Merge tag 'misc-habanalabs-next-2022-09-21' of...

Merge tag 'misc-habanalabs-next-2022-09-21' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux into char-misc-next

Oded writes:

  "This tag contains habanalabs driver changes for v6.1:

   - Support new notifier event for device state change through eventfd.
   - Add uAPI to retrieve device attestation information for Gaudi2.
   - Add uAPI to retrieve the h/w status of all h/w blocks.
   - Add uAPI to control the running mode of the engine cores in Gaudi2.
   - Expose whether the device runs with secured firmware through the INFO ioctl
     and sysfs.
   - Support trace events in DMA allocations and MMU map/unmap operations.
   - Notify firmware when the device was acquired by a user process and when it
     was released. This is done as part of the RAS that the f/w performs.
   - Multiple bug fixes, refactors and renames.
   - Cleanup of error messages, moving some to debug level.
   - Enhance log prints in case of h/w error events for Gaudi2."

* tag 'misc-habanalabs-next-2022-09-21' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux: (68 commits)
  habanalabs: eliminate aggregate use warning
  habanalabs/gaudi: use 8KB aligned address for TPC kernels
  habanalabs: remove some f/w descriptor validations
  habanalabs: build ASICs from new to old
  habanalabs/gaudi2: allow user to flush PCIE by read
  habanalabs: failure to open device due to reset is debug level
  habanalabs/gaudi2: Remove unnecessary (void*) conversions
  habanalabs/gaudi2: add secured attestation info uapi
  habanalabs/gaudi2: add handling to pmmu events in eqe handler
  habanalabs/gaudi: change TPC Assert to use TPC DEC instead of QMAN err
  habanalabs: rename error info structure
  habanalabs/gaudi2: get f/w reset status register dynamically
  habanalabs/gaudi2: increase hard-reset sleep time to 2 sec
  habanalabs/gaudi2: print RAZWI info upon PCIe access error
  habanalabs: MMU invalidation h/w is per device
  habanalabs: new notifier events for device state
  habanalabs/gaudi2: free event irq if init fails
  habanalabs: fix resetting the DRAM BAR
  habanalabs: add support for new cpucp return codes
  habanalabs/gaudi2: read F/W security indication after hard reset
  ...

Documentation/ABI/testing/sysfs-driver-habanalabs

@@ -16,7 +16,7 @@ Description:    Version of the application running on the device's CPU
 
 What:           /sys/class/habanalabs/hl<n>/clk_max_freq_mhz
 Date:           Jun 2019
-KernelVersion:  not yet upstreamed
+KernelVersion:  5.7
 Contact:        ogabbay@kernel.org
 Description:    Allows the user to set the maximum clock frequency, in MHz.
                 The device clock might be set to lower value than the maximum.
@@ -26,7 +26,7 @@ Description:    Allows the user to set the maximum clock frequency, in MHz.
 
 What:           /sys/class/habanalabs/hl<n>/clk_cur_freq_mhz
 Date:           Jun 2019
-KernelVersion:  not yet upstreamed
+KernelVersion:  5.7
 Contact:        ogabbay@kernel.org
 Description:    Displays the current frequency, in MHz, of the device clock.
                 This property is valid only for the Gaudi ASIC family
@@ -176,6 +176,12 @@ KernelVersion:  5.1
 Contact:        ogabbay@kernel.org
 Description:    Version of the device's preboot F/W code
 
+What:           /sys/class/habanalabs/hl<n>/security_enabled
+Date:           Oct 2022
+KernelVersion:  6.1
+Contact:        obitton@habana.ai
+Description:    Displays the device's security status
+
 What:           /sys/class/habanalabs/hl<n>/soft_reset
 Date:           Jan 2019
 KernelVersion:  5.1
@@ -230,6 +236,6 @@ Description:    Version of the u-boot running on the device's CPU
 
 What:           /sys/class/habanalabs/hl<n>/vrm_ver
 Date:           Jan 2022
-KernelVersion:  not yet upstreamed
+KernelVersion:  5.17
 Contact:        ogabbay@kernel.org
 Description:    Version of the Device's Voltage Regulator Monitor F/W code. N/A to GOYA and GAUDI

MAINTAINERS

@@ -8878,6 +8878,7 @@ T:	git https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux.git
 F:	Documentation/ABI/testing/debugfs-driver-habanalabs
 F:	Documentation/ABI/testing/sysfs-driver-habanalabs
 F:	drivers/misc/habanalabs/
+F:	include/trace/events/habanalabs.h
 F:	include/uapi/misc/habanalabs.h
 
 HACKRF MEDIA DRIVER

drivers/misc/habanalabs/Kconfig

@@ -10,6 +10,7 @@ config HABANA_AI
 	select HWMON
 	select DMA_SHARED_BUFFER
 	select CRC32
+	select FW_LOADER
 	help
 	  Enables PCIe card driver for Habana's AI Processors (AIP) that are
 	  designed to accelerate Deep Learning inference and training workloads.

drivers/misc/habanalabs/Makefile

@@ -8,13 +8,13 @@ obj-$(CONFIG_HABANA_AI) := habanalabs.o
 include $(src)/common/Makefile
 habanalabs-y += $(HL_COMMON_FILES)
 
-include $(src)/goya/Makefile
-habanalabs-y += $(HL_GOYA_FILES)
+include $(src)/gaudi2/Makefile
+habanalabs-y += $(HL_GAUDI2_FILES)
 
 include $(src)/gaudi/Makefile
 habanalabs-y += $(HL_GAUDI_FILES)
 
-include $(src)/gaudi2/Makefile
-habanalabs-y += $(HL_GAUDI2_FILES)
+include $(src)/goya/Makefile
+habanalabs-y += $(HL_GOYA_FILES)
 
 habanalabs-$(CONFIG_DEBUG_FS) += common/debugfs.o

drivers/misc/habanalabs/common/command_buffer.c

@@ -12,20 +12,18 @@
 #include <linux/slab.h>
 #include <linux/uaccess.h>
 
+#define CB_VA_POOL_SIZE		(4UL * SZ_1G)
+
 static int cb_map_mem(struct hl_ctx *ctx, struct hl_cb *cb)
 {
 	struct hl_device *hdev = ctx->hdev;
 	struct asic_fixed_properties *prop = &hdev->asic_prop;
-	struct hl_vm_va_block *va_block, *tmp;
-	dma_addr_t bus_addr;
-	u64 virt_addr;
 	u32 page_size = prop->pmmu.page_size;
-	s32 offset;
 	int rc;
 
 	if (!hdev->supports_cb_mapping) {
 		dev_err_ratelimited(hdev->dev,
-				"Cannot map CB because no VA range is allocated for CB mapping\n");
+				"Mapping a CB to the device's MMU is not supported\n");
 		return -EINVAL;
 	}
 
@@ -35,106 +33,45 @@ static int cb_map_mem(struct hl_ctx *ctx, struct hl_cb *cb)
 		return -EINVAL;
 	}
 
-	INIT_LIST_HEAD(&cb->va_block_list);
-
-	for (bus_addr = cb->bus_address;
-			bus_addr < cb->bus_address + cb->size;
-			bus_addr += page_size) {
-
-		virt_addr = (u64) gen_pool_alloc(ctx->cb_va_pool, page_size);
-		if (!virt_addr) {
-			dev_err(hdev->dev,
-				"Failed to allocate device virtual address for CB\n");
-			rc = -ENOMEM;
-			goto err_va_pool_free;
-		}
+	if (cb->is_mmu_mapped)
+		return 0;
 
-		va_block = kzalloc(sizeof(*va_block), GFP_KERNEL);
-		if (!va_block) {
-			rc = -ENOMEM;
-			gen_pool_free(ctx->cb_va_pool, virt_addr, page_size);
-			goto err_va_pool_free;
-		}
+	cb->roundup_size = roundup(cb->size, page_size);
 
-		va_block->start = virt_addr;
-		va_block->end = virt_addr + page_size - 1;
-		va_block->size = page_size;
-		list_add_tail(&va_block->node, &cb->va_block_list);
+	cb->virtual_addr = (u64) gen_pool_alloc(ctx->cb_va_pool, cb->roundup_size);
+	if (!cb->virtual_addr) {
+		dev_err(hdev->dev, "Failed to allocate device virtual address for CB\n");
+		return -ENOMEM;
 	}
 
-	mutex_lock(&ctx->mmu_lock);
-
-	bus_addr = cb->bus_address;
-	offset = 0;
-	list_for_each_entry(va_block, &cb->va_block_list, node) {
-		rc = hl_mmu_map_page(ctx, va_block->start, bus_addr,
-				va_block->size, list_is_last(&va_block->node,
-							&cb->va_block_list));
-		if (rc) {
-			dev_err(hdev->dev, "Failed to map VA %#llx to CB\n",
-				va_block->start);
-			goto err_va_umap;
-		}
-
-		bus_addr += va_block->size;
-		offset += va_block->size;
+	mutex_lock(&hdev->mmu_lock);
+	rc = hl_mmu_map_contiguous(ctx, cb->virtual_addr, cb->bus_address, cb->roundup_size);
+	if (rc) {
+		dev_err(hdev->dev, "Failed to map VA %#llx to CB\n", cb->virtual_addr);
+		goto err_va_umap;
 	}
-
 	rc = hl_mmu_invalidate_cache(hdev, false, MMU_OP_USERPTR | MMU_OP_SKIP_LOW_CACHE_INV);
-
-	mutex_unlock(&ctx->mmu_lock);
+	mutex_unlock(&hdev->mmu_lock);
 
 	cb->is_mmu_mapped = true;
-
 	return rc;
 
 err_va_umap:
-	list_for_each_entry(va_block, &cb->va_block_list, node) {
-		if (offset <= 0)
-			break;
-		hl_mmu_unmap_page(ctx, va_block->start, va_block->size,
-				offset <= va_block->size);
-		offset -= va_block->size;
-	}
-
-	rc = hl_mmu_invalidate_cache(hdev, true, MMU_OP_USERPTR);
-
-	mutex_unlock(&ctx->mmu_lock);
-
-err_va_pool_free:
-	list_for_each_entry_safe(va_block, tmp, &cb->va_block_list, node) {
-		gen_pool_free(ctx->cb_va_pool, va_block->start, va_block->size);
-		list_del(&va_block->node);
-		kfree(va_block);
-	}
-
+	mutex_unlock(&hdev->mmu_lock);
+	gen_pool_free(ctx->cb_va_pool, cb->virtual_addr, cb->roundup_size);
 	return rc;
 }
 
 static void cb_unmap_mem(struct hl_ctx *ctx, struct hl_cb *cb)
 {
 	struct hl_device *hdev = ctx->hdev;
-	struct hl_vm_va_block *va_block, *tmp;
-
-	mutex_lock(&ctx->mmu_lock);
-
-	list_for_each_entry(va_block, &cb->va_block_list, node)
-		if (hl_mmu_unmap_page(ctx, va_block->start, va_block->size,
-				list_is_last(&va_block->node,
-						&cb->va_block_list)))
-			dev_warn_ratelimited(hdev->dev,
-					"Failed to unmap CB's va 0x%llx\n",
-					va_block->start);
 
+	mutex_lock(&hdev->mmu_lock);
+	hl_mmu_unmap_contiguous(ctx, cb->virtual_addr, cb->roundup_size);
 	hl_mmu_invalidate_cache(hdev, true, MMU_OP_USERPTR);
+	mutex_unlock(&hdev->mmu_lock);
 
-	mutex_unlock(&ctx->mmu_lock);
-
-	list_for_each_entry_safe(va_block, tmp, &cb->va_block_list, node) {
-		gen_pool_free(ctx->cb_va_pool, va_block->start, va_block->size);
-		list_del(&va_block->node);
-		kfree(va_block);
-	}
+	gen_pool_free(ctx->cb_va_pool, cb->virtual_addr, cb->roundup_size);
 }
 
 static void cb_fini(struct hl_device *hdev, struct hl_cb *cb)
@@ -376,7 +313,6 @@ int hl_cb_destroy(struct hl_mem_mgr *mmg, u64 cb_handle)
 static int hl_cb_info(struct hl_mem_mgr *mmg,
 			u64 handle, u32 flags, u32 *usage_cnt, u64 *device_va)
 {
-	struct hl_vm_va_block *va_block;
 	struct hl_cb *cb;
 	int rc = 0;
 
@@ -388,9 +324,8 @@ static int hl_cb_info(struct hl_mem_mgr *mmg,
 	}
 
 	if (flags & HL_CB_FLAGS_GET_DEVICE_VA) {
-		va_block = list_first_entry(&cb->va_block_list, struct hl_vm_va_block, node);
-		if (va_block) {
-			*device_va = va_block->start;
+		if (cb->is_mmu_mapped) {
+			*device_va = cb->virtual_addr;
 		} else {
 			dev_err(mmg->dev, "CB is not mapped to the device's MMU\n");
 			rc = -EINVAL;
@@ -566,16 +501,23 @@ int hl_cb_va_pool_init(struct hl_ctx *ctx)
 		return -ENOMEM;
 	}
 
-	rc = gen_pool_add(ctx->cb_va_pool, prop->cb_va_start_addr,
-			prop->cb_va_end_addr - prop->cb_va_start_addr, -1);
+	ctx->cb_va_pool_base = hl_reserve_va_block(hdev, ctx, HL_VA_RANGE_TYPE_HOST,
+					CB_VA_POOL_SIZE, HL_MMU_VA_ALIGNMENT_NOT_NEEDED);
+	if (!ctx->cb_va_pool_base) {
+		rc = -ENOMEM;
+		goto err_pool_destroy;
+	}
+	rc = gen_pool_add(ctx->cb_va_pool, ctx->cb_va_pool_base, CB_VA_POOL_SIZE, -1);
 	if (rc) {
 		dev_err(hdev->dev,
 			"Failed to add memory to VA gen pool for CB mapping\n");
-		goto err_pool_destroy;
+		goto err_unreserve_va_block;
 	}
 
 	return 0;
 
+err_unreserve_va_block:
+	hl_unreserve_va_block(hdev, ctx, ctx->cb_va_pool_base, CB_VA_POOL_SIZE);
 err_pool_destroy:
 	gen_pool_destroy(ctx->cb_va_pool);
 
@@ -590,4 +532,5 @@ void hl_cb_va_pool_fini(struct hl_ctx *ctx)
 		return;
 
 	gen_pool_destroy(ctx->cb_va_pool);
+	hl_unreserve_va_block(hdev, ctx, ctx->cb_va_pool_base, CB_VA_POOL_SIZE);
 }

drivers/misc/habanalabs/common/command_submission.c

@@ -12,7 +12,9 @@
 #include <linux/slab.h>
 
 #define HL_CS_FLAGS_TYPE_MASK	(HL_CS_FLAGS_SIGNAL | HL_CS_FLAGS_WAIT | \
-					HL_CS_FLAGS_COLLECTIVE_WAIT)
+			HL_CS_FLAGS_COLLECTIVE_WAIT | HL_CS_FLAGS_RESERVE_SIGNALS_ONLY | \
+			HL_CS_FLAGS_UNRESERVE_SIGNALS_ONLY | HL_CS_FLAGS_ENGINE_CORE_COMMAND)
+
 
 #define MAX_TS_ITER_NUM 10
 
@@ -824,10 +826,10 @@ static void cs_timedout(struct work_struct *work)
 	}
 
 	/* Save only the first CS timeout parameters */
-	rc = atomic_cmpxchg(&hdev->last_error.cs_timeout.write_enable, 1, 0);
+	rc = atomic_cmpxchg(&hdev->captured_err_info.cs_timeout.write_enable, 1, 0);
 	if (rc) {
-		hdev->last_error.cs_timeout.timestamp = ktime_get();
-		hdev->last_error.cs_timeout.seq = cs->sequence;
+		hdev->captured_err_info.cs_timeout.timestamp = ktime_get();
+		hdev->captured_err_info.cs_timeout.seq = cs->sequence;
 
 		event_mask = device_reset ? (HL_NOTIFIER_EVENT_CS_TIMEOUT |
 				HL_NOTIFIER_EVENT_DEVICE_RESET) : HL_NOTIFIER_EVENT_CS_TIMEOUT;
@@ -1242,6 +1244,8 @@ static enum hl_cs_type hl_cs_get_cs_type(u32 cs_type_flags)
 		return CS_RESERVE_SIGNALS;
 	else if (cs_type_flags & HL_CS_FLAGS_UNRESERVE_SIGNALS_ONLY)
 		return CS_UNRESERVE_SIGNALS;
+	else if (cs_type_flags & HL_CS_FLAGS_ENGINE_CORE_COMMAND)
+		return CS_TYPE_ENGINE_CORE;
 	else
 		return CS_TYPE_DEFAULT;
 }
@@ -1253,6 +1257,7 @@ static int hl_cs_sanity_checks(struct hl_fpriv *hpriv, union hl_cs_args *args)
 	u32 cs_type_flags, num_chunks;
 	enum hl_device_status status;
 	enum hl_cs_type cs_type;
+	bool is_sync_stream;
 
 	if (!hl_device_operational(hdev, &status)) {
 		return -EBUSY;
@@ -1276,9 +1281,10 @@ static int hl_cs_sanity_checks(struct hl_fpriv *hpriv, union hl_cs_args *args)
 	cs_type = hl_cs_get_cs_type(cs_type_flags);
 	num_chunks = args->in.num_chunks_execute;
 
-	if (unlikely((cs_type == CS_TYPE_SIGNAL || cs_type == CS_TYPE_WAIT ||
-			cs_type == CS_TYPE_COLLECTIVE_WAIT) &&
-			!hdev->supports_sync_stream)) {
+	is_sync_stream = (cs_type == CS_TYPE_SIGNAL || cs_type == CS_TYPE_WAIT ||
+			cs_type == CS_TYPE_COLLECTIVE_WAIT);
+
+	if (unlikely(is_sync_stream && !hdev->supports_sync_stream)) {
 		dev_err(hdev->dev, "Sync stream CS is not supported\n");
 		return -EINVAL;
 	}
@@ -1288,7 +1294,7 @@ static int hl_cs_sanity_checks(struct hl_fpriv *hpriv, union hl_cs_args *args)
 			dev_err(hdev->dev, "Got execute CS with 0 chunks, context %d\n", ctx->asid);
 			return -EINVAL;
 		}
-	} else if (num_chunks != 1) {
+	} else if (is_sync_stream && num_chunks != 1) {
 		dev_err(hdev->dev,
 			"Sync stream CS mandates one chunk only, context %d\n",
 			ctx->asid);
@@ -1584,13 +1590,14 @@ static int hl_cs_ctx_switch(struct hl_fpriv *hpriv, union hl_cs_args *args,
 	struct hl_device *hdev = hpriv->hdev;
 	struct hl_ctx *ctx = hpriv->ctx;
 	bool need_soft_reset = false;
-	int rc = 0, do_ctx_switch;
+	int rc = 0, do_ctx_switch = 0;
 	void __user *chunks;
 	u32 num_chunks, tmp;
 	u16 sob_count;
 	int ret;
 
-	do_ctx_switch = atomic_cmpxchg(&ctx->thread_ctx_switch_token, 1, 0);
+	if (hdev->supports_ctx_switch)
+		do_ctx_switch = atomic_cmpxchg(&ctx->thread_ctx_switch_token, 1, 0);
 
 	if (do_ctx_switch || (args->in.cs_flags & HL_CS_FLAGS_FORCE_RESTORE)) {
 		mutex_lock(&hpriv->restore_phase_mutex);
@@ -1661,9 +1668,10 @@ static int hl_cs_ctx_switch(struct hl_fpriv *hpriv, union hl_cs_args *args,
 			}
 		}
 
-		ctx->thread_ctx_switch_wait_token = 1;
+		if (hdev->supports_ctx_switch)
+			ctx->thread_ctx_switch_wait_token = 1;
 
-	} else if (!ctx->thread_ctx_switch_wait_token) {
+	} else if (hdev->supports_ctx_switch && !ctx->thread_ctx_switch_wait_token) {
 		rc = hl_poll_timeout_memory(hdev,
 			&ctx->thread_ctx_switch_wait_token, tmp, (tmp == 1),
 			100, jiffies_to_usecs(hdev->timeout_jiffies), false);
@@ -2351,6 +2359,41 @@ static int cs_ioctl_signal_wait(struct hl_fpriv *hpriv, enum hl_cs_type cs_type,
 	return rc;
 }
 
+static int cs_ioctl_engine_cores(struct hl_fpriv *hpriv, u64 engine_cores,
+						u32 num_engine_cores, u32 core_command)
+{
+	int rc;
+	struct hl_device *hdev = hpriv->hdev;
+	void __user *engine_cores_arr;
+	u32 *cores;
+
+	if (!num_engine_cores || num_engine_cores > hdev->asic_prop.num_engine_cores) {
+		dev_err(hdev->dev, "Number of engine cores %d is invalid\n", num_engine_cores);
+		return -EINVAL;
+	}
+
+	if (core_command != HL_ENGINE_CORE_RUN && core_command != HL_ENGINE_CORE_HALT) {
+		dev_err(hdev->dev, "Engine core command is invalid\n");
+		return -EINVAL;
+	}
+
+	engine_cores_arr = (void __user *) (uintptr_t) engine_cores;
+	cores = kmalloc_array(num_engine_cores, sizeof(u32), GFP_KERNEL);
+	if (!cores)
+		return -ENOMEM;
+
+	if (copy_from_user(cores, engine_cores_arr, num_engine_cores * sizeof(u32))) {
+		dev_err(hdev->dev, "Failed to copy core-ids array from user\n");
+		kfree(cores);
+		return -EFAULT;
+	}
+
+	rc = hdev->asic_funcs->set_engine_cores(hdev, cores, num_engine_cores, core_command);
+	kfree(cores);
+
+	return rc;
+}
+
 int hl_cs_ioctl(struct hl_fpriv *hpriv, void *data)
 {
 	union hl_cs_args *args = data;
@@ -2403,6 +2446,10 @@ int hl_cs_ioctl(struct hl_fpriv *hpriv, void *data)
 		rc = cs_ioctl_unreserve_signals(hpriv,
 					args->in.encaps_sig_handle_id);
 		break;
+	case CS_TYPE_ENGINE_CORE:
+		rc = cs_ioctl_engine_cores(hpriv, args->in.engine_cores,
+				args->in.num_engine_cores, args->in.core_command);
+		break;
 	default:
 		rc = cs_ioctl_default(hpriv, chunks, num_chunks, &cs_seq,
 						args->in.cs_flags,
@@ -2524,7 +2571,7 @@ static int hl_cs_poll_fences(struct multi_cs_data *mcs_data, struct multi_cs_com
 	ktime_t max_ktime, first_cs_time;
 	enum hl_cs_wait_status status;
 
-	memset(fence_ptr, 0, arr_len * sizeof(*fence_ptr));
+	memset(fence_ptr, 0, arr_len * sizeof(struct hl_fence *));
 
 	/* get all fences under the same lock */
 	rc = hl_ctx_get_fences(mcs_data->ctx, seq_arr, fence_ptr, arr_len);
@@ -2826,7 +2873,7 @@ static int hl_multi_cs_wait_ioctl(struct hl_fpriv *hpriv, void *data)
 	}
 
 	/* allocate array for the fences */
-	fence_arr = kmalloc_array(seq_arr_len, sizeof(*fence_arr), GFP_KERNEL);
+	fence_arr = kmalloc_array(seq_arr_len, sizeof(struct hl_fence *), GFP_KERNEL);
 	if (!fence_arr) {
 		rc = -ENOMEM;
 		goto free_seq_arr;

drivers/misc/habanalabs/common/debugfs.c

@@ -291,14 +291,16 @@ static int vm_show(struct seq_file *s, void *data)
 		if (ctx->asid != HL_KERNEL_ASID_ID &&
 		    !list_empty(&ctx->hw_block_mem_list)) {
 			seq_puts(s, "\nhw_block mappings:\n\n");
-			seq_puts(s, "    virtual address    size    HW block id\n");
-			seq_puts(s, "-------------------------------------------\n");
+			seq_puts(s,
+				"    virtual address    block size    mapped size    HW block id\n");
+			seq_puts(s,
+				"---------------------------------------------------------------\n");
 			mutex_lock(&ctx->hw_block_list_lock);
-			list_for_each_entry(lnode, &ctx->hw_block_mem_list,
-					    node) {
+			list_for_each_entry(lnode, &ctx->hw_block_mem_list, node) {
 				seq_printf(s,
-					"    0x%-14lx   %-6u      %-9u\n",
-					lnode->vaddr, lnode->size, lnode->id);
+					"    0x%-14lx   %-6u        %-6u             %-9u\n",
+					lnode->vaddr, lnode->block_size, lnode->mapped_size,
+					lnode->id);
 			}
 			mutex_unlock(&ctx->hw_block_list_lock);
 		}
@@ -591,6 +593,7 @@ static int engines_show(struct seq_file *s, void *data)
 	struct hl_debugfs_entry *entry = s->private;
 	struct hl_dbg_device_entry *dev_entry = entry->dev_entry;
 	struct hl_device *hdev = dev_entry->hdev;
+	struct engines_data eng_data;
 
 	if (hdev->reset_info.in_reset) {
 		dev_warn_ratelimited(hdev->dev,
@@ -598,7 +601,25 @@ static int engines_show(struct seq_file *s, void *data)
 		return 0;
 	}
 
-	hdev->asic_funcs->is_device_idle(hdev, NULL, 0, s);
+	eng_data.actual_size = 0;
+	eng_data.allocated_buf_size = HL_ENGINES_DATA_MAX_SIZE;
+	eng_data.buf = vmalloc(eng_data.allocated_buf_size);
+	if (!eng_data.buf)
+		return -ENOMEM;
+
+	hdev->asic_funcs->is_device_idle(hdev, NULL, 0, &eng_data);
+
+	if (eng_data.actual_size > eng_data.allocated_buf_size) {
+		dev_err(hdev->dev,
+				"Engines data size (%d Bytes) is bigger than allocated size (%u Bytes)\n",
+				eng_data.actual_size, eng_data.allocated_buf_size);
+		vfree(eng_data.buf);
+		return -ENOMEM;
+	}
+
+	seq_write(s, eng_data.buf, eng_data.actual_size);
+
+	vfree(eng_data.buf);
 
 	return 0;
 }

drivers/misc/habanalabs/common/habanalabs_drv.c

@@ -14,6 +14,9 @@
 #include <linux/aer.h>
 #include <linux/module.h>
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/habanalabs.h>
+
 #define HL_DRIVER_AUTHOR	"HabanaLabs Kernel Driver Team"
 
 #define HL_DRIVER_DESC		"Driver for HabanaLabs's AI Accelerators"
@@ -27,7 +30,10 @@ static struct class *hl_class;
 static DEFINE_IDR(hl_devs_idr);
 static DEFINE_MUTEX(hl_devs_idr_lock);
 
-static int timeout_locked = 30;
+#define HL_DEFAULT_TIMEOUT_LOCKED	30	/* 30 seconds */
+#define GAUDI_DEFAULT_TIMEOUT_LOCKED	600	/* 10 minutes */
+
+static int timeout_locked = HL_DEFAULT_TIMEOUT_LOCKED;
 static int reset_on_lockup = 1;
 static int memory_scrub;
 static ulong boot_error_status_mask = ULONG_MAX;
@@ -55,14 +61,12 @@ MODULE_PARM_DESC(boot_error_status_mask,
 #define PCI_IDS_GAUDI_SEC		0x1010
 
 #define PCI_IDS_GAUDI2			0x1020
-#define PCI_IDS_GAUDI2_SEC		0x1030
 
 static const struct pci_device_id ids[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_HABANALABS, PCI_IDS_GOYA), },
 	{ PCI_DEVICE(PCI_VENDOR_ID_HABANALABS, PCI_IDS_GAUDI), },
 	{ PCI_DEVICE(PCI_VENDOR_ID_HABANALABS, PCI_IDS_GAUDI_SEC), },
 	{ PCI_DEVICE(PCI_VENDOR_ID_HABANALABS, PCI_IDS_GAUDI2), },
-	{ PCI_DEVICE(PCI_VENDOR_ID_HABANALABS, PCI_IDS_GAUDI2_SEC), },
 	{ 0, }
 };
 MODULE_DEVICE_TABLE(pci, ids);
@@ -92,9 +96,6 @@ static enum hl_asic_type get_asic_type(u16 device)
 	case PCI_IDS_GAUDI2:
 		asic_type = ASIC_GAUDI2;
 		break;
-	case PCI_IDS_GAUDI2_SEC:
-		asic_type = ASIC_GAUDI2_SEC;
-		break;
 	default:
 		asic_type = ASIC_INVALID;
 		break;
@@ -107,7 +108,6 @@ static bool is_asic_secured(enum hl_asic_type asic_type)
 {
 	switch (asic_type) {
 	case ASIC_GAUDI_SEC:
-	case ASIC_GAUDI2_SEC:
 		return true;
 	default:
 		return false;
@@ -161,7 +161,7 @@ int hl_device_open(struct inode *inode, struct file *filp)
 	mutex_lock(&hdev->fpriv_list_lock);
 
 	if (!hl_device_operational(hdev, &status)) {
-		dev_err_ratelimited(hdev->dev,
+		dev_dbg_ratelimited(hdev->dev,
 			"Can't open %s because it is %s\n",
 			dev_name(hdev->dev), hdev->status[status]);
 
@@ -207,11 +207,13 @@ int hl_device_open(struct inode *inode, struct file *filp)
 	list_add(&hpriv->dev_node, &hdev->fpriv_list);
 	mutex_unlock(&hdev->fpriv_list_lock);
 
+	hdev->asic_funcs->send_device_activity(hdev, true);
+
 	hl_debugfs_add_file(hpriv);
 
-	atomic_set(&hdev->last_error.cs_timeout.write_enable, 1);
-	atomic_set(&hdev->last_error.razwi.write_enable, 1);
-	hdev->last_error.undef_opcode.write_enable = true;
+	atomic_set(&hdev->captured_err_info.cs_timeout.write_enable, 1);
+	atomic_set(&hdev->captured_err_info.razwi.write_enable, 1);
+	hdev->captured_err_info.undef_opcode.write_enable = true;
 
 	hdev->open_counter++;
 	hdev->last_successful_open_jif = jiffies;
@@ -269,7 +271,7 @@ int hl_device_open_ctrl(struct inode *inode, struct file *filp)
 	mutex_lock(&hdev->fpriv_ctrl_list_lock);
 
 	if (!hl_device_operational(hdev, NULL)) {
-		dev_err_ratelimited(hdev->dev_ctrl,
+		dev_dbg_ratelimited(hdev->dev_ctrl,
 			"Can't open %s because it is disabled or in reset\n",
 			dev_name(hdev->dev_ctrl));
 		rc = -EPERM;
@@ -314,12 +316,22 @@ static void copy_kernel_module_params_to_device(struct hl_device *hdev)
 	hdev->boot_error_status_mask = boot_error_status_mask;
 }
 
-static void fixup_device_params_per_asic(struct hl_device *hdev)
+static void fixup_device_params_per_asic(struct hl_device *hdev, int timeout)
 {
 	switch (hdev->asic_type) {
-	case ASIC_GOYA:
 	case ASIC_GAUDI:
 	case ASIC_GAUDI_SEC:
+		/* If user didn't request a different timeout than the default one, we have
+		 * a different default timeout for Gaudi
+		 */
+		if (timeout == HL_DEFAULT_TIMEOUT_LOCKED)
+			hdev->timeout_jiffies = msecs_to_jiffies(GAUDI_DEFAULT_TIMEOUT_LOCKED *
+										MSEC_PER_SEC);
+
+		hdev->reset_upon_device_release = 0;
+		break;
+
+	case ASIC_GOYA:
 		hdev->reset_upon_device_release = 0;
 		break;
 
@@ -339,7 +351,7 @@ static int fixup_device_params(struct hl_device *hdev)
 	hdev->fw_comms_poll_interval_usec = HL_FW_STATUS_POLL_INTERVAL_USEC;
 
 	if (tmp_timeout)
-		hdev->timeout_jiffies = msecs_to_jiffies(tmp_timeout * 1000);
+		hdev->timeout_jiffies = msecs_to_jiffies(tmp_timeout * MSEC_PER_SEC);
 	else
 		hdev->timeout_jiffies = MAX_SCHEDULE_TIMEOUT;
 
@@ -360,7 +372,7 @@ static int fixup_device_params(struct hl_device *hdev)
 	if (!hdev->cpu_queues_enable)
 		hdev->heartbeat = 0;
 
-	fixup_device_params_per_asic(hdev);
+	fixup_device_params_per_asic(hdev, tmp_timeout);
 
 	return 0;
 }

drivers/misc/habanalabs/common/habanalabs_ioctl.c

@@ -14,6 +14,7 @@
 #include <linux/fs.h>
 #include <linux/uaccess.h>
 #include <linux/slab.h>
+#include <linux/vmalloc.h>
 
 static u32 hl_debug_struct_size[HL_DEBUG_OP_TIMESTAMP + 1] = {
 	[HL_DEBUG_OP_ETR] = sizeof(struct hl_debug_params_etr),
@@ -103,6 +104,7 @@ static int hw_ip_info(struct hl_device *hdev, struct hl_info_args *args)
 
 	hw_ip.edma_enabled_mask = prop->edma_enabled_mask;
 	hw_ip.server_type = prop->server_type;
+	hw_ip.security_enabled = prop->fw_security_enabled;
 
 	return copy_to_user(out, &hw_ip,
 		min((size_t) size, sizeof(hw_ip))) ? -EFAULT : 0;
@@ -591,8 +593,8 @@ static int cs_timeout_info(struct hl_fpriv *hpriv, struct hl_info_args *args)
 	if ((!max_size) || (!out))
 		return -EINVAL;
 
-	info.seq = hdev->last_error.cs_timeout.seq;
-	info.timestamp = ktime_to_ns(hdev->last_error.cs_timeout.timestamp);
+	info.seq = hdev->captured_err_info.cs_timeout.seq;
+	info.timestamp = ktime_to_ns(hdev->captured_err_info.cs_timeout.timestamp);
 
 	return copy_to_user(out, &info, min_t(size_t, max_size, sizeof(info))) ? -EFAULT : 0;
 }
@@ -607,12 +609,12 @@ static int razwi_info(struct hl_fpriv *hpriv, struct hl_info_args *args)
 	if ((!max_size) || (!out))
 		return -EINVAL;
 
-	info.timestamp = ktime_to_ns(hdev->last_error.razwi.timestamp);
-	info.addr = hdev->last_error.razwi.addr;
-	info.engine_id_1 = hdev->last_error.razwi.engine_id_1;
-	info.engine_id_2 = hdev->last_error.razwi.engine_id_2;
-	info.no_engine_id = hdev->last_error.razwi.non_engine_initiator;
-	info.error_type = hdev->last_error.razwi.type;
+	info.timestamp = ktime_to_ns(hdev->captured_err_info.razwi.timestamp);
+	info.addr = hdev->captured_err_info.razwi.addr;
+	info.engine_id_1 = hdev->captured_err_info.razwi.engine_id_1;
+	info.engine_id_2 = hdev->captured_err_info.razwi.engine_id_2;
+	info.no_engine_id = hdev->captured_err_info.razwi.non_engine_initiator;
+	info.error_type = hdev->captured_err_info.razwi.type;
 
 	return copy_to_user(out, &info, min_t(size_t, max_size, sizeof(info))) ? -EFAULT : 0;
 }
@@ -627,13 +629,13 @@ static int undefined_opcode_info(struct hl_fpriv *hpriv, struct hl_info_args *ar
 	if ((!max_size) || (!out))
 		return -EINVAL;
 
-	info.timestamp = ktime_to_ns(hdev->last_error.undef_opcode.timestamp);
-	info.engine_id = hdev->last_error.undef_opcode.engine_id;
-	info.cq_addr = hdev->last_error.undef_opcode.cq_addr;
-	info.cq_size = hdev->last_error.undef_opcode.cq_size;
-	info.stream_id = hdev->last_error.undef_opcode.stream_id;
-	info.cb_addr_streams_len = hdev->last_error.undef_opcode.cb_addr_streams_len;
-	memcpy(info.cb_addr_streams, hdev->last_error.undef_opcode.cb_addr_streams,
+	info.timestamp = ktime_to_ns(hdev->captured_err_info.undef_opcode.timestamp);
+	info.engine_id = hdev->captured_err_info.undef_opcode.engine_id;
+	info.cq_addr = hdev->captured_err_info.undef_opcode.cq_addr;
+	info.cq_size = hdev->captured_err_info.undef_opcode.cq_size;
+	info.stream_id = hdev->captured_err_info.undef_opcode.stream_id;
+	info.cb_addr_streams_len = hdev->captured_err_info.undef_opcode.cb_addr_streams_len;
+	memcpy(info.cb_addr_streams, hdev->captured_err_info.undef_opcode.cb_addr_streams,
 			sizeof(info.cb_addr_streams));
 
 	return copy_to_user(out, &info, min_t(size_t, max_size, sizeof(info))) ? -EFAULT : 0;
@@ -660,6 +662,55 @@ static int dev_mem_alloc_page_sizes_info(struct hl_fpriv *hpriv, struct hl_info_
 	return copy_to_user(out, &info, min_t(size_t, max_size, sizeof(info))) ? -EFAULT : 0;
 }
 
+static int sec_attest_info(struct hl_fpriv *hpriv, struct hl_info_args *args)
+{
+	void __user *out = (void __user *) (uintptr_t) args->return_pointer;
+	struct cpucp_sec_attest_info *sec_attest_info;
+	struct hl_info_sec_attest *info;
+	u32 max_size = args->return_size;
+	int rc;
+
+	if ((!max_size) || (!out))
+		return -EINVAL;
+
+	sec_attest_info = kmalloc(sizeof(*sec_attest_info), GFP_KERNEL);
+	if (!sec_attest_info)
+		return -ENOMEM;
+
+	info = kmalloc(sizeof(*info), GFP_KERNEL);
+	if (!info) {
+		rc = -ENOMEM;
+		goto free_sec_attest_info;
+	}
+
+	rc = hl_fw_get_sec_attest_info(hpriv->hdev, sec_attest_info, args->sec_attest_nonce);
+	if (rc)
+		goto free_info;
+
+	info->nonce = le32_to_cpu(sec_attest_info->nonce);
+	info->pcr_quote_len = le16_to_cpu(sec_attest_info->pcr_quote_len);
+	info->pub_data_len = le16_to_cpu(sec_attest_info->pub_data_len);
+	info->certificate_len = le16_to_cpu(sec_attest_info->certificate_len);
+	info->pcr_num_reg = sec_attest_info->pcr_num_reg;
+	info->pcr_reg_len = sec_attest_info->pcr_reg_len;
+	info->quote_sig_len = sec_attest_info->quote_sig_len;
+	memcpy(&info->pcr_data, &sec_attest_info->pcr_data, sizeof(info->pcr_data));
+	memcpy(&info->pcr_quote, &sec_attest_info->pcr_quote, sizeof(info->pcr_quote));
+	memcpy(&info->public_data, &sec_attest_info->public_data, sizeof(info->public_data));
+	memcpy(&info->certificate, &sec_attest_info->certificate, sizeof(info->certificate));
+	memcpy(&info->quote_sig, &sec_attest_info->quote_sig, sizeof(info->quote_sig));
+
+	rc = copy_to_user(out, info,
+				min_t(size_t, max_size, sizeof(*info))) ? -EFAULT : 0;
+
+free_info:
+	kfree(info);
+free_sec_attest_info:
+	kfree(sec_attest_info);
+
+	return rc;
+}
+
 static int eventfd_register(struct hl_fpriv *hpriv, struct hl_info_args *args)
 {
 	int rc;
@@ -697,6 +748,42 @@ static int eventfd_unregister(struct hl_fpriv *hpriv, struct hl_info_args *args)
 	return 0;
 }
 
+static int engine_status_info(struct hl_fpriv *hpriv, struct hl_info_args *args)
+{
+	void __user *out = (void __user *) (uintptr_t) args->return_pointer;
+	u32 status_buf_size = args->return_size;
+	struct hl_device *hdev = hpriv->hdev;
+	struct engines_data eng_data;
+	int rc;
+
+	if ((status_buf_size < SZ_1K) || (status_buf_size > HL_ENGINES_DATA_MAX_SIZE) || (!out))
+		return -EINVAL;
+
+	eng_data.actual_size = 0;
+	eng_data.allocated_buf_size = status_buf_size;
+	eng_data.buf = vmalloc(status_buf_size);
+	if (!eng_data.buf)
+		return -ENOMEM;
+
+	hdev->asic_funcs->is_device_idle(hdev, NULL, 0, &eng_data);
+
+	if (eng_data.actual_size > eng_data.allocated_buf_size) {
+		dev_err(hdev->dev,
+			"Engines data size (%d Bytes) is bigger than allocated size (%u Bytes)\n",
+			eng_data.actual_size, status_buf_size);
+		vfree(eng_data.buf);
+		return -ENOMEM;
+	}
+
+	args->user_buffer_actual_size = eng_data.actual_size;
+	rc = copy_to_user(out, eng_data.buf, min_t(size_t, status_buf_size, eng_data.actual_size)) ?
+				-EFAULT : 0;
+
+	vfree(eng_data.buf);
+
+	return rc;
+}
+
 static int _hl_info_ioctl(struct hl_fpriv *hpriv, void *data,
 				struct device *dev)
 {
@@ -806,12 +893,18 @@ static int _hl_info_ioctl(struct hl_fpriv *hpriv, void *data,
 	case HL_INFO_DRAM_PENDING_ROWS:
 		return dram_pending_rows_info(hpriv, args);
 
+	case HL_INFO_SECURED_ATTESTATION:
+		return sec_attest_info(hpriv, args);
+
 	case HL_INFO_REGISTER_EVENTFD:
 		return eventfd_register(hpriv, args);
 
 	case HL_INFO_UNREGISTER_EVENTFD:
 		return eventfd_unregister(hpriv, args);
 
+	case HL_INFO_ENGINE_STATUS:
+		return engine_status_info(hpriv, args);
+
 	default:
 		dev_err(dev, "Invalid request %d\n", args->op);
 		rc = -EINVAL;

drivers/misc/habanalabs/common/hw_queue.c

@@ -826,9 +826,7 @@ static int ext_and_cpu_queue_init(struct hl_device *hdev, struct hl_hw_queue *q,
 
 	q->kernel_address = p;
 
-	q->shadow_queue = kmalloc_array(HL_QUEUE_LENGTH,
-					sizeof(*q->shadow_queue),
-					GFP_KERNEL);
+	q->shadow_queue = kmalloc_array(HL_QUEUE_LENGTH, sizeof(struct hl_cs_job *), GFP_KERNEL);
 	if (!q->shadow_queue) {
 		dev_err(hdev->dev,
 			"Failed to allocate shadow queue for H/W queue %d\n",

drivers/misc/habanalabs/common/hwmon.c

@@ -194,7 +194,8 @@ int hl_build_hwmon_channel_info(struct hl_device *hdev, struct cpucp_sensor *sen
 		curr_arr[sensors_by_type_next_index[type]++] = flags;
 	}
 
-	channels_info = kcalloc(num_active_sensor_types + 1, sizeof(*channels_info), GFP_KERNEL);
+	channels_info = kcalloc(num_active_sensor_types + 1, sizeof(struct hwmon_channel_info *),
+				GFP_KERNEL);
 	if (!channels_info) {
 		rc = -ENOMEM;
 		goto channels_info_array_err;
@@ -910,3 +911,24 @@ void hl_hwmon_fini(struct hl_device *hdev)
 
 	hwmon_device_unregister(hdev->hwmon_dev);
 }
+
+void hl_hwmon_release_resources(struct hl_device *hdev)
+{
+	const struct hwmon_channel_info **channel_info_arr;
+	int i = 0;
+
+	if (!hdev->hl_chip_info->info)
+		return;
+
+	channel_info_arr = hdev->hl_chip_info->info;
+
+	while (channel_info_arr[i]) {
+		kfree(channel_info_arr[i]->config);
+		kfree(channel_info_arr[i]);
+		i++;
+	}
+
+	kfree(channel_info_arr);
+
+	hdev->hl_chip_info->info = NULL;
+}

drivers/misc/habanalabs/common/memory.c

@@ -457,7 +457,7 @@ static void merge_va_blocks_locked(struct hl_device *hdev,
 	prev = list_prev_entry(va_block, node);
 	if (&prev->node != va_list && prev->end + 1 == va_block->start) {
 		prev->end = va_block->end;
-		prev->size = prev->end - prev->start;
+		prev->size = prev->end - prev->start + 1;
 		list_del(&va_block->node);
 		kfree(va_block);
 		va_block = prev;
@@ -466,7 +466,7 @@ static void merge_va_blocks_locked(struct hl_device *hdev,
 	next = list_next_entry(va_block, node);
 	if (&next->node != va_list && va_block->end + 1 == next->start) {
 		next->start = va_block->start;
-		next->size = next->end - next->start;
+		next->size = next->end - next->start + 1;
 		list_del(&va_block->node);
 		kfree(va_block);
 	}
@@ -755,7 +755,7 @@ static u64 get_va_block(struct hl_device *hdev,
  * - Return the start address of the virtual block.
  */
 u64 hl_reserve_va_block(struct hl_device *hdev, struct hl_ctx *ctx,
-		enum hl_va_range_type type, u32 size, u32 alignment)
+		enum hl_va_range_type type, u64 size, u32 alignment)
 {
 	return get_va_block(hdev, ctx->va_range[type], size, 0,
 			max(alignment, ctx->va_range[type]->page_size),
@@ -1210,18 +1210,18 @@ static int map_device_va(struct hl_ctx *ctx, struct hl_mem_in *args, u64 *device
 		goto va_block_err;
 	}
 
-	mutex_lock(&ctx->mmu_lock);
+	mutex_lock(&hdev->mmu_lock);
 
 	rc = map_phys_pg_pack(ctx, ret_vaddr, phys_pg_pack);
 	if (rc) {
 		dev_err(hdev->dev, "mapping page pack failed for handle %u\n", handle);
-		mutex_unlock(&ctx->mmu_lock);
+		mutex_unlock(&hdev->mmu_lock);
 		goto map_err;
 	}
 
 	rc = hl_mmu_invalidate_cache_range(hdev, false, *vm_type | MMU_OP_SKIP_LOW_CACHE_INV,
 				ctx->asid, ret_vaddr, phys_pg_pack->total_size);
-	mutex_unlock(&ctx->mmu_lock);
+	mutex_unlock(&hdev->mmu_lock);
 	if (rc)
 		goto map_err;
 
@@ -1362,7 +1362,7 @@ static int unmap_device_va(struct hl_ctx *ctx, struct hl_mem_in *args,
 	else
 		vaddr &= ~(((u64) phys_pg_pack->page_size) - 1);
 
-	mutex_lock(&ctx->mmu_lock);
+	mutex_lock(&hdev->mmu_lock);
 
 	unmap_phys_pg_pack(ctx, vaddr, phys_pg_pack);
 
@@ -1375,7 +1375,7 @@ static int unmap_device_va(struct hl_ctx *ctx, struct hl_mem_in *args,
 		rc = hl_mmu_invalidate_cache_range(hdev, true, *vm_type, ctx->asid, vaddr,
 							phys_pg_pack->total_size);
 
-	mutex_unlock(&ctx->mmu_lock);
+	mutex_unlock(&hdev->mmu_lock);
 
 	/*
 	 * If the context is closing we don't need to check for the MMU cache
@@ -1418,18 +1418,23 @@ static int unmap_device_va(struct hl_ctx *ctx, struct hl_mem_in *args,
 	return rc;
 }
 
-static int map_block(struct hl_device *hdev, u64 address, u64 *handle,
-			u32 *size)
+static int map_block(struct hl_device *hdev, u64 address, u64 *handle, u32 *size)
 {
-	u32 block_id = 0;
+	u32 block_id;
 	int rc;
 
+	*handle = 0;
+	if (size)
+		*size = 0;
+
 	rc = hdev->asic_funcs->get_hw_block_id(hdev, address, size, &block_id);
+	if (rc)
+		return rc;
 
 	*handle = block_id | HL_MMAP_TYPE_BLOCK;
 	*handle <<= PAGE_SHIFT;
 
-	return rc;
+	return 0;
 }
 
 static void hw_block_vm_close(struct vm_area_struct *vma)
@@ -1437,6 +1442,13 @@ static void hw_block_vm_close(struct vm_area_struct *vma)
 	struct hl_vm_hw_block_list_node *lnode =
 		(struct hl_vm_hw_block_list_node *) vma->vm_private_data;
 	struct hl_ctx *ctx = lnode->ctx;
+	long new_mmap_size;
+
+	new_mmap_size = lnode->mapped_size - (vma->vm_end - vma->vm_start);
+	if (new_mmap_size > 0) {
+		lnode->mapped_size = new_mmap_size;
+		return;
+	}
 
 	mutex_lock(&ctx->hw_block_list_lock);
 	list_del(&lnode->node);
@@ -1487,23 +1499,23 @@ int hl_hw_block_mmap(struct hl_fpriv *hpriv, struct vm_area_struct *vma)
 	if (!lnode)
 		return -ENOMEM;
 
-	vma->vm_ops = &hw_block_vm_ops;
-	vma->vm_private_data = lnode;
-
-	hl_ctx_get(ctx);
-
 	rc = hdev->asic_funcs->hw_block_mmap(hdev, vma, block_id, block_size);
 	if (rc) {
-		hl_ctx_put(ctx);
 		kfree(lnode);
 		return rc;
 	}
 
+	hl_ctx_get(ctx);
+
 	lnode->ctx = ctx;
 	lnode->vaddr = vma->vm_start;
-	lnode->size = block_size;
+	lnode->block_size = block_size;
+	lnode->mapped_size = lnode->block_size;
 	lnode->id = block_id;
 
+	vma->vm_private_data = lnode;
+	vma->vm_ops = &hw_block_vm_ops;
+
 	mutex_lock(&ctx->hw_block_list_lock);
 	list_add_tail(&lnode->node, &ctx->hw_block_mem_list);
 	mutex_unlock(&ctx->hw_block_list_lock);
@@ -2296,8 +2308,7 @@ static int get_user_memory(struct hl_device *hdev, u64 addr, u64 size,
 		return -EFAULT;
 	}
 
-	userptr->pages = kvmalloc_array(npages, sizeof(*userptr->pages),
-					GFP_KERNEL);
+	userptr->pages = kvmalloc_array(npages, sizeof(struct page *), GFP_KERNEL);
 	if (!userptr->pages)
 		return -ENOMEM;
 
@@ -2759,13 +2770,13 @@ void hl_vm_ctx_fini(struct hl_ctx *ctx)
 		unmap_device_va(ctx, &args, true);
 	}
 
-	mutex_lock(&ctx->mmu_lock);
+	mutex_lock(&hdev->mmu_lock);
 
 	/* invalidate the cache once after the unmapping loop */
 	hl_mmu_invalidate_cache(hdev, true, MMU_OP_USERPTR);
 	hl_mmu_invalidate_cache(hdev, true, MMU_OP_PHYS_PACK);
 
-	mutex_unlock(&ctx->mmu_lock);
+	mutex_unlock(&hdev->mmu_lock);
 
 	INIT_LIST_HEAD(&free_list);
 

drivers/misc/habanalabs/common/memory_mgr.c

@@ -11,7 +11,7 @@
  * hl_mmap_mem_buf_get - increase the buffer refcount and return a pointer to
  *                        the buffer descriptor.
  *
- * @mmg: parent unifed memory manager
+ * @mmg: parent unified memory manager
  * @handle: requested buffer handle
  *
  * Find the buffer in the store and return a pointer to its descriptor.
@@ -104,7 +104,7 @@ int hl_mmap_mem_buf_put(struct hl_mmap_mem_buf *buf)
  * hl_mmap_mem_buf_put_handle - decrease the reference to the buffer with the
  *                              given handle.
  *
- * @mmg: parent unifed memory manager
+ * @mmg: parent unified memory manager
  * @handle: requested buffer handle
  *
  * Decrease the reference to the buffer, and release it if it was the last one.
@@ -137,7 +137,7 @@ int hl_mmap_mem_buf_put_handle(struct hl_mem_mgr *mmg, u64 handle)
 /**
  * hl_mmap_mem_buf_alloc - allocate a new mappable buffer
  *
- * @mmg: parent unifed memory manager
+ * @mmg: parent unified memory manager
  * @behavior: behavior object describing this buffer polymorphic behavior
  * @gfp: gfp flags to use for the memory allocations
  * @args: additional args passed to behavior->alloc
@@ -222,7 +222,7 @@ static const struct vm_operations_struct hl_mmap_mem_buf_vm_ops = {
 /**
  * hl_mem_mgr_mmap - map the given buffer to the user
  *
- * @mmg: unifed memory manager
+ * @mmg: unified memory manager
  * @vma: the vma object for which mmap was closed.
  * @args: additional args passed to behavior->mmap
  *
@@ -322,7 +322,7 @@ void hl_mem_mgr_init(struct device *dev, struct hl_mem_mgr *mmg)
 /**
  * hl_mem_mgr_fini - release unified memory manager
  *
- * @mmg: parent unifed memory manager
+ * @mmg: parent unified memory manager
  *
  * Release the unified memory manager. Shall be called from an interrupt context.
  */

drivers/misc/habanalabs/common/mmu/mmu.c

@@ -9,6 +9,8 @@
 
 #include "../habanalabs.h"
 
+#include <trace/events/habanalabs.h>
+
 /**
  * hl_mmu_get_funcs() - get MMU functions structure
  * @hdev: habanalabs device structure.
@@ -45,6 +47,8 @@ int hl_mmu_init(struct hl_device *hdev)
 	if (!hdev->mmu_enable)
 		return 0;
 
+	mutex_init(&hdev->mmu_lock);
+
 	if (hdev->mmu_func[MMU_DR_PGT].init != NULL) {
 		rc = hdev->mmu_func[MMU_DR_PGT].init(hdev);
 		if (rc)
@@ -86,6 +90,8 @@ void hl_mmu_fini(struct hl_device *hdev)
 
 	if (hdev->mmu_func[MMU_HR_PGT].fini != NULL)
 		hdev->mmu_func[MMU_HR_PGT].fini(hdev);
+
+	mutex_destroy(&hdev->mmu_lock);
 }
 
 /**
@@ -104,8 +110,6 @@ int hl_mmu_ctx_init(struct hl_ctx *ctx)
 	if (!hdev->mmu_enable)
 		return 0;
 
-	mutex_init(&ctx->mmu_lock);
-
 	if (hdev->mmu_func[MMU_DR_PGT].ctx_init != NULL) {
 		rc = hdev->mmu_func[MMU_DR_PGT].ctx_init(ctx);
 		if (rc)
@@ -149,8 +153,6 @@ void hl_mmu_ctx_fini(struct hl_ctx *ctx)
 
 	if (hdev->mmu_func[MMU_HR_PGT].ctx_fini != NULL)
 		hdev->mmu_func[MMU_HR_PGT].ctx_fini(ctx);
-
-	mutex_destroy(&ctx->mmu_lock);
 }
 
 /*
@@ -259,6 +261,9 @@ int hl_mmu_unmap_page(struct hl_ctx *ctx, u64 virt_addr, u32 page_size, bool flu
 	if (flush_pte)
 		mmu_funcs->flush(ctx);
 
+	if (trace_habanalabs_mmu_unmap_enabled() && !rc)
+		trace_habanalabs_mmu_unmap(hdev->dev, virt_addr, 0, page_size, flush_pte);
+
 	return rc;
 }
 
@@ -344,6 +349,8 @@ int hl_mmu_map_page(struct hl_ctx *ctx, u64 virt_addr, u64 phys_addr, u32 page_s
 	if (flush_pte)
 		mmu_funcs->flush(ctx);
 
+	trace_habanalabs_mmu_map(hdev->dev, virt_addr, phys_addr, page_size, flush_pte);
+
 	return 0;
 
 err:
@@ -403,6 +410,8 @@ int hl_mmu_map_contiguous(struct hl_ctx *ctx, u64 virt_addr,
 			dev_err(hdev->dev,
 				"Map failed for va 0x%llx to pa 0x%llx\n",
 				curr_va, curr_pa);
+			/* last mapping failed so don't try to unmap it - reduce off by page_size */
+			off -= page_size;
 			goto unmap;
 		}
 	}
@@ -600,9 +609,9 @@ int hl_mmu_get_tlb_info(struct hl_ctx *ctx, u64 virt_addr,
 	pgt_residency = mmu_prop->host_resident ? MMU_HR_PGT : MMU_DR_PGT;
 	mmu_funcs = hl_mmu_get_funcs(hdev, pgt_residency, is_dram_addr);
 
-	mutex_lock(&ctx->mmu_lock);
+	mutex_lock(&hdev->mmu_lock);
 	rc = mmu_funcs->get_tlb_info(ctx, virt_addr, hops);
-	mutex_unlock(&ctx->mmu_lock);
+	mutex_unlock(&hdev->mmu_lock);
 
 	if (rc)
 		return rc;
@@ -692,16 +701,16 @@ static void hl_mmu_prefetch_work_function(struct work_struct *work)
 {
 	struct hl_prefetch_work *pfw = container_of(work, struct hl_prefetch_work, pf_work);
 	struct hl_ctx *ctx = pfw->ctx;
+	struct hl_device *hdev = ctx->hdev;
 
-	if (!hl_device_operational(ctx->hdev, NULL))
+	if (!hl_device_operational(hdev, NULL))
 		goto put_ctx;
 
-	mutex_lock(&ctx->mmu_lock);
+	mutex_lock(&hdev->mmu_lock);
 
-	ctx->hdev->asic_funcs->mmu_prefetch_cache_range(ctx, pfw->flags, pfw->asid,
-								pfw->va, pfw->size);
+	hdev->asic_funcs->mmu_prefetch_cache_range(ctx, pfw->flags, pfw->asid, pfw->va, pfw->size);
 
-	mutex_unlock(&ctx->mmu_lock);
+	mutex_unlock(&hdev->mmu_lock);
 
 put_ctx:
 	/*

drivers/misc/habanalabs/common/sysfs.c

@@ -375,6 +375,14 @@ static ssize_t eeprom_read_handler(struct file *filp, struct kobject *kobj,
 	return max_size;
 }
 
+static ssize_t security_enabled_show(struct device *dev,
+				struct device_attribute *attr, char *buf)
+{
+	struct hl_device *hdev = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%d\n", hdev->asic_prop.fw_security_enabled);
+}
+
 static DEVICE_ATTR_RO(armcp_kernel_ver);
 static DEVICE_ATTR_RO(armcp_ver);
 static DEVICE_ATTR_RO(cpld_ver);
@@ -393,6 +401,7 @@ static DEVICE_ATTR_RO(status);
 static DEVICE_ATTR_RO(thermal_ver);
 static DEVICE_ATTR_RO(uboot_ver);
 static DEVICE_ATTR_RO(fw_os_ver);
+static DEVICE_ATTR_RO(security_enabled);
 
 static struct bin_attribute bin_attr_eeprom = {
 	.attr = {.name = "eeprom", .mode = (0444)},
@@ -417,6 +426,7 @@ static struct attribute *hl_dev_attrs[] = {
 	&dev_attr_thermal_ver.attr,
 	&dev_attr_uboot_ver.attr,
 	&dev_attr_fw_os_ver.attr,
+	&dev_attr_security_enabled.attr,
 	NULL,
 };
 

drivers/misc/habanalabs/gaudi2/gaudi2P.h

@@ -15,7 +15,6 @@
 #include "../include/gaudi2/gaudi2_packets.h"
 #include "../include/gaudi2/gaudi2_fw_if.h"
 #include "../include/gaudi2/gaudi2_async_events.h"
-#include "../include/gaudi2/gaudi2_async_virt_events.h"
 
 #define GAUDI2_LINUX_FW_FILE	"habanalabs/gaudi2/gaudi2-fit.itb"
 #define GAUDI2_BOOT_FIT_FILE	"habanalabs/gaudi2/gaudi2-boot-fit.itb"
@@ -140,9 +139,6 @@
 #define VA_HOST_SPACE_HPAGE_START		0xFFF0800000000000ull
 #define VA_HOST_SPACE_HPAGE_END			0xFFF1000000000000ull /* 140TB */
 
-#define VA_HOST_SPACE_USER_MAPPED_CB_START	0xFFF1000000000000ull
-#define VA_HOST_SPACE_USER_MAPPED_CB_END	0xFFF1000100000000ull /* 4GB */
-
 /* 140TB */
 #define VA_HOST_SPACE_PAGE_SIZE		(VA_HOST_SPACE_PAGE_END - VA_HOST_SPACE_PAGE_START)
 
@@ -458,7 +454,6 @@ struct dup_block_ctx {
  *                 the user can map.
  * @lfsr_rand_seeds: array of MME ACC random seeds to set.
  * @hw_queues_lock: protects the H/W queues from concurrent access.
- * @kdma_lock: protects the KDMA engine from concurrent access.
  * @scratchpad_kernel_address: general purpose PAGE_SIZE contiguous memory,
  *                             this memory region should be write-only.
  *                             currently used for HBW QMAN writes which is
@@ -510,9 +505,6 @@ struct dup_block_ctx {
  * @flush_db_fifo: flag to force flush DB FIFO after a write.
  * @hbm_cfg: HBM subsystem settings
  * @hw_queues_lock_mutex: used by simulator instead of hw_queues_lock.
- * @kdma_lock_mutex: used by simulator instead of kdma_lock.
- * @use_deprecated_event_mappings: use old event mappings which are about to be
- *                                 deprecated
  */
 struct gaudi2_device {
 	int (*cpucp_info_get)(struct hl_device *hdev);
@@ -521,7 +513,6 @@ struct gaudi2_device {
 	int				lfsr_rand_seeds[MME_NUM_OF_LFSR_SEEDS];
 
 	spinlock_t			hw_queues_lock;
-	spinlock_t			kdma_lock;
 
 	void				*scratchpad_kernel_address;
 	dma_addr_t			scratchpad_bus_address;
@@ -562,5 +553,6 @@ void gaudi2_pb_print_security_errors(struct hl_device *hdev, u32 block_addr, u32
 					u32 offended_addr);
 int gaudi2_init_security(struct hl_device *hdev);
 void gaudi2_ack_protection_bits_errors(struct hl_device *hdev);
+int gaudi2_send_device_activity(struct hl_device *hdev, bool open);
 
 #endif /* GAUDI2P_H_ */

drivers/misc/habanalabs/gaudi2/gaudi2_masks.h

@@ -51,12 +51,18 @@
 	(0x1F << PDMA0_QM_GLBL_CFG0_CP_EN_SHIFT)  | \
 	(0x1 << PDMA0_QM_GLBL_CFG0_ARC_CQF_EN_SHIFT))
 
-#define PDMA1_QMAN_ENABLE	\
+#define PDMA0_QMAN_ENABLE	\
 	((0x3 << PDMA0_QM_GLBL_CFG0_PQF_EN_SHIFT) | \
 	(0x1F << PDMA0_QM_GLBL_CFG0_CQF_EN_SHIFT) | \
 	(0x1F << PDMA0_QM_GLBL_CFG0_CP_EN_SHIFT)  | \
 	(0x1 << PDMA0_QM_GLBL_CFG0_ARC_CQF_EN_SHIFT))
 
+#define PDMA1_QMAN_ENABLE	\
+	((0x1 << PDMA0_QM_GLBL_CFG0_PQF_EN_SHIFT) | \
+	(0x1F << PDMA0_QM_GLBL_CFG0_CQF_EN_SHIFT) | \
+	(0x1F << PDMA0_QM_GLBL_CFG0_CP_EN_SHIFT)  | \
+	(0x1 << PDMA0_QM_GLBL_CFG0_ARC_CQF_EN_SHIFT))
+
 /* QM_IDLE_MASK is valid for all engines QM idle check */
 #define QM_IDLE_MASK	(DCORE0_EDMA0_QM_GLBL_STS0_PQF_IDLE_MASK | \
 			DCORE0_EDMA0_QM_GLBL_STS0_CQF_IDLE_MASK | \
@@ -138,4 +144,17 @@
 #define DCORE0_SYNC_MNGR_OBJS_SOB_OBJ_SIGN_SHIFT	15
 #define DCORE0_SYNC_MNGR_OBJS_SOB_OBJ_SIGN_MASK		0x8000
 
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_ERR_INTR_SHIFT		0
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_ERR_INTR_MASK		0x1
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_LBW_ERR_INTR_SHIFT		1
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_LBW_ERR_INTR_MASK		0x2
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_BAD_ACCESS_INTR_SHIFT		2
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_BAD_ACCESS_INTR_MASK		0x4
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_ERR_INTR_MASK_SHIFT		3
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_ERR_INTR_MASK_MASK		0x8
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_LBW_ERR_INTR_MASK_SHIFT	4
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_AXI_LBW_ERR_INTR_MASK_MASK	0x10
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_BAD_ACCESS_INTR_MASK_SHIFT	5
+#define PCIE_WRAP_PCIE_IC_SEI_INTR_IND_BAD_ACCESS_INTR_MASK_MASK	0x20
+
 #endif /* GAUDI2_MASKS_H_ */

drivers/misc/habanalabs/gaudi2/gaudi2_security.c

@@ -2559,6 +2559,10 @@ static const u32 gaudi2_pb_pcie[] = {
 	mmPCIE_WRAP_BASE,
 };
 
+static const u32 gaudi2_pb_pcie_unsecured_regs[] = {
+	mmPCIE_WRAP_SPECIAL_GLBL_SPARE_0,
+};
+
 static const u32 gaudi2_pb_thermal_sensor0[] = {
 	mmDCORE0_XFT_BASE,
 	mmDCORE0_TSTDVS_BASE,
@@ -2583,9 +2587,9 @@ struct gaudi2_tpc_pb_data {
 };
 
 static void gaudi2_config_tpcs_glbl_sec(struct hl_device *hdev, int dcore, int inst, u32 offset,
-					void *data)
+						struct iterate_module_ctx *ctx)
 {
-	struct gaudi2_tpc_pb_data *pb_data = (struct gaudi2_tpc_pb_data *)data;
+	struct gaudi2_tpc_pb_data *pb_data = ctx->data;
 
 	hl_config_glbl_sec(hdev, gaudi2_pb_dcr0_tpc0, pb_data->glbl_sec,
 					offset, pb_data->block_array_size);
@@ -2660,15 +2664,14 @@ static int gaudi2_init_pb_tpc(struct hl_device *hdev)
 struct gaudi2_tpc_arc_pb_data {
 	u32 unsecured_regs_arr_size;
 	u32 arc_regs_arr_size;
-	int rc;
 };
 
 static void gaudi2_config_tpcs_pb_ranges(struct hl_device *hdev, int dcore, int inst, u32 offset,
-					void *data)
+						struct iterate_module_ctx *ctx)
 {
-	struct gaudi2_tpc_arc_pb_data *pb_data = (struct gaudi2_tpc_arc_pb_data *)data;
+	struct gaudi2_tpc_arc_pb_data *pb_data = ctx->data;
 
-	pb_data->rc |= hl_init_pb_ranges(hdev, HL_PB_SHARED, HL_PB_NA, 1,
+	ctx->rc = hl_init_pb_ranges(hdev, HL_PB_SHARED, HL_PB_NA, 1,
 					offset, gaudi2_pb_dcr0_tpc0_arc,
 					pb_data->arc_regs_arr_size,
 					gaudi2_pb_dcr0_tpc0_arc_unsecured_regs,
@@ -2683,12 +2686,12 @@ static int gaudi2_init_pb_tpc_arc(struct hl_device *hdev)
 	tpc_arc_pb_data.arc_regs_arr_size = ARRAY_SIZE(gaudi2_pb_dcr0_tpc0_arc);
 	tpc_arc_pb_data.unsecured_regs_arr_size =
 			ARRAY_SIZE(gaudi2_pb_dcr0_tpc0_arc_unsecured_regs);
-	tpc_arc_pb_data.rc = 0;
+
 	tpc_iter.fn = &gaudi2_config_tpcs_pb_ranges;
 	tpc_iter.data = &tpc_arc_pb_data;
 	gaudi2_iterate_tpcs(hdev, &tpc_iter);
 
-	return tpc_arc_pb_data.rc;
+	return tpc_iter.rc;
 }
 
 static int gaudi2_init_pb_sm_objs(struct hl_device *hdev)
@@ -3419,7 +3422,8 @@ static int gaudi2_init_protection_bits(struct hl_device *hdev)
 	rc |= hl_init_pb(hdev, HL_PB_SHARED, HL_PB_NA,
 			HL_PB_SINGLE_INSTANCE, HL_PB_NA,
 			gaudi2_pb_pcie, ARRAY_SIZE(gaudi2_pb_pcie),
-			NULL, HL_PB_NA);
+			gaudi2_pb_pcie_unsecured_regs,
+			ARRAY_SIZE(gaudi2_pb_pcie_unsecured_regs));
 
 	/* Thermal Sensor.
 	 * Skip when security is enabled in F/W, because the blocks are protected by privileged RR.
@@ -3547,9 +3551,9 @@ struct gaudi2_ack_pb_tpc_data {
 };
 
 static void gaudi2_ack_pb_tpc_config(struct hl_device *hdev, int dcore, int inst, u32 offset,
-					void *data)
+					struct iterate_module_ctx *ctx)
 {
-	struct gaudi2_ack_pb_tpc_data *pb_data = (struct gaudi2_ack_pb_tpc_data *)data;
+	struct gaudi2_ack_pb_tpc_data *pb_data = ctx->data;
 
 	hl_ack_pb_single_dcore(hdev, offset, HL_PB_SINGLE_INSTANCE, HL_PB_NA,
 				gaudi2_pb_dcr0_tpc0, pb_data->tpc_regs_array_size);

drivers/misc/habanalabs/goya/goya.c

@@ -916,26 +916,11 @@ int goya_late_init(struct hl_device *hdev)
  */
 void goya_late_fini(struct hl_device *hdev)
 {
-	const struct hwmon_channel_info **channel_info_arr;
 	struct goya_device *goya = hdev->asic_specific;
-	int i = 0;
 
 	cancel_delayed_work_sync(&goya->goya_work->work_freq);
 
-	if (!hdev->hl_chip_info->info)
-		return;
-
-	channel_info_arr = hdev->hl_chip_info->info;
-
-	while (channel_info_arr[i]) {
-		kfree(channel_info_arr[i]->config);
-		kfree(channel_info_arr[i]);
-		i++;
-	}
-
-	kfree(channel_info_arr);
-
-	hdev->hl_chip_info->info = NULL;
+	hl_hwmon_release_resources(hdev);
 }
 
 static void goya_set_pci_memory_regions(struct hl_device *hdev)
@@ -1040,6 +1025,7 @@ static int goya_sw_init(struct hl_device *hdev)
 	hdev->asic_prop.supports_compute_reset = true;
 	hdev->asic_prop.allow_inference_soft_reset = true;
 	hdev->supports_wait_for_multi_cs = false;
+	hdev->supports_ctx_switch = true;
 
 	hdev->asic_funcs->set_pci_memory_regions(hdev);
 
@@ -4559,7 +4545,7 @@ static int goya_unmask_irq_arr(struct hl_device *hdev, u32 *irq_arr,
 	return rc;
 }
 
-static int goya_non_hard_reset_late_init(struct hl_device *hdev)
+static int goya_compute_reset_late_init(struct hl_device *hdev)
 {
 	/*
 	 * Unmask all IRQs since some could have been received
@@ -5137,8 +5123,8 @@ int goya_cpucp_info_get(struct hl_device *hdev)
 	return 0;
 }
 
-static bool goya_is_device_idle(struct hl_device *hdev, u64 *mask_arr,
-					u8 mask_len, struct seq_file *s)
+static bool goya_is_device_idle(struct hl_device *hdev, u64 *mask_arr, u8 mask_len,
+				struct engines_data *e)
 {
 	const char *fmt = "%-5d%-9s%#-14x%#-16x%#x\n";
 	const char *dma_fmt = "%-5d%-9s%#-14x%#x\n";
@@ -5149,9 +5135,9 @@ static bool goya_is_device_idle(struct hl_device *hdev, u64 *mask_arr,
 	u64 offset;
 	int i;
 
-	if (s)
-		seq_puts(s, "\nDMA  is_idle  QM_GLBL_STS0  DMA_CORE_STS0\n"
-				"---  -------  ------------  -------------\n");
+	if (e)
+		hl_engine_data_sprintf(e, "\nDMA  is_idle  QM_GLBL_STS0  DMA_CORE_STS0\n"
+					"---  -------  ------------  -------------\n");
 
 	offset = mmDMA_QM_1_GLBL_STS0 - mmDMA_QM_0_GLBL_STS0;
 
@@ -5164,13 +5150,13 @@ static bool goya_is_device_idle(struct hl_device *hdev, u64 *mask_arr,
 
 		if (mask && !is_eng_idle)
 			set_bit(GOYA_ENGINE_ID_DMA_0 + i, mask);
-		if (s)
-			seq_printf(s, dma_fmt, i, is_eng_idle ? "Y" : "N",
+		if (e)
+			hl_engine_data_sprintf(e, dma_fmt, i, is_eng_idle ? "Y" : "N",
 					qm_glbl_sts0, dma_core_sts0);
 	}
 
-	if (s)
-		seq_puts(s,
+	if (e)
+		hl_engine_data_sprintf(e,
 			"\nTPC  is_idle  QM_GLBL_STS0  CMDQ_GLBL_STS0  CFG_STATUS\n"
 			"---  -------  ------------  --------------  ----------\n");
 
@@ -5187,13 +5173,13 @@ static bool goya_is_device_idle(struct hl_device *hdev, u64 *mask_arr,
 
 		if (mask && !is_eng_idle)
 			set_bit(GOYA_ENGINE_ID_TPC_0 + i, mask);
-		if (s)
-			seq_printf(s, fmt, i, is_eng_idle ? "Y" : "N",
+		if (e)
+			hl_engine_data_sprintf(e, fmt, i, is_eng_idle ? "Y" : "N",
 				qm_glbl_sts0, cmdq_glbl_sts0, tpc_cfg_sts);
 	}
 
-	if (s)
-		seq_puts(s,
+	if (e)
+		hl_engine_data_sprintf(e,
 			"\nMME  is_idle  QM_GLBL_STS0  CMDQ_GLBL_STS0  ARCH_STATUS\n"
 			"---  -------  ------------  --------------  -----------\n");
 
@@ -5207,10 +5193,10 @@ static bool goya_is_device_idle(struct hl_device *hdev, u64 *mask_arr,
 
 	if (mask && !is_eng_idle)
 		set_bit(GOYA_ENGINE_ID_MME_0, mask);
-	if (s) {
-		seq_printf(s, fmt, 0, is_eng_idle ? "Y" : "N", qm_glbl_sts0,
+	if (e) {
+		hl_engine_data_sprintf(e, fmt, 0, is_eng_idle ? "Y" : "N", qm_glbl_sts0,
 				cmdq_glbl_sts0, mme_arch_sts);
-		seq_puts(s, "\n");
+		hl_engine_data_sprintf(e, "\n");
 	}
 
 	return is_idle;
@@ -5434,6 +5420,11 @@ static int goya_scrub_device_dram(struct hl_device *hdev, u64 val)
 	return -EOPNOTSUPP;
 }
 
+static int goya_send_device_activity(struct hl_device *hdev, bool open)
+{
+	return 0;
+}
+
 static const struct hl_asic_funcs goya_funcs = {
 	.early_init = goya_early_init,
 	.early_fini = goya_early_fini,
@@ -5478,11 +5469,9 @@ static const struct hl_asic_funcs goya_funcs = {
 	.send_heartbeat = goya_send_heartbeat,
 	.debug_coresight = goya_debug_coresight,
 	.is_device_idle = goya_is_device_idle,
-	.non_hard_reset_late_init = goya_non_hard_reset_late_init,
+	.compute_reset_late_init = goya_compute_reset_late_init,
 	.hw_queues_lock = goya_hw_queues_lock,
 	.hw_queues_unlock = goya_hw_queues_unlock,
-	.kdma_lock = NULL,
-	.kdma_unlock = NULL,
 	.get_pci_id = goya_get_pci_id,
 	.get_eeprom_data = goya_get_eeprom_data,
 	.get_monitor_dump = goya_get_monitor_dump,
@@ -5528,6 +5517,7 @@ static const struct hl_asic_funcs goya_funcs = {
 	.mmu_get_real_page_size = hl_mmu_get_real_page_size,
 	.access_dev_mem = hl_access_dev_mem,
 	.set_dram_bar_base = goya_set_ddr_bar_base,
+	.send_device_activity = goya_send_device_activity,
 };
 
 /*

drivers/misc/habanalabs/include/common/cpucp_if.h

@@ -629,6 +629,12 @@ enum pq_init_status {
  * CPUCP_PACKET_ENGINE_CORE_ASID_SET -
  *       Packet to perform engine core ASID configuration
  *
+ * CPUCP_PACKET_SEC_ATTEST_GET -
+ *       Get the attestaion data that is collected during various stages of the
+ *       boot sequence. the attestation data is also hashed with some unique
+ *       number (nonce) provided by the host to prevent replay attacks.
+ *       public key and certificate also provided as part of the FW response.
+ *
  * CPUCP_PACKET_MONITOR_DUMP_GET -
  *       Get monitors registers dump from the CpuCP kernel.
  *       The CPU will put the registers dump in the a buffer allocated by the driver
@@ -636,6 +642,10 @@ enum pq_init_status {
  *       passes the max size it allows the CpuCP to write to the structure, to prevent
  *       data corruption in case of mismatched driver/FW versions.
  *       Relevant only to Gaudi.
+ *
+ * CPUCP_PACKET_ACTIVE_STATUS_SET -
+ *       LKD sends FW indication whether device is free or in use, this indication is reported
+ *       also to the BMC.
  */
 
 enum cpucp_packet_id {
@@ -687,10 +697,17 @@ enum cpucp_packet_id {
 	CPUCP_PACKET_RESERVED,			/* not used */
 	CPUCP_PACKET_ENGINE_CORE_ASID_SET,	/* internal */
 	CPUCP_PACKET_RESERVED2,			/* not used */
+	CPUCP_PACKET_SEC_ATTEST_GET,		/* internal */
 	CPUCP_PACKET_RESERVED3,			/* not used */
 	CPUCP_PACKET_RESERVED4,			/* not used */
-	CPUCP_PACKET_RESERVED5,			/* not used */
 	CPUCP_PACKET_MONITOR_DUMP_GET,		/* debugfs */
+	CPUCP_PACKET_RESERVED5,			/* not used */
+	CPUCP_PACKET_RESERVED6,			/* not used */
+	CPUCP_PACKET_RESERVED7,			/* not used */
+	CPUCP_PACKET_RESERVED8,			/* not used */
+	CPUCP_PACKET_RESERVED9,			/* not used */
+	CPUCP_PACKET_ACTIVE_STATUS_SET,		/* internal */
+	CPUCP_PACKET_ID_MAX			/* must be last */
 };
 
 #define CPUCP_PACKET_FENCE_VAL	0xFE8CE7A5
@@ -783,6 +800,9 @@ struct cpucp_packet {
 		 * result cannot be used to hold general purpose data.
 		 */
 		__le32 status_mask;
+
+		/* random, used once number, for security packets */
+		__le32 nonce;
 	};
 
 	/* For NIC requests */
@@ -813,10 +833,25 @@ enum cpucp_led_index {
 	CPUCP_LED2_INDEX
 };
 
+/*
+ * enum cpucp_packet_rc - Error return code
+ * @cpucp_packet_success	-> in case of success.
+ * @cpucp_packet_invalid	-> this is to support Goya and Gaudi platform.
+ * @cpucp_packet_fault		-> in case of processing error like failing to
+ *                                 get device binding or semaphore etc.
+ * @cpucp_packet_invalid_pkt	-> when cpucp packet is un-supported. This is
+ *                                 supported Greco onwards.
+ * @cpucp_packet_invalid_params	-> when checking parameter like length of buffer
+ *				   or attribute value etc. Supported Greco onwards.
+ * @cpucp_packet_rc_max		-> It indicates size of enum so should be at last.
+ */
 enum cpucp_packet_rc {
 	cpucp_packet_success,
 	cpucp_packet_invalid,
-	cpucp_packet_fault
+	cpucp_packet_fault,
+	cpucp_packet_invalid_pkt,
+	cpucp_packet_invalid_params,
+	cpucp_packet_rc_max
 };
 
 /*
@@ -1193,6 +1228,70 @@ enum cpu_reset_status {
 	CPU_RST_STATUS_SOFT_RST_DONE = 1,
 };
 
+#define SEC_PCR_DATA_BUF_SZ	256
+#define SEC_PCR_QUOTE_BUF_SZ	510	/* (512 - 2) 2 bytes used for size */
+#define SEC_SIGNATURE_BUF_SZ	255	/* (256 - 1) 1 byte used for size */
+#define SEC_PUB_DATA_BUF_SZ	510	/* (512 - 2) 2 bytes used for size */
+#define SEC_CERTIFICATE_BUF_SZ	2046	/* (2048 - 2) 2 bytes used for size */
+
+/*
+ * struct cpucp_sec_attest_info - attestation report of the boot
+ * @pcr_data: raw values of the PCR registers
+ * @pcr_num_reg: number of PCR registers in the pcr_data array
+ * @pcr_reg_len: length of each PCR register in the pcr_data array (bytes)
+ * @nonce: number only used once. random number provided by host. this also
+ *	    passed to the quote command as a qualifying data.
+ * @pcr_quote_len: length of the attestation quote data (bytes)
+ * @pcr_quote: attestation report data structure
+ * @quote_sig_len: length of the attestation report signature (bytes)
+ * @quote_sig: signature structure of the attestation report
+ * @pub_data_len: length of the public data (bytes)
+ * @public_data: public key for the signed attestation
+ *		 (outPublic + name + qualifiedName)
+ * @certificate_len: length of the certificate (bytes)
+ * @certificate: certificate for the attestation signing key
+ */
+struct cpucp_sec_attest_info {
+	__u8 pcr_data[SEC_PCR_DATA_BUF_SZ];
+	__u8 pcr_num_reg;
+	__u8 pcr_reg_len;
+	__le16 pad0;
+	__le32 nonce;
+	__le16 pcr_quote_len;
+	__u8 pcr_quote[SEC_PCR_QUOTE_BUF_SZ];
+	__u8 quote_sig_len;
+	__u8 quote_sig[SEC_SIGNATURE_BUF_SZ];
+	__le16 pub_data_len;
+	__u8 public_data[SEC_PUB_DATA_BUF_SZ];
+	__le16 certificate_len;
+	__u8 certificate[SEC_CERTIFICATE_BUF_SZ];
+};
+
+/*
+ * struct cpucp_dev_info_signed - device information signed by a secured device
+ * @info: device information structure as defined above
+ * @nonce: number only used once. random number provided by host. this number is
+ *	   hashed and signed along with the device information.
+ * @info_sig_len: length of the attestation signature (bytes)
+ * @info_sig: signature of the info + nonce data.
+ * @pub_data_len: length of the public data (bytes)
+ * @public_data: public key info signed info data
+ *		 (outPublic + name + qualifiedName)
+ * @certificate_len: length of the certificate (bytes)
+ * @certificate: certificate for the signing key
+ */
+struct cpucp_dev_info_signed {
+	struct cpucp_info info;	/* assumed to be 64bit aligned */
+	__le32 nonce;
+	__le32 pad0;
+	__u8 info_sig_len;
+	__u8 info_sig[SEC_SIGNATURE_BUF_SZ];
+	__le16 pub_data_len;
+	__u8 public_data[SEC_PUB_DATA_BUF_SZ];
+	__le16 certificate_len;
+	__u8 certificate[SEC_CERTIFICATE_BUF_SZ];
+};
+
 /*
  * struct dcore_monitor_regs_data - DCORE monitor regs data.
  * the structure follows sync manager block layout. relevant only to Gaudi.

drivers/misc/habanalabs/include/common/hl_boot_if.h

@@ -34,6 +34,7 @@ enum cpu_boot_err {
 	CPU_BOOT_ERR_BINNING_FAIL = 19,
 	CPU_BOOT_ERR_TPM_FAIL = 20,
 	CPU_BOOT_ERR_TMP_THRESH_INIT_FAIL = 21,
+	CPU_BOOT_ERR_EEPROM_FAIL = 22,
 	CPU_BOOT_ERR_ENABLED = 31,
 	CPU_BOOT_ERR_SCND_EN = 63,
 	CPU_BOOT_ERR_LAST = 64 /* we have 2 registers of 32 bits */
@@ -115,6 +116,9 @@ enum cpu_boot_err {
  * CPU_BOOT_ERR0_TMP_THRESH_INIT_FAIL	Failed to set threshold for tmperature
  *					sensor.
  *
+ * CPU_BOOT_ERR_EEPROM_FAIL		Failed reading EEPROM data. Defaults
+ *					are used.
+ *
  * CPU_BOOT_ERR0_ENABLED		Error registers enabled.
  *					This is a main indication that the
  *					running FW populates the error
@@ -139,6 +143,7 @@ enum cpu_boot_err {
 #define CPU_BOOT_ERR0_BINNING_FAIL		(1 << CPU_BOOT_ERR_BINNING_FAIL)
 #define CPU_BOOT_ERR0_TPM_FAIL			(1 << CPU_BOOT_ERR_TPM_FAIL)
 #define CPU_BOOT_ERR0_TMP_THRESH_INIT_FAIL	(1 << CPU_BOOT_ERR_TMP_THRESH_INIT_FAIL)
+#define CPU_BOOT_ERR0_EEPROM_FAIL		(1 << CPU_BOOT_ERR_EEPROM_FAIL)
 #define CPU_BOOT_ERR0_ENABLED			(1 << CPU_BOOT_ERR_ENABLED)
 #define CPU_BOOT_ERR1_ENABLED			(1 << CPU_BOOT_ERR_ENABLED)
 
@@ -426,7 +431,9 @@ struct cpu_dyn_regs {
 	__le32 gic_host_ints_irq;
 	__le32 gic_host_soft_rst_irq;
 	__le32 gic_rot_qm_irq_ctrl;
-	__le32 reserved1[22];		/* reserve for future use */
+	__le32 cpu_rst_status;
+	__le32 eng_arc_irq_ctrl;
+	__le32 reserved1[20];		/* reserve for future use */
 };
 
 /* TODO: remove the desc magic after the code is updated to use message */
@@ -465,6 +472,26 @@ enum comms_msg_type {
 	HL_COMMS_BINNING_CONF_TYPE = 3,
 };
 
+/*
+ * Binning information shared between LKD and FW
+ * @tpc_mask - TPC binning information
+ * @dec_mask - Decoder binning information
+ * @hbm_mask - HBM binning information
+ * @edma_mask - EDMA binning information
+ * @mme_mask_l - MME binning information lower 32
+ * @mme_mask_h - MME binning information upper 32
+ * @reserved - reserved field for 64 bit alignment
+ */
+struct lkd_fw_binning_info {
+	__le64 tpc_mask;
+	__le32 dec_mask;
+	__le32 hbm_mask;
+	__le32 edma_mask;
+	__le32 mme_mask_l;
+	__le32 mme_mask_h;
+	__le32 reserved;
+};
+
 /* TODO: remove this struct after the code is updated to use message */
 /* this is the comms descriptor header - meta data */
 struct comms_desc_header {
@@ -525,13 +552,7 @@ struct lkd_fw_comms_msg {
 		struct {
 			__u8 fw_cfg_skip; /* 1 - skip, 0 - don't skip */
 		};
-		struct {
-			__le64 tpc_binning_conf;
-			__le32 dec_binning_conf;
-			__le32 hbm_binning_conf;
-			__le32 edma_binning_conf;
-			__le32 mme_redundancy_conf; /* use MME_REDUNDANT_COLUMN */
-		};
+		struct lkd_fw_binning_info binning_info;
 	};
 };
 

drivers/misc/habanalabs/include/gaudi2/asic_reg/gaudi2_regs.h

@@ -132,6 +132,7 @@
 #include "dcore0_mme_ctrl_lo_arch_tensor_a_regs.h"
 #include "dcore0_mme_ctrl_lo_arch_tensor_b_regs.h"
 #include "dcore0_mme_ctrl_lo_arch_tensor_cout_regs.h"
+#include "pcie_wrap_special_regs.h"
 
 #include "pdma0_qm_masks.h"
 #include "pdma0_core_masks.h"
@@ -239,6 +240,7 @@
 #define SFT_IF_RTR_OFFSET	(mmSFT0_HBW_RTR_IF1_RTR_H3_BASE - mmSFT0_HBW_RTR_IF0_RTR_H3_BASE)
 
 #define ARC_HALT_REQ_OFFSET	(mmARC_FARM_ARC0_AUX_RUN_HALT_REQ - mmARC_FARM_ARC0_AUX_BASE)
+#define ARC_HALT_ACK_OFFSET	(mmARC_FARM_ARC0_AUX_RUN_HALT_ACK - mmARC_FARM_ARC0_AUX_BASE)
 
 #define ARC_REGION_CFG_OFFSET(region) \
 	(mmARC_FARM_ARC0_AUX_ARC_REGION_CFG_0 + (region * 4) - mmARC_FARM_ARC0_AUX_BASE)

drivers/misc/habanalabs/include/gaudi2/asic_reg/pcie_wrap_special_regs.h

+/* SPDX-License-Identifier: GPL-2.0
+ *
+ * Copyright 2016-2020 HabanaLabs, Ltd.
+ * All Rights Reserved.
+ *
+ */
+
+/************************************
+ ** This is an auto-generated file **
+ **       DO NOT EDIT BELOW        **
+ ************************************/
+
+#ifndef ASIC_REG_PCIE_WRAP_SPECIAL_REGS_H_
+#define ASIC_REG_PCIE_WRAP_SPECIAL_REGS_H_
+
+/*
+ *****************************************
+ *   PCIE_WRAP_SPECIAL
+ *   (Prototype: SPECIAL_REGS)
+ *****************************************
+ */
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_0 0x4C01E80
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_1 0x4C01E84
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_2 0x4C01E88
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_3 0x4C01E8C
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_4 0x4C01E90
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_5 0x4C01E94
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_6 0x4C01E98
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_7 0x4C01E9C
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_8 0x4C01EA0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_9 0x4C01EA4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_10 0x4C01EA8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_11 0x4C01EAC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_12 0x4C01EB0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_13 0x4C01EB4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_14 0x4C01EB8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_15 0x4C01EBC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_16 0x4C01EC0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_17 0x4C01EC4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_18 0x4C01EC8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_19 0x4C01ECC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_20 0x4C01ED0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_21 0x4C01ED4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_22 0x4C01ED8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_23 0x4C01EDC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_24 0x4C01EE0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_25 0x4C01EE4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_26 0x4C01EE8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_27 0x4C01EEC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_28 0x4C01EF0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_29 0x4C01EF4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_30 0x4C01EF8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_PRIV_31 0x4C01EFC
+
+#define mmPCIE_WRAP_SPECIAL_MEM_GW_DATA 0x4C01F00
+
+#define mmPCIE_WRAP_SPECIAL_MEM_GW_REQ 0x4C01F04
+
+#define mmPCIE_WRAP_SPECIAL_MEM_NUMOF 0x4C01F0C
+
+#define mmPCIE_WRAP_SPECIAL_MEM_ECC_SEL 0x4C01F10
+
+#define mmPCIE_WRAP_SPECIAL_MEM_ECC_CTL 0x4C01F14
+
+#define mmPCIE_WRAP_SPECIAL_MEM_ECC_ERR_MASK 0x4C01F18
+
+#define mmPCIE_WRAP_SPECIAL_MEM_ECC_GLBL_ERR_MASK 0x4C01F1C
+
+#define mmPCIE_WRAP_SPECIAL_MEM_ECC_ERR_STS 0x4C01F20
+
+#define mmPCIE_WRAP_SPECIAL_MEM_ECC_ERR_ADDR 0x4C01F24
+
+#define mmPCIE_WRAP_SPECIAL_MEM_RM 0x4C01F28
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_ERR_MASK 0x4C01F40
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_ERR_ADDR 0x4C01F44
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_ERR_CAUSE 0x4C01F48
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SPARE_0 0x4C01F60
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SPARE_1 0x4C01F64
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SPARE_2 0x4C01F68
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SPARE_3 0x4C01F6C
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_0 0x4C01F80
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_1 0x4C01F84
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_2 0x4C01F88
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_3 0x4C01F8C
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_4 0x4C01F90
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_5 0x4C01F94
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_6 0x4C01F98
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_7 0x4C01F9C
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_8 0x4C01FA0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_9 0x4C01FA4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_10 0x4C01FA8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_11 0x4C01FAC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_12 0x4C01FB0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_13 0x4C01FB4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_14 0x4C01FB8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_15 0x4C01FBC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_16 0x4C01FC0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_17 0x4C01FC4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_18 0x4C01FC8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_19 0x4C01FCC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_20 0x4C01FD0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_21 0x4C01FD4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_22 0x4C01FD8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_23 0x4C01FDC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_24 0x4C01FE0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_25 0x4C01FE4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_26 0x4C01FE8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_27 0x4C01FEC
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_28 0x4C01FF0
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_29 0x4C01FF4
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_30 0x4C01FF8
+
+#define mmPCIE_WRAP_SPECIAL_GLBL_SEC_31 0x4C01FFC
+
+#endif /* ASIC_REG_PCIE_WRAP_SPECIAL_REGS_H_ */

drivers/misc/habanalabs/include/gaudi2/gaudi2_async_virt_events.h

-/* SPDX-License-Identifier: GPL-2.0
- *
- * Copyright 2022 HabanaLabs, Ltd.
- * All Rights Reserved.
- *
- */
-
-#ifndef __GAUDI2_ASYNC_VIRT_EVENTS_H_
-#define __GAUDI2_ASYNC_VIRT_EVENTS_H_
-
-enum gaudi2_async_virt_event_id {
-	GAUDI2_EVENT_NIC3_QM1_OLD = 1206,
-	GAUDI2_EVENT_NIC4_QM0_OLD = 1207,
-	GAUDI2_EVENT_NIC4_QM1_OLD = 1208,
-	GAUDI2_EVENT_NIC5_QM0_OLD = 1209,
-	GAUDI2_EVENT_NIC5_QM1_OLD = 1210,
-	GAUDI2_EVENT_NIC6_QM0_OLD = 1211,
-	GAUDI2_EVENT_NIC6_QM1_OLD = 1212,
-	GAUDI2_EVENT_NIC7_QM0_OLD = 1213,
-	GAUDI2_EVENT_NIC7_QM1_OLD = 1214,
-	GAUDI2_EVENT_NIC8_QM0_OLD = 1215,
-	GAUDI2_EVENT_NIC8_QM1_OLD = 1216,
-	GAUDI2_EVENT_NIC9_QM0_OLD = 1217,
-	GAUDI2_EVENT_NIC9_QM1_OLD = 1218,
-	GAUDI2_EVENT_NIC10_QM0_OLD = 1219,
-	GAUDI2_EVENT_NIC10_QM1_OLD = 1220,
-	GAUDI2_EVENT_NIC11_QM0_OLD = 1221,
-	GAUDI2_EVENT_NIC11_QM1_OLD = 1222,
-	GAUDI2_EVENT_CPU_PKT_SANITY_FAILED_OLD = 1223,
-	GAUDI2_EVENT_CPU0_STATUS_NIC0_ENG0_OLD = 1224,
-	GAUDI2_EVENT_CPU0_STATUS_NIC0_ENG1_OLD = 1225,
-	GAUDI2_EVENT_CPU1_STATUS_NIC1_ENG0_OLD = 1226,
-	GAUDI2_EVENT_CPU1_STATUS_NIC1_ENG1_OLD = 1227,
-	GAUDI2_EVENT_CPU2_STATUS_NIC2_ENG0_OLD = 1228,
-	GAUDI2_EVENT_CPU2_STATUS_NIC2_ENG1_OLD = 1229,
-	GAUDI2_EVENT_CPU3_STATUS_NIC3_ENG0_OLD = 1230,
-	GAUDI2_EVENT_CPU3_STATUS_NIC3_ENG1_OLD = 1231,
-	GAUDI2_EVENT_CPU4_STATUS_NIC4_ENG0_OLD = 1232,
-	GAUDI2_EVENT_CPU4_STATUS_NIC4_ENG1_OLD = 1233,
-	GAUDI2_EVENT_CPU5_STATUS_NIC5_ENG0_OLD = 1234,
-	GAUDI2_EVENT_CPU5_STATUS_NIC5_ENG1_OLD = 1235,
-	GAUDI2_EVENT_CPU6_STATUS_NIC6_ENG0_OLD = 1236,
-	GAUDI2_EVENT_CPU6_STATUS_NIC6_ENG1_OLD = 1237,
-	GAUDI2_EVENT_CPU7_STATUS_NIC7_ENG0_OLD = 1238,
-	GAUDI2_EVENT_CPU7_STATUS_NIC7_ENG1_OLD = 1239,
-	GAUDI2_EVENT_CPU8_STATUS_NIC8_ENG0_OLD = 1240,
-	GAUDI2_EVENT_CPU8_STATUS_NIC8_ENG1_OLD = 1241,
-	GAUDI2_EVENT_CPU9_STATUS_NIC9_ENG0_OLD = 1242,
-	GAUDI2_EVENT_CPU9_STATUS_NIC9_ENG1_OLD = 1243,
-	GAUDI2_EVENT_CPU10_STATUS_NIC10_ENG0_OLD = 1244,
-	GAUDI2_EVENT_CPU10_STATUS_NIC10_ENG1_OLD = 1245,
-	GAUDI2_EVENT_CPU11_STATUS_NIC11_ENG0_OLD = 1246,
-	GAUDI2_EVENT_CPU11_STATUS_NIC11_ENG1_OLD = 1247,
-	GAUDI2_EVENT_ARC_DCCM_FULL_OLD = 1248,
-};
-
-#endif /* __GAUDI2_ASYNC_VIRT_EVENTS_H_ */

include/trace/events/habanalabs.h

+/* SPDX-License-Identifier: GPL-2.0
+ *
+ * Copyright 2016-2021 HabanaLabs, Ltd.
+ * All Rights Reserved.
+ *
+ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM habanalabs
+
+#if !defined(_TRACE_HABANALABS_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_HABANALABS_H
+
+#include <linux/tracepoint.h>
+
+DECLARE_EVENT_CLASS(habanalabs_mmu_template,
+	TP_PROTO(struct device *dev, u64 virt_addr, u64 phys_addr, u32 page_size, bool flush_pte),
+
+	TP_ARGS(dev, virt_addr, phys_addr, page_size, flush_pte),
+
+	TP_STRUCT__entry(
+		__string(dname, dev_name(dev))
+		__field(u64, virt_addr)
+		__field(u64, phys_addr)
+		__field(u32, page_size)
+		__field(u8, flush_pte)
+	),
+
+	TP_fast_assign(
+		__assign_str(dname, dev_name(dev));
+		__entry->virt_addr = virt_addr;
+		__entry->phys_addr = phys_addr;
+		__entry->page_size = page_size;
+		__entry->flush_pte = flush_pte;
+	),
+
+	TP_printk("%s: vaddr: %#llx, paddr: %#llx, psize: %#x, flush: %s",
+		__get_str(dname),
+		__entry->virt_addr,
+		__entry->phys_addr,
+		__entry->page_size,
+		__entry->flush_pte ? "true" : "false")
+);
+
+DEFINE_EVENT(habanalabs_mmu_template, habanalabs_mmu_map,
+	TP_PROTO(struct device *dev, u64 virt_addr, u64 phys_addr, u32 page_size, bool flush_pte),
+	TP_ARGS(dev, virt_addr, phys_addr, page_size, flush_pte));
+
+DEFINE_EVENT(habanalabs_mmu_template, habanalabs_mmu_unmap,
+	TP_PROTO(struct device *dev, u64 virt_addr, u64 phys_addr, u32 page_size, bool flush_pte),
+	TP_ARGS(dev, virt_addr, phys_addr, page_size, flush_pte));
+
+DECLARE_EVENT_CLASS(habanalabs_dma_alloc_template,
+	TP_PROTO(struct device *dev, u64 cpu_addr, u64 dma_addr, size_t size, const char *caller),
+
+	TP_ARGS(dev, cpu_addr, dma_addr, size, caller),
+
+	TP_STRUCT__entry(
+		__string(dname, dev_name(dev))
+		__field(u64, cpu_addr)
+		__field(u64, dma_addr)
+		__field(u32, size)
+		__field(const char *, caller)
+	),
+
+	TP_fast_assign(
+		__assign_str(dname, dev_name(dev));
+		__entry->cpu_addr = cpu_addr;
+		__entry->dma_addr = dma_addr;
+		__entry->size = size;
+		__entry->caller = caller;
+	),
+
+	TP_printk("%s: cpu_addr: %#llx, dma_addr: %#llx, size: %#x, caller: %s",
+		__get_str(dname),
+		__entry->cpu_addr,
+		__entry->dma_addr,
+		__entry->size,
+		__entry->caller)
+);
+
+DEFINE_EVENT(habanalabs_dma_alloc_template, habanalabs_dma_alloc,
+	TP_PROTO(struct device *dev, u64 cpu_addr, u64 dma_addr, size_t size, const char *caller),
+	TP_ARGS(dev, cpu_addr, dma_addr, size, caller));
+
+DEFINE_EVENT(habanalabs_dma_alloc_template, habanalabs_dma_free,
+	TP_PROTO(struct device *dev, u64 cpu_addr, u64 dma_addr, size_t size, const char *caller),
+	TP_ARGS(dev, cpu_addr, dma_addr, size, caller));
+
+#endif /* if !defined(_TRACE_HABANALABS_H) || defined(TRACE_HEADER_MULTI_READ) */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>