[PATCH] drm/amdgpu: Fix S3 resume failre.

[harry.wentland@xxxxxxx (Harry Wentland)]

On 2018-07-19 11:25 AM, Christian König wrote:
> Am 19.07.2018 um 17:19 schrieb Andrey Grodzovsky:
>> Problem:
>> FB is still not unpinned during the first run of amdgpu_bo_evict_vram
>> and so it's left for the second run, but during second run the SDMA for
>> moving buffer around already disabled and you have to do
>> it with CPU, but FB is not in visible VRAM and hence the eviction failure
>> leading later to resume failure.
>>
>> Fix:
>> When DAL in use get a pointer to FB from crtc->primary->state rather
>> then from crtc->primary which is not set for DAL since it supports
>> atomic KMS.
> 

Good catch.

> Nice catch, but could we add a helper for unpinning them and just unpin both instead of checking if DC is enabled or not?
> 
> I think that would be a little bit cleaner.
> 
> Christian.
> 
>>
>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107065
>> Fixes e00fb85 drm: Stop updating plane->crtc/fb/old_fb on atomic drivers
>> Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky at amd.com>
>> ---
>> Â  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 ++-
>> Â  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> index 709e4a3..dd9ebf7 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> @@ -2642,7 +2642,8 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
>> Â Â Â Â Â  /* unpin the front buffers and cursors */
>> Â Â Â Â Â  list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
>> Â Â Â Â Â Â Â Â Â  struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
>> -Â Â Â Â Â Â Â  struct drm_framebuffer *fb = crtc->primary->fb;
>> +Â Â Â Â Â Â Â Â  struct drm_framebuffer *fb = amdgpu_device_has_dc_support(adev) ?
>> +Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  crtc->primary->state->fb : crtc->primary->fb;
Is that the only place or are there other places where DC is using fb from crtc, rather than crtc_state?

Harry

>> Â Â Â Â Â Â Â Â Â  struct amdgpu_bo *robj;
>> Â  Â Â Â Â Â Â Â Â Â  if (amdgpu_crtc->cursor_bo) {
>