[Intel-gfx] [PATCH 2/2] drm/i915: Make hangcheck logging more compact

[Tomas Elf]

On 08/05/2015 14:39, Mika Kuoppala wrote:
> With commit aaecdf611a05 ("drm/i915: Stop gathering error
> states for CS error interrupts") we only call i915_handle_error()
> on call sites where there is a stuck/hung gpu. So there is
> no more need to carry around extra information into dmesg.
>
> Emit one loud bang into dmesg with first hanging ring as
> culprit. Rest of the details will be in error state.
>
> Based-on-patch-by: Chris Wilson <chris at chris-wilson.co.uk>
> Signed-off-by: Mika Kuoppala <mika.kuoppala at intel.com>
> ---
>   drivers/gpu/drm/i915/i915_gpu_error.c |  4 +---
>   drivers/gpu/drm/i915/i915_irq.c       | 26 ++++++++------------------
>   2 files changed, 9 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 9c0db19..292cf1f 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -1251,9 +1251,7 @@ static void i915_error_capture_msg(struct drm_device *dev,
>   				 error->ring[ring_id].pid);
>
>   	scnprintf(error->error_msg + len, sizeof(error->error_msg) - len,
> -		  ", reason: %s, action: %s",
> -		  error_msg,
> -		  wedged ? "reset" : "continue");
> +		  ", %s", error_msg);
>   }
>

Once you've removed the reference to the wedged parameter from the 
scnprintf statement I can't see any other references to it anywhere else 
in the function. How about we remove that parameter entirely from the 
function signature?

Thanks,
Tomas

>   static void i915_capture_gen_state(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 a3244bd..a3b5001 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -2924,14 +2924,12 @@ static bool check_for_missed_irq(struct intel_engine_cs *ring)
>   	return true;
>   }
>
> -static bool hangcheck_handle_stuck_ring(struct intel_engine_cs *ring, u64 acthd)
> +static void hangcheck_handle_stuck_ring(struct intel_engine_cs *ring, u64 acthd)
>   {
>   #define BUSY 1
>   #define KICK 5
>   #define HUNG 20
> -
>   	struct intel_ring_hangcheck *hc = &ring->hangcheck;
> -	bool there_is_hope = true;
>
>   	/* We always increment the hangcheck score
>   	 * if the ring is busy and still processing
> @@ -2964,11 +2962,8 @@ static bool hangcheck_handle_stuck_ring(struct intel_engine_cs *ring, u64 acthd)
>   		break;
>   	case HANGCHECK_HUNG:
>   		hc->score += HUNG;
> -		there_is_hope = false;
>   		break;
>   	}
> -
> -	return there_is_hope;
>   }
>
>   /*
> @@ -2987,8 +2982,7 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
>   	struct drm_device *dev = dev_priv->dev;
>   	struct intel_engine_cs *ring;
>   	int i;
> -	int busy_count = 0, rings_hung = 0;
> -	bool stuck[I915_NUM_RINGS] = { 0 };
> +	int busy_count = 0, ring_hung = -1;
>
>   	if (!i915.enable_hangcheck)
>   		return;
> @@ -3043,19 +3037,15 @@ engine_check_done:
>   		hc->acthd = acthd;
>   		hc->start = start;
>   		busy_count += busy;
> -	}
>
> -	for_each_ring(ring, dev_priv, i) {
> -		if (ring->hangcheck.score >= HANGCHECK_SCORE_RING_HUNG) {
> -			DRM_INFO("%s on %s\n",
> -				 stuck[i] ? "stuck" : "no progress",
> -				 ring->name);
> -			rings_hung++;
> -		}
> +		if (ring_hung == -1 &&
> +		    ring->hangcheck.score >= HANGCHECK_SCORE_RING_HUNG)
> +			ring_hung = i;
>   	}
>
> -	if (rings_hung)
> -		return i915_handle_error(dev, true, "Ring hung");
> +	if (ring_hung != -1)
> +		return i915_handle_error(dev, true, "%s hung",
> +					 dev_priv->ring[ring_hung].name);
>
>   	if (busy_count)
>   		/* Reset timer case chip hangs without another request
>