[PATCH 32/48] drm/amdgpu/vcn: use per instance callbacks for idle work handler

Alex Deucher alexander.deucher at amd.com
Tue Dec 10 22:53:21 UTC 2024


Use the vcn instance power gating callbacks rather than
the IP powergating callback.  This limits power gating to
only the instance in use rather than all of the instances.

Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 27 ++++++++++++-------------
 1 file changed, 13 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 750cee0a8f184..04a95365b7095 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -424,8 +424,7 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
 	fences += fence[i];
 
 	if (!fences && !atomic_read(&vcn_inst->total_submission_cnt)) {
-		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
-						       AMD_PG_STATE_GATE);
+		vcn_inst->set_pg_state(vcn_inst, AMD_PG_STATE_GATE);
 		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
 						    false);
 		if (r)
@@ -438,45 +437,45 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
 void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *vcn_inst = &adev->vcn.inst[ring->me];
 	int r = 0;
 
-	atomic_inc(&adev->vcn.inst[ring->me].total_submission_cnt);
+	atomic_inc(&vcn_inst->total_submission_cnt);
 
-	if (!cancel_delayed_work_sync(&adev->vcn.inst[ring->me].idle_work)) {
+	if (!cancel_delayed_work_sync(&vcn_inst->idle_work)) {
 		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
 				true);
 		if (r)
 			dev_warn(adev->dev, "(%d) failed to switch to video power profile mode\n", r);
 	}
 
-	mutex_lock(&adev->vcn.inst[ring->me].vcn_pg_lock);
-	amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
-					       AMD_PG_STATE_UNGATE);
+	mutex_lock(&vcn_inst->vcn_pg_lock);
+	vcn_inst->set_pg_state(vcn_inst, AMD_PG_STATE_UNGATE);
 
 	/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
-	    !adev->vcn.inst[ring->me].using_unified_queue) {
+	    !vcn_inst->using_unified_queue) {
 		struct dpg_pause_state new_state;
 
 		if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) {
-			atomic_inc(&adev->vcn.inst[ring->me].dpg_enc_submission_cnt);
+			atomic_inc(&vcn_inst->dpg_enc_submission_cnt);
 			new_state.fw_based = VCN_DPG_STATE__PAUSE;
 		} else {
 			unsigned int fences = 0;
 			unsigned int i;
 
-			for (i = 0; i < adev->vcn.inst[ring->me].num_enc_rings; ++i)
-				fences += amdgpu_fence_count_emitted(&adev->vcn.inst[ring->me].ring_enc[i]);
+			for (i = 0; i < vcn_inst->num_enc_rings; ++i)
+				fences += amdgpu_fence_count_emitted(&vcn_inst->ring_enc[i]);
 
-			if (fences || atomic_read(&adev->vcn.inst[ring->me].dpg_enc_submission_cnt))
+			if (fences || atomic_read(&vcn_inst->dpg_enc_submission_cnt))
 				new_state.fw_based = VCN_DPG_STATE__PAUSE;
 			else
 				new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
 		}
 
-		adev->vcn.inst[ring->me].pause_dpg_mode(&adev->vcn.inst[ring->me], &new_state);
+		vcn_inst->pause_dpg_mode(vcn_inst, &new_state);
 	}
-	mutex_unlock(&adev->vcn.inst[ring->me].vcn_pg_lock);
+	mutex_unlock(&vcn_inst->vcn_pg_lock);
 }
 
 void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring)
-- 
2.47.1



More information about the amd-gfx mailing list