1 files changed, 36 insertions, 12 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index ab8f970b2849..64510898eedd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2928,6 +2928,14 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
+	/*
+	 * Per PMFW team's suggestion, driver needs to handle gfxoff
+	 * and df cstate features disablement for gpu reset(e.g. Mode1Reset)
+	 * scenario. Add the missing df cstate disablement here.
+	 */
+	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW))
+		dev_warn(adev->dev, "Failed to disallow df cstate");
+
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
@@ -3202,6 +3210,15 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
 			return r;
 		}
 		adev->ip_blocks[i].status.hw = true;
+
+		if (adev->in_s0ix && adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) {
+			/* disable gfxoff for IP resume. The gfxoff will be re-enabled in
+			 * amdgpu_device_resume() after IP resume.
+			 */
+			amdgpu_gfx_off_ctrl(adev, false);
+			DRM_DEBUG("will disable gfxoff for re-initializing other blocks\n");
+		}
+
 	}
 
 	return 0;
@@ -4043,15 +4060,18 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
  * at suspend time.
  *
  */
-static void amdgpu_device_evict_resources(struct amdgpu_device *adev)
+static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
 {
+	int ret;
+
 	/* No need to evict vram on APUs for suspend to ram or s2idle */
 	if ((adev->in_s3 || adev->in_s0ix) && (adev->flags & AMD_IS_APU))
-		return;
+		return 0;
 
-	if (amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM))
+	ret = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM);
+	if (ret)
 		DRM_WARN("evicting device resources failed\n");
-
+	return ret;
 }
 
 /*
@@ -4101,7 +4121,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 	if (!adev->in_s0ix)
 		amdgpu_amdkfd_suspend(adev, adev->in_runpm);
 
-	amdgpu_device_evict_resources(adev);
+	r = amdgpu_device_evict_resources(adev);
+	if (r)
+		return r;
 
 	amdgpu_fence_driver_hw_fini(adev);
 
@@ -4177,6 +4199,13 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 	/* Make sure IB tests flushed */
 	flush_delayed_work(&adev->delayed_init_work);
 
+	if (adev->in_s0ix) {
+		/* re-enable gfxoff after IP resume. This re-enables gfxoff after
+		 * it was disabled for IP resume in amdgpu_device_ip_resume_phase2().
+		 */
+		amdgpu_gfx_off_ctrl(adev, true);
+		DRM_DEBUG("will enable gfxoff for the mission mode\n");
+	}
 	if (fbcon)
 		drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, false);
 
@@ -5210,7 +5239,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 	reset_context->job = job;
 	reset_context->hive = hive;
-
 	/*
 	 * Build list of devices to reset.
 	 * In case we are in XGMI hive mode, resort the device list
@@ -5337,11 +5365,8 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 			amdgpu_ras_resume(adev);
 	} else {
 		r = amdgpu_do_asic_reset(device_list_handle, reset_context);
-		if (r && r == -EAGAIN) {
-			set_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context->flags);
-			adev->asic_reset_res = 0;
+		if (r && r == -EAGAIN)
 			goto retry;
-		}
 
 		if (!r && gpu_reset_for_dev_remove)
 			goto recover_end;
@@ -5377,7 +5402,7 @@ skip_hw_reset:
 			drm_sched_start(&ring->sched, !tmp_adev->asic_reset_res);
 		}
 
-		if (adev->enable_mes)
+		if (adev->enable_mes && adev->ip_versions[GC_HWIP][0] != IP_VERSION(11, 0, 3))
 			amdgpu_mes_self_test(tmp_adev);
 
 		if (!drm_drv_uses_atomic_modeset(adev_to_drm(tmp_adev)) && !job_signaled) {
@@ -5777,7 +5802,6 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev)
 	reset_context.reset_req_dev = adev;
 	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
 	set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags);
-	set_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags);
 
 	adev->no_hw_access = true;
 	r = amdgpu_device_pre_asic_reset(adev, &reset_context);