Merge tag 'amd-drm-next-5.15-2021-08-27' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-5.15-2021-08-27:

amdgpu:
- PLL fix for SI
- Misc code cleanups
- RAS fixes
- PSP cleanups
- Polaris UVD/VCE suspend fixes
- aldebaran fixes
- DCN3.x mclk fixes

amdkfd:
- CWSR fixes for arcturus and aldebaran
- SVM fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210827192336.4649-1-alexander.deucher@amd.com

40 files changed, 761 insertions, 143 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 0d814c957461..8d0748184a14 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -58,7 +58,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
 	amdgpu_vm_sdma.o amdgpu_discovery.o amdgpu_ras_eeprom.o amdgpu_nbio.o \
 	amdgpu_umc.o smu_v11_0_i2c.o amdgpu_fru_eeprom.o amdgpu_rap.o \
 	amdgpu_fw_attestation.o amdgpu_securedisplay.o amdgpu_hdp.o \
-	amdgpu_eeprom.o
+	amdgpu_eeprom.o amdgpu_mca.o
 
 amdgpu-$(CONFIG_PROC_FS) += amdgpu_fdinfo.o
 
@@ -189,6 +189,10 @@ amdgpu-y += \
 amdgpu-y += \
 	amdgpu_reset.o
 
+# add MCA block
+amdgpu-y += \
+	mca_v3_0.o
+
 # add amdkfd interfaces
 amdgpu-y += amdgpu_amdkfd.o
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 0f278cc3a5f4..dc3c6b3a00e5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -108,6 +108,7 @@
 #include "amdgpu_df.h"
 #include "amdgpu_smuio.h"
 #include "amdgpu_fdinfo.h"
+#include "amdgpu_mca.h"
 
 #define MAX_GPU_INSTANCE		16
 
@@ -1009,6 +1010,9 @@ struct amdgpu_device {
 	/* df */
 	struct amdgpu_df                df;
 
+	/* MCA */
+	struct amdgpu_mca               mca;
+
 	struct amdgpu_ip_block          ip_blocks[AMDGPU_MAX_IP_NUM];
 	uint32_t		        harvest_ip_mask;
 	int				num_ip_blocks;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index 260ba01d303e..4811b0faafd9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -1040,7 +1040,7 @@ void amdgpu_acpi_detect(void)
  */
 bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev)
 {
-#if IS_ENABLED(CONFIG_AMD_PMC) && IS_ENABLED(CONFIG_PM_SLEEP)
+#if IS_ENABLED(CONFIG_AMD_PMC) && IS_ENABLED(CONFIG_SUSPEND)
 	if (acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0) {
 		if (adev->flags & AMD_IS_APU)
 			return pm_suspend_target_state == PM_SUSPEND_TO_IDLE;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
index a5434b713856..46cd4ee6bafb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
@@ -44,4 +44,5 @@ const struct kfd2kgd_calls aldebaran_kfd2kgd = {
 	.get_atc_vmid_pasid_mapping_info =
 				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
+	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
index 6409d6b1b2df..5a7f680bcb3f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
@@ -305,5 +305,6 @@ const struct kfd2kgd_calls arcturus_kfd2kgd = {
 				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base =
 				kgd_gfx_v9_set_vm_context_page_table_base,
-	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy
+	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
+	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
index 154244916727..bcc1cbeb8799 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
@@ -882,7 +882,7 @@ void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
 				adev->gfx.cu_info.max_waves_per_simd;
 }
 
-static void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
+void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
                         uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
index e64deba8900f..c63591106879 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
@@ -65,3 +65,5 @@ void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd,
 			uint32_t vmid, uint64_t page_table_base);
 void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
 		int *pasid_wave_cnt, int *max_waves_per_cu);
+void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
+		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c
index b36405170ff3..76efd5f8950f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c
@@ -76,7 +76,7 @@ static int amdgpu_gart_dummy_page_init(struct amdgpu_device *adev)
 	if (adev->dummy_page_addr)
 		return 0;
 	adev->dummy_page_addr = dma_map_page(&adev->pdev->dev, dummy_page, 0,
-					     PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+					     PAGE_SIZE, DMA_BIDIRECTIONAL);
 	if (dma_mapping_error(&adev->pdev->dev, adev->dummy_page_addr)) {
 		dev_err(&adev->pdev->dev, "Failed to DMA MAP the dummy page\n");
 		adev->dummy_page_addr = 0;
@@ -96,8 +96,8 @@ void amdgpu_gart_dummy_page_fini(struct amdgpu_device *adev)
 {
 	if (!adev->dummy_page_addr)
 		return;
-	pci_unmap_page(adev->pdev, adev->dummy_page_addr,
-		       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+	dma_unmap_page(&adev->pdev->dev, adev->dummy_page_addr, PAGE_SIZE,
+		       DMA_BIDIRECTIONAL);
 	adev->dummy_page_addr = 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index c6f2fb9557ff..cb07cc3b06ed 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -904,7 +904,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 					 DIV_ROUND_UP(args->bpp, 8), 0);
 	args->size = (u64)args->pitch * args->height;
 	args->size = ALIGN(args->size, PAGE_SIZE);
-	domain = amdgpu_bo_get_preferred_pin_domain(adev,
+	domain = amdgpu_bo_get_preferred_domain(adev,
 				amdgpu_display_supported_domains(adev, flags));
 	r = amdgpu_gem_object_create(adev, args->size, 0, domain, flags,
 				     ttm_bo_type_device, NULL, &gobj);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index d0b8d415b63b..c7797eac83c3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -471,6 +471,27 @@ int amdgpu_gmc_ras_late_init(struct amdgpu_device *adev)
 			return r;
 	}
 
+	if (adev->mca.mp0.ras_funcs &&
+	    adev->mca.mp0.ras_funcs->ras_late_init) {
+		r = adev->mca.mp0.ras_funcs->ras_late_init(adev);
+		if (r)
+			return r;
+	}
+
+	if (adev->mca.mp1.ras_funcs &&
+	    adev->mca.mp1.ras_funcs->ras_late_init) {
+		r = adev->mca.mp1.ras_funcs->ras_late_init(adev);
+		if (r)
+			return r;
+	}
+
+	if (adev->mca.mpio.ras_funcs &&
+	    adev->mca.mpio.ras_funcs->ras_late_init) {
+		r = adev->mca.mpio.ras_funcs->ras_late_init(adev);
+		if (r)
+			return r;
+	}
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
index 8996cb4ed57a..9342aa23ebd2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
@@ -47,8 +47,6 @@ int amdgpu_jpeg_sw_fini(struct amdgpu_device *adev)
 {
 	int i;
 
-	cancel_delayed_work_sync(&adev->jpeg.idle_work);
-
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
 		if (adev->jpeg.harvest_config & (1 << i))
 			continue;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
new file mode 100644
index 000000000000..a2d3dbbf7d25
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu_ras.h"
+#include "amdgpu.h"
+#include "amdgpu_mca.h"
+
+#include "umc/umc_6_7_0_offset.h"
+#include "umc/umc_6_7_0_sh_mask.h"
+
+void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev,
+					      uint64_t mc_status_addr,
+					      unsigned long *error_count)
+{
+	uint64_t mc_status = RREG64_PCIE(mc_status_addr * 4);
+
+	if (REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)
+		*error_count += 1;
+}
+
+void amdgpu_mca_query_uncorrectable_error_count(struct amdgpu_device *adev,
+						uint64_t mc_status_addr,
+						unsigned long *error_count)
+{
+	uint64_t mc_status = RREG64_PCIE(mc_status_addr * 4);
+
+	if ((REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1))
+		*error_count += 1;
+}
+
+void amdgpu_mca_reset_error_count(struct amdgpu_device *adev,
+				  uint64_t mc_status_addr)
+{
+	WREG64_PCIE(mc_status_addr * 4, 0x0ULL);
+}
+
+void amdgpu_mca_query_ras_error_count(struct amdgpu_device *adev,
+				      uint64_t mc_status_addr,
+				      void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+
+	amdgpu_mca_query_correctable_error_count(adev, mc_status_addr, &(err_data->ce_count));
+	amdgpu_mca_query_uncorrectable_error_count(adev, mc_status_addr, &(err_data->ue_count));
+
+	amdgpu_mca_reset_error_count(adev, mc_status_addr);
+}
+
+int amdgpu_mca_ras_late_init(struct amdgpu_device *adev,
+			     struct amdgpu_mca_ras *mca_dev)
+{
+	int r;
+	struct ras_ih_if ih_info = {
+		.cb = NULL,
+	};
+	struct ras_fs_if fs_info = {
+		.sysfs_name = mca_dev->ras_funcs->sysfs_name,
+	};
+
+	if (!mca_dev->ras_if) {
+		mca_dev->ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
+		if (!mca_dev->ras_if)
+			return -ENOMEM;
+		mca_dev->ras_if->block = mca_dev->ras_funcs->ras_block;
+		mca_dev->ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+		mca_dev->ras_if->sub_block_index = 0;
+	}
+	ih_info.head = fs_info.head = *mca_dev->ras_if;
+	r = amdgpu_ras_late_init(adev, mca_dev->ras_if,
+				 &fs_info, &ih_info);
+	if (r || !amdgpu_ras_is_supported(adev, mca_dev->ras_if->block)) {
+		kfree(mca_dev->ras_if);
+		mca_dev->ras_if = NULL;
+	}
+
+	return r;
+}
+
+void amdgpu_mca_ras_fini(struct amdgpu_device *adev,
+			 struct amdgpu_mca_ras *mca_dev)
+{
+	struct ras_ih_if ih_info = {
+		.cb = NULL,
+	};
+
+	if (!mca_dev->ras_if)
+		return;
+
+	amdgpu_ras_late_fini(adev, mca_dev->ras_if, &ih_info);
+	kfree(mca_dev->ras_if);
+	mca_dev->ras_if = NULL;
+}
\ No newline at end of file
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
new file mode 100644
index 000000000000..f860f2f0e296
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
@@ -0,0 +1,72 @@
+/*
+ * Copyright (C) 2021  Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#ifndef __AMDGPU_MCA_H__
+#define __AMDGPU_MCA_H__
+
+struct amdgpu_mca_ras_funcs {
+	int (*ras_late_init)(struct amdgpu_device *adev);
+	void (*ras_fini)(struct amdgpu_device *adev);
+	void (*query_ras_error_count)(struct amdgpu_device *adev,
+				      void *ras_error_status);
+	void (*query_ras_error_address)(struct amdgpu_device *adev,
+					void *ras_error_status);
+	uint32_t ras_block;
+	const char* sysfs_name;
+};
+
+struct amdgpu_mca_ras {
+	struct ras_common_if *ras_if;
+	const struct amdgpu_mca_ras_funcs *ras_funcs;
+};
+
+struct amdgpu_mca_funcs {
+	void (*init)(struct amdgpu_device *adev);
+};
+
+struct amdgpu_mca {
+	const struct amdgpu_mca_funcs *funcs;
+	struct amdgpu_mca_ras mp0;
+	struct amdgpu_mca_ras mp1;
+	struct amdgpu_mca_ras mpio;
+};
+
+void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev,
+					      uint64_t mc_status_addr,
+					      unsigned long *error_count);
+
+void amdgpu_mca_query_uncorrectable_error_count(struct amdgpu_device *adev,
+						uint64_t mc_status_addr,
+						unsigned long *error_count);
+
+void amdgpu_mca_reset_error_count(struct amdgpu_device *adev,
+				  uint64_t mc_status_addr);
+
+void amdgpu_mca_query_ras_error_count(struct amdgpu_device *adev,
+				      uint64_t mc_status_addr,
+				      void *ras_error_status);
+
+int amdgpu_mca_ras_late_init(struct amdgpu_device *adev,
+			     struct amdgpu_mca_ras *mca_dev);
+
+void amdgpu_mca_ras_fini(struct amdgpu_device *adev,
+			 struct amdgpu_mca_ras *mca_dev);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 7734c10ae74e..01a78c786536 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -947,7 +947,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	/* This assumes only APU display buffers are pinned with (VRAM|GTT).
 	 * See function amdgpu_display_supported_domains()
 	 */
-	domain = amdgpu_bo_get_preferred_pin_domain(adev, domain);
+	domain = amdgpu_bo_get_preferred_domain(adev, domain);
 
 	if (bo->tbo.base.import_attach)
 		dma_buf_pin(bo->tbo.base.import_attach);
@@ -1518,14 +1518,14 @@ u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo)
 }
 
 /**
- * amdgpu_bo_get_preferred_pin_domain - get preferred domain for scanout
+ * amdgpu_bo_get_preferred_domain - get preferred domain
  * @adev: amdgpu device object
  * @domain: allowed :ref:`memory domains <amdgpu_memory_domains>`
  *
  * Returns:
- * Which of the allowed domains is preferred for pinning the BO for scanout.
+ * Which of the allowed domains is preferred for allocating the BO.
  */
-uint32_t amdgpu_bo_get_preferred_pin_domain(struct amdgpu_device *adev,
+uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev,
 					    uint32_t domain)
 {
 	if (domain == (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT)) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index e72f329e7f18..9d6c001c15f8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -333,7 +333,7 @@ void amdgpu_bo_get_memory(struct amdgpu_bo *bo, uint64_t *vram_mem,
 void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo);
 int amdgpu_bo_restore_shadow(struct amdgpu_bo *shadow,
 			     struct dma_fence **fence);
-uint32_t amdgpu_bo_get_preferred_pin_domain(struct amdgpu_device *adev,
+uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev,
 					    uint32_t domain);
 
 /*
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c
index f2e20666c9c1..4eaec446b49d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c
@@ -80,12 +80,17 @@ static void amdgpu_pll_reduce_ratio(unsigned *nom, unsigned *den,
  * Calculate feedback and reference divider for a given post divider. Makes
  * sure we stay within the limits.
  */
-static void amdgpu_pll_get_fb_ref_div(unsigned nom, unsigned den, unsigned post_div,
-				      unsigned fb_div_max, unsigned ref_div_max,
-				      unsigned *fb_div, unsigned *ref_div)
+static void amdgpu_pll_get_fb_ref_div(struct amdgpu_device *adev, unsigned int nom,
+				      unsigned int den, unsigned int post_div,
+				      unsigned int fb_div_max, unsigned int ref_div_max,
+				      unsigned int *fb_div, unsigned int *ref_div)
 {
+
 	/* limit reference * post divider to a maximum */
-	ref_div_max = min(128 / post_div, ref_div_max);
+	if (adev->family == AMDGPU_FAMILY_SI)
+		ref_div_max = min(100 / post_div, ref_div_max);
+	else
+		ref_div_max = min(128 / post_div, ref_div_max);
 
 	/* get matching reference and feedback divider */
 	*ref_div = min(max(DIV_ROUND_CLOSEST(den, post_div), 1u), ref_div_max);
@@ -112,7 +117,8 @@ static void amdgpu_pll_get_fb_ref_div(unsigned nom, unsigned den, unsigned post_
  * Try to calculate the PLL parameters to generate the given frequency:
  * dot_clock = (ref_freq * feedback_div) / (ref_div * post_div)
  */
-void amdgpu_pll_compute(struct amdgpu_pll *pll,
+void amdgpu_pll_compute(struct amdgpu_device *adev,
+			struct amdgpu_pll *pll,
 			u32 freq,
 			u32 *dot_clock_p,
 			u32 *fb_div_p,
@@ -199,7 +205,7 @@ void amdgpu_pll_compute(struct amdgpu_pll *pll,
 
 	for (post_div = post_div_min; post_div <= post_div_max; ++post_div) {
 		unsigned diff;
-		amdgpu_pll_get_fb_ref_div(nom, den, post_div, fb_div_max,
+		amdgpu_pll_get_fb_ref_div(adev, nom, den, post_div, fb_div_max,
 					  ref_div_max, &fb_div, &ref_div);
 		diff = abs(target_clock - (pll->reference_freq * fb_div) /
 			(ref_div * post_div));
@@ -214,7 +220,7 @@ void amdgpu_pll_compute(struct amdgpu_pll *pll,
 	post_div = post_div_best;
 
 	/* get the feedback and reference divider for the optimal value */
-	amdgpu_pll_get_fb_ref_div(nom, den, post_div, fb_div_max, ref_div_max,
+	amdgpu_pll_get_fb_ref_div(adev, nom, den, post_div, fb_div_max, ref_div_max,
 				  &fb_div, &ref_div);
 
 	/* reduce the numbers to a simpler ratio once more */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.h
index db6136f68b82..44a583d6c9b4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.h
@@ -24,7 +24,8 @@
 #ifndef __AMDGPU_PLL_H__
 #define __AMDGPU_PLL_H__
 
-void amdgpu_pll_compute(struct amdgpu_pll *pll,
+void amdgpu_pll_compute(struct amdgpu_device *adev,
+			 struct amdgpu_pll *pll,
 			 u32 freq,
 			 u32 *dot_clock_p,
 			 u32 *fb_div_p,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index a78a832d8fea..23efdc672502 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -899,23 +899,37 @@ static void psp_prep_ta_load_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 	cmd->cmd.cmd_load_ta.cmd_buf_len	 = ta_shared_size;
 }
 
-static int psp_xgmi_init_shared_buf(struct psp_context *psp)
+static int psp_ta_init_shared_buf(struct psp_context *psp,
+				  struct ta_mem_context *mem_ctx,
+				  uint32_t shared_mem_size)
 {
 	int ret;
 
 	/*
-	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
-	 * physical) for xgmi ta <-> Driver
-	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_XGMI_SHARED_MEM_SIZE,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->xgmi_context.context.mem_context.shared_bo,
-				      &psp->xgmi_context.context.mem_context.shared_mc_addr,
-				      &psp->xgmi_context.context.mem_context.shared_buf);
+	* Allocate 16k memory aligned to 4k from Frame Buffer (local
+	* physical) for ta to host memory
+	*/
+	ret = amdgpu_bo_create_kernel(psp->adev, shared_mem_size, PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM,
+				      &mem_ctx->shared_bo,
+				      &mem_ctx->shared_mc_addr,
+				      &mem_ctx->shared_buf);
 
 	return ret;
 }
 
+static void psp_ta_free_shared_buf(struct ta_mem_context *mem_ctx)
+{
+	amdgpu_bo_free_kernel(&mem_ctx->shared_bo, &mem_ctx->shared_mc_addr,
+			      &mem_ctx->shared_buf);
+}
+
+static int psp_xgmi_init_shared_buf(struct psp_context *psp)
+{
+	return psp_ta_init_shared_buf(psp, &psp->xgmi_context.context.mem_context,
+				      PSP_XGMI_SHARED_MEM_SIZE);
+}
+
 static void psp_prep_ta_invoke_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 				       uint32_t ta_cmd_id,
 				       uint32_t session_id)
@@ -1020,9 +1034,7 @@ int psp_xgmi_terminate(struct psp_context *psp)
 	psp->xgmi_context.context.initialized = false;
 
 	/* free xgmi shared memory */
-	amdgpu_bo_free_kernel(&psp->xgmi_context.context.mem_context.shared_bo,
-			&psp->xgmi_context.context.mem_context.shared_mc_addr,
-			&psp->xgmi_context.context.mem_context.shared_buf);
+	psp_ta_free_shared_buf(&psp->xgmi_context.context.mem_context);
 
 	return 0;
 }
@@ -1270,19 +1282,8 @@ int psp_xgmi_set_topology_info(struct psp_context *psp,
 // ras begin
 static int psp_ras_init_shared_buf(struct psp_context *psp)
 {
-	int ret;
-
-	/*
-	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
-	 * physical) for ras ta <-> Driver
-	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_RAS_SHARED_MEM_SIZE,
-			PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-			&psp->ras_context.context.mem_context.shared_bo,
-			&psp->ras_context.context.mem_context.shared_mc_addr,
-			&psp->ras_context.context.mem_context.shared_buf);
-
-	return ret;
+	return psp_ta_init_shared_buf(psp, &psp->ras_context.context.mem_context,
+				      PSP_RAS_SHARED_MEM_SIZE);
 }
 
 static int psp_ras_load(struct psp_context *psp)
@@ -1466,9 +1467,7 @@ static int psp_ras_terminate(struct psp_context *psp)
 	psp->ras_context.context.initialized = false;
 
 	/* free ras shared memory */
-	amdgpu_bo_free_kernel(&psp->ras_context.context.mem_context.shared_bo,
-			&psp->ras_context.context.mem_context.shared_mc_addr,
-			&psp->ras_context.context.mem_context.shared_buf);
+	psp_ta_free_shared_buf(&psp->ras_context.context.mem_context);
 
 	return 0;
 }
@@ -1576,19 +1575,8 @@ int psp_ras_trigger_error(struct psp_context *psp,
 // HDCP start
 static int psp_hdcp_init_shared_buf(struct psp_context *psp)
 {
-	int ret;
-
-	/*
-	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
-	 * physical) for hdcp ta <-> Driver
-	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_HDCP_SHARED_MEM_SIZE,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->hdcp_context.context.mem_context.shared_bo,
-				      &psp->hdcp_context.context.mem_context.shared_mc_addr,
-				      &psp->hdcp_context.context.mem_context.shared_buf);
-
-	return ret;
+	return psp_ta_init_shared_buf(psp, &psp->hdcp_context.context.mem_context,
+				      PSP_HDCP_SHARED_MEM_SIZE);
 }
 
 static int psp_hdcp_load(struct psp_context *psp)
@@ -1712,9 +1700,7 @@ static int psp_hdcp_terminate(struct psp_context *psp)
 
 out:
 	/* free hdcp shared memory */
-	amdgpu_bo_free_kernel(&psp->hdcp_context.context.mem_context.shared_bo,
-			      &psp->hdcp_context.context.mem_context.shared_mc_addr,
-			      &psp->hdcp_context.context.mem_context.shared_buf);
+	psp_ta_free_shared_buf(&psp->hdcp_context.context.mem_context);
 
 	return 0;
 }
@@ -1723,19 +1709,8 @@ out:
 // DTM start
 static int psp_dtm_init_shared_buf(struct psp_context *psp)
 {
-	int ret;
-
-	/*
-	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
-	 * physical) for dtm ta <-> Driver
-	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_DTM_SHARED_MEM_SIZE,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->dtm_context.context.mem_context.shared_bo,
-				      &psp->dtm_context.context.mem_context.shared_mc_addr,
-				      &psp->dtm_context.context.mem_context.shared_buf);
-
-	return ret;
+	return psp_ta_init_shared_buf(psp, &psp->dtm_context.context.mem_context,
+				      PSP_DTM_SHARED_MEM_SIZE);
 }
 
 static int psp_dtm_load(struct psp_context *psp)
@@ -1858,10 +1833,8 @@ static int psp_dtm_terminate(struct psp_context *psp)
 	psp->dtm_context.context.initialized = false;
 
 out:
-	/* free hdcp shared memory */
-	amdgpu_bo_free_kernel(&psp->dtm_context.context.mem_context.shared_bo,
-			      &psp->dtm_context.context.mem_context.shared_mc_addr,
-			      &psp->dtm_context.context.mem_context.shared_buf);
+	/* free dtm shared memory */
+	psp_ta_free_shared_buf(&psp->dtm_context.context.mem_context);
 
 	return 0;
 }
@@ -1870,19 +1843,8 @@ out:
 // RAP start
 static int psp_rap_init_shared_buf(struct psp_context *psp)
 {
-	int ret;
-
-	/*
-	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
-	 * physical) for rap ta <-> Driver
-	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_RAP_SHARED_MEM_SIZE,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->rap_context.context.mem_context.shared_bo,
-				      &psp->rap_context.context.mem_context.shared_mc_addr,
-				      &psp->rap_context.context.mem_context.shared_buf);
-
-	return ret;
+	return psp_ta_init_shared_buf(psp, &psp->rap_context.context.mem_context,
+				      PSP_RAP_SHARED_MEM_SIZE);
 }
 
 static int psp_rap_load(struct psp_context *psp)
@@ -1958,9 +1920,7 @@ static int psp_rap_initialize(struct psp_context *psp)
 	if (ret || status != TA_RAP_STATUS__SUCCESS) {
 		psp_rap_unload(psp);
 
-		amdgpu_bo_free_kernel(&psp->rap_context.context.mem_context.shared_bo,
-			      &psp->rap_context.context.mem_context.shared_mc_addr,
-			      &psp->rap_context.context.mem_context.shared_buf);
+		psp_ta_free_shared_buf(&psp->rap_context.context.mem_context);
 
 		psp->rap_context.context.initialized = false;
 
@@ -1985,9 +1945,7 @@ static int psp_rap_terminate(struct psp_context *psp)
 	psp->rap_context.context.initialized = false;
 
 	/* free rap shared memory */
-	amdgpu_bo_free_kernel(&psp->rap_context.context.mem_context.shared_bo,
-			      &psp->rap_context.context.mem_context.shared_mc_addr,
-			      &psp->rap_context.context.mem_context.shared_buf);
+	psp_ta_free_shared_buf(&psp->rap_context.context.mem_context);
 
 	return ret;
 }
@@ -2030,19 +1988,9 @@ out_unlock:
 /* securedisplay start */
 static int psp_securedisplay_init_shared_buf(struct psp_context *psp)
 {
-	int ret;
-
-	/*
-	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
-	 * physical) for sa ta <-> Driver
-	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_SECUREDISPLAY_SHARED_MEM_SIZE,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->securedisplay_context.context.mem_context.shared_bo,
-				      &psp->securedisplay_context.context.mem_context.shared_mc_addr,
-				      &psp->securedisplay_context.context.mem_context.shared_buf);
-
-	return ret;
+	return psp_ta_init_shared_buf(
+		psp, &psp->securedisplay_context.context.mem_context,
+		PSP_SECUREDISPLAY_SHARED_MEM_SIZE);
 }
 
 static int psp_securedisplay_load(struct psp_context *psp)
@@ -2120,9 +2068,7 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	if (ret) {
 		psp_securedisplay_unload(psp);
 
-		amdgpu_bo_free_kernel(&psp->securedisplay_context.context.mem_context.shared_bo,
-			      &psp->securedisplay_context.context.mem_context.shared_mc_addr,
-			      &psp->securedisplay_context.context.mem_context.shared_buf);
+		psp_ta_free_shared_buf(&psp->securedisplay_context.context.mem_context);
 
 		psp->securedisplay_context.context.initialized = false;
 
@@ -2159,9 +2105,7 @@ static int psp_securedisplay_terminate(struct psp_context *psp)
 	psp->securedisplay_context.context.initialized = false;
 
 	/* free securedisplay shared memory */
-	amdgpu_bo_free_kernel(&psp->securedisplay_context.context.mem_context.shared_bo,
-			      &psp->securedisplay_context.context.mem_context.shared_mc_addr,
-			      &psp->securedisplay_context.context.mem_context.shared_buf);
+	psp_ta_free_shared_buf(&psp->securedisplay_context.context.mem_context);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index abc5710898e8..eae604fd90b8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -49,6 +49,7 @@ enum amdgpu_ras_block {
 	AMDGPU_RAS_BLOCK__MP0,
 	AMDGPU_RAS_BLOCK__MP1,
 	AMDGPU_RAS_BLOCK__FUSE,
+	AMDGPU_RAS_BLOCK__MPIO,
 
 	AMDGPU_RAS_BLOCK__LAST
 };
@@ -309,6 +310,7 @@ struct ras_common_if {
 	enum amdgpu_ras_block block;
 	enum amdgpu_ras_error_type type;
 	uint32_t sub_block_index;
+	char name[32];
 };
 
 struct amdgpu_ras {
@@ -419,7 +421,7 @@ struct ras_badpage {
 /* interfaces for IP */
 struct ras_fs_if {
 	struct ras_common_if head;
-	char sysfs_name[32];
+	const char* sysfs_name;
 	char debugfs_name[32];
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 0f576f294d8a..d451c359606a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -326,7 +326,6 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 {
 	int i, j;
 
-	cancel_delayed_work_sync(&adev->uvd.idle_work);
 	drm_sched_entity_destroy(&adev->uvd.entity);
 
 	for (j = 0; j < adev->uvd.num_uvd_inst; ++j) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
index 1ae7f824adc7..8e8dee9fac9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
@@ -218,7 +218,6 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev)
 	if (adev->vce.vcpu_bo == NULL)
 		return 0;
 
-	cancel_delayed_work_sync(&adev->vce.idle_work);
 	drm_sched_entity_destroy(&adev->vce.entity);
 
 	amdgpu_bo_free_kernel(&adev->vce.vcpu_bo, &adev->vce.gpu_addr,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 6780df0fb265..008a308a4eca 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -258,8 +258,6 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)
 {
 	int i, j;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
-
 	for (j = 0; j < adev->vcn.num_vcn_inst; ++j) {
 		if (adev->vcn.harvest_config & (1 << j))
 			continue;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 2af8860d74cc..6b15cad78de9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -926,7 +926,7 @@ static int amdgpu_vm_pt_create(struct amdgpu_device *adev,
 	bp.size = amdgpu_vm_bo_size(adev, level);
 	bp.byte_align = AMDGPU_GPU_PAGE_SIZE;
 	bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
-	bp.domain = amdgpu_bo_get_preferred_pin_domain(adev, bp.domain);
+	bp.domain = amdgpu_bo_get_preferred_domain(adev, bp.domain);
 	bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
 		AMDGPU_GEM_CREATE_CPU_GTT_USWC;
 
@@ -3345,12 +3345,13 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
  * @adev: amdgpu device pointer
  * @pasid: PASID of the VM
  * @addr: Address of the fault
+ * @write_fault: true is write fault, false is read fault
  *
  * Try to gracefully handle a VM fault. Return true if the fault was handled and
  * shouldn't be reported any more.
  */
 bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
-			    uint64_t addr)
+			    uint64_t addr, bool write_fault)
 {
 	bool is_compute_context = false;
 	struct amdgpu_bo *root;
@@ -3375,7 +3376,7 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
 	addr /= AMDGPU_GPU_PAGE_SIZE;
 
 	if (is_compute_context &&
-	    !svm_range_restore_pages(adev, pasid, addr)) {
+	    !svm_range_restore_pages(adev, pasid, addr, write_fault)) {
 		amdgpu_bo_unref(&root);
 		return true;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 80cc9ab2c1d0..85fcfb8c5efd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -448,7 +448,7 @@ void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);
 void amdgpu_vm_get_task_info(struct amdgpu_device *adev, u32 pasid,
 			     struct amdgpu_task_info *task_info);
 bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
-			    uint64_t addr);
+			    uint64_t addr, bool write_fault);
 
 void amdgpu_vm_set_task_info(struct amdgpu_vm *vm);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index dda4f0c5c4e7..978ac927ac11 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -32,6 +32,10 @@
 #include "wafl/wafl2_4_0_0_smn.h"
 #include "wafl/wafl2_4_0_0_sh_mask.h"
 
+#define smnPCS_XGMI23_PCS_ERROR_STATUS   0x11a01210
+#define smnPCS_XGMI3X16_PCS_ERROR_STATUS 0x11a0020c
+#define smnPCS_GOPX1_PCS_ERROR_STATUS    0x12200210
+
 static DEFINE_MUTEX(xgmi_mutex);
 
 #define AMDGPU_MAX_XGMI_DEVICE_PER_HIVE		4
@@ -63,6 +67,33 @@ static const int wafl_pcs_err_status_reg_arct[] = {
 	smnPCS_GOPX1_0_PCS_GOPX1_PCS_ERROR_STATUS + 0x100000,
 };
 
+static const int xgmi23_pcs_err_status_reg_aldebaran[] = {
+	smnPCS_XGMI23_PCS_ERROR_STATUS,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x100000,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x200000,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x300000,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x400000,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x500000,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x600000,
+	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x700000
+};
+
+static const int xgmi3x16_pcs_err_status_reg_aldebaran[] = {
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x100000,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x200000,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x300000,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x400000,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x500000,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x600000,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x700000
+};
+
+static const int walf_pcs_err_status_reg_aldebaran[] = {
+	smnPCS_GOPX1_PCS_ERROR_STATUS,
+	smnPCS_GOPX1_PCS_ERROR_STATUS + 0x100000
+};
+
 static const struct amdgpu_pcs_ras_field xgmi_pcs_ras_fields[] = {
 	{"XGMI PCS DataLossErr",
 	 SOC15_REG_FIELD(XGMI0_PCS_GOPX16_PCS_ERROR_STATUS, DataLossErr)},
@@ -771,6 +802,17 @@ static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
 			pcs_clear_status(adev,
 					 xgmi_pcs_err_status_reg_vg20[i]);
 		break;
+	case CHIP_ALDEBARAN:
+		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
+			pcs_clear_status(adev,
+					 xgmi23_pcs_err_status_reg_aldebaran[i]);
+		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
+			pcs_clear_status(adev,
+					 xgmi23_pcs_err_status_reg_aldebaran[i]);
+		for (i = 0; i < ARRAY_SIZE(walf_pcs_err_status_reg_aldebaran); i++)
+			pcs_clear_status(adev,
+					 walf_pcs_err_status_reg_aldebaran[i]);
+		break;
 	default:
 		break;
 	}
@@ -848,7 +890,6 @@ static int amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev,
 		}
 		break;
 	case CHIP_VEGA20:
-	default:
 		/* check xgmi pcs error */
 		for (i = 0; i < ARRAY_SIZE(xgmi_pcs_err_status_reg_vg20); i++) {
 			data = RREG32_PCIE(xgmi_pcs_err_status_reg_vg20[i]);
@@ -864,6 +905,32 @@ static int amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev,
 						data, &ue_cnt, &ce_cnt, false);
 		}
 		break;
+	case CHIP_ALDEBARAN:
+		/* check xgmi23 pcs error */
+		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++) {
+			data = RREG32_PCIE(xgmi23_pcs_err_status_reg_aldebaran[i]);
+			if (data)
+				amdgpu_xgmi_query_pcs_error_status(adev,
+						data, &ue_cnt, &ce_cnt, true);
+		}
+		/* check xgmi3x16 pcs error */
+		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_aldebaran); i++) {
+			data = RREG32_PCIE(xgmi3x16_pcs_err_status_reg_aldebaran[i]);
+			if (data)
+				amdgpu_xgmi_query_pcs_error_status(adev,
+						data, &ue_cnt, &ce_cnt, true);
+		}
+		/* check wafl pcs error */
+		for (i = 0; i < ARRAY_SIZE(walf_pcs_err_status_reg_aldebaran); i++) {
+			data = RREG32_PCIE(walf_pcs_err_status_reg_aldebaran[i]);
+			if (data)
+				amdgpu_xgmi_query_pcs_error_status(adev,
+						data, &ue_cnt, &ce_cnt, false);
+		}
+		break;
+	default:
+		dev_warn(adev->dev, "XGMI RAS error query not supported");
+		break;
 	}
 
 	adev->gmc.xgmi.ras_funcs->reset_ras_error_count(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c b/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
index 159a2a4385a1..afad094f84c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
@@ -851,7 +851,7 @@ void amdgpu_atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode
 	pll->reference_div = amdgpu_crtc->pll_reference_div;
 	pll->post_div = amdgpu_crtc->pll_post_div;
 
-	amdgpu_pll_compute(pll, amdgpu_crtc->adjusted_clock, &pll_clock,
+	amdgpu_pll_compute(adev, pll, amdgpu_crtc->adjusted_clock, &pll_clock,
 			    &fb_div, &frac_fb_div, &ref_div, &post_div);
 
 	amdgpu_atombios_crtc_program_ss(adev, ATOM_DISABLE, amdgpu_crtc->pll_id,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 24b781e90bef..41c3a0d70b7c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -93,6 +93,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 				       struct amdgpu_iv_entry *entry)
 {
 	bool retry_fault = !!(entry->src_data[1] & 0x80);
+	bool write_fault = !!(entry->src_data[1] & 0x20);
 	struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src];
 	struct amdgpu_task_info task_info;
 	uint32_t status = 0;
@@ -121,7 +122,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 		/* Try to handle the recoverable page faults by filling page
 		 * tables
 		 */
-		if (amdgpu_vm_handle_fault(adev, entry->pasid, addr))
+		if (amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault))
 			return 1;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 097230b5e946..d90c16a6b2b8 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -55,6 +55,7 @@
 #include "umc_v6_0.h"
 #include "umc_v6_7.h"
 #include "hdp_v4_0.h"
+#include "mca_v3_0.h"
 
 #include "ivsrcid/vmc/irqsrcs_vmc_1_0.h"
 
@@ -506,6 +507,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 				      struct amdgpu_iv_entry *entry)
 {
 	bool retry_fault = !!(entry->src_data[1] & 0x80);
+	bool write_fault = !!(entry->src_data[1] & 0x20);
 	uint32_t status = 0, cid = 0, rw = 0;
 	struct amdgpu_task_info task_info;
 	struct amdgpu_vmhub *hub;
@@ -536,7 +538,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 		/* Try to handle the recoverable page faults by filling page
 		 * tables
 		 */
-		if (amdgpu_vm_handle_fault(adev, entry->pasid, addr))
+		if (amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault))
 			return 1;
 	}
 
@@ -1229,6 +1231,18 @@ static void gmc_v9_0_set_hdp_ras_funcs(struct amdgpu_device *adev)
 	adev->hdp.ras_funcs = &hdp_v4_0_ras_funcs;
 }
 
+static void gmc_v9_0_set_mca_funcs(struct amdgpu_device *adev)
+{
+	switch (adev->asic_type) {
+	case CHIP_ALDEBARAN:
+		if (!adev->gmc.xgmi.connected_to_cpu)
+			adev->mca.funcs = &mca_v3_0_funcs;
+		break;
+	default:
+		break;
+	}
+}
+
 static int gmc_v9_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -1250,6 +1264,7 @@ static int gmc_v9_0_early_init(void *handle)
 	gmc_v9_0_set_mmhub_ras_funcs(adev);
 	gmc_v9_0_set_gfxhub_funcs(adev);
 	gmc_v9_0_set_hdp_ras_funcs(adev);
+	gmc_v9_0_set_mca_funcs(adev);
 
 	adev->gmc.shared_aperture_start = 0x2000000000000000ULL;
 	adev->gmc.shared_aperture_end =
@@ -1461,6 +1476,8 @@ static int gmc_v9_0_sw_init(void *handle)
 	adev->gfxhub.funcs->init(adev);
 
 	adev->mmhub.funcs->init(adev);
+	if (adev->mca.funcs)
+		adev->mca.funcs->init(adev);
 
 	spin_lock_init(&adev->gmc.invalidate_lock);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c
new file mode 100644
index 000000000000..058b65730a84
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c
@@ -0,0 +1,125 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu_ras.h"
+#include "amdgpu.h"
+#include "amdgpu_mca.h"
+
+#define smnMCMP0_STATUST0 	0x03830408
+#define smnMCMP1_STATUST0 	0x03b30408
+#define smnMCMPIO_STATUST0 	0x0c930408
+
+
+static void mca_v3_0_mp0_query_ras_error_count(struct amdgpu_device *adev,
+					       void *ras_error_status)
+{
+	amdgpu_mca_query_ras_error_count(adev,
+				         smnMCMP0_STATUST0,
+				         ras_error_status);
+}
+
+static int mca_v3_0_mp0_ras_late_init(struct amdgpu_device *adev)
+{
+	return amdgpu_mca_ras_late_init(adev, &adev->mca.mp0);
+}
+
+static void mca_v3_0_mp0_ras_fini(struct amdgpu_device *adev)
+{
+	amdgpu_mca_ras_fini(adev, &adev->mca.mp0);
+}
+
+const struct amdgpu_mca_ras_funcs mca_v3_0_mp0_ras_funcs = {
+	.ras_late_init = mca_v3_0_mp0_ras_late_init,
+	.ras_fini = mca_v3_0_mp0_ras_fini,
+	.query_ras_error_count = mca_v3_0_mp0_query_ras_error_count,
+	.query_ras_error_address = NULL,
+	.ras_block = AMDGPU_RAS_BLOCK__MP0,
+	.sysfs_name = "mp0_err_count",
+};
+
+static void mca_v3_0_mp1_query_ras_error_count(struct amdgpu_device *adev,
+					       void *ras_error_status)
+{
+	amdgpu_mca_query_ras_error_count(adev,
+				         smnMCMP1_STATUST0,
+				         ras_error_status);
+}
+
+static int mca_v3_0_mp1_ras_late_init(struct amdgpu_device *adev)
+{
+	return amdgpu_mca_ras_late_init(adev, &adev->mca.mp1);
+}
+
+static void mca_v3_0_mp1_ras_fini(struct amdgpu_device *adev)
+{
+	amdgpu_mca_ras_fini(adev, &adev->mca.mp1);
+}
+
+const struct amdgpu_mca_ras_funcs mca_v3_0_mp1_ras_funcs = {
+	.ras_late_init = mca_v3_0_mp1_ras_late_init,
+	.ras_fini = mca_v3_0_mp1_ras_fini,
+	.query_ras_error_count = mca_v3_0_mp1_query_ras_error_count,
+	.query_ras_error_address = NULL,
+	.ras_block = AMDGPU_RAS_BLOCK__MP1,
+	.sysfs_name = "mp1_err_count",
+};
+
+static void mca_v3_0_mpio_query_ras_error_count(struct amdgpu_device *adev,
+					       void *ras_error_status)
+{
+	amdgpu_mca_query_ras_error_count(adev,
+				         smnMCMPIO_STATUST0,
+				         ras_error_status);
+}
+
+static int mca_v3_0_mpio_ras_late_init(struct amdgpu_device *adev)
+{
+	return amdgpu_mca_ras_late_init(adev, &adev->mca.mpio);
+}
+
+static void mca_v3_0_mpio_ras_fini(struct amdgpu_device *adev)
+{
+	amdgpu_mca_ras_fini(adev, &adev->mca.mpio);
+}
+
+const struct amdgpu_mca_ras_funcs mca_v3_0_mpio_ras_funcs = {
+	.ras_late_init = mca_v3_0_mpio_ras_late_init,
+	.ras_fini = mca_v3_0_mpio_ras_fini,
+	.query_ras_error_count = mca_v3_0_mpio_query_ras_error_count,
+	.query_ras_error_address = NULL,
+	.ras_block = AMDGPU_RAS_BLOCK__MPIO,
+	.sysfs_name = "mpio_err_count",
+};
+
+
+static void mca_v3_0_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_mca *mca = &adev->mca;
+
+	mca->mp0.ras_funcs = &mca_v3_0_mp0_ras_funcs;
+	mca->mp1.ras_funcs = &mca_v3_0_mp1_ras_funcs;
+	mca->mpio.ras_funcs = &mca_v3_0_mpio_ras_funcs;
+}
+
+const struct amdgpu_mca_funcs mca_v3_0_funcs = {
+	.init = mca_v3_0_init,
+};
\ No newline at end of file
diff --git a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h
new file mode 100644
index 000000000000..b899b86194c2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (C) 2021  Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#ifndef __MCA_V3_0_H__
+#define __MCA_V3_0_H__
+
+extern const struct amdgpu_mca_funcs mca_v3_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
index 1c94a14fc18d..ba1d3ab869c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
@@ -85,6 +85,11 @@
 #define mmRCC_DEV0_EPF0_STRAP0_ALDE			0x0015
 #define mmRCC_DEV0_EPF0_STRAP0_ALDE_BASE_IDX		2
 
+#define mmBIF_DOORBELL_INT_CNTL_ALDE 			0x3878
+#define mmBIF_DOORBELL_INT_CNTL_ALDE_BASE_IDX 		2
+#define BIF_DOORBELL_INT_CNTL_ALDE__DOORBELL_INTERRUPT_DISABLE__SHIFT	0x18
+#define BIF_DOORBELL_INT_CNTL_ALDE__DOORBELL_INTERRUPT_DISABLE_MASK	0x01000000L
+
 static void nbio_v7_4_query_ras_error_count(struct amdgpu_device *adev,
 					void *ras_error_status);
 
@@ -346,14 +351,21 @@ static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device
 	struct ras_err_data err_data = {0, 0, 0, NULL};
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
-	bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL);
+	if (adev->asic_type == CHIP_ALDEBARAN)
+		bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL_ALDE);
+	else
+		bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL);
+
 	if (REG_GET_FIELD(bif_doorbell_intr_cntl,
 		BIF_DOORBELL_INT_CNTL, RAS_CNTLR_INTERRUPT_STATUS)) {
 		/* driver has to clear the interrupt status when bif ring is disabled */
 		bif_doorbell_intr_cntl = REG_SET_FIELD(bif_doorbell_intr_cntl,
 						BIF_DOORBELL_INT_CNTL,
 						RAS_CNTLR_INTERRUPT_CLEAR, 1);
-		WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
+		if (adev->asic_type == CHIP_ALDEBARAN)
+			WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL_ALDE, bif_doorbell_intr_cntl);
+		else
+			WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
 
 		if (!ras->disable_ras_err_cnt_harvest) {
 			/*
@@ -395,14 +407,22 @@ static void nbio_v7_4_handle_ras_err_event_athub_intr_no_bifring(struct amdgpu_d
 {
 	uint32_t bif_doorbell_intr_cntl;
 
-	bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL);
+	if (adev->asic_type == CHIP_ALDEBARAN)
+		bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL_ALDE);
+	else
+		bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL);
+
 	if (REG_GET_FIELD(bif_doorbell_intr_cntl,
 		BIF_DOORBELL_INT_CNTL, RAS_ATHUB_ERR_EVENT_INTERRUPT_STATUS)) {
 		/* driver has to clear the interrupt status when bif ring is disabled */
 		bif_doorbell_intr_cntl = REG_SET_FIELD(bif_doorbell_intr_cntl,
 						BIF_DOORBELL_INT_CNTL,
 						RAS_ATHUB_ERR_EVENT_INTERRUPT_CLEAR, 1);
-		WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
+
+		if (adev->asic_type == CHIP_ALDEBARAN)
+			WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL_ALDE, bif_doorbell_intr_cntl);
+		else
+			WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
 
 		amdgpu_ras_global_ras_isr(adev);
 	}
@@ -572,7 +592,11 @@ static void nbio_v7_4_query_ras_error_count(struct amdgpu_device *adev,
 static void nbio_v7_4_enable_doorbell_interrupt(struct amdgpu_device *adev,
 						bool enable)
 {
-	WREG32_FIELD15(NBIO, 0, BIF_DOORBELL_INT_CNTL,
+	if (adev->asic_type == CHIP_ALDEBARAN)
+		WREG32_FIELD15(NBIO, 0, BIF_DOORBELL_INT_CNTL_ALDE,
+		       DOORBELL_INTERRUPT_DISABLE, enable ? 0 : 1);
+	else
+		WREG32_FIELD15(NBIO, 0, BIF_DOORBELL_INT_CNTL,
 		       DOORBELL_INTERRUPT_DISABLE, enable ? 0 : 1);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index f7b56a746c15..0fc97c364fd7 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -1353,8 +1353,6 @@ static int soc15_common_early_init(void *handle)
 		adev->asic_funcs = &vega20_asic_funcs;
 		adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
 			AMD_CG_SUPPORT_GFX_MGLS |
-			AMD_CG_SUPPORT_GFX_CGCG |
-			AMD_CG_SUPPORT_GFX_CGLS |
 			AMD_CG_SUPPORT_GFX_CP_LS |
 			AMD_CG_SUPPORT_HDP_LS |
 			AMD_CG_SUPPORT_SDMA_MGCG |
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
index 6c0e91495365..7232241e3bfb 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
@@ -698,6 +698,30 @@ static int uvd_v3_1_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->uvd.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_uvd(adev, false);
+	} else {
+		amdgpu_asic_set_uvd_clocks(adev, 0, 0);
+		/* shutdown the UVD block */
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_CG_STATE_GATE);
+	}
+
 	if (RREG32(mmUVD_STATUS) != 0)
 		uvd_v3_1_stop(adev);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
index a301518e4957..52d6de969f46 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
@@ -212,6 +212,30 @@ static int uvd_v4_2_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->uvd.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_uvd(adev, false);
+	} else {
+		amdgpu_asic_set_uvd_clocks(adev, 0, 0);
+		/* shutdown the UVD block */
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_CG_STATE_GATE);
+	}
+
 	if (RREG32(mmUVD_STATUS) != 0)
 		uvd_v4_2_stop(adev);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
index a4d5bd21c83c..db6d06758e4d 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
@@ -210,6 +210,30 @@ static int uvd_v5_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->uvd.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_uvd(adev, false);
+	} else {
+		amdgpu_asic_set_uvd_clocks(adev, 0, 0);
+		/* shutdown the UVD block */
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_CG_STATE_GATE);
+	}
+
 	if (RREG32(mmUVD_STATUS) != 0)
 		uvd_v5_0_stop(adev);
 
@@ -224,7 +248,6 @@ static int uvd_v5_0_suspend(void *handle)
 	r = uvd_v5_0_hw_fini(adev);
 	if (r)
 		return r;
-	uvd_v5_0_set_clockgating_state(adev, AMD_CG_STATE_GATE);
 
 	return amdgpu_uvd_suspend(adev);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
index cf3803f8f075..bc571833632e 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
@@ -543,6 +543,30 @@ static int uvd_v6_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->uvd.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_uvd(adev, false);
+	} else {
+		amdgpu_asic_set_uvd_clocks(adev, 0, 0);
+		/* shutdown the UVD block */
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_CG_STATE_GATE);
+	}
+
 	if (RREG32(mmUVD_STATUS) != 0)
 		uvd_v6_0_stop(adev);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index 939bcfa2a4ec..b6e82d75561f 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -606,6 +606,30 @@ static int uvd_v7_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->uvd.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_uvd(adev, false);
+	} else {
+		amdgpu_asic_set_uvd_clocks(adev, 0, 0);
+		/* shutdown the UVD block */
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+						       AMD_CG_STATE_GATE);
+	}
+
 	if (!amdgpu_sriov_vf(adev))
 		uvd_v7_0_stop(adev);
 	else {
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
index c7d28c169be5..b70c17f0c52e 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
@@ -477,6 +477,31 @@ static int vce_v2_0_hw_init(void *handle)
 
 static int vce_v2_0_hw_fini(void *handle)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->vce.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_vce(adev, false);
+	} else {
+		amdgpu_asic_set_vce_clocks(adev, 0, 0);
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_CG_STATE_GATE);
+	}
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
index 3b82fb289ef6..9de66893ccd6 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
@@ -490,6 +490,29 @@ static int vce_v3_0_hw_fini(void *handle)
 	int r;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->vce.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_vce(adev, false);
+	} else {
+		amdgpu_asic_set_vce_clocks(adev, 0, 0);
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_CG_STATE_GATE);
+	}
+
 	r = vce_v3_0_wait_for_idle(handle);
 	if (r)
 		return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
index 90910d19db12..fec902b800c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
@@ -542,6 +542,29 @@ static int vce_v4_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->vce.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_vce(adev, false);
+	} else {
+		amdgpu_asic_set_vce_clocks(adev, 0, 0);
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_CG_STATE_GATE);
+	}
+
 	if (!amdgpu_sriov_vf(adev)) {
 		/* vce_v4_0_wait_for_idle(handle); */
 		vce_v4_0_stop(adev);