Re: [PATCH 2/3] drm/i915: Record pid/comm of hanging task

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Feb 10, 2014 at 04:30:50PM +0200, Mika Kuoppala wrote:
> From: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> 
> After finding the guilty batch and request, we can use it to find the
> process that submitted the batch and then add the culprit into the error
> state.
> 
> This is a slightly different approach from Ben's in that instead of
> adding the extra information into the struct i915_hw_context, we use the
> information already captured in struct drm_file which is then referenced
> from the request.
> 
> v2: Also capture the workaround buffer for gen2, so that we can compare
>     its contents against the intended batch for the active request.
> 
> v3: Rebase (Mika)
> 
> Link: http://lists.freedesktop.org/archives/intel-gfx/2013-August/032280.html
> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> (v2)
> Signed-off-by: Mika Kuoppala <mika.kuoppala@xxxxxxxxx> (v3)
> Cc: Ben Widawsky <ben@xxxxxxxxxxxx>
> ---
>  drivers/gpu/drm/i915/i915_drv.h       |    6 +-
>  drivers/gpu/drm/i915/i915_gem.c       |    1 +
>  drivers/gpu/drm/i915/i915_gpu_error.c |  121 ++++++++++++++++++---------------
>  3 files changed, 74 insertions(+), 54 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index eb260bc..2a61a29 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -356,7 +356,7 @@ struct drm_i915_error_state {
>  			int page_count;
>  			u32 gtt_offset;
>  			u32 *pages[0];
> -		} *ringbuffer, *batchbuffer, *ctx, *hws_page;
> +		} *ringbuffer, *batchbuffer, *wa_batchbuffer, *ctx, *hws_page;
>  
>  		struct drm_i915_error_request {
>  			long jiffies;
> @@ -371,6 +371,9 @@ struct drm_i915_error_state {
>  				u32 pp_dir_base;
>  			};
>  		} vm_info;
> +
> +		pid_t pid;
> +		char comm[TASK_COMM_LEN];
>  	} ring[I915_NUM_RINGS];
>  	struct drm_i915_error_buffer {
>  		u32 size;
> @@ -1786,6 +1789,7 @@ struct drm_i915_gem_request {
>  
>  struct drm_i915_file_private {
>  	struct drm_i915_private *dev_priv;
> +	struct drm_file *file;
>  
>  	struct {
>  		spinlock_t lock;
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 20e55ef..efbd1dc 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -4883,6 +4883,7 @@ int i915_gem_open(struct drm_device *dev, struct drm_file *file)
>  
>  	file->driver_priv = file_priv;
>  	file_priv->dev_priv = dev->dev_private;
> +	file_priv->file = file;
>  
>  	spin_lock_init(&file_priv->mm.lock);
>  	INIT_LIST_HEAD(&file_priv->mm.request_list);
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 5bd075a..a90971a 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -301,13 +301,28 @@ void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
>  	va_end(args);
>  }
>  
> +static void print_error_obj(struct drm_i915_error_state_buf *m,
> +			    struct drm_i915_error_object *obj)
> +{
> +	int page, offset, elt;
> +
> +	for (page = offset = 0; page < obj->page_count; page++) {
> +		for (elt = 0; elt < PAGE_SIZE/4; elt++) {
> +			err_printf(m, "%08x :  %08x\n", offset,
> +				   obj->pages[page][elt]);
> +			offset += 4;
> +		}
> +	}
> +}
> +
>  int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
>  			    const struct i915_error_state_file_priv *error_priv)
>  {
>  	struct drm_device *dev = error_priv->dev;
>  	drm_i915_private_t *dev_priv = dev->dev_private;
>  	struct drm_i915_error_state *error = error_priv->error;
> -	int i, j, page, offset, elt;
> +	int i, j, offset, elt;
> +	int max_hangcheck_score;
>  
>  	if (!error) {
>  		err_printf(m, "no error state collected\n");
> @@ -317,6 +332,20 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
>  	err_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
>  		   error->time.tv_usec);
>  	err_printf(m, "Kernel: " UTS_RELEASE "\n");
> +	max_hangcheck_score = 0;
> +	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
> +		if (error->ring[i].hangcheck_score > max_hangcheck_score)
> +			max_hangcheck_score = error->ring[i].hangcheck_score;
> +	}
> +	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
> +		if (error->ring[i].hangcheck_score == max_hangcheck_score &&
> +		    error->ring[i].pid != -1) {
> +			err_printf(m, "Active process (on ring %s): %s [%d]\n",
> +				   ring_str(i),
> +				   error->ring[i].comm,
> +				   error->ring[i].pid);
> +		}
> +	}
>  	err_printf(m, "PCI ID: 0x%04x\n", dev->pdev->device);
>  	err_printf(m, "EIR: 0x%08x\n", error->eir);
>  	err_printf(m, "IER: 0x%08x\n", error->ier);
> @@ -360,17 +389,19 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
>  		struct drm_i915_error_object *obj;
>  
>  		if ((obj = error->ring[i].batchbuffer)) {
> -			err_printf(m, "%s --- gtt_offset = 0x%08x\n",
> -				   dev_priv->ring[i].name,
> +			err_puts(m, dev_priv->ring[i].name);
> +			if (error->ring[i].pid != -1)
> +				err_printf(m, " (submitted by %s [%d])",
> +					   error->ring[i].comm, error->ring[i].pid);
> +			err_printf(m, " --- gtt_offset = 0x%08x\n",
>  				   obj->gtt_offset);
> -			offset = 0;
> -			for (page = 0; page < obj->page_count; page++) {
> -				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
> -					err_printf(m, "%08x :  %08x\n", offset,
> -						   obj->pages[page][elt]);
> -					offset += 4;
> -				}
> -			}
> +			print_error_obj(m, obj);
> +		}
> +
> +		if ((obj = error->ring[i].wa_batchbuffer)) {
> +			err_printf(m, "%s (w/a) --- gtt_offset = 0x%08x\n",
> +				   dev_priv->ring[i].name, obj->gtt_offset);
> +			print_error_obj(m, obj);
>  		}
>  
>  		if (error->ring[i].num_requests) {
> @@ -389,15 +420,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
>  			err_printf(m, "%s --- ringbuffer = 0x%08x\n",
>  				   dev_priv->ring[i].name,
>  				   obj->gtt_offset);
> -			offset = 0;
> -			for (page = 0; page < obj->page_count; page++) {
> -				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
> -					err_printf(m, "%08x :  %08x\n",
> -						   offset,
> -						   obj->pages[page][elt]);
> -					offset += 4;
> -				}
> -			}
> +			print_error_obj(m, obj);
>  		}
>  
>  		if ((obj = error->ring[i].hws_page)) {
> @@ -713,37 +736,6 @@ static void i915_gem_record_fences(struct drm_device *dev,
>  	}
>  }
>  
> -static struct drm_i915_error_object *
> -i915_error_first_batchbuffer(struct drm_i915_private *dev_priv,
> -			     struct intel_ring_buffer *ring)
> -{
> -	struct drm_i915_gem_request *request;
> -
> -	if (HAS_BROKEN_CS_TLB(dev_priv->dev)) {
> -		struct drm_i915_gem_object *obj;
> -		u32 acthd = I915_READ(ACTHD);
> -
> -		if (WARN_ON(ring->id != RCS))
> -			return NULL;
> -
> -		obj = ring->scratch.obj;
> -		if (obj != NULL &&
> -		    acthd >= i915_gem_obj_ggtt_offset(obj) &&
> -		    acthd < i915_gem_obj_ggtt_offset(obj) + obj->base.size)
> -			return i915_error_ggtt_object_create(dev_priv, obj);
> -	}
> -
> -	request = i915_gem_find_active_request(ring);
> -	if (request) {
> -		/* We need to copy these to an anonymous buffer as the simplest
> -		 * method to avoid being overwritten by userspace.
> -		 */
> -		return i915_error_object_create(dev_priv, request->batch_obj, request->ctx->vm);
> -	}
> -
> -	return NULL;
> -}
> -
>  static void i915_record_ring_state(struct drm_device *dev,
>  				   struct intel_ring_buffer *ring,
>  				   struct drm_i915_error_ring *ering)
> @@ -892,8 +884,31 @@ static void i915_gem_record_rings(struct drm_device *dev,
>  
>  		i915_record_ring_state(dev, ring, &error->ring[i]);
>  
> -		error->ring[i].batchbuffer =
> -			i915_error_first_batchbuffer(dev_priv, ring);
> +		error->ring[i].pid = -1;
> +		request = i915_gem_find_active_request(ring);
> +		if (request) {
> +			/* We need to copy these to an anonymous buffer as the simplest
> +			 * method to avoid being overwritten by userspace.
> +			 */
> +			error->ring[i].batchbuffer =
> +				i915_error_object_create(dev_priv, request->batch_obj, request->ctx->vm);
> +
> +			if (HAS_BROKEN_CS_TLB(dev_priv->dev) && ring->scratch.obj)
> +				error->ring[i].wa_batchbuffer =
> +					i915_error_ggtt_object_create(dev_priv, ring->scratch.obj);
> +
> +			if (request->file_priv) {
> +				struct task_struct *task;
> +
> +				rcu_read_lock();
> +				task = pid_task(request->file_priv->file->pid, PIDTYPE_PID);
> +				if (task) {
> +					strcpy(error->ring[i].comm, task->comm);
> +					error->ring[i].pid = task->pid;
> +				}
> +				rcu_read_unlock();
> +			}

I still like my solution which does the strcpy on context creation. Your
solution is deferred in the usual case since the hangcheck is
asynchronous. The process could be long dead and gone.

Also, I think my solution is simpler, and with full PPGTT should get the
same information.

> +		}
>  
>  		error->ring[i].ringbuffer =
>  			i915_error_ggtt_object_create(dev_priv, ring->obj);

-- 
Ben Widawsky, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx




[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux