drm/amd/display: Recreate private_obj->state during S3 resume
authorLeo Li <sunpeng.li@amd.com>
Tue, 19 Mar 2019 23:29:16 +0000 (19:29 -0400)
committerAlex Deucher <alexander.deucher@amd.com>
Wed, 10 Apr 2019 18:53:28 +0000 (13:53 -0500)
commit113b7a010872117e90b16a3e411558db5cb5979e
tree6f806d720f725f50ad78e02e84e48192bd8fd21b
parentfbd06ec492fc6e078a62664c5b4a0d74b90285be
drm/amd/display: Recreate private_obj->state during S3 resume

[Why]

When entering S3, amdgpu first calls DRM to cache the current atomic
state, then commit the 'all-disabled' state. This sets dc->current_state
to point at the private atomic object's dm_atomic_state->context, as
any regular atomic commit would.

Afterwards, amdgpu_dm calls dc_set_power_state() with S3 power state.
This invalidates dc->current_state by wiping it to 0, consequently
wiping dm_atomic_state->context.

During resume, the cached atomic state is restored. When getting the
private object however, the dm_atomic_state - containing the wiped
context - is duplicated into the atomic state. This causes DC validation
to fail during atomic check, as necessary function pointers in dc_state
are now NULL.

[How]

Recreate the private object's dm_atomic_state->context during resume,
restoring any static values such as function pointers.

A TODO item is added to move static read-only values out of dc_state -
they shouldn't be there anyways.

Signed-off-by: Leo Li <sunpeng.li@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c