[Intel-gfx] [PATCH 2/3] drm/i915/tdr: Prepare error handler to accept mask of hung engines

[Mika Kuoppala]

Arun Siluvery <arun.siluvery at linux.intel.com> writes:

> [ text/plain ]
> In preparation for engine reset, the wedged argument of i915_handle_error()
> is extended to reflect as a mask of engines that are hung. This is further
> passed down to error state capture functions which are also updated.
>
> Engine reset recovery mechanism uses this mask and schedules recovery work
> for those particular engines.
>
> Cc: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Mika Kuoppala <mika.kuoppala at intel.com>
> Signed-off-by: Tomas Elf <tomas.elf at intel.com>
> Signed-off-by: Arun Siluvery <arun.siluvery at linux.intel.com>
> ---
>  drivers/gpu/drm/i915/i915_drv.h       |  4 ++--
>  drivers/gpu/drm/i915/i915_gpu_error.c |  8 ++++----
>  drivers/gpu/drm/i915/i915_irq.c       | 16 ++++++++--------
>  3 files changed, 14 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index 549a232..49ac065 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -2735,7 +2735,7 @@ bool intel_hpd_pin_to_port(enum hpd_pin pin, enum port *port);
>  /* i915_irq.c */
>  void i915_queue_hangcheck(struct drm_device *dev);
>  __printf(3, 4)
> -void i915_handle_error(struct drm_device *dev, bool wedged,
> +void i915_handle_error(struct drm_device *dev, u32 engine_mask,
>  		       const char *fmt, ...);
>  
>  extern void intel_irq_init(struct drm_i915_private *dev_priv);
> @@ -3321,7 +3321,7 @@ static inline void i915_error_state_buf_release(
>  {
>  	kfree(eb->buf);
>  }
> -void i915_capture_error_state(struct drm_device *dev, bool wedge,
> +void i915_capture_error_state(struct drm_device *dev, u32 engine_mask,
>  			      const char *error_msg);
>  void i915_error_state_get(struct drm_device *dev,
>  			  struct i915_error_state_file_priv *error_priv);
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index db8600a..1f8ff06 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -1301,7 +1301,7 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
>  
>  static void i915_error_capture_msg(struct drm_device *dev,
>  				   struct drm_i915_error_state *error,
> -				   bool wedged,
> +				   u32 engine_mask,
>  				   const char *error_msg)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
> @@ -1324,7 +1324,7 @@ static void i915_error_capture_msg(struct drm_device *dev,
>  	scnprintf(error->error_msg + len, sizeof(error->error_msg) - len,
>  		  ", reason: %s, action: %s",
>  		  error_msg,
> -		  wedged ? "reset" : "continue");
> +		  engine_mask ? "reset" : "continue");
>  }
>  
>  static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
> @@ -1347,7 +1347,7 @@ static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
>   * out a structure which becomes available in debugfs for user level tools
>   * to pick up.
>   */
> -void i915_capture_error_state(struct drm_device *dev, bool wedged,
> +void i915_capture_error_state(struct drm_device *dev, u32 engine_mask,
>  			      const char *error_msg)
>  {
>  	static bool warned;
> @@ -1375,7 +1375,7 @@ void i915_capture_error_state(struct drm_device *dev, bool wedged,
>  	error->overlay = intel_overlay_capture_error_state(dev);
>  	error->display = intel_display_capture_error_state(dev);
>  
> -	i915_error_capture_msg(dev, error, wedged, error_msg);
> +	i915_error_capture_msg(dev, error, engine_mask, error_msg);
>  	DRM_INFO("%s\n", error->error_msg);
>  
>  	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index 8f3e330..a55a7cc 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -2653,14 +2653,14 @@ static void i915_report_and_clear_eir(struct drm_device *dev)
>  /**
>   * i915_handle_error - handle a gpu error
>   * @dev: drm device
> - *
> + * @engine_mask: mask representing engines that are hung
>   * Do some basic checking of register state at error time and
>   * dump it to the syslog.  Also call i915_capture_error_state() to make
>   * sure we get a record and make it available in debugfs.  Fire a uevent
>   * so userspace knows something bad happened (should trigger collection
>   * of a ring dump etc.).
>   */
> -void i915_handle_error(struct drm_device *dev, bool wedged,
> +void i915_handle_error(struct drm_device *dev, u32 engine_mask,
>  		       const char *fmt, ...)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
> @@ -2671,10 +2671,10 @@ void i915_handle_error(struct drm_device *dev, bool wedged,
>  	vscnprintf(error_msg, sizeof(error_msg), fmt, args);
>  	va_end(args);
>  
> -	i915_capture_error_state(dev, wedged, error_msg);
> +	i915_capture_error_state(dev, engine_mask, error_msg);
>  	i915_report_and_clear_eir(dev);
>  
> -	if (wedged) {
> +	if (engine_mask) {
>  		atomic_or(I915_RESET_IN_PROGRESS_FLAG,
>  				&dev_priv->gpu_error.reset_counter);
>  
> @@ -3033,7 +3033,7 @@ ring_stuck(struct intel_engine_cs *engine, u64 acthd)
>  	 */
>  	tmp = I915_READ_CTL(engine);
>  	if (tmp & RING_WAIT) {
> -		i915_handle_error(dev, false,
> +		i915_handle_error(dev, 0,
>  				  "Kicking stuck wait on %s",
>  				  engine->name);
>  		I915_WRITE_CTL(engine, tmp);
> @@ -3045,7 +3045,7 @@ ring_stuck(struct intel_engine_cs *engine, u64 acthd)
>  		default:
>  			return HANGCHECK_HUNG;
>  		case 1:
> -			i915_handle_error(dev, false,
> +			i915_handle_error(dev, 0,
>  					  "Kicking stuck semaphore on %s",
>  					  engine->name);
>  			I915_WRITE_CTL(engine, tmp);
> @@ -3189,12 +3189,12 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
>  			DRM_INFO("%s on %s\n",
>  				 stuck[i] ? "stuck" : "no progress",
>  				 engine->name);
> -			rings_hung++;
> +			rings_hung |= intel_engine_flag(engine);

We can change the int to u32 when we rename rings_hung to engines_hung.

Reviewed-by: Mika Kuoppala <mika.kuoppala at intel.com>

>  		}
>  	}
>  
>  	if (rings_hung) {
> -		i915_handle_error(dev, true, "Ring hung");
> +		i915_handle_error(dev, rings_hung, "Engine(s) hung");
>  		goto out;
>  	}
>  
> -- 
> 1.9.1