Merge tag 'amd-drm-next-6.1-2022-09-30' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.1-2022-09-30:

amdgpu:
- RLC FW code cleanup
- RLC fixes for GC 11.x
- SMU 13.x fixes
- CP FW code cleanup
- SDMA FW code cleanup
- GC 11.x fixes
- DCN 3.2.x fixes
- DCN 3.1.4 fixes
- Misc fixes
- RAS fixes
- SR-IOV fixes
- VCN 4.x fixes

amdkfd:
- GC 11.x fixes
- Xnack fixes
- UBSAN warning fix

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220930162012.5823-1-alexander.deucher@amd.com

82 files changed, 2116 insertions, 1987 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index 3da27436922c..b14800ac179e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -1053,6 +1053,10 @@ bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev)
 {
 	if (adev->flags & AMD_IS_APU)
 		return false;
+
+	if (amdgpu_sriov_vf(adev))
+		return false;
+
 	return pm_suspend_target_state != PM_SUSPEND_TO_IDLE;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index c7b1a2dfde13..9e98f3866edc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -756,11 +756,7 @@ void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, bo
 {
 	struct ras_err_data err_data = {0, 0, 0, NULL};
 
-	/* CPU MCA will handle page retirement if connected_to_cpu is 1 */
-	if (!adev->gmc.xgmi.connected_to_cpu)
-		amdgpu_umc_poison_handler(adev, &err_data, reset);
-	else if (reset)
-		amdgpu_amdkfd_gpu_reset(adev);
+	amdgpu_umc_poison_handler(adev, &err_data, reset);
 }
 
 bool amdgpu_amdkfd_ras_query_utcl2_poison_status(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c04ea7f1e819..ab8f970b2849 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -3154,7 +3154,8 @@ static int amdgpu_device_ip_resume_phase1(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 && amdgpu_sriov_vf(adev))) {
 
 			r = adev->ip_blocks[i].version->funcs->resume(adev);
 			if (r) {
@@ -4069,12 +4070,20 @@ static void amdgpu_device_evict_resources(struct amdgpu_device *adev)
 int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
+	int r = 0;
 
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
 	adev->in_suspend = true;
 
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_fini_data_exchange(adev);
+		r = amdgpu_virt_request_full_gpu(adev, false);
+		if (r)
+			return r;
+	}
+
 	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D3))
 		DRM_WARN("smart shift update failed\n");
 
@@ -4098,6 +4107,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 
 	amdgpu_device_ip_suspend_phase2(adev);
 
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_release_full_gpu(adev, false);
+
 	return 0;
 }
 
@@ -4116,6 +4128,12 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r = 0;
 
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_request_full_gpu(adev, true);
+		if (r)
+			return r;
+	}
+
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
@@ -4130,6 +4148,13 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 	}
 
 	r = amdgpu_device_ip_resume(adev);
+
+	/* no matter what r is, always need to properly release full GPU */
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_init_data_exchange(adev);
+		amdgpu_virt_release_full_gpu(adev, true);
+	}
+
 	if (r) {
 		dev_err(adev->dev, "amdgpu_device_ip_resume failed (%d).\n", r);
 		return r;
@@ -5576,9 +5601,9 @@ bool amdgpu_device_is_peer_accessible(struct amdgpu_device *adev,
 		~*peer_adev->dev->dma_mask : ~((1ULL << 32) - 1);
 	resource_size_t aper_limit =
 		adev->gmc.aper_base + adev->gmc.aper_size - 1;
-	bool p2p_access = !adev->gmc.xgmi.connected_to_cpu &&
-			  !(pci_p2pdma_distance_many(adev->pdev,
-					&peer_adev->dev, 1, true) < 0);
+	bool p2p_access =
+		!adev->gmc.xgmi.connected_to_cpu &&
+		!(pci_p2pdma_distance(adev->pdev, peer_adev->dev, false) < 0);
 
 	return pcie_p2p && p2p_access && (adev->gmc.visible_vram_size &&
 		adev->gmc.real_vram_size == adev->gmc.visible_vram_size &&
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 9fa2a5ceb77d..3993e6134914 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -229,7 +229,7 @@ static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, ui
 		return r;
 	}
 
-	memcpy((u8 *)binary, (u8 *)fw->data, adev->mman.discovery_tmr_size);
+	memcpy((u8 *)binary, (u8 *)fw->data, fw->size);
 	release_firmware(fw);
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 782cbca37538..7bd8e33b14be 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -58,7 +58,7 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	int r;
 
-	if (pci_p2pdma_distance_many(adev->pdev, &attach->dev, 1, true) < 0)
+	if (pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
 		attach->peer2peer = false;
 
 	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index 8adeb7469f1e..d0d99ed607dd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -400,7 +400,6 @@ unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring)
 	/* We are not protected by ring lock when reading the last sequence
 	 * but it's ok to report slightly wrong fence count here.
 	 */
-	amdgpu_fence_process(ring);
 	emitted = 0x100000000ull;
 	emitted -= atomic_read(&ring->fence_drv.last_seq);
 	emitted += READ_ONCE(ring->fence_drv.sync_seq);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index ceb91469958a..9546adc8a76f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -23,6 +23,7 @@
  *
  */
 
+#include <linux/firmware.h>
 #include "amdgpu.h"
 #include "amdgpu_gfx.h"
 #include "amdgpu_rlc.h"
@@ -865,3 +866,142 @@ int amdgpu_gfx_get_num_kcq(struct amdgpu_device *adev)
 	}
 	return amdgpu_num_kcq;
 }
+
+void amdgpu_gfx_cp_init_microcode(struct amdgpu_device *adev,
+				  uint32_t ucode_id)
+{
+	const struct gfx_firmware_header_v1_0 *cp_hdr;
+	const struct gfx_firmware_header_v2_0 *cp_hdr_v2_0;
+	struct amdgpu_firmware_info *info = NULL;
+	const struct firmware *ucode_fw;
+	unsigned int fw_size;
+
+	switch (ucode_id) {
+	case AMDGPU_UCODE_ID_CP_PFP:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.pfp_fw->data;
+		adev->gfx.pfp_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.pfp_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.pfp_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.pfp_fw->data;
+		adev->gfx.pfp_fw_version =
+			le32_to_cpu(cp_hdr_v2_0->header.ucode_version);
+		adev->gfx.pfp_feature_version =
+			le32_to_cpu(cp_hdr_v2_0->ucode_feature_version);
+		ucode_fw = adev->gfx.pfp_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.pfp_fw->data;
+		ucode_fw = adev->gfx.pfp_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->data_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_ME:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.me_fw->data;
+		adev->gfx.me_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.me_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.me_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.me_fw->data;
+		adev->gfx.me_fw_version =
+			le32_to_cpu(cp_hdr_v2_0->header.ucode_version);
+		adev->gfx.me_feature_version =
+			le32_to_cpu(cp_hdr_v2_0->ucode_feature_version);
+		ucode_fw = adev->gfx.me_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.me_fw->data;
+		ucode_fw = adev->gfx.me_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->data_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_CE:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.ce_fw->data;
+		adev->gfx.ce_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.ce_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.ce_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC1:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec_fw->data;
+		adev->gfx.mec_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.mec_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes) -
+			  le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC1_JT:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec_fw->data;
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC2:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec2_fw->data;
+		adev->gfx.mec2_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.mec2_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.mec2_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes) -
+			  le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC2_JT:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec2_fw->data;
+		ucode_fw = adev->gfx.mec2_fw;
+		fw_size = le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.mec_fw->data;
+		adev->gfx.mec_fw_version =
+			le32_to_cpu(cp_hdr_v2_0->header.ucode_version);
+		adev->gfx.mec_feature_version =
+			le32_to_cpu(cp_hdr_v2_0->ucode_feature_version);
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.mec_fw->data;
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->data_size_bytes);
+		break;
+	default:
+		break;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {