Skip to contentSkip to conversation

drm/i915: Created a sized object error dump

PATCH In review Ben Widawsky 3 participants 8 messages Feb 25, 2013 – Mar 5, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/3 drm/i915: Created a sized object error dump

v2: Actually use num_pages (Chris)

---
 drivers/gpu/drm/i915/i915_irq.c | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/i915_irq.c +10 -8
index 29037e0..420911c 100644
@@ -952,24 +952,23 @@ static void i915_get_extra_instdone(struct drm_device *dev,
 
 #ifdef CONFIG_DEBUG_FS
 static struct drm_i915_error_object *
-i915_error_object_create(struct drm_i915_private *dev_priv,
-			 struct drm_i915_gem_object *src)
+i915_error_object_create_sized(struct drm_i915_private *dev_priv,
+			       struct drm_i915_gem_object *src,
+			       const int num_pages)
 {
 	struct drm_i915_error_object *dst;
-	int i, count;
+	int i;
 	u32 reloc_offset;
 
 	if (src == NULL || src->pages == NULL)
 		return NULL;
 
-	count = src->base.size / PAGE_SIZE;
-
-	dst = kmalloc(sizeof(*dst) + count * sizeof(u32 *), GFP_ATOMIC);
+	dst = kmalloc(sizeof(*dst) + num_pages * sizeof(u32 *), GFP_ATOMIC);
 	if (dst == NULL)
 		return NULL;
 
 	reloc_offset = src->gtt_offset;
-	for (i = 0; i < count; i++) {
+	for (i = 0; i < num_pages; i++) {
 		unsigned long flags;
 		void *d;
 
@@ -1019,7 +1018,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 
 		reloc_offset += PAGE_SIZE;
 	}
-	dst->page_count = count;
+	dst->page_count = num_pages;
 	dst->gtt_offset = src->gtt_offset;
 
 	return dst;
@@ -1030,6 +1029,9 @@ unwind:
 	kfree(dst);
 	return NULL;
 }
+#define i915_error_object_create(dev_priv, src) \
+	i915_error_object_create_sized((dev_priv), (src), \
+				       (src)->base.size>>PAGE_SHIFT)
 
 static void
 i915_error_object_free(struct drm_i915_error_object *obj)
Signature · 1 line
1.8.1.4
#1 lore raw

PATCH v2 3/3 drm/i915: Capture current context on error

On error, this represents the state of the currently running context at the time it was loaded.

Unfortunately, since we're hung and can't switch out the context this may not tell us too much about the most current state of the context, but does give clues about what has happened since loading.

Thanks to recent doc updates, we have a little more confidence regarding what is actually in this memory, and perhaps it will help us gain more insight into certain bugs. AFAICT, the most interesting info is in the first page. To save space, we only capture the first page. In the future, we might want to dump more.

Sample of the relevant part of error state:
render ring --- HW Context = 0x01b20000
[0000] 00000000 1100105f 00002028 ffff0880
[0010] 0000209c feff4040 000020c0 efdf0080
[0020] 00002178 00000001 0000217c 00145855
[0030] 00002310 00000000 00002314 00000000

v2: Move error collection to the ring error code Change format of dump to not confuse intel_error_decode (Chris) Put the context error object with the others (Chris) Don't search bound_list instead of active_list (chris)

References: https://bugs.freedesktop.org/show_bug.cgi?id=55845

---
 drivers/gpu/drm/i915/i915_debugfs.c | 15 +++++++++++++++
 drivers/gpu/drm/i915/i915_drv.h     |  2 +-
 drivers/gpu/drm/i915/i915_irq.c     | 10 ++++++++++
 3 files changed, 26 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_debugfs.c +15 -0
index 7c65ab8..9b4f3d7 100644
@@ -772,6 +772,21 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
+		if ((obj = error->ring[i].ctx)) {
+			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+				   dev_priv->ring[i].name,
+				   obj->gtt_offset);
+			offset = 0;
+			for (elt = 0; elt < PAGE_SIZE/16; elt+=4) {
+				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+					   offset,
+					   obj->pages[0][elt],
+					   obj->pages[0][elt+1],
+					   obj->pages[0][elt+2],
+					   obj->pages[0][elt+3]);
+					offset += 16;
+			}
+		}
 	}
 
 	if (error->overlay)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index e95337c..bedabcd 100644
@@ -243,7 +243,7 @@ struct drm_i915_error_state {
 			int page_count;
 			u32 gtt_offset;
 			u32 *pages[0];
-		} *ringbuffer, *batchbuffer;
+		} *ringbuffer, *batchbuffer, *ctx;
 		struct drm_i915_error_request {
 			long jiffies;
 			u32 seqno;
drivers/gpu/drm/i915/i915_irq.c +10 -0
index 6a328b8..125b7d8 100644
@@ -1244,6 +1244,7 @@ static void i915_gem_record_rings(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring;
 	struct drm_i915_gem_request *request;
+	struct drm_i915_gem_object *obj;
 	int i, count;
 
 	for_each_ring(ring, dev_priv, i) {
@@ -1255,6 +1256,15 @@ static void i915_gem_record_rings(struct drm_device *dev,
 		error->ring[i].ringbuffer =
 			i915_error_object_create(dev_priv, ring->obj);
 
+		/* Currently render ring is the only HW context user */
+		if ((ring->id == RCS) && error->ccid) {
+			list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list)
+			if ((error->ccid & PAGE_MASK) == obj->gtt_offset)
+				error->ring[i].ctx =
+					i915_error_object_create_sized(dev_priv,
+								       obj, 1);
+		}
+
 		count = 0;
 		list_for_each_entry(request, &ring->request_list, list)
 			count++;
Signature · 1 line
1.8.1.4
Reviewed-by on v2 3/3
#2 lore raw

Re: [PATCH 3/3] [v2] drm/i915: Capture current context on error

Quoted 20 lines · Ben Widawsky

On error, this represents the state of the currently running context at the time it was loaded.

Unfortunately, since we're hung and can't switch out the context this may not tell us too much about the most current state of the context, but does give clues about what has happened since loading.

Thanks to recent doc updates, we have a little more confidence regarding what is actually in this memory, and perhaps it will help us gain more insight into certain bugs. AFAICT, the most interesting info is in the first page. To save space, we only capture the first page. In the future, we might want to dump more.

Sample of the relevant part of error state:
render ring --- HW Context = 0x01b20000
[0000] 00000000 1100105f 00002028 ffff0880
[0010] 0000209c feff4040 000020c0 efdf0080
[0020] 00002178 00000001 0000217c 00145855
[0030] 00002310 00000000 00002314 00000000

Quoted 8 lines

v2: Move error collection to the ring error code Change format of dump to not confuse intel_error_decode (Chris) Put the context error object with the others (Chris) Don't search bound_list instead of active_list (chris)

References: https://bugs.freedesktop.org/show_bug.cgi?id=55845

Can't spot anything else to nag about so all 3 are

-Chris

Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
#3 lore raw

Re: [PATCH 3/3] [v2] drm/i915: Capture current context on error

In reply to Ben Widawsky
Quoted 86 lines · Ben Widawsky

On error, this represents the state of the currently running context at the time it was loaded.

Unfortunately, since we're hung and can't switch out the context this may not tell us too much about the most current state of the context, but does give clues about what has happened since loading.

Thanks to recent doc updates, we have a little more confidence regarding what is actually in this memory, and perhaps it will help us gain more insight into certain bugs. AFAICT, the most interesting info is in the first page. To save space, we only capture the first page. In the future, we might want to dump more.

Sample of the relevant part of error state:
render ring --- HW Context = 0x01b20000
[0000] 00000000 1100105f 00002028 ffff0880
[0010] 0000209c feff4040 000020c0 efdf0080
[0020] 00002178 00000001 0000217c 00145855
[0030] 00002310 00000000 00002314 00000000

v2: Move error collection to the ring error code Change format of dump to not confuse intel_error_decode (Chris) Put the context error object with the others (Chris) Don't search bound_list instead of active_list (chris)

References: https://bugs.freedesktop.org/show_bug.cgi?id=55845

---
 drivers/gpu/drm/i915/i915_debugfs.c | 15 +++++++++++++++
 drivers/gpu/drm/i915/i915_drv.h     |  2 +-
 drivers/gpu/drm/i915/i915_irq.c     | 10 ++++++++++
 3 files changed, 26 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_debugfs.c +15 -0
index 7c65ab8..9b4f3d7 100644
@@ -772,6 +772,21 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
+		if ((obj = error->ring[i].ctx)) {
+			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+				   dev_priv->ring[i].name,
+				   obj->gtt_offset);
+			offset = 0;
+			for (elt = 0; elt < PAGE_SIZE/16; elt+=4) {
+				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+					   offset,
+					   obj->pages[0][elt],
+					   obj->pages[0][elt+1],
+					   obj->pages[0][elt+2],
+					   obj->pages[0][elt+3]);
+					offset += 16;
+			}
+		}
 	}
 
 	if (error->overlay)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index e95337c..bedabcd 100644
@@ -243,7 +243,7 @@ struct drm_i915_error_state {
 			int page_count;
 			u32 gtt_offset;
 			u32 *pages[0];
-		} *ringbuffer, *batchbuffer;
+		} *ringbuffer, *batchbuffer, *ctx;
 		struct drm_i915_error_request {
 			long jiffies;
 			u32 seqno;
drivers/gpu/drm/i915/i915_irq.c +1 -0
index 6a328b8..125b7d8 100644
@@ -1244,6 +1244,7 @@ static void i915_gem_record_rings(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring;
 	struct drm_i915_gem_request *request;
+	struct drm_i915_gem_object *obj;
 	int i, count;
 
 	for_each_ring(ring, dev_priv, i) {

drivers/gpu/drm/i915/i915_irq.c static void i915_gem_record_rings(struct drm_device *dev, · line 1265

Quoted 10 lines
drivers/gpu/drm/i915/i915_irq.c +7 -0
@@ -1255,6 +1256,15 @@ static void i915_gem_record_rings(struct drm_device *dev,
 		error->ring[i].ringbuffer =
 			i915_error_object_create(dev_priv, ring->obj);
 
+		/* Currently render ring is the only HW context user */
+		if ((ring->id == RCS) && error->ccid) {
+			list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list)
+			if ((error->ccid & PAGE_MASK) == obj->gtt_offset)
+				error->ring[i].ctx =
+					i915_error_object_create_sized(dev_priv,
+								       obj, 1);

checkpatch is a bit unhappy about your hunk in i915_debugfs.c. And it complains about the overtly long line here. I wanted to quickly fix this up until I've noticed that the indentation here is a bit ... artistic. Fixing it up indents the code way too much, I think this should be extracted into a tiny helper function, which allows us to flatten the logic with early return;s and continue;s.

First two patches of this series merged to dinq, thanks. -Daniel

Quoted 12 lines
drivers/gpu/drm/i915/i915_irq.c +2 -0
+		}
+
 		count = 0;
 		list_for_each_entry(request, &ring->request_list, list)
 			count++;
Signature · 6 lines
1.8.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#4 lore raw

PATCH v3 1/1 drm/i915: Capture current context on error

In reply to Ben Widawsky

On error, this represents the state of the currently running context at the time it was loaded.

Unfortunately, since we're hung and can't switch out the context this may not tell us too much about the most current state of the context, but does give clues about what has happened since loading.

Thanks to recent doc updates, we have a little more confidence regarding what is actually in this memory, and perhaps it will help us gain more insight into certain bugs. AFAICT, the most interesting info is in the first page. To save space, we only capture the first page. In the future, we might want to dump more.

Sample of the relevant part of error state:
render ring --- HW Context = 0x01b20000
[0000] 00000000 1100105f 00002028 ffff0880
[0010] 0000209c feff4040 000020c0 efdf0080
[0020] 00002178 00000001 0000217c 00145855
[0030] 00002310 00000000 00002314 00000000

v2: Move error collection to the ring error code Change format of dump to not confuse intel_error_decode (Chris) Put the context error object with the others (Chris) Don't search bound_list instead of active_list (chris)

v3: extract and flatten context recording (daniel) checkpatch related fixes for the copypasta in debugfs

References: https://bugs.freedesktop.org/show_bug.cgi?id=55845
Reviewed-by (v2): Chris Wilson <[email protected]>

---
 drivers/gpu/drm/i915/i915_debugfs.c | 17 +++++++++++++++++
 drivers/gpu/drm/i915/i915_drv.h     |  2 +-
 drivers/gpu/drm/i915/i915_irq.c     | 23 +++++++++++++++++++++++
 3 files changed, 41 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_debugfs.c +17 -0
index 7c65ab8..c92ae7f 100644
@@ -772,6 +772,23 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
+
+		obj = error->ring[i].ctx;
+		if (obj) {
+			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+				   dev_priv->ring[i].name,
+				   obj->gtt_offset);
+			offset = 0;
+			for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
+				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+					   offset,
+					   obj->pages[0][elt],
+					   obj->pages[0][elt+1],
+					   obj->pages[0][elt+2],
+					   obj->pages[0][elt+3]);
+					offset += 16;
+			}
+		}
 	}
 
 	if (error->overlay)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index e95337c..bedabcd 100644
@@ -243,7 +243,7 @@ struct drm_i915_error_state {
 			int page_count;
 			u32 gtt_offset;
 			u32 *pages[0];
-		} *ringbuffer, *batchbuffer;
+		} *ringbuffer, *batchbuffer, *ctx;
 		struct drm_i915_error_request {
 			long jiffies;
 			u32 seqno;
drivers/gpu/drm/i915/i915_irq.c +23 -0
index 6a328b8..ac2ad0f 100644
@@ -1238,6 +1238,26 @@ static void i915_record_ring_state(struct drm_device *dev,
 	error->cpu_ring_tail[ring->id] = ring->tail;
 }
 
+
+static void i915_gem_record_active_context(struct intel_ring_buffer *ring,
+					   struct drm_i915_error_state *error,
+					   struct drm_i915_error_ring *ering)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_i915_gem_object *obj;
+
+	/* Currently render ring is the only HW context user */
+	if ((ring->id == RCS) && error->ccid)
+		return;
+
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+		if ((error->ccid & PAGE_MASK) == obj->gtt_offset) {
+			ering->ctx = i915_error_object_create_sized(dev_priv,
+								    obj, 1);
+		}
+	}
+}
+
 static void i915_gem_record_rings(struct drm_device *dev,
 				  struct drm_i915_error_state *error)
 {
@@ -1255,6 +1275,9 @@ static void i915_gem_record_rings(struct drm_device *dev,
 		error->ring[i].ringbuffer =
 			i915_error_object_create(dev_priv, ring->obj);
 
+
+		i915_gem_record_active_context(ring, error, &error->ring[i]);
+
 		count = 0;
 		list_for_each_entry(request, &ring->request_list, list)
 			count++;
Signature · 1 line
1.8.1.5
#5 lore raw

PATCH v4 1/1 drm/i915: Capture current context on error

In reply to Ben Widawsky

On error, this represents the state of the currently running context at the time it was loaded.

Unfortunately, since we're hung and can't switch out the context this may not tell us too much about the most current state of the context, but does give clues about what has happened since loading.

Thanks to recent doc updates, we have a little more confidence regarding what is actually in this memory, and perhaps it will help us gain more insight into certain bugs. AFAICT, the most interesting info is in the first page. To save space, we only capture the first page. In the future, we might want to dump more.

Sample of the relevant part of error state:
render ring --- HW Context = 0x01b20000
[0000] 00000000 1100105f 00002028 ffff0880
[0010] 0000209c feff4040 000020c0 efdf0080
[0020] 00002178 00000001 0000217c 00145855
[0030] 00002310 00000000 00002314 00000000

v2: Move error collection to the ring error code Change format of dump to not confuse intel_error_decode (Chris) Put the context error object with the others (Chris) Don't search bound_list instead of active_list (chris)

v3: extract and flatten context recording (daniel) checkpatch related fixes for the copypasta in debugfs

v4: bug in v3 (Daniel)
- if ((ring->id == RCS) && error->ccid)
+ if ((ring->id != RCS) || !error->ccid)

References: https://bugs.freedesktop.org/show_bug.cgi?id=55845
Reviewed-by (v2): Chris Wilson <[email protected]>

---
 drivers/gpu/drm/i915/i915_debugfs.c | 17 +++++++++++++++++
 drivers/gpu/drm/i915/i915_drv.h     |  2 +-
 drivers/gpu/drm/i915/i915_irq.c     | 23 +++++++++++++++++++++++
 3 files changed, 41 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_debugfs.c +17 -0
index 7c65ab8..c92ae7f 100644
@@ -772,6 +772,23 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
+
+		obj = error->ring[i].ctx;
+		if (obj) {
+			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+				   dev_priv->ring[i].name,
+				   obj->gtt_offset);
+			offset = 0;
+			for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
+				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+					   offset,
+					   obj->pages[0][elt],
+					   obj->pages[0][elt+1],
+					   obj->pages[0][elt+2],
+					   obj->pages[0][elt+3]);
+					offset += 16;
+			}
+		}
 	}
 
 	if (error->overlay)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index e95337c..bedabcd 100644
@@ -243,7 +243,7 @@ struct drm_i915_error_state {
 			int page_count;
 			u32 gtt_offset;
 			u32 *pages[0];
-		} *ringbuffer, *batchbuffer;
+		} *ringbuffer, *batchbuffer, *ctx;
 		struct drm_i915_error_request {
 			long jiffies;
 			u32 seqno;
drivers/gpu/drm/i915/i915_irq.c +23 -0
index 6a328b8..dd95e82 100644
@@ -1238,6 +1238,26 @@ static void i915_record_ring_state(struct drm_device *dev,
 	error->cpu_ring_tail[ring->id] = ring->tail;
 }
 
+
+static void i915_gem_record_active_context(struct intel_ring_buffer *ring,
+					   struct drm_i915_error_state *error,
+					   struct drm_i915_error_ring *ering)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_i915_gem_object *obj;
+
+	/* Currently render ring is the only HW context user */
+	if ((ring->id != RCS) || !error->ccid)
+		return;
+
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+		if ((error->ccid & PAGE_MASK) == obj->gtt_offset) {
+			ering->ctx = i915_error_object_create_sized(dev_priv,
+								    obj, 1);
+		}
+	}
+}
+
 static void i915_gem_record_rings(struct drm_device *dev,
 				  struct drm_i915_error_state *error)
 {
@@ -1255,6 +1275,9 @@ static void i915_gem_record_rings(struct drm_device *dev,
 		error->ring[i].ringbuffer =
 			i915_error_object_create(dev_priv, ring->obj);
 
+
+		i915_gem_record_active_context(ring, error, &error->ring[i]);
+
 		count = 0;
 		list_for_each_entry(request, &ring->request_list, list)
 			count++;
Signature · 1 line
1.8.1.5
#6 lore raw

Re: [PATCH] [v4] drm/i915: Capture current context on error

Quoted 29 lines · Ben Widawsky

On error, this represents the state of the currently running context at the time it was loaded.

Unfortunately, since we're hung and can't switch out the context this may not tell us too much about the most current state of the context, but does give clues about what has happened since loading.

Thanks to recent doc updates, we have a little more confidence regarding what is actually in this memory, and perhaps it will help us gain more insight into certain bugs. AFAICT, the most interesting info is in the first page. To save space, we only capture the first page. In the future, we might want to dump more.

Sample of the relevant part of error state:
render ring --- HW Context = 0x01b20000
[0000] 00000000 1100105f 00002028 ffff0880
[0010] 0000209c feff4040 000020c0 efdf0080
[0020] 00002178 00000001 0000217c 00145855
[0030] 00002310 00000000 00002314 00000000

v2: Move error collection to the ring error code Change format of dump to not confuse intel_error_decode (Chris) Put the context error object with the others (Chris) Don't search bound_list instead of active_list (chris)

v3: extract and flatten context recording (daniel) checkpatch related fixes for the copypasta in debugfs

v4: bug in v3 (Daniel)

Review

Quoted 2 lines
(quoted diff) +1 -1
-       if ((ring->id == RCS) && error->ccid)
+       if ((ring->id != RCS) || !error->ccid)

Still a redundant () pair here ... I've killed it.

Quoted 4 lines

References: https://bugs.freedesktop.org/show_bug.cgi?id=55845
Reviewed-by (v2): Chris Wilson <[email protected]>

Queued for -next, thanks for the patch.
-Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#7 lore raw

PATCH 2/3 drm/i915: exclude CCID for platforms without it

In reply to Ben Widawsky
---
 drivers/gpu/drm/i915/i915_irq.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_irq.c +2 -1
index 420911c..6a328b8 100644
@@ -1317,7 +1317,8 @@ static void i915_capture_error_state(struct drm_device *dev)
 	kref_init(&error->ref);
 	error->eir = I915_READ(EIR);
 	error->pgtbl_er = I915_READ(PGTBL_ER);
-	error->ccid = I915_READ(CCID);
+	if (HAS_HW_CONTEXTS(dev))
+		error->ccid = I915_READ(CCID);
 
 	if (HAS_PCH_SPLIT(dev))
 		error->ier = I915_READ(DEIER) | I915_READ(GTIER);
Signature · 1 line
1.8.1.4