[PATCH 2/4] drm/ttm: zero allocation only for CPU domain

[zhoucm1]

Thanks for reply, this patch is mainly for performance reason. since TTM 
is to clear page for all non-fixed memory (GTT and CPU domains), which 
takes much more CPU overhead. The patch is to remove 'clear page' for 
GTT domain.
btw: Micheal has rejected this patch.

Regards,
David Zhou (Chunming)

On 2016年01月05日 22:32, Lucas Stach wrote:
> Hi David,
>
> why is this patch 2/4 if it's just a single patch?
>
> Am Dienstag, den 05.01.2016, 11:55 +0800 schrieb david1.zhou at amd.com:
>> From: Chunming Zhou <David1.Zhou at amd.com>
>>
>> Change-Id: I653d61e32faaa11c39bb978bf90301513a0c35b8
> This change-id is just pure noise when you submit something upstream.
>
> Please add a proper change log on why the change below is
> necessary/helpful.
>
> Regards,
> Lucas
>
>> Signed-off-by: Chunming Zhou <David1.Zhou at amd.com>
>> ---
>>   drivers/gpu/drm/ttm/ttm_bo.c | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c
>> b/drivers/gpu/drm/ttm/ttm_bo.c
>> index 745e996..0633b65 100644
>> --- a/drivers/gpu/drm/ttm/ttm_bo.c
>> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
>> @@ -301,6 +301,7 @@ static int ttm_bo_handle_move_mem(struct
>> ttm_buffer_object *bo,
>>      if (!(new_man->flags & TTM_MEMTYPE_FLAG_FIXED)) {
>>              if (bo->ttm == NULL) {
>>                      bool zero = !(old_man->flags &
>> TTM_MEMTYPE_FLAG_FIXED);
>> +                    zero = (mem->mem_type == TTM_PL_SYSTEM) ?
>> zero : false;
>>                      ret = ttm_bo_add_ttm(bo, zero);
>>                      if (ret)
>>                              goto out_err;