Merge branch 'drm-fixes-5.0' of git://people.freedesktop.org/~agd5f/linux into drm-fixes

- Powerplay fixes
- Virtual display pinning fixes
- Golden register updates for vega
- Pitch and gem size validation fixes
- Fix for error case in sr-iov init
- Disable page tables in system memory on RV due to issues with IOMMU
  reported on some platforms

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190109204336.3315-1-alexander.deucher@amd.com

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

@@ -1701,8 +1701,10 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 		amdgpu_xgmi_add_device(adev);
 	amdgpu_amdkfd_device_init(adev);
 
-	if (amdgpu_sriov_vf(adev))
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_init_data_exchange(adev);
 		amdgpu_virt_release_full_gpu(adev, true);
+	}
 
 	return 0;
 }
@@ -2632,9 +2634,6 @@ fence_driver_init:
 		goto failed;
 	}
 
-	if (amdgpu_sriov_vf(adev))
-		amdgpu_virt_init_data_exchange(adev);
-
 	amdgpu_fbdev_init(adev);
 
 	r = amdgpu_pm_sysfs_init(adev);
@@ -2798,7 +2797,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
 			struct drm_framebuffer *fb = crtc->primary->fb;
 			struct amdgpu_bo *robj;
 
-			if (amdgpu_crtc->cursor_bo) {
+			if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) {
 				struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
 				r = amdgpu_bo_reserve(aobj, true);
 				if (r == 0) {
@@ -2906,7 +2905,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 			struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 
-			if (amdgpu_crtc->cursor_bo) {
+			if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) {
 				struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
 				r = amdgpu_bo_reserve(aobj, true);
 				if (r == 0) {
@@ -3226,6 +3225,7 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 	r = amdgpu_ib_ring_tests(adev);
 
 error:
+	amdgpu_virt_init_data_exchange(adev);
 	amdgpu_virt_release_full_gpu(adev, true);
 	if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) {
 		atomic_inc(&adev->vram_lost_counter);

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

@@ -188,10 +188,12 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc,
 		goto cleanup;
 	}
 
-	r = amdgpu_bo_pin(new_abo, amdgpu_display_supported_domains(adev));
-	if (unlikely(r != 0)) {
-		DRM_ERROR("failed to pin new abo buffer before flip\n");
-		goto unreserve;
+	if (!adev->enable_virtual_display) {
+		r = amdgpu_bo_pin(new_abo, amdgpu_display_supported_domains(adev));
+		if (unlikely(r != 0)) {
+			DRM_ERROR("failed to pin new abo buffer before flip\n");
+			goto unreserve;
+		}
 	}
 
 	r = amdgpu_ttm_alloc_gart(&new_abo->tbo);
@@ -211,7 +213,8 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc,
 	amdgpu_bo_get_tiling_flags(new_abo, &tiling_flags);
 	amdgpu_bo_unreserve(new_abo);
 
-	work->base = amdgpu_bo_gpu_offset(new_abo);
+	if (!adev->enable_virtual_display)
+		work->base = amdgpu_bo_gpu_offset(new_abo);
 	work->target_vblank = target - (uint32_t)drm_crtc_vblank_count(crtc) +
 		amdgpu_get_vblank_counter_kms(dev, work->crtc_id);
 
@@ -242,9 +245,10 @@ pflip_cleanup:
 		goto cleanup;
 	}
 unpin:
-	if (unlikely(amdgpu_bo_unpin(new_abo) != 0)) {
-		DRM_ERROR("failed to unpin new abo in error path\n");
-	}
+	if (!adev->enable_virtual_display)
+		if (unlikely(amdgpu_bo_unpin(new_abo) != 0))
+			DRM_ERROR("failed to unpin new abo in error path\n");
+
 unreserve:
 	amdgpu_bo_unreserve(new_abo);
 
@@ -527,6 +531,17 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 	struct drm_gem_object *obj;
 	struct amdgpu_framebuffer *amdgpu_fb;
 	int ret;
+	int height;
+	struct amdgpu_device *adev = dev->dev_private;
+	int cpp = drm_format_plane_cpp(mode_cmd->pixel_format, 0);
+	int pitch = mode_cmd->pitches[0] / cpp;
+
+	pitch = amdgpu_align_pitch(adev, pitch, cpp, false);
+	if (mode_cmd->pitches[0] != pitch) {
+		DRM_DEBUG_KMS("Invalid pitch: expecting %d but got %d\n",
+			      pitch, mode_cmd->pitches[0]);
+		return ERR_PTR(-EINVAL);
+	}
 
 	obj = drm_gem_object_lookup(file_priv, mode_cmd->handles[0]);
 	if (obj ==  NULL) {
@@ -541,6 +556,13 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 		return ERR_PTR(-EINVAL);
 	}
 
+	height = ALIGN(mode_cmd->height, 8);
+	if (obj->size < pitch * height) {
+		DRM_DEBUG_KMS("Invalid GEM size: expecting >= %d but got %zu\n",
+			      pitch * height, obj->size);
+		return ERR_PTR(-EINVAL);
+	}
+
 	amdgpu_fb = kzalloc(sizeof(*amdgpu_fb), GFP_KERNEL);
 	if (amdgpu_fb == NULL) {
 		drm_gem_object_put_unlocked(obj);

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

@@ -2008,6 +2008,7 @@ void amdgpu_pm_print_power_states(struct amdgpu_device *adev)
 
 int amdgpu_pm_sysfs_init(struct amdgpu_device *adev)
 {
+	struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle;
 	int ret;
 
 	if (adev->pm.sysfs_initialized)
@@ -2091,12 +2092,14 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev)
 				"pp_power_profile_mode\n");
 		return ret;
 	}
-	ret = device_create_file(adev->dev,
-			&dev_attr_pp_od_clk_voltage);
-	if (ret) {
-		DRM_ERROR("failed to create device file	"
-				"pp_od_clk_voltage\n");
-		return ret;
+	if (hwmgr->od_enabled) {
+		ret = device_create_file(adev->dev,
+				&dev_attr_pp_od_clk_voltage);
+		if (ret) {
+			DRM_ERROR("failed to create device file	"
+					"pp_od_clk_voltage\n");
+			return ret;
+		}
 	}
 	ret = device_create_file(adev->dev,
 			&dev_attr_gpu_busy_percent);
@@ -2118,6 +2121,8 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev)
 
 void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev)
 {
+	struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle;
+
 	if (adev->pm.dpm_enabled == 0)
 		return;
 
@@ -2138,8 +2143,9 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev)
 	device_remove_file(adev->dev, &dev_attr_pp_mclk_od);
 	device_remove_file(adev->dev,
 			&dev_attr_pp_power_profile_mode);
-	device_remove_file(adev->dev,
-			&dev_attr_pp_od_clk_voltage);
+	if (hwmgr->od_enabled)
+		device_remove_file(adev->dev,
+				&dev_attr_pp_od_clk_voltage);
 	device_remove_file(adev->dev, &dev_attr_gpu_busy_percent);
 }
 

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

@@ -847,9 +847,6 @@ static void amdgpu_vm_bo_param(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	bp->size = amdgpu_vm_bo_size(adev, level);
 	bp->byte_align = AMDGPU_GPU_PAGE_SIZE;
 	bp->domain = AMDGPU_GEM_DOMAIN_VRAM;
-	if (bp->size <= PAGE_SIZE && adev->asic_type >= CHIP_VEGA10 &&
-	    adev->flags & AMD_IS_APU)
-		bp->domain |= AMDGPU_GEM_DOMAIN_GTT;
 	bp->domain = amdgpu_bo_get_preferred_pin_domain(adev, bp->domain);
 	bp->flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
 		AMDGPU_GEM_CREATE_CPU_GTT_USWC;

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

@@ -167,19 +167,6 @@ static void dce_virtual_crtc_disable(struct drm_crtc *crtc)
 	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 
 	dce_virtual_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
-	if (crtc->primary->fb) {
-		int r;
-		struct amdgpu_bo *abo;
-
-		abo = gem_to_amdgpu_bo(crtc->primary->fb->obj[0]);
-		r = amdgpu_bo_reserve(abo, true);
-		if (unlikely(r))
-			DRM_ERROR("failed to reserve abo before unpin\n");
-		else {
-			amdgpu_bo_unpin(abo);
-			amdgpu_bo_unreserve(abo);
-		}
-	}
 
 	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
 	amdgpu_crtc->encoder = NULL;
@@ -692,7 +679,9 @@ static int dce_virtual_pageflip(struct amdgpu_device *adev,
 	spin_unlock_irqrestore(&adev->ddev->event_lock, flags);
 
 	drm_crtc_vblank_put(&amdgpu_crtc->base);
-	schedule_work(&works->unpin_work);
+	amdgpu_bo_unref(&works->old_abo);
+	kfree(works->shared);
+	kfree(works);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

@@ -4233,7 +4233,6 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev)
 	u32 tmp;
 	u32 rb_bufsz;
 	u64 rb_addr, rptr_addr, wptr_gpu_addr;
-	int r;
 
 	/* Set the write pointer delay */
 	WREG32(mmCP_RB_WPTR_DELAY, 0);
@@ -4278,9 +4277,8 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev)
 	amdgpu_ring_clear_ring(ring);
 	gfx_v8_0_cp_gfx_start(adev);
 	ring->sched.ready = true;
-	r = amdgpu_ring_test_helper(ring);
 
-	return r;
+	return 0;
 }
 
 static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
@@ -4369,10 +4367,9 @@ static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
 		amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr));
 	}
 
-	r = amdgpu_ring_test_helper(kiq_ring);
-	if (r)
-		DRM_ERROR("KCQ enable failed\n");
-	return r;
+	amdgpu_ring_commit(kiq_ring);
+
+	return 0;
 }
 
 static int gfx_v8_0_deactivate_hqd(struct amdgpu_device *adev, u32 req)
@@ -4709,18 +4706,34 @@ static int gfx_v8_0_kcq_resume(struct amdgpu_device *adev)
 	if (r)
 		goto done;
 
-	/* Test KCQs - reversing the order of rings seems to fix ring test failure
-	 * after GPU reset
-	 */
-	for (i = adev->gfx.num_compute_rings - 1; i >= 0; i--) {
-		ring = &adev->gfx.compute_ring[i];
-		r = amdgpu_ring_test_helper(ring);
-	}
-
 done:
 	return r;
 }
 
+static int gfx_v8_0_cp_test_all_rings(struct amdgpu_device *adev)
+{
+	int r, i;
+	struct amdgpu_ring *ring;
+
+	/* collect all the ring_tests here, gfx, kiq, compute */
+	ring = &adev->gfx.gfx_ring[0];
+	r = amdgpu_ring_test_helper(ring);
+	if (r)
+		return r;
+
+	ring = &adev->gfx.kiq.ring;
+	r = amdgpu_ring_test_helper(ring);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i];
+		amdgpu_ring_test_helper(ring);
+	}
+
+	return 0;
+}
+
 static int gfx_v8_0_cp_resume(struct amdgpu_device *adev)
 {
 	int r;
@@ -4739,6 +4752,11 @@ static int gfx_v8_0_cp_resume(struct amdgpu_device *adev)
 	r = gfx_v8_0_kcq_resume(adev);
 	if (r)
 		return r;
+
+	r = gfx_v8_0_cp_test_all_rings(adev);
+	if (r)
+		return r;
+
 	gfx_v8_0_enable_gui_idle_interrupt(adev, true);
 
 	return 0;
@@ -5086,6 +5104,8 @@ static int gfx_v8_0_post_soft_reset(void *handle)
 	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_GFX))
 		gfx_v8_0_cp_gfx_resume(adev);
 
+	gfx_v8_0_cp_test_all_rings(adev);
+
 	adev->gfx.rlc.funcs->start(adev);
 
 	return 0;

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

@@ -113,7 +113,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_0[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x4a2c0e68),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0xb5d3f197),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_CACHE_INVALIDATION, 0x3fff3af3, 0x19200000),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000003ff)
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000003ff),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000)
 };
 
 static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] =
@@ -135,10 +138,7 @@ static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmRMI_UTCL1_CNTL2, 0x00030000, 0x00020000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0x0000000f, 0x01000107),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x00001800, 0x00000800),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000)
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080)
 };
 
 static const struct soc15_reg_golden golden_settings_gc_9_0_vg20[] =
@@ -3587,6 +3587,8 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 {
 	uint32_t data, def;
 
+	amdgpu_gfx_rlc_enter_safe_mode(adev);
+
 	/* It is disabled by HW by default */
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
 		/* 1 - RLC_CGTT_MGCG_OVERRIDE */
@@ -3651,6 +3653,8 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 			WREG32_SOC15(GC, 0, mmCP_MEM_SLP_CNTL, data);
 		}
 	}
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev);
 }
 
 static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,

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

@@ -174,7 +174,7 @@ static int xgpu_ai_send_access_requests(struct amdgpu_device *adev,
 			return r;
 		}
 		/* Retrieve checksum from mailbox2 */
-		if (req == IDH_REQ_GPU_INIT_ACCESS) {
+		if (req == IDH_REQ_GPU_INIT_ACCESS || req == IDH_REQ_GPU_RESET_ACCESS) {
 			adev->virt.fw_reserve.checksum_key =
 				RREG32_NO_KIQ(SOC15_REG_OFFSET(NBIO, 0,
 					mmBIF_BX_PF0_MAILBOX_MSGBUF_RCV_DW2));

drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c

@@ -78,7 +78,6 @@ static const struct soc15_reg_golden golden_settings_sdma_4[] = {
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC1_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000),
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_PAGE, 0x000003ff, 0x000003c0),
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_WATERMK, 0xfc000000, 0x00000000),
-	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CLK_CTRL, 0xffffffff, 0x3f000100),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GFX_IB_CNTL, 0x800f0100, 0x00000100),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GFX_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000),
@@ -96,6 +95,7 @@ static const struct soc15_reg_golden golden_settings_sdma_4[] = {
 static const struct soc15_reg_golden golden_settings_sdma_vg10[] = {
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0018773f, 0x00104002),
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104002),
+	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831d07),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0018773f, 0x00104002),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104002)
 };
@@ -103,6 +103,7 @@ static const struct soc15_reg_golden golden_settings_sdma_vg10[] = {
 static const struct soc15_reg_golden golden_settings_sdma_vg12[] = {
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0018773f, 0x00104001),
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104001),
+	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831d07),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0018773f, 0x00104001),
 	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104001)
 };