[PATCH v2] drm/amdgpu/gfx9: fix ngg enablement to clear gds reserved memory (v2)

[ckoenig.leichtzumerken@xxxxxxxxx (Christian König)]

Can't judge the other changes, but replacing the SOC15 macro with a 
register read looks like a valid fix to me.

For now the patch is Acked-by: Christian König <christian.koenig at amd.com>.

Regards,
Christian.

Am 11.01.2018 um 03:57 schrieb Zhang, Jerry (Junwei):
> On 01/11/2018 10:50 AM, Chunming Zhou wrote:
>> Could I how to verify this is valid fix?
>
> For now, check if there is no vm fault or any other side effect when 
> loading amdgpu with ngg=1.
>
> Later development will be implemented with OGL team together.
>
> Jerry
>
>>
>> Regards,
>>
>> David Zhou
>>
>>
>> On 2018å¹´01æ??11æ?¥ 10:30, Junwei Zhang wrote:
>>> v2: fix register access
>>>
>>> Signed-off-by: Junwei Zhang <Jerry.Zhang at amd.com>
>>> ---
>>> Â  drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 9 +++++----
>>> Â  1 file changed, 5 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
>>> b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
>>> index ae976e3..6db3645 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
>>> @@ -1068,8 +1068,8 @@ static int gfx_v9_0_ngg_init(struct 
>>> amdgpu_device *adev)
>>> Â Â Â Â Â  adev->gfx.ngg.gds_reserve_size = ALIGN(5 * 4, 0x40);
>>> Â Â Â Â Â  adev->gds.mem.total_size -= adev->gfx.ngg.gds_reserve_size;
>>> Â Â Â Â Â  adev->gds.mem.gfx_partition_size -= 
>>> adev->gfx.ngg.gds_reserve_size;
>>> -Â Â Â  adev->gfx.ngg.gds_reserve_addr = SOC15_REG_OFFSET(GC, 0, 
>>> mmGDS_VMID0_BASE);
>>> -Â Â Â  adev->gfx.ngg.gds_reserve_addr += 
>>> adev->gds.mem.gfx_partition_size;
>>> +Â Â Â  adev->gfx.ngg.gds_reserve_addr = RREG32_SOC15(GC, 0, 
>>> mmGDS_VMID0_BASE);
>>> +Â Â Â  adev->gfx.ngg.gds_reserve_addr += RREG32_SOC15(GC, 0, 
>>> mmGDS_VMID0_SIZE);
>>> Â Â Â Â Â  /* Primitive Buffer */
>>> Â Â Â Â Â  r = gfx_v9_0_ngg_create_buf(adev, &adev->gfx.ngg.buf[NGG_PRIM],
>>> @@ -1181,13 +1181,14 @@ static int gfx_v9_0_ngg_en(struct 
>>> amdgpu_device *adev)
>>> Â Â Â Â Â  amdgpu_ring_write(ring, PACKET3(PACKET3_DMA_DATA, 5));
>>> Â Â Â Â Â  amdgpu_ring_write(ring, (PACKET3_DMA_DATA_CP_SYNC |
>>> +Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  PACKET3_DMA_DATA_DST_SEL(1) |
>>> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  PACKET3_DMA_DATA_SRC_SEL(2)));
>>> Â Â Â Â Â  amdgpu_ring_write(ring, 0);
>>> Â Â Â Â Â  amdgpu_ring_write(ring, 0);
>>> Â Â Â Â Â  amdgpu_ring_write(ring, adev->gfx.ngg.gds_reserve_addr);
>>> Â Â Â Â Â  amdgpu_ring_write(ring, 0);
>>> -Â Â Â  amdgpu_ring_write(ring, adev->gfx.ngg.gds_reserve_size);
>>> -
>>> +Â Â Â  amdgpu_ring_write(ring, PACKET3_DMA_DATA_CMD_RAW_WAIT |
>>> +Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  adev->gfx.ngg.gds_reserve_size);
>>> Â Â Â Â Â  gfx_v9_0_write_data_to_reg(ring, 0, false,
>>> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  SOC15_REG_OFFSET(GC, 0, mmGDS_VMID0_SIZE), 0);
>>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx