drm/amdgpu: add save restore list cntl gpm and srm firmware support

RLC save/restore list cntl/gpm_mem/srm_mem ucodes are used for CGPG and gfxoff
function.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -774,9 +774,18 @@ struct amdgpu_rlc {
 	u32 starting_offsets_start;
 	u32 reg_list_format_size_bytes;
 	u32 reg_list_size_bytes;
+	u32 reg_list_format_direct_reg_list_length;
+	u32 save_restore_list_cntl_size_bytes;
+	u32 save_restore_list_gpm_size_bytes;
+	u32 save_restore_list_srm_size_bytes;
 
 	u32 *register_list_format;
 	u32 *register_restore;
+	u8 *save_restore_list_cntl;
+	u8 *save_restore_list_gpm;
+	u8 *save_restore_list_srm;
+
+	bool is_rlc_v2_1;
 };
 
 #define AMDGPU_MAX_COMPUTE_QUEUES KGD_MAX_QUEUES
@@ -943,6 +952,12 @@ struct amdgpu_gfx {
 	uint32_t			ce_feature_version;
 	uint32_t			pfp_feature_version;
 	uint32_t			rlc_feature_version;
+	uint32_t			rlc_srlc_fw_version;
+	uint32_t			rlc_srlc_feature_version;
+	uint32_t			rlc_srlg_fw_version;
+	uint32_t			rlc_srlg_feature_version;
+	uint32_t			rlc_srls_fw_version;
+	uint32_t			rlc_srls_feature_version;
 	uint32_t			mec_feature_version;
 	uint32_t			mec2_feature_version;
 	struct amdgpu_ring		gfx_ring[AMDGPU_MAX_GFX_RINGS];

drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c

@@ -215,6 +215,18 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 		fw_info->ver = adev->gfx.rlc_fw_version;
 		fw_info->feature = adev->gfx.rlc_feature_version;
 		break;
+	case AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_CNTL:
+		fw_info->ver = adev->gfx.rlc_srlc_fw_version;
+		fw_info->feature = adev->gfx.rlc_srlc_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_GPM_MEM:
+		fw_info->ver = adev->gfx.rlc_srlg_fw_version;
+		fw_info->feature = adev->gfx.rlc_srlg_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_SRM_MEM:
+		fw_info->ver = adev->gfx.rlc_srls_fw_version;
+		fw_info->feature = adev->gfx.rlc_srls_feature_version;
+		break;
 	case AMDGPU_INFO_FW_GFX_MEC:
 		if (query_fw->index == 0) {
 			fw_info->ver = adev->gfx.mec_fw_version;
@@ -1149,6 +1161,30 @@ static int amdgpu_debugfs_firmware_info(struct seq_file *m, void *data)
 	seq_printf(m, "RLC feature version: %u, firmware version: 0x%08x\n",
 		   fw_info.feature, fw_info.ver);
 
+	/* RLC SAVE RESTORE LIST CNTL */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_CNTL;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "RLC SRLC feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* RLC SAVE RESTORE LIST GPM MEM */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_GPM_MEM;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "RLC SRLG feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* RLC SAVE RESTORE LIST SRM MEM */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_SRM_MEM;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "RLC SRLS feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
 	/* MEC */
 	query_fw.fw_type = AMDGPU_INFO_FW_GFX_MEC;
 	query_fw.index = 0;

drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c

@@ -337,7 +337,10 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 	    (ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC1 &&
 	     ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC2 &&
 	     ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC1_JT &&
-	     ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC2_JT)) {
+	     ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC2_JT &&
+	     ucode->ucode_id != AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL &&
+	     ucode->ucode_id != AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM &&
+	     ucode->ucode_id != AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM)) {
 		ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes);
 
 		memcpy(ucode->kaddr, (void *)((uint8_t *)ucode->fw->data +
@@ -359,6 +362,18 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 					      le32_to_cpu(header->ucode_array_offset_bytes) +
 					      le32_to_cpu(cp_hdr->jt_offset) * 4),
 		       ucode->ucode_size);
+	} else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL) {
+		ucode->ucode_size = adev->gfx.rlc.save_restore_list_cntl_size_bytes;
+		memcpy(ucode->kaddr, adev->gfx.rlc.save_restore_list_cntl,
+		       ucode->ucode_size);
+	} else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM) {
+		ucode->ucode_size = adev->gfx.rlc.save_restore_list_gpm_size_bytes;
+		memcpy(ucode->kaddr, adev->gfx.rlc.save_restore_list_gpm,
+		       ucode->ucode_size);
+	} else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM) {
+		ucode->ucode_size = adev->gfx.rlc.save_restore_list_srm_size_bytes;
+		memcpy(ucode->kaddr, adev->gfx.rlc.save_restore_list_srm,
+		       ucode->ucode_size);
 	}
 
 	return 0;

drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h

@@ -187,6 +187,9 @@ enum AMDGPU_UCODE_ID {
 	AMDGPU_UCODE_ID_CP_MEC2,
 	AMDGPU_UCODE_ID_CP_MEC2_JT,
 	AMDGPU_UCODE_ID_RLC_G,
+	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL,
+	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM,
+	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM,
 	AMDGPU_UCODE_ID_STORAGE,
 	AMDGPU_UCODE_ID_SMC,
 	AMDGPU_UCODE_ID_UVD,

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -41,7 +41,6 @@
 #define GFX9_MEC_HPD_SIZE 2048
 #define RLCG_UCODE_LOADING_START_ADDRESS 0x00002000L
 #define RLC_SAVE_RESTORE_ADDR_STARTING_OFFSET 0x00000000L
-#define GFX9_RLC_FORMAT_DIRECT_REG_LIST_LENGTH 34
 
 #define mmPWR_MISC_CNTL_STATUS					0x0183
 #define mmPWR_MISC_CNTL_STATUS_BASE_IDX				0
@@ -401,6 +400,27 @@ static void gfx_v9_0_free_microcode(struct amdgpu_device *adev)
 	kfree(adev->gfx.rlc.register_list_format);
 }
 
+static void gfx_v9_0_init_rlc_ext_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_1 *rlc_hdr;
+
+	rlc_hdr = (const struct rlc_firmware_header_v2_1 *)adev->gfx.rlc_fw->data;
+	adev->gfx.rlc_srlc_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_ucode_ver);
+	adev->gfx.rlc_srlc_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_feature_ver);
+	adev->gfx.rlc.save_restore_list_cntl_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_cntl_size_bytes);
+	adev->gfx.rlc.save_restore_list_cntl = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_cntl_offset_bytes);
+	adev->gfx.rlc_srlg_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_ucode_ver);
+	adev->gfx.rlc_srlg_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_feature_ver);
+	adev->gfx.rlc.save_restore_list_gpm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_gpm_size_bytes);
+	adev->gfx.rlc.save_restore_list_gpm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_gpm_offset_bytes);
+	adev->gfx.rlc_srls_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_ucode_ver);
+	adev->gfx.rlc_srls_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_feature_ver);
+	adev->gfx.rlc.save_restore_list_srm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_srm_size_bytes);
+	adev->gfx.rlc.save_restore_list_srm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_srm_offset_bytes);
+	adev->gfx.rlc.reg_list_format_direct_reg_list_length =
+			le32_to_cpu(rlc_hdr->reg_list_format_direct_reg_list_length);
+}
+
 static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
@@ -412,6 +432,8 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 	const struct rlc_firmware_header_v2_0 *rlc_hdr;
 	unsigned int *tmp = NULL;
 	unsigned int i = 0;
+	uint16_t version_major;
+	uint16_t version_minor;
 
 	DRM_DEBUG("\n");
 
@@ -468,6 +490,12 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 		goto out;
 	err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
 	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+
+	version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+	version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+	if (version_major == 2 && version_minor == 1)
+		adev->gfx.rlc.is_rlc_v2_1 = true;
+
 	adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
 	adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
 	adev->gfx.rlc.save_and_restore_offset =
@@ -508,6 +536,9 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 	for (i = 0 ; i < (rlc_hdr->reg_list_size_bytes >> 2); i++)
 		adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
 
+	if (adev->gfx.rlc.is_rlc_v2_1)
+		gfx_v9_0_init_rlc_ext_microcode(adev);
+
 	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
 	err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
 	if (err)
@@ -566,6 +597,26 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 		adev->firmware.fw_size +=
 			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
 
+		if (adev->gfx.rlc.is_rlc_v2_1) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.save_restore_list_cntl_size_bytes, PAGE_SIZE);
+
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.save_restore_list_gpm_size_bytes, PAGE_SIZE);
+
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.save_restore_list_srm_size_bytes, PAGE_SIZE);
+		}
+
 		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1];
 		info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1;
 		info->fw = adev->gfx.mec_fw;
@@ -1781,7 +1832,7 @@ static int gfx_v9_0_init_rlc_save_restore_list(struct amdgpu_device *adev)
 
 	/* setup unique_indirect_regs array and indirect_start_offsets array */
 	gfx_v9_0_parse_ind_reg_list(register_list_format,
-				GFX9_RLC_FORMAT_DIRECT_REG_LIST_LENGTH,
+				adev->gfx.rlc.reg_list_format_direct_reg_list_length,
 				adev->gfx.rlc.reg_list_format_size_bytes >> 2,
 				unique_indirect_regs,
 				&unique_indirect_reg_count,

drivers/gpu/drm/amd/amdgpu/psp_v10_0.c

@@ -70,6 +70,15 @@ psp_v10_0_get_fw_type(struct amdgpu_firmware_info *ucode, enum psp_gfx_fw_type *
 	case AMDGPU_UCODE_ID_RLC_G:
 		*type = GFX_FW_TYPE_RLC_G;
 		break;
+	case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL:
+		*type = GFX_FW_TYPE_RLC_RESTORE_LIST_CNTL;
+		break;
+	case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM:
+		*type = GFX_FW_TYPE_RLC_RESTORE_LIST_GPM_MEM;
+		break;
+	case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM:
+		*type = GFX_FW_TYPE_RLC_RESTORE_LIST_SRM_MEM;
+		break;
 	case AMDGPU_UCODE_ID_SMC:
 		*type = GFX_FW_TYPE_SMU;
 		break;

include/uapi/drm/amdgpu_drm.h

@@ -630,6 +630,12 @@ struct drm_amdgpu_cs_chunk_data {
 	#define AMDGPU_INFO_FW_ASD		0x0d
 	/* Subquery id: Query VCN firmware version */
 	#define AMDGPU_INFO_FW_VCN		0x0e
+	/* Subquery id: Query GFX RLC SRLC firmware version */
+	#define AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_CNTL 0x0f
+	/* Subquery id: Query GFX RLC SRLG firmware version */
+	#define AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_GPM_MEM 0x10
+	/* Subquery id: Query GFX RLC SRLS firmware version */
+	#define AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_SRM_MEM 0x11
 /* number of bytes moved for TTM migration */
 #define AMDGPU_INFO_NUM_BYTES_MOVED		0x0f
 /* the used VRAM size */