1 files changed, 510 insertions, 363 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 6e5284e6c028..1e4dd09a5072 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -25,6 +25,7 @@
  *          Alex Deucher
  *          Jerome Glisse
  */
+#include <linux/power_supply.h>
 #include <linux/kthread.h>
 #include <linux/console.h>
 #include <linux/slab.h>
@@ -61,6 +62,8 @@
 MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/raven_gpu_info.bin");
+MODULE_FIRMWARE("amdgpu/picasso_gpu_info.bin");
+MODULE_FIRMWARE("amdgpu/raven2_gpu_info.bin");
 
 #define AMDGPU_RESUME_MS		2000
 
@@ -651,73 +654,6 @@ void amdgpu_device_wb_free(struct amdgpu_device *adev, u32 wb)
 }
 
 /**
- * amdgpu_device_vram_location - try to find VRAM location
- *
- * @adev: amdgpu device structure holding all necessary informations
- * @mc: memory controller structure holding memory informations
- * @base: base address at which to put VRAM
- *
- * Function will try to place VRAM at base address provided
- * as parameter.
- */
-void amdgpu_device_vram_location(struct amdgpu_device *adev,
-				 struct amdgpu_gmc *mc, u64 base)
-{
-	uint64_t limit = (uint64_t)amdgpu_vram_limit << 20;
-
-	mc->vram_start = base;
-	mc->vram_end = mc->vram_start + mc->mc_vram_size - 1;
-	if (limit && limit < mc->real_vram_size)
-		mc->real_vram_size = limit;
-	dev_info(adev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n",
-			mc->mc_vram_size >> 20, mc->vram_start,
-			mc->vram_end, mc->real_vram_size >> 20);
-}
-
-/**
- * amdgpu_device_gart_location - try to find GTT location
- *
- * @adev: amdgpu device structure holding all necessary informations
- * @mc: memory controller structure holding memory informations
- *
- * Function will place try to place GTT before or after VRAM.
- *
- * If GTT size is bigger than space left then we ajust GTT size.
- * Thus function will never fails.
- *
- * FIXME: when reducing GTT size align new size on power of 2.
- */
-void amdgpu_device_gart_location(struct amdgpu_device *adev,
-				 struct amdgpu_gmc *mc)
-{
-	u64 size_af, size_bf;
-
-	mc->gart_size += adev->pm.smu_prv_buffer_size;
-
-	size_af = adev->gmc.mc_mask - mc->vram_end;
-	size_bf = mc->vram_start;
-	if (size_bf > size_af) {
-		if (mc->gart_size > size_bf) {
-			dev_warn(adev->dev, "limiting GTT\n");
-			mc->gart_size = size_bf;
-		}
-		mc->gart_start = 0;
-	} else {
-		if (mc->gart_size > size_af) {
-			dev_warn(adev->dev, "limiting GTT\n");
-			mc->gart_size = size_af;
-		}
-		/* VCE doesn't like it when BOs cross a 4GB segment, so align
-		 * the GART base on a 4GB boundary as well.
-		 */
-		mc->gart_start = ALIGN(mc->vram_end + 1, 0x100000000ULL);
-	}
-	mc->gart_end = mc->gart_start + mc->gart_size - 1;
-	dev_info(adev->dev, "GTT: %lluM 0x%016llX - 0x%016llX\n",
-			mc->gart_size >> 20, mc->gart_start, mc->gart_end);
-}
-
-/**
  * amdgpu_device_resize_fb_bar - try to resize FB BAR
  *
  * @adev: amdgpu_device pointer
@@ -1077,7 +1013,7 @@ static const struct vga_switcheroo_client_ops amdgpu_switcheroo_ops = {
 /**
  * amdgpu_device_ip_set_clockgating_state - set the CG state
  *
- * @adev: amdgpu_device pointer
+ * @dev: amdgpu_device pointer
  * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.)
  * @state: clockgating state (gate or ungate)
  *
@@ -1111,7 +1047,7 @@ int amdgpu_device_ip_set_clockgating_state(void *dev,
 /**
  * amdgpu_device_ip_set_powergating_state - set the PG state
  *
- * @adev: amdgpu_device pointer
+ * @dev: amdgpu_device pointer
  * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.)
  * @state: powergating state (gate or ungate)
  *
@@ -1222,7 +1158,7 @@ bool amdgpu_device_ip_is_idle(struct amdgpu_device *adev,
  * amdgpu_device_ip_get_ip_block - get a hw IP pointer
  *
  * @adev: amdgpu_device pointer
- * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.)
+ * @type: Type of hardware IP (SMU, GFX, UVD, etc.)
  *
  * Returns a pointer to the hardware IP block structure
  * if it exists for the asic, otherwise NULL.
@@ -1398,7 +1334,12 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
 		chip_name = "vega12";
 		break;
 	case CHIP_RAVEN:
-		chip_name = "raven";
+		if (adev->rev_id >= 8)
+			chip_name = "raven2";
+		else if (adev->pdev->device == 0x15d8)
+			chip_name = "picasso";
+		else
+			chip_name = "raven";
 		break;
 	}
 
@@ -1552,6 +1493,8 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 	}
 
 	adev->powerplay.pp_feature = amdgpu_pp_feature_mask;
+	if (amdgpu_sriov_vf(adev))
+		adev->powerplay.pp_feature &= ~PP_GFXOFF_MASK;
 
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if ((amdgpu_ip_block_mask & (1 << i)) == 0) {
@@ -1582,6 +1525,92 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int amdgpu_device_ip_hw_init_phase1(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_blocks[i].status.sw)
+			continue;
+		if (adev->ip_blocks[i].status.hw)
+			continue;
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) {
+			r = adev->ip_blocks[i].version->funcs->hw_init(adev);
+			if (r) {
+				DRM_ERROR("hw_init of IP block <%s> failed %d\n",
+					  adev->ip_blocks[i].version->funcs->name, r);
+				return r;
+			}
+			adev->ip_blocks[i].status.hw = true;
+		}
+	}
+
+	return 0;
+}
+
+static int amdgpu_device_ip_hw_init_phase2(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_blocks[i].status.sw)
+			continue;
+		if (adev->ip_blocks[i].status.hw)
+			continue;
+		r = adev->ip_blocks[i].version->funcs->hw_init(adev);
+		if (r) {
+			DRM_ERROR("hw_init of IP block <%s> failed %d\n",
+				  adev->ip_blocks[i].version->funcs->name, r);
+			return r;
+		}
+		adev->ip_blocks[i].status.hw = true;
+	}
+
+	return 0;
+}
+
+static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
+{
+	int r = 0;
+	int i;
+
+	if (adev->asic_type >= CHIP_VEGA10) {
+		for (i = 0; i < adev->num_ip_blocks; i++) {
+			if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP) {
+				if (adev->in_gpu_reset || adev->in_suspend) {
+					if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset)
+						break; /* sriov gpu reset, psp need to do hw_init before IH because of hw limit */
+					r = adev->ip_blocks[i].version->funcs->resume(adev);
+					if (r) {
+						DRM_ERROR("resume of IP block <%s> failed %d\n",
+							  adev->ip_blocks[i].version->funcs->name, r);
+						return r;
+					}
+				} else {
+					r = adev->ip_blocks[i].version->funcs->hw_init(adev);
+					if (r) {
+						DRM_ERROR("hw_init of IP block <%s> failed %d\n",
+						  adev->ip_blocks[i].version->funcs->name, r);
+						return r;
+					}
+				}
+				adev->ip_blocks[i].status.hw = true;
+			}
+		}
+	}
+
+	if (adev->powerplay.pp_funcs->load_firmware) {
+		r = adev->powerplay.pp_funcs->load_firmware(adev->powerplay.pp_handle);
+		if (r) {
+			pr_err("firmware loading failed\n");
+			return r;
+		}
+	}
+
+	return 0;
+}
+
 /**
  * amdgpu_device_ip_init - run init for hardware IPs
  *
@@ -1638,20 +1667,23 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 		}
 	}
 
-	for (i = 0; i < adev->num_ip_blocks; i++) {
-		if (!adev->ip_blocks[i].status.sw)
-			continue;
-		if (adev->ip_blocks[i].status.hw)
-			continue;
-		r = adev->ip_blocks[i].version->funcs->hw_init((void *)adev);
-		if (r) {
-			DRM_ERROR("hw_init of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
-			return r;
-		}
-		adev->ip_blocks[i].status.hw = true;
-	}
+	r = amdgpu_ucode_create_bo(adev); /* create ucode bo when sw_init complete*/
+	if (r)
+		return r;
+
+	r = amdgpu_device_ip_hw_init_phase1(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_device_fw_loading(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_device_ip_hw_init_phase2(adev);
+	if (r)
+		return r;
 
+	amdgpu_xgmi_add_device(adev);
 	amdgpu_amdkfd_device_init(adev);
 
 	if (amdgpu_sriov_vf(adev))
@@ -1691,29 +1723,28 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev)
 }
 
 /**
- * amdgpu_device_ip_late_set_cg_state - late init for clockgating
+ * amdgpu_device_set_cg_state - set clockgating for amdgpu device
  *
  * @adev: amdgpu_device pointer
  *
- * Late initialization pass enabling clockgating for hardware IPs.
  * The list of all the hardware IPs that make up the asic is walked and the
- * set_clockgating_state callbacks are run.  This stage is run late
- * in the init process.
+ * set_clockgating_state callbacks are run.
+ * Late initialization pass enabling clockgating for hardware IPs.
+ * Fini or suspend, pass disabling clockgating for hardware IPs.
  * Returns 0 on success, negative error code on failure.
  */
-static int amdgpu_device_ip_late_set_cg_state(struct amdgpu_device *adev)
+
+static int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
+						enum amd_clockgating_state state)
 {
-	int i = 0, r;
+	int i, j, r;
 
 	if (amdgpu_emu_mode == 1)
 		return 0;
 
-	r = amdgpu_ib_ring_tests(adev);
-	if (r)
-		DRM_ERROR("ib ring test failed (%d).\n", r);
-
-	for (i = 0; i < adev->num_ip_blocks; i++) {
-		if (!adev->ip_blocks[i].status.valid)
+	for (j = 0; j < adev->num_ip_blocks; j++) {
+		i = state == AMD_CG_STATE_GATE ? j : adev->num_ip_blocks - j - 1;
+		if (!adev->ip_blocks[i].status.late_initialized)
 			continue;
 		/* skip CG for VCE/UVD, it's handled specially */
 		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
@@ -1722,7 +1753,7 @@ static int amdgpu_device_ip_late_set_cg_state(struct amdgpu_device *adev)
 		    adev->ip_blocks[i].version->funcs->set_clockgating_state) {
 			/* enable clockgating to save power */
 			r = adev->ip_blocks[i].version->funcs->set_clockgating_state((void *)adev,
-										     AMD_CG_STATE_GATE);
+										     state);
 			if (r) {
 				DRM_ERROR("set_clockgating_state(gate) of IP block <%s> failed %d\n",
 					  adev->ip_blocks[i].version->funcs->name, r);
@@ -1731,17 +1762,35 @@ static int amdgpu_device_ip_late_set_cg_state(struct amdgpu_device *adev)
 		}
 	}
 
-	if (adev->powerplay.pp_feature & PP_GFXOFF_MASK) {
-		/* enable gfx powergating */
-		amdgpu_device_ip_set_powergating_state(adev,
-						       AMD_IP_BLOCK_TYPE_GFX,
-						       AMD_PG_STATE_GATE);
-		/* enable gfxoff */
-		amdgpu_device_ip_set_powergating_state(adev,
-						       AMD_IP_BLOCK_TYPE_SMC,
-						       AMD_PG_STATE_GATE);
-	}
+	return 0;
+}
+
+static int amdgpu_device_set_pg_state(struct amdgpu_device *adev, enum amd_powergating_state state)
+{
+	int i, j, r;
+
+	if (amdgpu_emu_mode == 1)
+		return 0;
 
+	for (j = 0; j < adev->num_ip_blocks; j++) {
+		i = state == AMD_PG_STATE_GATE ? j : adev->num_ip_blocks - j - 1;
+		if (!adev->ip_blocks[i].status.late_initialized)
+			continue;
+		/* skip CG for VCE/UVD, it's handled specially */
+		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
+		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCE &&
+		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCN &&
+		    adev->ip_blocks[i].version->funcs->set_powergating_state) {
+			/* enable powergating to save power */
+			r = adev->ip_blocks[i].version->funcs->set_powergating_state((void *)adev,
+											state);
+			if (r) {
+				DRM_ERROR("set_powergating_state(gate) of IP block <%s> failed %d\n",
+					  adev->ip_blocks[i].version->funcs->name, r);
+				return r;
+			}
+		}
+	}
 	return 0;
 }
 
@@ -1762,7 +1811,7 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 	int i = 0, r;
 
 	for (i = 0; i < adev->num_ip_blocks; i++) {
-		if (!adev->ip_blocks[i].status.valid)
+		if (!adev->ip_blocks[i].status.hw)
 			continue;
 		if (adev->ip_blocks[i].version->funcs->late_init) {
 			r = adev->ip_blocks[i].version->funcs->late_init((void *)adev);
@@ -1771,10 +1820,13 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 					  adev->ip_blocks[i].version->funcs->name, r);
 				return r;
 			}
-			adev->ip_blocks[i].status.late_initialized = true;
 		}
+		adev->ip_blocks[i].status.late_initialized = true;
 	}
 
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
+
 	queue_delayed_work(system_wq, &adev->late_init_work,
 			   msecs_to_jiffies(AMDGPU_RESUME_MS));
 
@@ -1799,20 +1851,15 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 	int i, r;
 
 	amdgpu_amdkfd_device_fini(adev);
+
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
+
 	/* need to disable SMC first */
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if (!adev->ip_blocks[i].status.hw)
 			continue;
-		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC &&
-			adev->ip_blocks[i].version->funcs->set_clockgating_state) {
-			/* ungate blocks before hw fini so that we can shutdown the blocks safely */
-			r = adev->ip_blocks[i].version->funcs->set_clockgating_state((void *)adev,
-										     AMD_CG_STATE_UNGATE);
-			if (r) {
-				DRM_ERROR("set_clockgating_state(ungate) of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
-				return r;
-			}
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) {
 			r = adev->ip_blocks[i].version->funcs->hw_fini((void *)adev);
 			/* XXX handle errors */
 			if (r) {
@@ -1828,20 +1875,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 		if (!adev->ip_blocks[i].status.hw)
 			continue;
 
-		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
-			adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCE &&
-			adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCN &&
-			adev->ip_blocks[i].version->funcs->set_clockgating_state) {
-			/* ungate blocks before hw fini so that we can shutdown the blocks safely */
-			r = adev->ip_blocks[i].version->funcs->set_clockgating_state((void *)adev,
-										     AMD_CG_STATE_UNGATE);
-			if (r) {
-				DRM_ERROR("set_clockgating_state(ungate) of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
-				return r;
-			}
-		}
-
 		r = adev->ip_blocks[i].version->funcs->hw_fini((void *)adev);
 		/* XXX handle errors */
 		if (r) {
@@ -1858,6 +1891,7 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 			continue;
 
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) {
+			amdgpu_ucode_free_bo(adev);
 			amdgpu_free_static_csa(adev);
 			amdgpu_device_wb_fini(adev);
 			amdgpu_device_vram_scratch_fini(adev);
@@ -1888,24 +1922,78 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int amdgpu_device_enable_mgpu_fan_boost(void)
+{
+	struct amdgpu_gpu_instance *gpu_ins;
+	struct amdgpu_device *adev;
+	int i, ret = 0;
+
+	mutex_lock(&mgpu_info.mutex);
+
+	/*
+	 * MGPU fan boost feature should be enabled
+	 * only when there are two or more dGPUs in
+	 * the system
+	 */
+	if (mgpu_info.num_dgpu < 2)
+		goto out;
+
+	for (i = 0; i < mgpu_info.num_dgpu; i++) {
+		gpu_ins = &(mgpu_info.gpu_ins[i]);
+		adev = gpu_ins->adev;
+		if (!(adev->flags & AMD_IS_APU) &&
+		    !gpu_ins->mgpu_fan_enabled &&
+		    adev->powerplay.pp_funcs &&
+		    adev->powerplay.pp_funcs->enable_mgpu_fan_boost) {
+			ret = amdgpu_dpm_enable_mgpu_fan_boost(adev);
+			if (ret)
+				break;
+
+			gpu_ins->mgpu_fan_enabled = 1;
+		}
+	}
+
+out:
+	mutex_unlock(&mgpu_info.mutex);
+
+	return ret;
+}
+
 /**
- * amdgpu_device_ip_late_init_func_handler - work handler for clockgating
- *
- * @work: work_struct
+ * amdgpu_device_ip_late_init_func_handler - work handler for ib test
  *
- * Work handler for amdgpu_device_ip_late_set_cg_state.  We put the
- * clockgating setup into a worker thread to speed up driver init and
- * resume from suspend.
+ * @work: work_struct.
  */
 static void amdgpu_device_ip_late_init_func_handler(struct work_struct *work)
 {
 	struct amdgpu_device *adev =
 		container_of(work, struct amdgpu_device, late_init_work.work);
-	amdgpu_device_ip_late_set_cg_state(adev);
+	int r;
+
+	r = amdgpu_ib_ring_tests(adev);
+	if (r)
+		DRM_ERROR("ib ring test failed (%d).\n", r);
+
+	r = amdgpu_device_enable_mgpu_fan_boost();
+	if (r)
+		DRM_ERROR("enable mgpu fan boost failed (%d).\n", r);
+}
+
+static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
+{
+	struct amdgpu_device *adev =
+		container_of(work, struct amdgpu_device, gfx.gfx_off_delay_work.work);
+
+	mutex_lock(&adev->gfx.gfx_off_mutex);
+	if (!adev->gfx.gfx_off_state && !adev->gfx.gfx_off_req_count) {
+		if (!amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, true))
+			adev->gfx.gfx_off_state = true;
+	}
+	mutex_unlock(&adev->gfx.gfx_off_mutex);
 }
 
 /**
- * amdgpu_device_ip_suspend - run suspend for hardware IPs
+ * amdgpu_device_ip_suspend_phase1 - run suspend for hardware IPs (phase 1)
  *
  * @adev: amdgpu_device pointer
  *
@@ -1915,39 +2003,52 @@ static void amdgpu_device_ip_late_init_func_handler(struct work_struct *work)
  * in each IP into a state suitable for suspend.
  * Returns 0 on success, negative error code on failure.
  */
-int amdgpu_device_ip_suspend(struct amdgpu_device *adev)
+static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 {
 	int i, r;
 
-	if (amdgpu_sriov_vf(adev))
-		amdgpu_virt_request_full_gpu(adev, false);
-
-	/* ungate SMC block powergating */
-	if (adev->powerplay.pp_feature & PP_GFXOFF_MASK)
-		amdgpu_device_ip_set_powergating_state(adev,
-						       AMD_IP_BLOCK_TYPE_SMC,
-						       AMD_CG_STATE_UNGATE);
-
-	/* ungate SMC block first */
-	r = amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_SMC,
-						   AMD_CG_STATE_UNGATE);
-	if (r) {
-		DRM_ERROR("set_clockgating_state(ungate) SMC failed %d\n", r);
-	}
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
-		/* ungate blocks so that suspend can properly shut them down */
-		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_SMC &&
-			adev->ip_blocks[i].version->funcs->set_clockgating_state) {
-			r = adev->ip_blocks[i].version->funcs->set_clockgating_state((void *)adev,
-										     AMD_CG_STATE_UNGATE);
+		/* displays are handled separately */
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE) {
+			/* XXX handle errors */
+			r = adev->ip_blocks[i].version->funcs->suspend(adev);
+			/* XXX handle errors */
 			if (r) {
-				DRM_ERROR("set_clockgating_state(ungate) of IP block <%s> failed %d\n",
+				DRM_ERROR("suspend of IP block <%s> failed %d\n",
 					  adev->ip_blocks[i].version->funcs->name, r);
 			}
 		}
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_device_ip_suspend_phase2 - run suspend for hardware IPs (phase 2)
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Main suspend function for hardware IPs.  The list of all the hardware
+ * IPs that make up the asic is walked, clockgating is disabled and the
+ * suspend callbacks are run.  suspend puts the hardware and software state
+ * in each IP into a state suitable for suspend.
+ * Returns 0 on success, negative error code on failure.
+ */
+static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!adev->ip_blocks[i].status.valid)
+			continue;
+		/* displays are handled in phase1 */
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE)
+			continue;
 		/* XXX handle errors */
 		r = adev->ip_blocks[i].version->funcs->suspend(adev);
 		/* XXX handle errors */
@@ -1957,10 +2058,36 @@ int amdgpu_device_ip_suspend(struct amdgpu_device *adev)
 		}
 	}
 
+	return 0;
+}
+
+/**
+ * amdgpu_device_ip_suspend - run suspend for hardware IPs
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Main suspend function for hardware IPs.  The list of all the hardware
+ * IPs that make up the asic is walked, clockgating is disabled and the
+ * suspend callbacks are run.  suspend puts the hardware and software state
+ * in each IP into a state suitable for suspend.
+ * Returns 0 on success, negative error code on failure.
+ */
+int amdgpu_device_ip_suspend(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_request_full_gpu(adev, false);
+
+	r = amdgpu_device_ip_suspend_phase1(adev);
+	if (r)
+		return r;
+	r = amdgpu_device_ip_suspend_phase2(adev);
+
 	if (amdgpu_sriov_vf(adev))
 		amdgpu_virt_release_full_gpu(adev, false);
 
-	return 0;
+	return r;
 }
 
 static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev)
@@ -1970,6 +2097,7 @@ static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev)
 	static enum amd_ip_block_type ip_order[] = {
 		AMD_IP_BLOCK_TYPE_GMC,
 		AMD_IP_BLOCK_TYPE_COMMON,
+		AMD_IP_BLOCK_TYPE_PSP,
 		AMD_IP_BLOCK_TYPE_IH,
 	};
 
@@ -1985,7 +2113,7 @@ static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev)
 				continue;
 
 			r = block->version->funcs->hw_init(adev);
-			DRM_INFO("RE-INIT: %s %s\n", block->version->funcs->name, r?"failed":"successed");
+			DRM_INFO("RE-INIT: %s %s\n", block->version->funcs->name, r?"failed":"succeeded");
 			if (r)
 				return r;
 		}
@@ -2000,7 +2128,6 @@ static int amdgpu_device_ip_reinit_late_sriov(struct amdgpu_device *adev)
 
 	static enum amd_ip_block_type ip_order[] = {
 		AMD_IP_BLOCK_TYPE_SMC,
-		AMD_IP_BLOCK_TYPE_PSP,
 		AMD_IP_BLOCK_TYPE_DCE,
 		AMD_IP_BLOCK_TYPE_GFX,
 		AMD_IP_BLOCK_TYPE_SDMA,
@@ -2020,7 +2147,7 @@ static int amdgpu_device_ip_reinit_late_sriov(struct amdgpu_device *adev)
 				continue;
 
 			r = block->version->funcs->hw_init(adev);
-			DRM_INFO("RE-INIT: %s %s\n", block->version->funcs->name, r?"failed":"successed");
+			DRM_INFO("RE-INIT: %s %s\n", block->version->funcs->name, r?"failed":"succeeded");
 			if (r)
 				return r;
 		}
@@ -2085,7 +2212,8 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
 		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
-		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH)
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH ||
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)
 			continue;
 		r = adev->ip_blocks[i].version->funcs->resume(adev);
 		if (r) {
@@ -2117,6 +2245,11 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev)
 	r = amdgpu_device_ip_resume_phase1(adev);
 	if (r)
 		return r;
+
+	r = amdgpu_device_fw_loading(adev);
+	if (r)
+		return r;
+
 	r = amdgpu_device_ip_resume_phase2(adev);
 
 	return r;
@@ -2210,7 +2343,7 @@ bool amdgpu_device_has_dc_support(struct amdgpu_device *adev)
  * amdgpu_device_init - initialize the driver
  *
  * @adev: amdgpu_device pointer
- * @pdev: drm dev pointer
+ * @ddev: drm dev pointer
  * @pdev: pci dev pointer
  * @flags: driver flags
  *
@@ -2242,7 +2375,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->mman.buffer_funcs = NULL;
 	adev->mman.buffer_funcs_ring = NULL;
 	adev->vm_manager.vm_pte_funcs = NULL;
-	adev->vm_manager.vm_pte_num_rings = 0;
+	adev->vm_manager.vm_pte_num_rqs = 0;
 	adev->gmc.gmc_funcs = NULL;
 	adev->fence_context = dma_fence_context_alloc(AMDGPU_MAX_RINGS);
 	bitmap_zero(adev->gfx.pipe_reserve_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
@@ -2274,6 +2407,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->gfx.gpu_clock_mutex);
 	mutex_init(&adev->srbm_mutex);
 	mutex_init(&adev->gfx.pipe_reserve_mutex);
+	mutex_init(&adev->gfx.gfx_off_mutex);
 	mutex_init(&adev->grbm_idx_mutex);
 	mutex_init(&adev->mn_lock);
 	mutex_init(&adev->virt.vf_errors.lock);
@@ -2300,6 +2434,11 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 
 	INIT_DELAYED_WORK(&adev->late_init_work,
 			  amdgpu_device_ip_late_init_func_handler);
+	INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work,
+			  amdgpu_device_delay_enable_gfx_off);
+
+	adev->gfx.gfx_off_req_count = 1;
+	adev->pm.ac_power = power_supply_is_system_supplied() > 0 ? true : false;
 
 	/* Registers mapping */
 	/* TODO: block userspace mapping of io register */
@@ -2581,8 +2720,9 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
 /**
  * amdgpu_device_suspend - initiate device suspend
  *
- * @pdev: drm dev pointer
- * @state: suspend state
+ * @dev: drm dev pointer
+ * @suspend: suspend state
+ * @fbcon : notify the fbdev of suspend
  *
  * Puts the hw in the suspend state (all asics).
  * Returns 0 for success or an error on failure.
@@ -2604,8 +2744,14 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
+	adev->in_suspend = true;
 	drm_kms_helper_poll_disable(dev);
 
+	if (fbcon)
+		amdgpu_fbdev_set_suspend(adev, 1);
+
+	cancel_delayed_work_sync(&adev->late_init_work);
+
 	if (!amdgpu_device_has_dc_support(adev)) {
 		/* turn off display hw */
 		drm_modeset_lock_all(dev);
@@ -2613,44 +2759,46 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
 			drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
 		}
 		drm_modeset_unlock_all(dev);
-	}
-
-	amdgpu_amdkfd_suspend(adev);
-
-	/* unpin the front buffers and cursors */
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
-		struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-		struct drm_framebuffer *fb = crtc->primary->fb;
-		struct amdgpu_bo *robj;
+			/* unpin the front buffers and cursors */
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+			struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+			struct drm_framebuffer *fb = crtc->primary->fb;
+			struct amdgpu_bo *robj;
 
-		if (amdgpu_crtc->cursor_bo) {
-			struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
-			r = amdgpu_bo_reserve(aobj, true);
-			if (r == 0) {
-				amdgpu_bo_unpin(aobj);
-				amdgpu_bo_unreserve(aobj);
+			if (amdgpu_crtc->cursor_bo) {
+				struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
+				r = amdgpu_bo_reserve(aobj, true);
+				if (r == 0) {
+					amdgpu_bo_unpin(aobj);
+					amdgpu_bo_unreserve(aobj);
+				}
 			}
-		}
 
-		if (fb == NULL || fb->obj[0] == NULL) {
-			continue;
-		}
-		robj = gem_to_amdgpu_bo(fb->obj[0]);
-		/* don't unpin kernel fb objects */
-		if (!amdgpu_fbdev_robj_is_fb(adev, robj)) {
-			r = amdgpu_bo_reserve(robj, true);
-			if (r == 0) {
-				amdgpu_bo_unpin(robj);
-				amdgpu_bo_unreserve(robj);
+			if (fb == NULL || fb->obj[0] == NULL) {
+				continue;
+			}
+			robj = gem_to_amdgpu_bo(fb->obj[0]);
+			/* don't unpin kernel fb objects */
+			if (!amdgpu_fbdev_robj_is_fb(adev, robj)) {
+				r = amdgpu_bo_reserve(robj, true);
+				if (r == 0) {
+					amdgpu_bo_unpin(robj);
+					amdgpu_bo_unreserve(robj);
+				}
 			}
 		}
 	}
+
+	amdgpu_amdkfd_suspend(adev);
+
+	r = amdgpu_device_ip_suspend_phase1(adev);
+
 	/* evict vram memory */
 	amdgpu_bo_evict_vram(adev);
 
 	amdgpu_fence_driver_suspend(adev);
 
-	r = amdgpu_device_ip_suspend(adev);
+	r = amdgpu_device_ip_suspend_phase2(adev);
 
 	/* evict remaining vram memory
 	 * This second call to evict vram is to evict the gart page table
@@ -2669,18 +2817,15 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
 			DRM_ERROR("amdgpu asic reset failed\n");
 	}
 
-	if (fbcon) {
-		console_lock();
-		amdgpu_fbdev_set_suspend(adev, 1);
-		console_unlock();
-	}
 	return 0;
 }
 
 /**
  * amdgpu_device_resume - initiate device resume
  *
- * @pdev: drm dev pointer
+ * @dev: drm dev pointer
+ * @resume: resume state
+ * @fbcon : notify the fbdev of resume
  *
  * Bring the hw back to operating state (all asics).
  * Returns 0 for success or an error on failure.
@@ -2696,15 +2841,12 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	if (fbcon)
-		console_lock();
-
 	if (resume) {
 		pci_set_power_state(dev->pdev, PCI_D0);
 		pci_restore_state(dev->pdev);
 		r = pci_enable_device(dev->pdev);
 		if (r)
-			goto unlock;
+			return r;
 	}
 
 	/* post card */
@@ -2717,29 +2859,30 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 	r = amdgpu_device_ip_resume(adev);
 	if (r) {
 		DRM_ERROR("amdgpu_device_ip_resume failed (%d).\n", r);
-		goto unlock;
+		return r;
 	}
 	amdgpu_fence_driver_resume(adev);
 
 
 	r = amdgpu_device_ip_late_init(adev);
 	if (r)
-		goto unlock;
+		return r;
 
-	/* pin cursors */
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
-		struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	if (!amdgpu_device_has_dc_support(adev)) {
+		/* pin cursors */
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+			struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 
-		if (amdgpu_crtc->cursor_bo) {
-			struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
-			r = amdgpu_bo_reserve(aobj, true);
-			if (r == 0) {
-				r = amdgpu_bo_pin(aobj,
-						  AMDGPU_GEM_DOMAIN_VRAM,
-						  &amdgpu_crtc->cursor_addr);
-				if (r != 0)
-					DRM_ERROR("Failed to pin cursor BO (%d)\n", r);
-				amdgpu_bo_unreserve(aobj);
+			if (amdgpu_crtc->cursor_bo) {
+				struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
+				r = amdgpu_bo_reserve(aobj, true);
+				if (r == 0) {
+					r = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM);
+					if (r != 0)
+						DRM_ERROR("Failed to pin cursor BO (%d)\n", r);
+					amdgpu_crtc->cursor_addr = amdgpu_bo_gpu_offset(aobj);
+					amdgpu_bo_unreserve(aobj);
+				}
 			}
 		}
 	}
@@ -2747,6 +2890,9 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 	if (r)
 		return r;
 
+	/* Make sure IB tests flushed */
+	flush_delayed_work(&adev->late_init_work);
+
 	/* blat the mode back in */
 	if (fbcon) {
 		if (!amdgpu_device_has_dc_support(adev)) {
@@ -2760,6 +2906,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 			}
 			drm_modeset_unlock_all(dev);
 		}
+		amdgpu_fbdev_set_suspend(adev, 0);
 	}
 
 	drm_kms_helper_poll_enable(dev);
@@ -2783,15 +2930,9 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 #ifdef CONFIG_PM
 	dev->dev->power.disable_depth--;
 #endif
+	adev->in_suspend = false;
 
-	if (fbcon)
-		amdgpu_fbdev_set_suspend(adev, 0);
-
-unlock:
-	if (fbcon)
-		console_unlock();
-
-	return r;
+	return 0;
 }
 
 /**
@@ -2949,117 +3090,64 @@ static int amdgpu_device_ip_post_soft_reset(struct amdgpu_device *adev)
 }
 
 /**
- * amdgpu_device_recover_vram_from_shadow - restore shadowed VRAM buffers
- *
- * @adev: amdgpu_device pointer
- * @ring: amdgpu_ring for the engine handling the buffer operations
- * @bo: amdgpu_bo buffer whose shadow is being restored
- * @fence: dma_fence associated with the operation
- *
- * Restores the VRAM buffer contents from the shadow in GTT.  Used to
- * restore things like GPUVM page tables after a GPU reset where
- * the contents of VRAM might be lost.
- * Returns 0 on success, negative error code on failure.
- */
-static int amdgpu_device_recover_vram_from_shadow(struct amdgpu_device *adev,
-						  struct amdgpu_ring *ring,
-						  struct amdgpu_bo *bo,
-						  struct dma_fence **fence)
-{
-	uint32_t domain;
-	int r;
-
-	if (!bo->shadow)
-		return 0;
-
-	r = amdgpu_bo_reserve(bo, true);
-	if (r)
-		return r;
-	domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
-	/* if bo has been evicted, then no need to recover */
-	if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
-		r = amdgpu_bo_validate(bo->shadow);
-		if (r) {
-			DRM_ERROR("bo validate failed!\n");
-			goto err;
-		}
-
-		r = amdgpu_bo_restore_from_shadow(adev, ring, bo,
-						 NULL, fence, true);
-		if (r) {
-			DRM_ERROR("recover page table failed!\n");
-			goto err;
-		}
-	}
-err:
-	amdgpu_bo_unreserve(bo);
-	return r;
-}
-
-/**
- * amdgpu_device_handle_vram_lost - Handle the loss of VRAM contents
+ * amdgpu_device_recover_vram - Recover some VRAM contents
  *
  * @adev: amdgpu_device pointer
  *
  * Restores the contents of VRAM buffers from the shadows in GTT.  Used to
  * restore things like GPUVM page tables after a GPU reset where
  * the contents of VRAM might be lost.
- * Returns 0 on success, 1 on failure.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
  */
-static int amdgpu_device_handle_vram_lost(struct amdgpu_device *adev)
+static int amdgpu_device_recover_vram(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	struct amdgpu_bo *bo, *tmp;
 	struct dma_fence *fence = NULL, *next = NULL;
-	long r = 1;
-	int i = 0;
-	long tmo;
+	struct amdgpu_bo *shadow;
+	long r = 1, tmo;
 
 	if (amdgpu_sriov_runtime(adev))
-		tmo = msecs_to_jiffies(amdgpu_lockup_timeout);
+		tmo = msecs_to_jiffies(8000);
 	else
 		tmo = msecs_to_jiffies(100);
 
 	DRM_INFO("recover vram bo from shadow start\n");
 	mutex_lock(&adev->shadow_list_lock);
-	list_for_each_entry_safe(bo, tmp, &adev->shadow_list, shadow_list) {
-		next = NULL;
-		amdgpu_device_recover_vram_from_shadow(adev, ring, bo, &next);
+	list_for_each_entry(shadow, &adev->shadow_list, shadow_list) {
+
+		/* No need to recover an evicted BO */
+		if (shadow->tbo.mem.mem_type != TTM_PL_TT ||
+		    shadow->parent->tbo.mem.mem_type != TTM_PL_VRAM)
+			continue;
+
+		r = amdgpu_bo_restore_shadow(shadow, &next);
+		if (r)
+			break;
+
 		if (fence) {
 			r = dma_fence_wait_timeout(fence, false, tmo);
-			if (r == 0)
-				pr_err("wait fence %p[%d] timeout\n", fence, i);
-			else if (r < 0)
-				pr_err("wait fence %p[%d] interrupted\n", fence, i);
-			if (r < 1) {
-				dma_fence_put(fence);
-				fence = next;
+			dma_fence_put(fence);
+			fence = next;
+			if (r <= 0)
 				break;
-			}
-			i++;
+		} else {
+			fence = next;
 		}
-
-		dma_fence_put(fence);
-		fence = next;
 	}
 	mutex_unlock(&adev->shadow_list_lock);
 
-	if (fence) {
-		r = dma_fence_wait_timeout(fence, false, tmo);
-		if (r == 0)
-			pr_err("wait fence %p[%d] timeout\n", fence, i);
-		else if (r < 0)
-			pr_err("wait fence %p[%d] interrupted\n", fence, i);
-
-	}
+	if (fence)
+		tmo = dma_fence_wait_timeout(fence, false, tmo);
 	dma_fence_put(fence);
 
-	if (r > 0)
-		DRM_INFO("recover vram bo from shadow done\n");
-	else
+	if (r <= 0 || tmo <= 0) {
 		DRM_ERROR("recover vram bo from shadow failed\n");
+		return -EIO;
+	}
 
-	return (r > 0) ? 0 : 1;
+	DRM_INFO("recover vram bo from shadow done\n");
+	return 0;
 }
 
 /**
@@ -3068,7 +3156,7 @@ static int amdgpu_device_handle_vram_lost(struct amdgpu_device *adev)
  * @adev: amdgpu device pointer
  *
  * attempt to do soft-reset or full-reset and reinitialize Asic
- * return 0 means successed otherwise failed
+ * return 0 means succeeded otherwise failed
  */
 static int amdgpu_device_reset(struct amdgpu_device *adev)
 {
@@ -3112,6 +3200,10 @@ retry:
 			if (r)
 				goto out;
 
+			r = amdgpu_device_fw_loading(adev);
+			if (r)
+				return r;
+
 			r = amdgpu_device_ip_resume_phase2(adev);
 			if (r)
 				goto out;
@@ -3133,8 +3225,8 @@ out:
 		}
 	}
 
-	if (!r && ((need_full_reset && !(adev->flags & AMD_IS_APU)) || vram_lost))
-		r = amdgpu_device_handle_vram_lost(adev);
+	if (!r)
+		r = amdgpu_device_recover_vram(adev);
 
 	return r;
 }
@@ -3143,9 +3235,10 @@ out:
  * amdgpu_device_reset_sriov - reset ASIC for SR-IOV vf
  *
  * @adev: amdgpu device pointer
+ * @from_hypervisor: request from hypervisor
  *
  * do VF FLR and reinitialize Asic
- * return 0 means successed otherwise failed
+ * return 0 means succeeded otherwise failed
  */
 static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 				     bool from_hypervisor)
@@ -3167,6 +3260,10 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 	/* we need recover gart prior to run SMC/CP/SDMA resume */
 	amdgpu_gtt_mgr_recover(&adev->mman.bdev.man[TTM_PL_TT]);
 
+	r = amdgpu_device_fw_loading(adev);
+	if (r)
+		return r;
+
 	/* now we are okay to resume SMC/CP/SDMA */
 	r = amdgpu_device_ip_reinit_late_sriov(adev);
 	if (r)
@@ -3179,44 +3276,59 @@ error:
 	amdgpu_virt_release_full_gpu(adev, true);
 	if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) {
 		atomic_inc(&adev->vram_lost_counter);
-		r = amdgpu_device_handle_vram_lost(adev);
+		r = amdgpu_device_recover_vram(adev);
 	}
 
 	return r;
 }
 
 /**
+ * amdgpu_device_should_recover_gpu - check if we should try GPU recovery
+ *
+ * @adev: amdgpu device pointer
+ *
+ * Check amdgpu_gpu_recovery and SRIOV status to see if we should try to recover
+ * a hung GPU.
+ */
+bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev)
+{
+	if (!amdgpu_device_ip_check_soft_reset(adev)) {
+		DRM_INFO("Timeout, but no hardware hang detected.\n");
+		return false;
+	}
+
+	if (amdgpu_gpu_recovery == 0 || (amdgpu_gpu_recovery == -1  &&
+					 !amdgpu_sriov_vf(adev))) {
+		DRM_INFO("GPU recovery disabled.\n");
+		return false;
+	}
+
+	return true;
+}
+
+/**
  * amdgpu_device_gpu_recover - reset the asic and recover scheduler
  *
  * @adev: amdgpu device pointer
  * @job: which job trigger hang
- * @force forces reset regardless of amdgpu_gpu_recovery
  *
  * Attempt to reset the GPU if it has hung (all asics).
  * Returns 0 for success or an error on failure.
  */
 int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
-			      struct amdgpu_job *job, bool force)
+			      struct amdgpu_job *job)
 {
 	int i, r, resched;
 
-	if (!force && !amdgpu_device_ip_check_soft_reset(adev)) {
-		DRM_INFO("No hardware hang detected. Did some blocks stall?\n");
-		return 0;
-	}
-
-	if (!force && (amdgpu_gpu_recovery == 0 ||
-			(amdgpu_gpu_recovery == -1  && !amdgpu_sriov_vf(adev)))) {
-		DRM_INFO("GPU recovery disabled.\n");
-		return 0;
-	}
-
 	dev_info(adev->dev, "GPU reset begin!\n");
 
 	mutex_lock(&adev->lock_reset);
 	atomic_inc(&adev->gpu_reset_counter);
 	adev->in_gpu_reset = 1;
 
+	/* Block kfd */
+	amdgpu_amdkfd_pre_reset(adev);
+
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);
 
@@ -3229,10 +3341,10 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 		kthread_park(ring->sched.thread);
 
-		if (job && job->ring->idx != i)
+		if (job && job->base.sched == &ring->sched)
 			continue;
 
-		drm_sched_hw_job_reset(&ring->sched, &job->base);
+		drm_sched_hw_job_reset(&ring->sched, job ? &job->base : NULL);
 
 		/* after all hw jobs are reset, hw fence is meaningless, so force_completion */
 		amdgpu_fence_driver_force_completion(ring);
@@ -3253,7 +3365,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		 * or all rings (in the case @job is NULL)
 		 * after above amdgpu_reset accomplished
 		 */
-		if ((!job || job->ring->idx == i) && !r)
+		if ((!job || job->base.sched == &ring->sched) && !r)
 			drm_sched_job_recovery(&ring->sched);
 
 		kthread_unpark(ring->sched.thread);
@@ -3270,9 +3382,11 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		dev_info(adev->dev, "GPU reset(%d) failed\n", atomic_read(&adev->gpu_reset_counter));
 		amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_GPU_RESET_FAIL, 0, r);
 	} else {
-		dev_info(adev->dev, "GPU reset(%d) successed!\n",atomic_read(&adev->gpu_reset_counter));
+		dev_info(adev->dev, "GPU reset(%d) succeeded!\n",atomic_read(&adev->gpu_reset_counter));
 	}
 
+	/*unlock kfd */
+	amdgpu_amdkfd_post_reset(adev);
 	amdgpu_vf_error_trans_all(adev);
 	adev->in_gpu_reset = 0;
 	mutex_unlock(&adev->lock_reset);
@@ -3290,8 +3404,9 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
  */
 static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 {
-	u32 mask;
-	int ret;
+	struct pci_dev *pdev;
+	enum pci_bus_speed speed_cap;
+	enum pcie_link_width link_width;
 
 	if (amdgpu_pcie_gen_cap)
 		adev->pm.pcie_gen_mask = amdgpu_pcie_gen_cap;
@@ -3309,27 +3424,61 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 	}
 
 	if (adev->pm.pcie_gen_mask == 0) {
-		ret = drm_pcie_get_speed_cap_mask(adev->ddev, &mask);
-		if (!ret) {
-			adev->pm.pcie_gen_mask = (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+		/* asic caps */
+		pdev = adev->pdev;
+		speed_cap = pcie_get_speed_cap(pdev);
+		if (speed_cap == PCI_SPEED_UNKNOWN) {
+			adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
 						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
 						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3);
-
-			if (mask & DRM_PCIE_SPEED_25)
-				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1;
-			if (mask & DRM_PCIE_SPEED_50)
-				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2;
-			if (mask & DRM_PCIE_SPEED_80)
-				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3;
 		} else {
-			adev->pm.pcie_gen_mask = AMDGPU_DEFAULT_PCIE_GEN_MASK;
+			if (speed_cap == PCIE_SPEED_16_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN4);
+			else if (speed_cap == PCIE_SPEED_8_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3);
+			else if (speed_cap == PCIE_SPEED_5_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2);
+			else
+				adev->pm.pcie_gen_mask |= CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1;
+		}
+		/* platform caps */
+		pdev = adev->ddev->pdev->bus->self;
+		speed_cap = pcie_get_speed_cap(pdev);
+		if (speed_cap == PCI_SPEED_UNKNOWN) {
+			adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+						   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2);
+		} else {
+			if (speed_cap == PCIE_SPEED_16_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4);
+			else if (speed_cap == PCIE_SPEED_8_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3);
+			else if (speed_cap == PCIE_SPEED_5_0GT)
+				adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 |
+							   CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2);
+			else
+				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1;
+
 		}
 	}
 	if (adev->pm.pcie_mlw_mask == 0) {
-		ret = drm_pcie_get_max_link_width(adev->ddev, &mask);
-		if (!ret) {
-			switch (mask) {
-			case 32:
+		pdev = adev->ddev->pdev->bus->self;
+		link_width = pcie_get_width_cap(pdev);
+		if (link_width == PCIE_LNK_WIDTH_UNKNOWN) {
+			adev->pm.pcie_mlw_mask |= AMDGPU_DEFAULT_PCIE_MLW_MASK;
+		} else {
+			switch (link_width) {
+			case PCIE_LNK_X32:
 				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
@@ -3338,7 +3487,7 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
-			case 16:
+			case PCIE_LNK_X16:
 				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
@@ -3346,36 +3495,34 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
-			case 12:
+			case PCIE_LNK_X12:
 				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
-			case 8:
+			case PCIE_LNK_X8:
 				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
-			case 4:
+			case PCIE_LNK_X4:
 				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
-			case 2:
+			case PCIE_LNK_X2:
 				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
 							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
-			case 1:
+			case PCIE_LNK_X1:
 				adev->pm.pcie_mlw_mask = CAIL_PCIE_LINK_WIDTH_SUPPORT_X1;
 				break;
 			default:
 				break;
 			}
-		} else {
-			adev->pm.pcie_mlw_mask = AMDGPU_DEFAULT_PCIE_MLW_MASK;
 		}
 	}
 }