[PATCH 27/57] drm/amdgpu: Add vega20 soc init sequence on emulator (v3)

Grazvydas Ignotas notasas at gmail.com
Wed May 16 12:23:06 UTC 2018


On Wed, May 16, 2018 at 2:11 PM, Grazvydas Ignotas <notasas at gmail.com> wrote:
> On Tue, May 15, 2018 at 5:59 PM, Alex Deucher <alexdeucher at gmail.com> wrote:
>> From: Shaoyun Liu <Shaoyun.Liu at amd.com>
>>
>> v2: cleanups (Alex)
>> v3: make it vega20 only (Alex)
>>
>> Signed-off-by: Shaoyun Liu <Shaoyun.Liu at amd.com>
>> Acked-by: Alex Deucher <alexander.deucher at amd.com>
>> Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
>> ---
>>  drivers/gpu/drm/amd/amdgpu/emu_soc.c | 10091 +++++++++++++++++++++++++++++++++
>>  1 file changed, 10091 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/emu_soc.c b/drivers/gpu/drm/amd/amdgpu/emu_soc.c
>> index d72c25c1b987..91f00fbe550a 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/emu_soc.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/emu_soc.c
>> @@ -26,8 +26,10099 @@
>>  #include "soc15_common.h"
>>  #include "soc15_hw_ip.h"
>>
>> +static void wreg32_idx_byteoffset(struct amdgpu_device *adev, u32 offset, u32 value) {
>> +
>> +       static u32 maxoffset = 0;
>> +       static int count = 0;
>> +
>> +       WREG32(0xc, offset);
>> +       RREG32(0xc);
>> +       WREG32(0xd, value);
>> +       RREG32(0xd);
>> +
>> +       if (offset > maxoffset)
>> +               maxoffset = offset;
>> +
>> +       count++;
>> +       if (count % 100 == 0) {
>> +               DRM_INFO("%5d registers written, max offset %08x\n", count, maxoffset);
>> +               msleep(1);
>> +       }
>> +
>> +}
>> +
>> +static void vg20_lsd_soc_init_with_umc(struct amdgpu_device *adev)
>> +{
>> +
>> +               wreg32_idx_byteoffset(adev, 0x10131800, 0x40a40);
>> +               wreg32_idx_byteoffset(adev, 0x10141010, 0x10000000);
>> +               wreg32_idx_byteoffset(adev, 0x10134008, 0xfa042021);
>> +               wreg32_idx_byteoffset(adev, 0x387C, 0x3);
>
> Indentation (double tabs).
> Could this be done in a similar way as
> amdgpu_device_program_register_sequence() instead of thousands of
> function calls?

Actually this will bloat the amdgpu module for everyone, while almost
nobody needs this, perhaps hide this under Kconfig option at least?

GraÅžvydas


More information about the amd-gfx mailing list