4 files changed, 53 insertions, 21 deletions

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 72e2be7a6c8..ec690ca40af 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -719,24 +719,19 @@ int i915_reset(struct drm_device *dev)
 
 	simulated = dev_priv->gpu_error.stop_rings != 0;
 
-	if (!simulated && get_seconds() - dev_priv->gpu_error.last_reset < 5) {
-		DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
-		ret = -ENODEV;
-	} else {
-		ret = intel_gpu_reset(dev);
-
-		/* Also reset the gpu hangman. */
-		if (simulated) {
-			DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
-			dev_priv->gpu_error.stop_rings = 0;
-			if (ret == -ENODEV) {
-				DRM_ERROR("Reset not implemented, but ignoring "
-					  "error for simulated gpu hangs\n");
-				ret = 0;
-			}
-		} else
-			dev_priv->gpu_error.last_reset = get_seconds();
+	ret = intel_gpu_reset(dev);
+
+	/* Also reset the gpu hangman. */
+	if (simulated) {
+		DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
+		dev_priv->gpu_error.stop_rings = 0;
+		if (ret == -ENODEV) {
+			DRM_ERROR("Reset not implemented, but ignoring "
+				  "error for simulated gpu hangs\n");
+			ret = 0;
+		}
 	}
+
 	if (ret) {
 		DRM_ERROR("Failed to reset chip.\n");
 		mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e357995a6aa..c5f0abaa9a2 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -586,6 +586,12 @@ struct i915_ctx_hang_stats {
 
 	/* This context had batch active when hang was declared */
 	unsigned batch_active;
+
+	/* Time when this context was last blamed for a GPU reset */
+	unsigned long guilty_ts;
+
+	/* This context is banned to submit more work */
+	bool banned;
 };
 
 /* This must match up with the value previously used for execbuf2.rsvd1. */
@@ -987,6 +993,9 @@ struct i915_gpu_error {
 	/* For hangcheck timer */
 #define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
 #define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
+	/* Hang gpu twice in this window and your context gets banned */
+#define DRM_I915_CTX_BAN_PERIOD DIV_ROUND_UP(8*DRM_I915_HANGCHECK_PERIOD, 1000)
+
 	struct timer_list hangcheck_timer;
 
 	/* For reset and error_state handling. */
@@ -995,8 +1004,6 @@ struct i915_gpu_error {
 	struct drm_i915_error_state *first_error;
 	struct work_struct work;
 
-	unsigned long last_reset;
-
 	/**
 	 * State variable and reset counter controlling the reset flow
 	 *
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 399e159016e..04e810c59d6 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2188,6 +2188,21 @@ static bool i915_request_guilty(struct drm_i915_gem_request *request,
 	return false;
 }
 
+static bool i915_context_is_banned(const struct i915_ctx_hang_stats *hs)
+{
+	const unsigned long elapsed = get_seconds() - hs->guilty_ts;
+
+	if (hs->banned)
+		return true;
+
+	if (elapsed <= DRM_I915_CTX_BAN_PERIOD) {
+		DRM_ERROR("context hanging too fast, declaring banned!\n");
+		return true;
+	}
+
+	return false;
+}
+
 static void i915_set_reset_status(struct intel_ring_buffer *ring,
 				  struct drm_i915_gem_request *request,
 				  u32 acthd)
@@ -2224,10 +2239,13 @@ static void i915_set_reset_status(struct intel_ring_buffer *ring,
 		hs = &request->file_priv->hang_stats;
 
 	if (hs) {
-		if (guilty)
+		if (guilty) {
+			hs->banned = i915_context_is_banned(hs);
 			hs->batch_active++;
-		else
+			hs->guilty_ts = get_seconds();
+		} else {
 			hs->batch_pending++;
+		}
 	}
 }
 
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index e519f9f6e5c..c8a01c14164 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -929,6 +929,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	struct drm_i915_gem_object *batch_obj;
 	struct drm_clip_rect *cliprects = NULL;
 	struct intel_ring_buffer *ring;
+	struct i915_ctx_hang_stats *hs;
 	u32 ctx_id = i915_execbuffer2_get_context_id(*args);
 	u32 exec_start, exec_len;
 	u32 mask, flags;
@@ -1118,6 +1119,17 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	if (ret)
 		goto err;
 
+	hs = i915_gem_context_get_hang_stats(dev, file, ctx_id);
+	if (IS_ERR(hs)) {
+		ret = PTR_ERR(hs);
+		goto err;
+	}
+
+	if (hs->banned) {
+		ret = -EIO;
+		goto err;
+	}
+
 	ret = i915_switch_context(ring, file, ctx_id);
 	if (ret)
 		goto err;