[amd-gfx] [PATCH 2/3] drm/amdgpu/gfx7: set USER_SHADER_ARRAY_CONFIG based on disable_cu parameter

Nicolai Hähnle nhaehnle at gmail.com
Fri Jun 17 17:32:08 UTC 2016


On 17.06.2016 17:51, Alex Deucher wrote:
> On Fri, Jun 17, 2016 at 9:17 AM, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
>> From: Nicolai Hähnle <nicolai.haehnle at amd.com>
>>
>> Signed-off-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c | 17 +++++++++++++++++
>>   1 file changed, 17 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
>> index 8c6ad1e..b2dce60 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
>> @@ -3867,6 +3867,17 @@ static void gfx_v7_0_enable_gfx_cgpg(struct amdgpu_device *adev,
>>          }
>>   }
>>
>> +static void gfx_v7_0_set_user_cu_inactive_bitmap(struct amdgpu_device *adev,
>> +                                                u32 bitmap)
>> +{
>> +       u32 data;
>> +
>
> Maybe bail early if there is no override?  Saves a reg write and
> doesn't overwrite the user value if someone else (e.g., vbios) set it
> previously.  Same with patch 3.

Makes sense, I'm sending a v2.

Nicolai

>
> Alex
>
>> +       data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
>> +       data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
>> +
>> +       WREG32(mmGC_USER_SHADER_ARRAY_CONFIG, data);
>> +}
>> +
>>   static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev)
>>   {
>>          u32 data, mask;
>> @@ -5032,9 +5043,12 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
>>          int i, j, k, counter, active_cu_number = 0;
>>          u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0;
>>          struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info;
>> +       unsigned disable_masks[4 * 2];
>>
>>          memset(cu_info, 0, sizeof(*cu_info));
>>
>> +       amdgpu_gfx_parse_disable_cu(disable_masks, 4, 2);
>> +
>>          mutex_lock(&adev->grbm_idx_mutex);
>>          for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
>>                  for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
>> @@ -5042,6 +5056,9 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
>>                          ao_bitmap = 0;
>>                          counter = 0;
>>                          gfx_v7_0_select_se_sh(adev, i, j);
>> +                       if (i < 4 && j < 2)
>> +                               gfx_v7_0_set_user_cu_inactive_bitmap(
>> +                                       adev, disable_masks[i * 2 + j]);
>>                          bitmap = gfx_v7_0_get_cu_active_bitmap(adev);
>>                          cu_info->bitmap[i][j] = bitmap;
>>
>> --
>> 2.7.4
>>
>> _______________________________________________
>> amd-gfx mailing list
>> amd-gfx at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/amd-gfx


More information about the amd-gfx mailing list