On 10/24/2017 7:48 PM, Andrey Grodzovsky wrote:
>
>
> On 10/24/2017 09:51 AM, S, Shirish wrote:
>> From: Shirish S <shirish.s at amd.com>
>>
>> On some systems amdgpu_bo_gpu_offset seems to be
>> called before the BO is pinned, leading to page faults
>> as the GPU address is still subject to change.
>>
>> This patch fixes the above issue.
>>
>> Signed-off-by: Shirish S <shirish.s at amd.com>
>> ---
>> Â drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 7 +++++--
>> Â 1 file changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
>> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
>> index 974ee97..b251b7e 100644
>> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
>> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
>> @@ -1786,8 +1786,11 @@ static int get_fb_info(const struct
>> amdgpu_framebuffer *amdgpu_fb,
>> Â Â Â Â Â Â Â Â Â return r;
>> Â Â Â Â Â }
>> Â -Â Â Â if (fb_location)
>> -Â Â Â Â Â Â Â *fb_location = amdgpu_bo_gpu_offset(rbo);
>> +Â Â Â r = amdgpu_bo_pin(rbo, AMDGPU_GEM_DOMAIN_VRAM, fb_location);
>> +Â Â Â if (unlikely(r != 0)) {
>> +Â Â Â Â Â Â Â amdgpu_bo_unreserve(rbo);
>> +Â Â Â Â Â Â Â return -EINVAL;
>> +Â Â Â }
>
> Where the unpin would happen then ? Today DAL calls pin from
> dm_plane_helper_prepare_fb while the unpin from
> dm_plane_helper_cleanup_fb.
>
Good point, am trying to address the issue Michel is facing, if this
patch works, we can think of unpinning the bo appropriately.
Ideally it should be at the end of amdgpu_dm_atomic_check(). Can you
think of any better place?
Regards,
Shirish S
> Thanks,
> Andrey
>
>> Â Â Â Â Â Â if (tiling_flags)
>> Â Â Â Â Â Â Â Â Â amdgpu_bo_get_tiling_flags(rbo, tiling_flags);
>
