Merge tag 'amd-drm-fixes-6.4-2023-06-07' of...

Merge tag 'amd-drm-fixes-6.4-2023-06-07' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-6.4-2023-06-07:

amdgpu:
- S0ix fixes
- GPU reset fixes
- SMU13 fixes
- SMU11 fixes
- Misc Display fixes
- Revert RV/RV2/PCO clock counter changes
- Fix Stoney xclk value
- Fix reserved vram debug info

radeon:
- Fix a potential use after free
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230607213740.7723-1-alexander.deucher@amd.com

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

@@ -1092,16 +1092,20 @@ bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev)
 	 * S0ix even though the system is suspending to idle, so return false
 	 * in that case.
 	 */
-	if (!(acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0))
-		dev_warn_once(adev->dev,
+	if (!(acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0)) {
+		dev_err_once(adev->dev,
 			      "Power consumption will be higher as BIOS has not been configured for suspend-to-idle.\n"
 			      "To use suspend-to-idle change the sleep mode in BIOS setup.\n");
+		return false;
+	}
 
 #if !IS_ENABLED(CONFIG_AMD_PMC)
-	dev_warn_once(adev->dev,
+	dev_err_once(adev->dev,
 		      "Power consumption will be higher as the kernel has not been compiled with CONFIG_AMD_PMC.\n");
-#endif /* CONFIG_AMD_PMC */
+	return false;
+#else
 	return true;
+#endif /* CONFIG_AMD_PMC */
 }
 
 #endif /* CONFIG_SUSPEND */

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

@@ -79,9 +79,10 @@ static void amdgpu_bo_user_destroy(struct ttm_buffer_object *tbo)
 static void amdgpu_bo_vm_destroy(struct ttm_buffer_object *tbo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev);
-	struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo);
+	struct amdgpu_bo *shadow_bo = ttm_to_amdgpu_bo(tbo), *bo;
 	struct amdgpu_bo_vm *vmbo;
 
+	bo = shadow_bo->parent;
 	vmbo = to_amdgpu_bo_vm(bo);
 	/* in case amdgpu_device_recover_vram got NULL of bo->parent */
 	if (!list_empty(&vmbo->shadow_list)) {
@@ -694,11 +695,6 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev,
 		return r;
 
 	*vmbo_ptr = to_amdgpu_bo_vm(bo_ptr);
-	INIT_LIST_HEAD(&(*vmbo_ptr)->shadow_list);
-	/* Set destroy callback to amdgpu_bo_vm_destroy after vmbo->shadow_list
-	 * is initialized.
-	 */
-	bo_ptr->tbo.destroy = &amdgpu_bo_vm_destroy;
 	return r;
 }
 
@@ -715,6 +711,8 @@ void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo)
 
 	mutex_lock(&adev->shadow_list_lock);
 	list_add_tail(&vmbo->shadow_list, &adev->shadow_list);
+	vmbo->shadow->parent = amdgpu_bo_ref(&vmbo->bo);
+	vmbo->shadow->tbo.destroy = &amdgpu_bo_vm_destroy;
 	mutex_unlock(&adev->shadow_list_lock);
 }
 

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

@@ -564,7 +564,6 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		return r;
 	}
 
-	(*vmbo)->shadow->parent = amdgpu_bo_ref(bo);
 	amdgpu_bo_add_to_shadow_list(*vmbo);
 
 	return 0;

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

@@ -800,7 +800,7 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man,
 {
 	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct drm_buddy *mm = &mgr->mm;
-	struct drm_buddy_block *block;
+	struct amdgpu_vram_reservation *rsv;
 
 	drm_printf(printer, "  vis usage:%llu\n",
 		   amdgpu_vram_mgr_vis_usage(mgr));
@@ -812,8 +812,9 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man,
 	drm_buddy_print(mm, printer);
 
 	drm_printf(printer, "reserved:\n");
-	list_for_each_entry(block, &mgr->reserved_pages, link)
-		drm_buddy_block_print(mm, block, printer);
+	list_for_each_entry(rsv, &mgr->reserved_pages, blocks)
+		drm_printf(printer, "%#018llx-%#018llx: %llu\n",
+			rsv->start, rsv->start + rsv->size, rsv->size);
 	mutex_unlock(&mgr->lock);
 }
 

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

@@ -149,16 +149,6 @@ MODULE_FIRMWARE("amdgpu/aldebaran_sjt_mec2.bin");
 #define mmGOLDEN_TSC_COUNT_LOWER_Renoir                0x0026
 #define mmGOLDEN_TSC_COUNT_LOWER_Renoir_BASE_IDX       1
 
-#define mmGOLDEN_TSC_COUNT_UPPER_Raven   0x007a
-#define mmGOLDEN_TSC_COUNT_UPPER_Raven_BASE_IDX 0
-#define mmGOLDEN_TSC_COUNT_LOWER_Raven   0x007b
-#define mmGOLDEN_TSC_COUNT_LOWER_Raven_BASE_IDX 0
-
-#define mmGOLDEN_TSC_COUNT_UPPER_Raven2   0x0068
-#define mmGOLDEN_TSC_COUNT_UPPER_Raven2_BASE_IDX 0
-#define mmGOLDEN_TSC_COUNT_LOWER_Raven2   0x0069
-#define mmGOLDEN_TSC_COUNT_LOWER_Raven2_BASE_IDX 0
-
 enum ta_ras_gfx_subblock {
 	/*CPC*/
 	TA_RAS_BLOCK__GFX_CPC_INDEX_START = 0,
@@ -4004,31 +3994,6 @@ static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 		preempt_enable();
 		clock = clock_lo | (clock_hi << 32ULL);
 		break;
-	case IP_VERSION(9, 1, 0):
-	case IP_VERSION(9, 2, 2):
-		preempt_disable();
-		if (adev->rev_id >= 0x8) {
-			clock_hi = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven2);
-			clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven2);
-			hi_check = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven2);
-		} else {
-			clock_hi = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven);
-			clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven);
-			hi_check = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven);
-		}
-		/* The PWR TSC clock frequency is 100MHz, which sets 32-bit carry over
-		* roughly every 42 seconds.
-		*/
-		if (hi_check != clock_hi) {
-			if (adev->rev_id >= 0x8)
-				clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven2);
-			else
-				clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven);
-			clock_hi = hi_check;
-		}
-		preempt_enable();
-		clock = clock_lo | (clock_hi << 32ULL);
-		break;
 	default:
 		amdgpu_gfx_off_ctrl(adev, false);
 		mutex_lock(&adev->gfx.gpu_clock_mutex);

drivers/gpu/drm/amd/amdgpu/soc15.c

@@ -301,10 +301,11 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev)
 	u32 reference_clock = adev->clock.spll.reference_freq;
 
 	if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 0) ||
-	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 1) ||
-	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 0) ||
-	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 1))
+	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 1))
 		return 10000;
+	if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 0) ||
+	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 1))
+		return reference_clock / 4;
 
 	return reference_clock;
 }

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

@@ -542,8 +542,15 @@ static u32 vi_get_xclk(struct amdgpu_device *adev)
 	u32 reference_clock = adev->clock.spll.reference_freq;
 	u32 tmp;
 
-	if (adev->flags & AMD_IS_APU)
-		return reference_clock;
+	if (adev->flags & AMD_IS_APU) {
+		switch (adev->asic_type) {
+		case CHIP_STONEY:
+			/* vbios says 48Mhz, but the actual freq is 100Mhz */
+			return 10000;
+		default:
+			return reference_clock;
+		}
+	}
 
 	tmp = RREG32_SMC(ixCG_CLKPIN_CNTL_2);
 	if (REG_GET_FIELD(tmp, CG_CLKPIN_CNTL_2, MUX_TCLK_TO_XCLK))

drivers/gpu/drm/amd/display/dc/core/dc.c

@@ -1981,6 +1981,9 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 	return result;
 }
 
+static bool commit_minimal_transition_state(struct dc *dc,
+		struct dc_state *transition_base_context);
+
 /**
  * dc_commit_streams - Commit current stream state
  *
@@ -2002,6 +2005,8 @@ enum dc_status dc_commit_streams(struct dc *dc,
 	struct dc_state *context;
 	enum dc_status res = DC_OK;
 	struct dc_validation_set set[MAX_STREAMS] = {0};
+	struct pipe_ctx *pipe;
+	bool handle_exit_odm2to1 = false;
 
 	if (dc->ctx->dce_environment == DCE_ENV_VIRTUAL_HW)
 		return res;
@@ -2026,6 +2031,22 @@ enum dc_status dc_commit_streams(struct dc *dc,
 		}
 	}
 
+	/* Check for case where we are going from odm 2:1 to max
+	 *  pipe scenario.  For these cases, we will call
+	 *  commit_minimal_transition_state() to exit out of odm 2:1
+	 *  first before processing new streams
+	 */
+	if (stream_count == dc->res_pool->pipe_count) {
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+			if (pipe->next_odm_pipe)
+				handle_exit_odm2to1 = true;
+		}
+	}
+
+	if (handle_exit_odm2to1)
+		res = commit_minimal_transition_state(dc, dc->current_state);
+
 	context = dc_create_state(dc);
 	if (!context)
 		goto context_alloc_fail;
@@ -3872,6 +3893,7 @@ static bool commit_minimal_transition_state(struct dc *dc,
 	unsigned int i, j;
 	unsigned int pipe_in_use = 0;
 	bool subvp_in_use = false;
+	bool odm_in_use = false;
 
 	if (!transition_context)
 		return false;
@@ -3900,6 +3922,18 @@ static bool commit_minimal_transition_state(struct dc *dc,
 		}
 	}
 
+	/* If ODM is enabled and we are adding or removing planes from any ODM
+	 * pipe, we must use the minimal transition.
+	 */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream && pipe->next_odm_pipe) {
+			odm_in_use = true;
+			break;
+		}
+	}
+
 	/* When the OS add a new surface if we have been used all of pipes with odm combine
 	 * and mpc split feature, it need use commit_minimal_transition_state to transition safely.
 	 * After OS exit MPO, it will back to use odm and mpc split with all of pipes, we need
@@ -3908,7 +3942,7 @@ static bool commit_minimal_transition_state(struct dc *dc,
 	 * Reduce the scenarios to use dc_commit_state_no_check in the stage of flip. Especially
 	 * enter/exit MPO when DCN still have enough resources.
 	 */
-	if (pipe_in_use != dc->res_pool->pipe_count && !subvp_in_use) {
+	if (pipe_in_use != dc->res_pool->pipe_count && !subvp_in_use && !odm_in_use) {
 		dc_release_state(transition_context);
 		return true;
 	}

drivers/gpu/drm/amd/display/dc/core/dc_resource.c

@@ -1444,6 +1444,26 @@ static int acquire_first_split_pipe(
 			split_pipe->plane_res.mpcc_inst = pool->dpps[i]->inst;
 			split_pipe->pipe_idx = i;
 
+			split_pipe->stream = stream;
+			return i;
+		} else if (split_pipe->prev_odm_pipe &&
+				split_pipe->prev_odm_pipe->plane_state == split_pipe->plane_state) {
+			split_pipe->prev_odm_pipe->next_odm_pipe = split_pipe->next_odm_pipe;
+			if (split_pipe->next_odm_pipe)
+				split_pipe->next_odm_pipe->prev_odm_pipe = split_pipe->prev_odm_pipe;
+
+			if (split_pipe->prev_odm_pipe->plane_state)
+				resource_build_scaling_params(split_pipe->prev_odm_pipe);
+
+			memset(split_pipe, 0, sizeof(*split_pipe));
+			split_pipe->stream_res.tg = pool->timing_generators[i];
+			split_pipe->plane_res.hubp = pool->hubps[i];
+			split_pipe->plane_res.ipp = pool->ipps[i];
+			split_pipe->plane_res.dpp = pool->dpps[i];
+			split_pipe->stream_res.opp = pool->opps[i];
+			split_pipe->plane_res.mpcc_inst = pool->dpps[i]->inst;
+			split_pipe->pipe_idx = i;
+
 			split_pipe->stream = stream;
 			return i;
 		}

drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c

@@ -138,7 +138,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_2_soc = {
 	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
 	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
 	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_sdp_bw_after_urgent = 100.0,
+	.pct_ideal_sdp_bw_after_urgent = 90.0,
 	.pct_ideal_fabric_bw_after_urgent = 67.0,
 	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 20.0,
 	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0, // N/A, for now keep as is until DML implemented

drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c

@@ -2067,33 +2067,94 @@ static int sienna_cichlid_display_disable_memory_clock_switch(struct smu_context
 	return ret;
 }
 
+static void sienna_cichlid_get_override_pcie_settings(struct smu_context *smu,
+						      uint32_t *gen_speed_override,
+						      uint32_t *lane_width_override)
+{
+	struct amdgpu_device *adev = smu->adev;
+
+	*gen_speed_override = 0xff;
+	*lane_width_override = 0xff;
+
+	switch (adev->pdev->device) {
+	case 0x73A0:
+	case 0x73A1:
+	case 0x73A2:
+	case 0x73A3:
+	case 0x73AB:
+	case 0x73AE:
+		/* Bit 7:0: PCIE lane width, 1 to 7 corresponds is x1 to x32 */
+		*lane_width_override = 6;
+		break;
+	case 0x73E0:
+	case 0x73E1:
+	case 0x73E3:
+		*lane_width_override = 4;
+		break;
+	case 0x7420:
+	case 0x7421:
+	case 0x7422:
+	case 0x7423:
+	case 0x7424:
+		*lane_width_override = 3;
+		break;
+	default:
+		break;
+	}
+}
+
+#define MAX(a, b)	((a) > (b) ? (a) : (b))
+
 static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu,
 					 uint32_t pcie_gen_cap,
 					 uint32_t pcie_width_cap)
 {
 	struct smu_11_0_dpm_context *dpm_context = smu->smu_dpm.dpm_context;
-
-	uint32_t smu_pcie_arg;
+	struct smu_11_0_pcie_table *pcie_table = &dpm_context->dpm_tables.pcie_table;
+	uint32_t gen_speed_override, lane_width_override;
 	uint8_t *table_member1, *table_member2;
+	uint32_t min_gen_speed, max_gen_speed;
+	uint32_t min_lane_width, max_lane_width;
+	uint32_t smu_pcie_arg;
 	int ret, i;
 
 	GET_PPTABLE_MEMBER(PcieGenSpeed, &table_member1);
 	GET_PPTABLE_MEMBER(PcieLaneCount, &table_member2);
 
-	/* lclk dpm table setup */
-	for (i = 0; i < MAX_PCIE_CONF; i++) {
-		dpm_context->dpm_tables.pcie_table.pcie_gen[i] = table_member1[i];
-		dpm_context->dpm_tables.pcie_table.pcie_lane[i] = table_member2[i];
+	sienna_cichlid_get_override_pcie_settings(smu,
+						  &gen_speed_override,
+						  &lane_width_override);
+
+	/* PCIE gen speed override */
+	if (gen_speed_override != 0xff) {
+		min_gen_speed = MIN(pcie_gen_cap, gen_speed_override);
+		max_gen_speed = MIN(pcie_gen_cap, gen_speed_override);
+	} else {
+		min_gen_speed = MAX(0, table_member1[0]);
+		max_gen_speed = MIN(pcie_gen_cap, table_member1[1]);
+		min_gen_speed = min_gen_speed > max_gen_speed ?
+				max_gen_speed : min_gen_speed;
 	}
+	pcie_table->pcie_gen[0] = min_gen_speed;
+	pcie_table->pcie_gen[1] = max_gen_speed;
+
+	/* PCIE lane width override */
+	if (lane_width_override != 0xff) {
+		min_lane_width = MIN(pcie_width_cap, lane_width_override);
+		max_lane_width = MIN(pcie_width_cap, lane_width_override);
+	} else {
+		min_lane_width = MAX(1, table_member2[0]);
+		max_lane_width = MIN(pcie_width_cap, table_member2[1]);
+		min_lane_width = min_lane_width > max_lane_width ?
+				 max_lane_width : min_lane_width;
+	}
+	pcie_table->pcie_lane[0] = min_lane_width;
+	pcie_table->pcie_lane[1] = max_lane_width;
 
 	for (i = 0; i < NUM_LINK_LEVELS; i++) {
-		smu_pcie_arg = (i << 16) |
-			((table_member1[i] <= pcie_gen_cap) ?
-			 (table_member1[i] << 8) :
-			 (pcie_gen_cap << 8)) |
-			((table_member2[i] <= pcie_width_cap) ?
-			 table_member2[i] :
-			 pcie_width_cap);
+		smu_pcie_arg = (i << 16 |
+				pcie_table->pcie_gen[i] << 8 |
+				pcie_table->pcie_lane[i]);
 
 		ret = smu_cmn_send_smc_msg_with_param(smu,
 				SMU_MSG_OverridePcieParameters,
@@ -2101,11 +2162,6 @@ static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu,
 				NULL);
 		if (ret)
 			return ret;
-
-		if (table_member1[i] > pcie_gen_cap)
-			dpm_context->dpm_tables.pcie_table.pcie_gen[i] = pcie_gen_cap;
-		if (table_member2[i] > pcie_width_cap)
-			dpm_context->dpm_tables.pcie_table.pcie_lane[i] = pcie_width_cap;
 	}
 
 	return 0;

drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c

@@ -573,11 +573,11 @@ int smu_v13_0_init_power(struct smu_context *smu)
 	if (smu_power->power_context || smu_power->power_context_size != 0)
 		return -EINVAL;
 
-	smu_power->power_context = kzalloc(sizeof(struct smu_13_0_dpm_context),
+	smu_power->power_context = kzalloc(sizeof(struct smu_13_0_power_context),
 					   GFP_KERNEL);
 	if (!smu_power->power_context)
 		return -ENOMEM;
-	smu_power->power_context_size = sizeof(struct smu_13_0_dpm_context);
+	smu_power->power_context_size = sizeof(struct smu_13_0_power_context);
 
 	return 0;
 }

drivers/gpu/drm/radeon/radeon_gem.c

@@ -459,7 +459,6 @@ int radeon_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 	struct radeon_device *rdev = dev->dev_private;
 	struct drm_radeon_gem_set_domain *args = data;
 	struct drm_gem_object *gobj;
-	struct radeon_bo *robj;
 	int r;
 
 	/* for now if someone requests domain CPU -
@@ -472,13 +471,12 @@ int radeon_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 		up_read(&rdev->exclusive_lock);
 		return -ENOENT;
 	}
-	robj = gem_to_radeon_bo(gobj);
 
 	r = radeon_gem_set_domain(gobj, args->read_domains, args->write_domain);
 
 	drm_gem_object_put(gobj);
 	up_read(&rdev->exclusive_lock);
-	r = radeon_gem_handle_lockup(robj->rdev, r);
+	r = radeon_gem_handle_lockup(rdev, r);
 	return r;
 }