drm/i915: The return of i915_gpu_info to debugfs

Once upon a time before we had automated GPU state capture upon hangs,
we had intel_gpu_dump. Now we come almost full circle and reinstate that
view of the current GPU queues and registers by using the error capture
facility to snapshot the GPU state when debugfs/.../i915_gpu_info is
opened - which should provided useful debugging to both the error
capture routines (without having to cause a hang and avoid the error
state being eaten by igt) and generally.

v2: Rename drm_i915_error_state to i915_gpu_state to alleviate some name
collisions between the error state dump and inspecting the gpu state.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170214164611.11381-1-chris@chris-wilson.co.uk

1 files changed, 88 insertions, 77 deletions

diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index c28ccfef84ab..3a3c7c3c4931 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -342,7 +342,7 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
 }
 
 static void error_print_instdone(struct drm_i915_error_state_buf *m,
-				 struct drm_i915_error_engine *ee)
+				 const struct drm_i915_error_engine *ee)
 {
 	int slice;
 	int subslice;
@@ -372,7 +372,7 @@ static void error_print_instdone(struct drm_i915_error_state_buf *m,
 
 static void error_print_request(struct drm_i915_error_state_buf *m,
 				const char *prefix,
-				struct drm_i915_error_request *erq)
+				const struct drm_i915_error_request *erq)
 {
 	if (!erq->seqno)
 		return;
@@ -386,7 +386,7 @@ static void error_print_request(struct drm_i915_error_state_buf *m,
 
 static void error_print_context(struct drm_i915_error_state_buf *m,
 				const char *header,
-				struct drm_i915_error_context *ctx)
+				const struct drm_i915_error_context *ctx)
 {
 	err_printf(m, "%s%s[%d] user_handle %d hw_id %d, ban score %d guilty %d active %d\n",
 		   header, ctx->comm, ctx->pid, ctx->handle, ctx->hw_id,
@@ -394,7 +394,7 @@ static void error_print_context(struct drm_i915_error_state_buf *m,
 }
 
 static void error_print_engine(struct drm_i915_error_state_buf *m,
-			       struct drm_i915_error_engine *ee)
+			       const struct drm_i915_error_engine *ee)
 {
 	err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
 	err_printf(m, "  START: 0x%08x\n", ee->start);
@@ -569,21 +569,32 @@ static void err_print_params(struct drm_i915_error_state_buf *m,
 #undef PRINT
 }
 
+static void err_print_pciid(struct drm_i915_error_state_buf *m,
+			    struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = i915->drm.pdev;
+
+	err_printf(m, "PCI ID: 0x%04x\n", pdev->device);
+	err_printf(m, "PCI Revision: 0x%02x\n", pdev->revision);
+	err_printf(m, "PCI Subsystem: %04x:%04x\n",
+		   pdev->subsystem_vendor,
+		   pdev->subsystem_device);
+}
+
 int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
-			    const struct i915_error_state_file_priv *error_priv)
+			    const struct i915_gpu_state *error)
 {
-	struct drm_i915_private *dev_priv = error_priv->i915;
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	struct drm_i915_error_state *error = error_priv->error;
+	struct drm_i915_private *dev_priv = m->i915;
 	struct drm_i915_error_object *obj;
 	int i, j;
 
 	if (!error) {
-		err_printf(m, "no error state collected\n");
-		goto out;
+		err_printf(m, "No error state collected\n");
+		return 0;
 	}
 
-	err_printf(m, "%s\n", error->error_msg);
+	if (*error->error_msg)
+		err_printf(m, "%s\n", error->error_msg);
 	err_printf(m, "Kernel: " UTS_RELEASE "\n");
 	err_printf(m, "Time: %ld s %ld us\n",
 		   error->time.tv_sec, error->time.tv_usec);
@@ -605,11 +616,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 	err_printf(m, "Reset count: %u\n", error->reset_count);
 	err_printf(m, "Suspend count: %u\n", error->suspend_count);
 	err_printf(m, "Platform: %s\n", intel_platform_name(error->device_info.platform));
-	err_printf(m, "PCI ID: 0x%04x\n", pdev->device);
-	err_printf(m, "PCI Revision: 0x%02x\n", pdev->revision);
-	err_printf(m, "PCI Subsystem: %04x:%04x\n",
-		   pdev->subsystem_vendor,
-		   pdev->subsystem_device);
+	err_print_pciid(m, error->i915);
 
 	err_printf(m, "IOMMU enabled?: %d\n", error->iommu);
 
@@ -625,19 +632,15 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 
 	err_printf(m, "EIR: 0x%08x\n", error->eir);
 	err_printf(m, "IER: 0x%08x\n", error->ier);
-	if (INTEL_GEN(dev_priv) >= 8) {
-		for (i = 0; i < 4; i++)
-			err_printf(m, "GTIER gt %d: 0x%08x\n", i,
-				   error->gtier[i]);
-	} else if (HAS_PCH_SPLIT(dev_priv) || IS_VALLEYVIEW(dev_priv))
-		err_printf(m, "GTIER: 0x%08x\n", error->gtier[0]);
+	for (i = 0; i < error->ngtier; i++)
+		err_printf(m, "GTIER[%d]: 0x%08x\n", i, error->gtier[i]);
 	err_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
 	err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
 	err_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
 	err_printf(m, "CCID: 0x%08x\n", error->ccid);
 	err_printf(m, "Missed interrupts: 0x%08lx\n", dev_priv->gpu_error.missed_irq_rings);
 
-	for (i = 0; i < dev_priv->num_fence_regs; i++)
+	for (i = 0; i < error->nfence; i++)
 		err_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
 
 	if (INTEL_GEN(dev_priv) >= 6) {
@@ -686,7 +689,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 			    error->pinned_bo_count);
 
 	for (i = 0; i < ARRAY_SIZE(error->engine); i++) {
-		struct drm_i915_error_engine *ee = &error->engine[i];
+		const struct drm_i915_error_engine *ee = &error->engine[i];
 
 		obj = ee->batchbuffer;
 		if (obj) {
@@ -751,12 +754,11 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		intel_overlay_print_error_state(m, error->overlay);
 
 	if (error->display)
-		intel_display_print_error_state(m, dev_priv, error->display);
+		intel_display_print_error_state(m, error->display);
 
 	err_print_capabilities(m, &error->device_info);
 	err_print_params(m, &error->params);
 
-out:
 	if (m->bytes == 0 && m->err)
 		return m->err;
 
@@ -808,10 +810,10 @@ static void i915_error_object_free(struct drm_i915_error_object *obj)
 	kfree(obj);
 }
 
-static void i915_error_state_free(struct kref *error_ref)
+void __i915_gpu_state_free(struct kref *error_ref)
 {
-	struct drm_i915_error_state *error = container_of(error_ref,
-							  typeof(*error), ref);
+	struct i915_gpu_state *error =
+		container_of(error_ref, typeof(*error), ref);
 	int i;
 
 	for (i = 0; i < ARRAY_SIZE(error->engine); i++) {
@@ -976,7 +978,7 @@ static u32 capture_error_bo(struct drm_i915_error_buffer *err,
  * It's only a small step better than a random number in its current form.
  */
 static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv,
-					 struct drm_i915_error_state *error,
+					 struct i915_gpu_state *error,
 					 int *engine_id)
 {
 	uint32_t error_code = 0;
@@ -1001,20 +1003,21 @@ static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv,
 }
 
 static void i915_gem_record_fences(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error)
+				   struct i915_gpu_state *error)
 {
 	int i;
 
-	if (IS_GEN3(dev_priv) || IS_GEN2(dev_priv)) {
+	if (INTEL_GEN(dev_priv) >= 6) {
 		for (i = 0; i < dev_priv->num_fence_regs; i++)
-			error->fence[i] = I915_READ(FENCE_REG(i));
-	} else if (IS_GEN5(dev_priv) || IS_GEN4(dev_priv)) {
+			error->fence[i] = I915_READ64(FENCE_REG_GEN6_LO(i));
+	} else if (INTEL_GEN(dev_priv) >= 4) {
 		for (i = 0; i < dev_priv->num_fence_regs; i++)
 			error->fence[i] = I915_READ64(FENCE_REG_965_LO(i));
-	} else if (INTEL_GEN(dev_priv) >= 6) {
+	} else {
 		for (i = 0; i < dev_priv->num_fence_regs; i++)
-			error->fence[i] = I915_READ64(FENCE_REG_GEN6_LO(i));
+			error->fence[i] = I915_READ(FENCE_REG(i));
 	}
+	error->nfence = i;
 }
 
 static inline u32
@@ -1038,7 +1041,7 @@ gen8_engine_sync_index(struct intel_engine_cs *engine,
 	return idx;
 }
 
-static void gen8_record_semaphore_state(struct drm_i915_error_state *error,
+static void gen8_record_semaphore_state(struct i915_gpu_state *error,
 					struct intel_engine_cs *engine,
 					struct drm_i915_error_engine *ee)
 {
@@ -1131,7 +1134,7 @@ static void error_record_engine_waiters(struct intel_engine_cs *engine,
 	spin_unlock_irq(&b->lock);
 }
 
-static void error_record_engine_registers(struct drm_i915_error_state *error,
+static void error_record_engine_registers(struct i915_gpu_state *error,
 					  struct intel_engine_cs *engine,
 					  struct drm_i915_error_engine *ee)
 {
@@ -1328,7 +1331,7 @@ static void record_context(struct drm_i915_error_context *e,
 }
 
 static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
-				  struct drm_i915_error_state *error)
+				  struct i915_gpu_state *error)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	int i;
@@ -1404,7 +1407,7 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 }
 
 static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
-				struct drm_i915_error_state *error,
+				struct i915_gpu_state *error,
 				struct i915_address_space *vm,
 				int idx)
 {
@@ -1430,7 +1433,7 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
 }
 
 static void i915_capture_active_buffers(struct drm_i915_private *dev_priv,
-					struct drm_i915_error_state *error)
+					struct i915_gpu_state *error)
 {
 	int cnt = 0, i, j;
 
@@ -1455,7 +1458,7 @@ static void i915_capture_active_buffers(struct drm_i915_private *dev_priv,
 }
 
 static void i915_capture_pinned_buffers(struct drm_i915_private *dev_priv,
-					struct drm_i915_error_state *error)
+					struct i915_gpu_state *error)
 {
 	struct i915_address_space *vm = &dev_priv->ggtt.base;
 	struct drm_i915_error_buffer *bo;
@@ -1486,7 +1489,7 @@ static void i915_capture_pinned_buffers(struct drm_i915_private *dev_priv,
 }
 
 static void i915_gem_capture_guc_log_buffer(struct drm_i915_private *dev_priv,
-					    struct drm_i915_error_state *error)
+					    struct i915_gpu_state *error)
 {
 	/* Capturing log buf contents won't be useful if logging was disabled */
 	if (!dev_priv->guc.log.vma || (i915.guc_log_level < 0))
@@ -1498,7 +1501,7 @@ static void i915_gem_capture_guc_log_buffer(struct drm_i915_private *dev_priv,
 
 /* Capture all registers which don't fit into another category. */
 static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error)
+				   struct i915_gpu_state *error)
 {
 	int i;
 
@@ -1555,9 +1558,11 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 		error->ier = I915_READ(GEN8_DE_MISC_IER);
 		for (i = 0; i < 4; i++)
 			error->gtier[i] = I915_READ(GEN8_GT_IER(i));
+		error->ngtier = 4;
 	} else if (HAS_PCH_SPLIT(dev_priv)) {
 		error->ier = I915_READ(DEIER);
 		error->gtier[0] = I915_READ(GTIER);
+		error->ngtier = 1;
 	} else if (IS_GEN2(dev_priv)) {
 		error->ier = I915_READ16(IER);
 	} else if (!IS_VALLEYVIEW(dev_priv)) {
@@ -1568,7 +1573,7 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 }
 
 static void i915_error_capture_msg(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error,
+				   struct i915_gpu_state *error,
 				   u32 engine_mask,
 				   const char *error_msg)
 {
@@ -1595,7 +1600,7 @@ static void i915_error_capture_msg(struct drm_i915_private *dev_priv,
 }
 
 static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error)
+				   struct i915_gpu_state *error)
 {
 	error->iommu = -1;
 #ifdef CONFIG_INTEL_IOMMU
@@ -1611,7 +1616,7 @@ static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
 
 static int capture(void *data)
 {
-	struct drm_i915_error_state *error = data;
+	struct i915_gpu_state *error = data;
 
 	do_gettimeofday(&error->time);
 	error->boottime = ktime_to_timeval(ktime_get_boottime());
@@ -1637,6 +1642,23 @@ static int capture(void *data)
 
 #define DAY_AS_SECONDS(x) (24 * 60 * 60 * (x))
 
+struct i915_gpu_state *
+i915_capture_gpu_state(struct drm_i915_private *i915)
+{
+	struct i915_gpu_state *error;
+
+	error = kzalloc(sizeof(*error), GFP_ATOMIC);
+	if (!error)
+		return NULL;
+
+	kref_init(&error->ref);
+	error->i915 = i915;
+
+	stop_machine(capture, error, NULL);
+
+	return error;
+}
+
 /**
  * i915_capture_error_state - capture an error record for later analysis
  * @dev: drm device
@@ -1651,7 +1673,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 			      const char *error_msg)
 {
 	static bool warned;
-	struct drm_i915_error_state *error;
+	struct i915_gpu_state *error;
 	unsigned long flags;
 
 	if (!i915.error_capture)
@@ -1660,18 +1682,12 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	if (READ_ONCE(dev_priv->gpu_error.first_error))
 		return;
 
-	/* Account for pipe specific data like PIPE*STAT */
-	error = kzalloc(sizeof(*error), GFP_ATOMIC);
+	error = i915_capture_gpu_state(dev_priv);
 	if (!error) {
 		DRM_DEBUG_DRIVER("out of memory, not capturing error state\n");
 		return;
 	}
 
-	kref_init(&error->ref);
-	error->i915 = dev_priv;
-
-	stop_machine(capture, error, NULL);
-
 	i915_error_capture_msg(dev_priv, error, engine_mask, error_msg);
 	DRM_INFO("%s\n", error->error_msg);
 
@@ -1685,7 +1701,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	}
 
 	if (error) {
-		i915_error_state_free(&error->ref);
+		__i915_gpu_state_free(&error->ref);
 		return;
 	}
 
@@ -1701,33 +1717,28 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	}
 }
 
-void i915_error_state_get(struct drm_device *dev,
-			  struct i915_error_state_file_priv *error_priv)
+struct i915_gpu_state *
+i915_first_error_state(struct drm_i915_private *i915)
 {
-	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct i915_gpu_state *error;
 
-	spin_lock_irq(&dev_priv->gpu_error.lock);
-	error_priv->error = dev_priv->gpu_error.first_error;
-	if (error_priv->error)
-		kref_get(&error_priv->error->ref);
-	spin_unlock_irq(&dev_priv->gpu_error.lock);
-}
+	spin_lock_irq(&i915->gpu_error.lock);
+	error = i915->gpu_error.first_error;
+	if (error)
+		i915_gpu_state_get(error);
+	spin_unlock_irq(&i915->gpu_error.lock);
 
-void i915_error_state_put(struct i915_error_state_file_priv *error_priv)
-{
-	if (error_priv->error)
-		kref_put(&error_priv->error->ref, i915_error_state_free);
+	return error;
 }
 
-void i915_destroy_error_state(struct drm_i915_private *dev_priv)
+void i915_reset_error_state(struct drm_i915_private *i915)
 {
-	struct drm_i915_error_state *error;
+	struct i915_gpu_state *error;
 
-	spin_lock_irq(&dev_priv->gpu_error.lock);
-	error = dev_priv->gpu_error.first_error;
-	dev_priv->gpu_error.first_error = NULL;
-	spin_unlock_irq(&dev_priv->gpu_error.lock);
+	spin_lock_irq(&i915->gpu_error.lock);
+	error = i915->gpu_error.first_error;
+	i915->gpu_error.first_error = NULL;
+	spin_unlock_irq(&i915->gpu_error.lock);
 
-	if (error)
-		kref_put(&error->ref, i915_error_state_free);
+	i915_gpu_state_put(error);
 }