[Intel-gfx] [PATCH 07/20] drm/i915: Watchdog timeout: Hang detection integration into error handler

[Chris Wilson]

On Wed, Jan 13, 2016 at 05:28:19PM +0000, Arun Siluvery wrote:
>  /* i915_irq.c */
>  void i915_queue_hangcheck(struct drm_device *dev);
> -__printf(4, 5)
> -void i915_handle_error(struct drm_device *dev, u32 engine_mask, bool wedged,
> -		       const char *fmt, ...);
> +__printf(5, 6)
> +void i915_handle_error(struct drm_device *dev, u32 engine_mask,
> +		       bool watchdog, bool wedged, const char *fmt, ...);
>  
>  extern void intel_irq_init(struct drm_i915_private *dev_priv);
>  int intel_irq_install(struct drm_i915_private *dev_priv);
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index 8937c82..0710724 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -2726,6 +2726,7 @@ static void i915_report_and_clear_eir(struct drm_device *dev)
>   *			If a previous engine reset was attempted too recently
>   *			or if one of the current engine resets fails we fall
>   *			back to legacy full GPU reset.
> + * @watchdog: 		true = Engine hang detected by hardware watchdog.
>   * @wedged: 		true = Hang detected, invoke hang recovery.

A bitmask and 2 booleans? Whilst this isn't going to be the most widely
used of functions, those parameters are just inviting trouble.

>   * @fmt, ...: 		Error message describing reason for error.
>   *
> @@ -2737,8 +2738,8 @@ static void i915_report_and_clear_eir(struct drm_device *dev)
>   * reset the associated engine. Failing that, try to fall back to legacy
>   * full GPU reset recovery mode.
>   */
> -void i915_handle_error(struct drm_device *dev, u32 engine_mask, bool wedged,
> -		       const char *fmt, ...)
> +void i915_handle_error(struct drm_device *dev, u32 engine_mask,
> +                       bool watchdog, bool wedged, const char *fmt, ...)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
>  	va_list args;
> @@ -2776,20 +2777,27 @@ void i915_handle_error(struct drm_device *dev, u32 engine_mask, bool wedged,
>  			u32 i;
>  
>  			for_each_ring(engine, dev_priv, i) {
> -				u32 now, last_engine_reset_timediff;

Oops skipped a patch, I'll be back.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre