21)
>   #define TTM_PL_FLAG_TOPDOWN     (1 << 22)
>   
> @@ -82,14 +73,4 @@
>   
>   #define TTM_PL_MASK_MEMTYPE     (TTM_PL_MASK_MEM | TTM_PL_MASK_CACHING)
>   
> -/*
> - * Access flags to be used for CPU- and GPU- mappings.
> - * The idea is that the TTM synchronization mechanism will
> - * allow concurrent READ access and exclusive write access.
> - * Currently GPU- and CPU accesses are exclusive.
> - */
> -
> -#define TTM_ACCESS_READ         (1 << 0)
> -#define TTM_ACCESS_WRITE        (1 << 1)
> -
>   #endif



More information about the amd-gfx mailing list