Merge branch 'etnaviv/next' of https://git.pengutronix.de/git/lst/linux into drm-next

- fix for potential out-of-bounds reads in the perfmon ioctl
  implementation from Christian
- override to expose proper feature flags for the GC400 found on the
  STM32MP1 SoC, also from Christian
- Guido fixed an issue where we would spuriously fail to enter
  runtime suspend due to a new GPU engine status bit on GC7000
- tree-wide change from Gustavo to get rid of zero-length arrays
- fix for missed TS cache flush on GC7000, leading to spurious
  MMU faults from me
- request pages from DMA32 zone on systems where we can't address
  all present memory from me

Signed-off-by: Dave Airlie <[email protected]>
From: Lucas Stach <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]

Author: airlied

drivers/gpu/drm/etnaviv/etnaviv_buffer.c

@@ -12,6 +12,7 @@
 
 #include "common.xml.h"
 #include "state.xml.h"
+#include "state_blt.xml.h"
 #include "state_hi.xml.h"
 #include "state_3d.xml.h"
 #include "cmdstream.xml.h"
@@ -233,6 +234,8 @@ void etnaviv_buffer_end(struct etnaviv_gpu *gpu)
 	struct etnaviv_cmdbuf *buffer = &gpu->buffer;
 	unsigned int waitlink_offset = buffer->user_size - 16;
 	u32 link_target, flush = 0;
+	bool has_blt = !!(gpu->identity.minor_features5 &
+			  chipMinorFeatures5_BLT_ENGINE);
 
 	lockdep_assert_held(&gpu->lock);
 
@@ -248,16 +251,38 @@ void etnaviv_buffer_end(struct etnaviv_gpu *gpu)
 	if (flush) {
 		unsigned int dwords = 7;
 
+		if (has_blt)
+			dwords += 10;
+
 		link_target = etnaviv_buffer_reserve(gpu, buffer, dwords);
 
 		CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
 		CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
+		if (has_blt) {
+			CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
+			CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
+			CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
+			CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
+		}
 		CMD_LOAD_STATE(buffer, VIVS_GL_FLUSH_CACHE, flush);
-		if (gpu->exec_state == ETNA_PIPE_3D)
-			CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
-				       VIVS_TS_FLUSH_CACHE_FLUSH);
+		if (gpu->exec_state == ETNA_PIPE_3D) {
+			if (has_blt) {
+				CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
+				CMD_LOAD_STATE(buffer, VIVS_BLT_SET_COMMAND, 0x1);
+				CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
+			} else {
+				CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
+					       VIVS_TS_FLUSH_CACHE_FLUSH);
+			}
+		}
 		CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
 		CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
+		if (has_blt) {
+			CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
+			CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
+			CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
+			CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
+		}
 		CMD_END(buffer);
 
 		etnaviv_buffer_replace_wait(buffer, waitlink_offset,
@@ -323,6 +348,8 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state,
 	bool switch_mmu_context = gpu->mmu_context != mmu_context;
 	unsigned int new_flush_seq = READ_ONCE(gpu->mmu_context->flush_seq);
 	bool need_flush = switch_mmu_context || gpu->flush_seq != new_flush_seq;
+	bool has_blt = !!(gpu->identity.minor_features5 &
+			  chipMinorFeatures5_BLT_ENGINE);
 
 	lockdep_assert_held(&gpu->lock);
 
@@ -433,6 +460,15 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state,
 	 * 2 semaphore stall + 1 event + 1 wait + 1 link.
 	 */
 	return_dwords = 7;
+
+	/*
+	 * When the BLT engine is present we need 6 more dwords in the return
+	 * target: 3 enable/flush/disable + 4 enable/semaphore stall/disable,
+	 * but we don't need the normal TS flush state.
+	 */
+	if (has_blt)
+		return_dwords += 6;
+
 	return_target = etnaviv_buffer_reserve(gpu, buffer, return_dwords);
 	CMD_LINK(cmdbuf, return_dwords, return_target);
 
@@ -447,11 +483,25 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state,
 		CMD_LOAD_STATE(buffer, VIVS_GL_FLUSH_CACHE,
 				       VIVS_GL_FLUSH_CACHE_DEPTH |
 				       VIVS_GL_FLUSH_CACHE_COLOR);
-		CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
-				       VIVS_TS_FLUSH_CACHE_FLUSH);
+		if (has_blt) {
+			CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
+			CMD_LOAD_STATE(buffer, VIVS_BLT_SET_COMMAND, 0x1);
+			CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
+		} else {
+			CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
+					       VIVS_TS_FLUSH_CACHE_FLUSH);
+		}
 	}
 	CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
 	CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
+
+	if (has_blt) {
+		CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
+		CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
+		CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
+		CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
+	}
+
 	CMD_LOAD_STATE(buffer, VIVS_GL_EVENT, VIVS_GL_EVENT_EVENT_ID(event) |
 		       VIVS_GL_EVENT_FROM_PE);
 	CMD_WAIT(buffer);

drivers/gpu/drm/etnaviv/etnaviv_drv.c

@@ -551,6 +551,7 @@ static int etnaviv_bind(struct device *dev)
 	mutex_init(&priv->gem_lock);
 	INIT_LIST_HEAD(&priv->gem_list);
 	priv->num_gpus = 0;
+	priv->shm_gfp_mask = GFP_HIGHUSER | __GFP_RETRY_MAYFAIL | __GFP_NOWARN;
 
 	priv->cmdbuf_suballoc = etnaviv_cmdbuf_suballoc_new(drm->dev);
 	if (IS_ERR(priv->cmdbuf_suballoc)) {

drivers/gpu/drm/etnaviv/etnaviv_drv.h

@@ -35,6 +35,7 @@ struct etnaviv_drm_private {
 	int num_gpus;
 	struct device_dma_parameters dma_parms;
 	struct etnaviv_gpu *gpu[ETNA_MAX_PIPES];
+	gfp_t shm_gfp_mask;
 
 	struct etnaviv_cmdbuf_suballoc *cmdbuf_suballoc;
 	struct etnaviv_iommu_global *mmu_global;

drivers/gpu/drm/etnaviv/etnaviv_gem.c

@@ -602,6 +602,7 @@ static int etnaviv_gem_new_impl(struct drm_device *dev, u32 size, u32 flags,
 int etnaviv_gem_new_handle(struct drm_device *dev, struct drm_file *file,
 	u32 size, u32 flags, u32 *handle)
 {
+	struct etnaviv_drm_private *priv = dev->dev_private;
 	struct drm_gem_object *obj = NULL;
 	int ret;
 
@@ -624,8 +625,7 @@ int etnaviv_gem_new_handle(struct drm_device *dev, struct drm_file *file,
 	 * above new_inode() why this is required _and_ expected if you're
 	 * going to pin these pages.
 	 */
-	mapping_set_gfp_mask(obj->filp->f_mapping, GFP_HIGHUSER |
-			     __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	mapping_set_gfp_mask(obj->filp->f_mapping, priv->shm_gfp_mask);
 
 	etnaviv_gem_obj_add(dev, obj);
 

drivers/gpu/drm/etnaviv/etnaviv_gem.h

@@ -105,7 +105,7 @@ struct etnaviv_gem_submit {
 	unsigned int nr_pmrs;
 	struct etnaviv_perfmon_request *pmrs;
 	unsigned int nr_bos;
-	struct etnaviv_gem_submit_bo bos[0];
+	struct etnaviv_gem_submit_bo bos[];
 	/* No new members here, the previous one is variable-length! */
 };
 

drivers/gpu/drm/etnaviv/etnaviv_gpu.c

@@ -333,9 +333,13 @@ static void etnaviv_hw_identify(struct etnaviv_gpu *gpu)
 		gpu->identity.revision = etnaviv_field(chipIdentity,
 					 VIVS_HI_CHIP_IDENTITY_REVISION);
 	} else {
+		u32 chipDate = gpu_read(gpu, VIVS_HI_CHIP_DATE);
 
 		gpu->identity.model = gpu_read(gpu, VIVS_HI_CHIP_MODEL);
 		gpu->identity.revision = gpu_read(gpu, VIVS_HI_CHIP_REV);
+		gpu->identity.product_id = gpu_read(gpu, VIVS_HI_CHIP_PRODUCT_ID);
+		gpu->identity.customer_id = gpu_read(gpu, VIVS_HI_CHIP_CUSTOMER_ID);
+		gpu->identity.eco_id = gpu_read(gpu, VIVS_HI_CHIP_ECO_ID);
 
 		/*
 		 * !!!! HACK ALERT !!!!
@@ -350,7 +354,6 @@ static void etnaviv_hw_identify(struct etnaviv_gpu *gpu)
 
 		/* Another special case */
 		if (etnaviv_is_model_rev(gpu, GC300, 0x2201)) {
-			u32 chipDate = gpu_read(gpu, VIVS_HI_CHIP_DATE);
 			u32 chipTime = gpu_read(gpu, VIVS_HI_CHIP_TIME);
 
 			if (chipDate == 0x20080814 && chipTime == 0x12051100) {
@@ -373,6 +376,12 @@ static void etnaviv_hw_identify(struct etnaviv_gpu *gpu)
 			gpu->identity.model = chipModel_GC3000;
 			gpu->identity.revision &= 0xffff;
 		}
+
+		if (etnaviv_is_model_rev(gpu, GC1000, 0x5037) && (chipDate == 0x20120617))
+			gpu->identity.eco_id = 1;
+
+		if (etnaviv_is_model_rev(gpu, GC320, 0x5303) && (chipDate == 0x20140511))
+			gpu->identity.eco_id = 1;
 	}
 
 	dev_info(gpu->dev, "model: GC%x, revision: %x\n",
@@ -506,7 +515,7 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu)
 		/* read idle register. */
 		idle = gpu_read(gpu, VIVS_HI_IDLE_STATE);
 
-		/* try reseting again if FE it not idle */
+		/* try resetting again if FE is not idle */
 		if ((idle & VIVS_HI_IDLE_STATE_FE) == 0) {
 			dev_dbg(gpu->dev, "FE is not idle\n");
 			continue;
@@ -772,6 +781,14 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu)
 		gpu->identity.features &= ~chipFeatures_FAST_CLEAR;
 	}
 
+	/*
+	 * If the GPU is part of a system with DMA addressing limitations,
+	 * request pages for our SHM backend buffers from the DMA32 zone to
+	 * hopefully avoid performance killing SWIOTLB bounce buffering.
+	 */
+	if (dma_addressing_limited(gpu->dev))
+		priv->shm_gfp_mask |= GFP_DMA32;
+
 	/* Create buffer: */
 	ret = etnaviv_cmdbuf_init(priv->cmdbuf_suballoc, &gpu->buffer,
 				  PAGE_SIZE);
@@ -851,6 +868,13 @@ int etnaviv_gpu_debugfs(struct etnaviv_gpu *gpu, struct seq_file *m)
 
 	verify_dma(gpu, &debug);
 
+	seq_puts(m, "\tidentity\n");
+	seq_printf(m, "\t model: 0x%x\n", gpu->identity.model);
+	seq_printf(m, "\t revision: 0x%x\n", gpu->identity.revision);
+	seq_printf(m, "\t product_id: 0x%x\n", gpu->identity.product_id);
+	seq_printf(m, "\t customer_id: 0x%x\n", gpu->identity.customer_id);
+	seq_printf(m, "\t eco_id: 0x%x\n", gpu->identity.eco_id);
+
 	seq_puts(m, "\tfeatures\n");
 	seq_printf(m, "\t major_features: 0x%08x\n",
 		   gpu->identity.features);
@@ -930,6 +954,20 @@ int etnaviv_gpu_debugfs(struct etnaviv_gpu *gpu, struct seq_file *m)
 		seq_puts(m, "\t FP is not idle\n");
 	if ((idle & VIVS_HI_IDLE_STATE_TS) == 0)
 		seq_puts(m, "\t TS is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_BL) == 0)
+		seq_puts(m, "\t BL is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_ASYNCFE) == 0)
+		seq_puts(m, "\t ASYNCFE is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_MC) == 0)
+		seq_puts(m, "\t MC is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_PPA) == 0)
+		seq_puts(m, "\t PPA is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_WD) == 0)
+		seq_puts(m, "\t WD is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_NN) == 0)
+		seq_puts(m, "\t NN is not idle\n");
+	if ((idle & VIVS_HI_IDLE_STATE_TP) == 0)
+		seq_puts(m, "\t TP is not idle\n");
 	if (idle & VIVS_HI_IDLE_STATE_AXI_LP)
 		seq_puts(m, "\t AXI low power mode\n");
 
@@ -1805,11 +1843,15 @@ static int etnaviv_gpu_rpm_suspend(struct device *dev)
 	if (atomic_read(&gpu->sched.hw_rq_count))
 		return -EBUSY;
 
-	/* Check whether the hardware (except FE) is idle */
-	mask = gpu->idle_mask & ~VIVS_HI_IDLE_STATE_FE;
+	/* Check whether the hardware (except FE and MC) is idle */
+	mask = gpu->idle_mask & ~(VIVS_HI_IDLE_STATE_FE |
+				  VIVS_HI_IDLE_STATE_MC);
 	idle = gpu_read(gpu, VIVS_HI_IDLE_STATE) & mask;
-	if (idle != mask)
+	if (idle != mask) {
+		dev_warn_ratelimited(dev, "GPU not yet idle, mask: 0x%08x\n",
+				     idle);
 		return -EBUSY;
+	}
 
 	return etnaviv_gpu_hw_suspend(gpu);
 }

drivers/gpu/drm/etnaviv/etnaviv_gpu.h

@@ -15,11 +15,11 @@ struct etnaviv_gem_submit;
 struct etnaviv_vram_mapping;
 
 struct etnaviv_chip_identity {
-	/* Chip model. */
 	u32 model;
-
-	/* Revision value.*/
 	u32 revision;
+	u32 product_id;
+	u32 customer_id;
+	u32 eco_id;
 
 	/* Supported feature fields. */
 	u32 features;

drivers/gpu/drm/etnaviv/etnaviv_hwdb.c

@@ -6,9 +6,43 @@
 #include "etnaviv_gpu.h"
 
 static const struct etnaviv_chip_identity etnaviv_chip_identities[] = {
+	{
+		.model = 0x400,
+		.revision = 0x4652,
+		.product_id = 0x70001,
+		.customer_id = 0x100,
+		.eco_id = 0,
+		.stream_count = 4,
+		.register_max = 64,
+		.thread_count = 128,
+		.shader_core_count = 1,
+		.vertex_cache_size = 8,
+		.vertex_output_buffer_size = 1024,
+		.pixel_pipes = 1,
+		.instruction_count = 256,
+		.num_constants = 320,
+		.buffer_size = 0,
+		.varyings_count = 8,
+		.features = 0xa0e9e004,
+		.minor_features0 = 0xe1299fff,
+		.minor_features1 = 0xbe13b219,
+		.minor_features2 = 0xce110010,
+		.minor_features3 = 0x8000001,
+		.minor_features4 = 0x20102,
+		.minor_features5 = 0x120000,
+		.minor_features6 = 0x0,
+		.minor_features7 = 0x0,
+		.minor_features8 = 0x0,
+		.minor_features9 = 0x0,
+		.minor_features10 = 0x0,
+		.minor_features11 = 0x0,
+	},
 	{
 		.model = 0x7000,
 		.revision = 0x6214,
+		.product_id = ~0U,
+		.customer_id = ~0U,
+		.eco_id = ~0U,
 		.stream_count = 16,
 		.register_max = 64,
 		.thread_count = 1024,
@@ -43,7 +77,13 @@ bool etnaviv_fill_identity_from_hwdb(struct etnaviv_gpu *gpu)
 
 	for (i = 0; i < ARRAY_SIZE(etnaviv_chip_identities); i++) {
 		if (etnaviv_chip_identities[i].model == ident->model &&
-		    etnaviv_chip_identities[i].revision == ident->revision) {
+		    etnaviv_chip_identities[i].revision == ident->revision &&
+		    (etnaviv_chip_identities[i].product_id == ident->product_id ||
+			 etnaviv_chip_identities[i].product_id == ~0U) &&
+		    (etnaviv_chip_identities[i].customer_id == ident->customer_id ||
+			 etnaviv_chip_identities[i].customer_id == ~0U) &&
+		    (etnaviv_chip_identities[i].eco_id == ident->eco_id ||
+			 etnaviv_chip_identities[i].eco_id == ~0U)) {
 			memcpy(ident, &etnaviv_chip_identities[i],
 			       sizeof(*ident));
 			return true;