- various code cleanups

- enhancements for NPU and MRT support
 -----BEGIN PGP SIGNATURE-----
 
 iQFLBAABCAA1FiEEz9DmVLy+XdLQMNIiz8wpfG0KEgsFAmXpksIXHGwuc3RhY2hA
 cGVuZ3V0cm9uaXguZGUACgkQz8wpfG0KEgterAgAnvMM2QRG0qUXHmUJAqwQorU0
 tErsfxsGrSnsQU2sCqN+D1GUqbIwvNUXX3xvj8r7LKCxkFzC+c7/rcKOHUHDsgOU
 WvsDkJOJkx7In2NsiPSYOzUiyRpMwKa2MdFq6GdLB6ZT63GomK+rwL5/9+b23z8P
 NbvI0zNMbZnWDpGnHcdTWCmkVd/sLtAnOa7MMqwR9r2RzdsP9ssDpSFXDPfViVxp
 c8E6a11mhW+rI3KDRvGpQDRhd+7pDYS9+75aeAsDOhhiJCNVEUGnIyL71fb1FzDJ
 oFypU2lNKEdLPUcdSOJSyhC98yupcmQ1j9lWzfqcGcpt8kfl1MlWTwj98hiRfw==
 =3Frs
 -----END PGP SIGNATURE-----

Merge tag 'drm-etnaviv-next-2024-03-07' of https://git.pengutronix.de/git/lst/linux into drm-next

- various code cleanups
- enhancements for NPU and MRT support

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Lucas Stach <l.stach@pengutronix.de>
Link: https://patchwork.freedesktop.org/patch/msgid/72a783cd98d60f6ebb43b90a6b453eea87224409.camel@pengutronix.de

drivers/gpu/drm/etnaviv/etnaviv_cmd_parser.c

@@ -54,6 +54,7 @@ static const struct {
 	ST(0x1480, 8),
 	ST(0x1500, 8),
 	ST(0x1520, 8),
+	ST(0x1540, 8),
 	ST(0x1608, 1),
 	ST(0x1610, 1),
 	ST(0x1658, 1),

drivers/gpu/drm/etnaviv/etnaviv_drv.c

@@ -29,6 +29,17 @@
  * DRM operations:
  */
 
+static struct device_node *etnaviv_of_first_available_node(void)
+{
+	struct device_node *np;
+
+	for_each_compatible_node(np, NULL, "vivante,gc") {
+		if (of_device_is_available(np))
+			return np;
+	}
+
+	return NULL;
+}
 
 static void load_gpu(struct drm_device *dev)
 {
@@ -79,7 +90,7 @@ static int etnaviv_open(struct drm_device *dev, struct drm_file *file)
 			drm_sched_entity_init(&ctx->sched_entity[i],
 					      DRM_SCHED_PRIORITY_NORMAL, &sched,
 					      1, NULL);
-			}
+		}
 	}
 
 	file->driver_priv = ctx;
@@ -233,11 +244,11 @@ static int show_each_gpu(struct seq_file *m, void *arg)
 }
 
 static struct drm_info_list etnaviv_debugfs_list[] = {
-		{"gpu", show_each_gpu, 0, etnaviv_gpu_debugfs},
-		{"gem", show_unlocked, 0, etnaviv_gem_show},
-		{ "mm", show_unlocked, 0, etnaviv_mm_show },
-		{"mmu", show_each_gpu, 0, etnaviv_mmu_show},
-		{"ring", show_each_gpu, 0, etnaviv_ring_show},
+	{"gpu", show_each_gpu, 0, etnaviv_gpu_debugfs},
+	{"gem", show_unlocked, 0, etnaviv_gem_show},
+	{ "mm", show_unlocked, 0, etnaviv_mm_show },
+	{"mmu", show_each_gpu, 0, etnaviv_mmu_show},
+	{"ring", show_each_gpu, 0, etnaviv_ring_show},
 };
 
 static void etnaviv_debugfs_init(struct drm_minor *minor)
@@ -494,7 +505,7 @@ static const struct drm_driver etnaviv_drm_driver = {
 	.desc               = "etnaviv DRM",
 	.date               = "20151214",
 	.major              = 1,
-	.minor              = 3,
+	.minor              = 4,
 };
 
 /*
@@ -597,9 +608,6 @@ static int etnaviv_pdev_probe(struct platform_device *pdev)
 			if (!of_device_is_available(core_node))
 				continue;
 
-			if (!first_node)
-				first_node = core_node;
-
 			drm_of_component_match_add(&pdev->dev, &match,
 						   component_compare_of, core_node);
 		}
@@ -634,8 +642,11 @@ static int etnaviv_pdev_probe(struct platform_device *pdev)
 	 * device as the GPU we found. This assumes that all Vivante
 	 * GPUs in the system share the same DMA constraints.
 	 */
-	if (first_node)
+	first_node = etnaviv_of_first_available_node();
+	if (first_node) {
 		of_dma_configure(&pdev->dev, first_node, true);
+		of_node_put(first_node);
+	}
 
 	return component_master_add_with_match(dev, &etnaviv_master_ops, match);
 }
@@ -653,11 +664,43 @@ static struct platform_driver etnaviv_platform_driver = {
 	},
 };
 
+static int etnaviv_create_platform_device(const char *name,
+					  struct platform_device **ppdev)
+{
+	struct platform_device *pdev;
+	int ret;
+
+	pdev = platform_device_alloc(name, PLATFORM_DEVID_NONE);
+	if (!pdev)
+		return -ENOMEM;
+
+	ret = platform_device_add(pdev);
+	if (ret) {
+		platform_device_put(pdev);
+		return ret;
+	}
+
+	*ppdev = pdev;
+
+	return 0;
+}
+
+static void etnaviv_destroy_platform_device(struct platform_device **ppdev)
+{
+	struct platform_device *pdev = *ppdev;
+
+	if (!pdev)
+		return;
+
+	platform_device_unregister(pdev);
+
+	*ppdev = NULL;
+}
+
 static struct platform_device *etnaviv_drm;
 
 static int __init etnaviv_init(void)
 {
-	struct platform_device *pdev;
 	int ret;
 	struct device_node *np;
 
@@ -675,27 +718,13 @@ static int __init etnaviv_init(void)
 	 * If the DT contains at least one available GPU device, instantiate
 	 * the DRM platform device.
 	 */
-	for_each_compatible_node(np, NULL, "vivante,gc") {
-		if (!of_device_is_available(np))
-			continue;
-
-		pdev = platform_device_alloc("etnaviv", PLATFORM_DEVID_NONE);
-		if (!pdev) {
-			ret = -ENOMEM;
-			of_node_put(np);
-			goto unregister_platform_driver;
-		}
-
-		ret = platform_device_add(pdev);
-		if (ret) {
-			platform_device_put(pdev);
-			of_node_put(np);
-			goto unregister_platform_driver;
-		}
-
-		etnaviv_drm = pdev;
+	np = etnaviv_of_first_available_node();
+	if (np) {
 		of_node_put(np);
-		break;
+
+		ret = etnaviv_create_platform_device("etnaviv", &etnaviv_drm);
+		if (ret)
+			goto unregister_platform_driver;
 	}
 
 	return 0;
@@ -710,7 +739,7 @@ module_init(etnaviv_init);
 
 static void __exit etnaviv_exit(void)
 {
-	platform_device_unregister(etnaviv_drm);
+	etnaviv_destroy_platform_device(&etnaviv_drm);
 	platform_driver_unregister(&etnaviv_platform_driver);
 	platform_driver_unregister(&etnaviv_gpu_driver);
 }

drivers/gpu/drm/etnaviv/etnaviv_gem.c

@@ -100,11 +100,10 @@ struct page **etnaviv_gem_get_pages(struct etnaviv_gem_object *etnaviv_obj)
 
 	if (!etnaviv_obj->sgt) {
 		struct drm_device *dev = etnaviv_obj->base.dev;
-		int npages = etnaviv_obj->base.size >> PAGE_SHIFT;
+		unsigned int npages = etnaviv_obj->base.size >> PAGE_SHIFT;
 		struct sg_table *sgt;
 
-		sgt = drm_prime_pages_to_sg(etnaviv_obj->base.dev,
-					    etnaviv_obj->pages, npages);
+		sgt = drm_prime_pages_to_sg(dev, etnaviv_obj->pages, npages);
 		if (IS_ERR(sgt)) {
 			dev_err(dev->dev, "failed to allocate sgt: %ld\n",
 				PTR_ERR(sgt));
@@ -542,7 +541,7 @@ static const struct drm_gem_object_funcs etnaviv_gem_object_funcs = {
 	.vm_ops = &vm_ops,
 };
 
-static int etnaviv_gem_new_impl(struct drm_device *dev, u32 size, u32 flags,
+static int etnaviv_gem_new_impl(struct drm_device *dev, u32 flags,
 	const struct etnaviv_gem_ops *ops, struct drm_gem_object **obj)
 {
 	struct etnaviv_gem_object *etnaviv_obj;
@@ -591,8 +590,7 @@ int etnaviv_gem_new_handle(struct drm_device *dev, struct drm_file *file,
 
 	size = PAGE_ALIGN(size);
 
-	ret = etnaviv_gem_new_impl(dev, size, flags,
-				   &etnaviv_gem_shmem_ops, &obj);
+	ret = etnaviv_gem_new_impl(dev, flags, &etnaviv_gem_shmem_ops, &obj);
 	if (ret)
 		goto fail;
 
@@ -627,7 +625,7 @@ int etnaviv_gem_new_private(struct drm_device *dev, size_t size, u32 flags,
 	struct drm_gem_object *obj;
 	int ret;
 
-	ret = etnaviv_gem_new_impl(dev, size, flags, ops, &obj);
+	ret = etnaviv_gem_new_impl(dev, flags, ops, &obj);
 	if (ret)
 		return ret;
 

drivers/gpu/drm/etnaviv/etnaviv_gpu.c

@@ -164,6 +164,26 @@ int etnaviv_gpu_get_param(struct etnaviv_gpu *gpu, u32 param, u64 *value)
 		*value = gpu->identity.eco_id;
 		break;
 
+	case ETNAVIV_PARAM_GPU_NN_CORE_COUNT:
+		*value = gpu->identity.nn_core_count;
+		break;
+
+	case ETNAVIV_PARAM_GPU_NN_MAD_PER_CORE:
+		*value = gpu->identity.nn_mad_per_core;
+		break;
+
+	case ETNAVIV_PARAM_GPU_TP_CORE_COUNT:
+		*value = gpu->identity.tp_core_count;
+		break;
+
+	case ETNAVIV_PARAM_GPU_ON_CHIP_SRAM_SIZE:
+		*value = gpu->identity.on_chip_sram_size;
+		break;
+
+	case ETNAVIV_PARAM_GPU_AXI_SRAM_SIZE:
+		*value = gpu->identity.axi_sram_size;
+		break;
+
 	default:
 		DBG("%s: invalid param: %u", dev_name(gpu->dev), param);
 		return -EINVAL;
@@ -513,8 +533,19 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu)
 	timeout = jiffies + msecs_to_jiffies(1000);
 
 	while (time_is_after_jiffies(timeout)) {
-		/* enable clock */
 		unsigned int fscale = 1 << (6 - gpu->freq_scale);
+		u32 pulse_eater = 0x01590880;
+
+		/* disable clock gating */
+		gpu_write_power(gpu, VIVS_PM_POWER_CONTROLS, 0x0);
+
+		/* disable pulse eater */
+		pulse_eater |= BIT(17);
+		gpu_write_power(gpu, VIVS_PM_PULSE_EATER, pulse_eater);
+		pulse_eater |= BIT(0);
+		gpu_write_power(gpu, VIVS_PM_PULSE_EATER, pulse_eater);
+
+		/* enable clock */
 		control = VIVS_HI_CLOCK_CONTROL_FSCALE_VAL(fscale);
 		etnaviv_gpu_load_clock(gpu, control);
 

drivers/gpu/drm/etnaviv/etnaviv_gpu.h

@@ -54,6 +54,18 @@ struct etnaviv_chip_identity {
 	/* Number of Neural Network cores. */
 	u32 nn_core_count;
 
+	/* Number of MAD units per Neural Network core. */
+	u32 nn_mad_per_core;
+
+	/* Number of Tensor Processing cores. */
+	u32 tp_core_count;
+
+	/* Size in bytes of the SRAM inside the NPU. */
+	u32 on_chip_sram_size;
+
+	/* Size in bytes of the SRAM across the AXI bus. */
+	u32 axi_sram_size;
+
 	/* Size of the vertex cache. */
 	u32 vertex_cache_size;
 

drivers/gpu/drm/etnaviv/etnaviv_hwdb.c

@@ -17,6 +17,10 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.thread_count = 128,
 		.shader_core_count = 1,
 		.nn_core_count = 0,
+		.nn_mad_per_core = 0,
+		.tp_core_count = 0,
+		.on_chip_sram_size = 0,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 8,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 1,
@@ -48,6 +52,11 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.register_max = 64,
 		.thread_count = 256,
 		.shader_core_count = 1,
+		.nn_core_count = 0,
+		.nn_mad_per_core = 0,
+		.tp_core_count = 0,
+		.on_chip_sram_size = 0,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 8,
 		.vertex_output_buffer_size = 512,
 		.pixel_pipes = 1,
@@ -80,6 +89,10 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.thread_count = 512,
 		.shader_core_count = 2,
 		.nn_core_count = 0,
+		.nn_mad_per_core = 0,
+		.tp_core_count = 0,
+		.on_chip_sram_size = 0,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 16,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 1,
@@ -112,6 +125,10 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.thread_count = 512,
 		.shader_core_count = 2,
 		.nn_core_count = 0,
+		.nn_mad_per_core = 0,
+		.tp_core_count = 0,
+		.on_chip_sram_size = 0,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 16,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 1,
@@ -143,6 +160,11 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.register_max = 64,
 		.thread_count = 512,
 		.shader_core_count = 2,
+		.nn_core_count = 0,
+		.nn_mad_per_core = 0,
+		.tp_core_count = 0,
+		.on_chip_sram_size = 0,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 16,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 1,
@@ -175,6 +197,10 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.thread_count = 1024,
 		.shader_core_count = 4,
 		.nn_core_count = 0,
+		.nn_mad_per_core = 0,
+		.tp_core_count = 0,
+		.on_chip_sram_size = 0,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 16,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 2,
@@ -207,6 +233,10 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.thread_count = 256,
 		.shader_core_count = 1,
 		.nn_core_count = 8,
+		.nn_mad_per_core = 64,
+		.tp_core_count = 4,
+		.on_chip_sram_size = 524288,
+		.axi_sram_size = 1048576,
 		.vertex_cache_size = 16,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 1,
@@ -239,6 +269,10 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 		.thread_count = 256,
 		.shader_core_count = 1,
 		.nn_core_count = 6,
+		.nn_mad_per_core = 64,
+		.tp_core_count = 3,
+		.on_chip_sram_size = 262144,
+		.axi_sram_size = 0,
 		.vertex_cache_size = 16,
 		.vertex_output_buffer_size = 1024,
 		.pixel_pipes = 1,
@@ -265,6 +299,9 @@ static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
 bool etnaviv_fill_identity_from_hwdb(struct etnaviv_gpu *gpu)
 {
 	struct etnaviv_chip_identity *ident = &gpu->identity;
+	const u32 product_id = ident->product_id;
+	const u32 customer_id = ident->customer_id;
+	const u32 eco_id = ident->eco_id;
 	int i;
 
 	for (i = 0; i < ARRAY_SIZE(etnaviv_chip_identities); i++) {
@@ -278,6 +315,12 @@ bool etnaviv_fill_identity_from_hwdb(struct etnaviv_gpu *gpu)
 			 etnaviv_chip_identities[i].eco_id == ~0U)) {
 			memcpy(ident, &etnaviv_chip_identities[i],
 			       sizeof(*ident));
+
+			/* Restore some id values as ~0U aka 'don't care' might been used. */
+			ident->product_id = product_id;
+			ident->customer_id = customer_id;
+			ident->eco_id = eco_id;
+
 			return true;
 		}
 	}

drivers/gpu/drm/etnaviv/etnaviv_mmu.c

@@ -70,7 +70,7 @@ static int etnaviv_context_map(struct etnaviv_iommu_context *context,
 }
 
 static int etnaviv_iommu_map(struct etnaviv_iommu_context *context, u32 iova,
-			     struct sg_table *sgt, unsigned len, int prot)
+			     struct sg_table *sgt, int prot)
 {	struct scatterlist *sg;
 	unsigned int da = iova;
 	unsigned int i;
@@ -314,7 +314,7 @@ int etnaviv_iommu_map_gem(struct etnaviv_iommu_context *context,
 		goto unlock;
 
 	mapping->iova = node->start;
-	ret = etnaviv_iommu_map(context, node->start, sgt, etnaviv_obj->base.size,
+	ret = etnaviv_iommu_map(context, node->start, sgt,
 				ETNAVIV_PROT_READ | ETNAVIV_PROT_WRITE);
 
 	if (ret < 0) {

drivers/gpu/drm/etnaviv/etnaviv_perfmon.c

@@ -511,7 +511,7 @@ int etnaviv_pm_query_dom(struct etnaviv_gpu *gpu,
 
 	domain->id = domain->iter;
 	domain->nr_signals = dom->nr_signals;
-	strncpy(domain->name, dom->name, sizeof(domain->name));
+	strscpy_pad(domain->name, dom->name, sizeof(domain->name));
 
 	domain->iter++;
 	if (domain->iter == nr_domains)
@@ -540,7 +540,7 @@ int etnaviv_pm_query_sig(struct etnaviv_gpu *gpu,
 	sig = &dom->signal[signal->iter];
 
 	signal->id = signal->iter;
-	strncpy(signal->name, sig->name, sizeof(signal->name));
+	strscpy_pad(signal->name, sig->name, sizeof(signal->name));
 
 	signal->iter++;
 	if (signal->iter == dom->nr_signals)

include/uapi/drm/etnaviv_drm.h

@@ -77,6 +77,11 @@ struct drm_etnaviv_timespec {
 #define ETNAVIV_PARAM_GPU_PRODUCT_ID                0x1c
 #define ETNAVIV_PARAM_GPU_CUSTOMER_ID               0x1d
 #define ETNAVIV_PARAM_GPU_ECO_ID                    0x1e
+#define ETNAVIV_PARAM_GPU_NN_CORE_COUNT             0x1f
+#define ETNAVIV_PARAM_GPU_NN_MAD_PER_CORE           0x20
+#define ETNAVIV_PARAM_GPU_TP_CORE_COUNT             0x21
+#define ETNAVIV_PARAM_GPU_ON_CHIP_SRAM_SIZE         0x22
+#define ETNAVIV_PARAM_GPU_AXI_SRAM_SIZE             0x23
 
 #define ETNA_MAX_PIPES 4