drm/amdgpu/vcn3.0: add software ring share memory support

Support software ring share memory with vcn firmware.
Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c

@@ -170,6 +170,7 @@ static int vcn_v3_0_sw_init(void *handle)
 	}
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		volatile struct amdgpu_fw_shared *fw_shared;
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
@@ -234,6 +235,10 @@ static int vcn_v3_0_sw_init(void *handle)
 			if (r)
 				return r;
 		}
+
+		fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
+		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_SW_RING_FLAG);
+		fw_shared->sw_ring.is_enabled = cpu_to_le32(DEC_SW_RING_ENABLED);
 	}
 
 	if (amdgpu_sriov_vf(adev)) {
@@ -257,7 +262,17 @@ static int vcn_v3_0_sw_init(void *handle)
 static int vcn_v3_0_sw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int r;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		volatile struct amdgpu_fw_shared *fw_shared;
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+		fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
+		fw_shared->present_flag_0 = 0;
+		fw_shared->sw_ring.is_enabled = false;
+	}
 
 	if (amdgpu_sriov_vf(adev))
 		amdgpu_virt_free_mm_table(adev);
@@ -467,6 +482,15 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
 		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
 	WREG32_SOC15(VCN, inst, mmUVD_VCPU_CACHE_OFFSET2, 0);
 	WREG32_SOC15(VCN, inst, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].fw_shared_gpu_addr));
+	WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].fw_shared_gpu_addr));
+	WREG32_SOC15(VCN, inst, mmUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(VCN, inst, mmUVD_VCPU_NONCACHE_SIZE0,
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)));
 }
 
 static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
@@ -549,13 +573,16 @@ static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 
 	/* non-cache window */
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
-			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW), 0, 0, indirect);
+			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
-			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH), 0, 0, indirect);
+			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
-			VCN, inst_idx, mmUVD_VCPU_NONCACHE_SIZE0), 0, 0, indirect);
+			VCN, inst_idx, mmUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)), 0, indirect);
 }
 
 static void vcn_v3_0_disable_static_power_gating(struct amdgpu_device *adev, int inst)