[Intel-gfx] [PATCH] drm/i915: Display current hangcheck status in debugfs
Dave Gordon
david.s.gordon at intel.com
Fri Nov 21 16:36:23 CET 2014
On 20/11/14 19:13, Chris Wilson wrote:
> For example,
>
> /sys/kernel/debug/dri/0/i915_hangcheck_info:
>
> Hangcheck active, fires in 15887800ms
> render ring:
> seqno = -4059 [current -583]
> action = 2
> score = 0
> ACTHD = 1ee8 [current 21f980]
> max ACTHD = 0
>
> v2: Include expiration ETA. Can anyone spot a problem?
Well, the time-until-the-hangcheck-fires in your example above looks
pretty bogus -- over 4 hours until next check seems rather unlikely, not
to say useless. The seq_printf format should be %u rather than %d, but I
don't think that's really the reason.
[[ Aside: the signatures of the jiffy-converting functions are:
unsigned int jiffies_to_msecs(const unsigned long j);
unsigned int jiffies_to_usecs(const unsigned long j);
which don't seem very sensible, as
(a) they return a possibly-smaller type than their argument, despite
probably needing a larger numeric range (if HZ <= MSEC_PER_SEC, hence
interval-in-ms is a larger number than interval-in-jiffies, and even
more so for usec), and
(b) they're defined only for non-negative quantities, whereas time
intervals can meaningfully be signed. Even if jiffies are regarded as
unsigned-ticks-since-boot, the subtraction of two unsigned quantities
yields a signed difference, requiring extra checks if you don't know
whether a certain absolute time has passed or not.
(c) 'ktime_t's are so much nicer than jiffies ]]
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Mika Kuoppala <mika.kuoppala at intel.com>
> ---
> drivers/gpu/drm/i915/i915_debugfs.c | 35 +++++++++++++++++++++++++++++++++++
> 1 file changed, 35 insertions(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
> index f91e7f7c92af..8c8743dec20e 100644
> --- a/drivers/gpu/drm/i915/i915_debugfs.c
> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
> @@ -1165,6 +1165,40 @@ out:
> return ret;
> }
>
> +static int i915_hangcheck_info(struct seq_file *m, void *unused)
> +{
> + struct drm_info_node *node = m->private;
> + struct drm_i915_private *dev_priv = to_i915(node->minor->dev);
> + struct intel_engine_cs *ring;
> + int i;
> +
> + if (!i915.enable_hangcheck) {
> + seq_printf(m, "Hangcheck disabled\n");
> + return 0;
> + }
> +
> + if (timer_pending(&dev_priv->gpu_error.hangcheck_timer)) {
> + seq_printf(m, "Hangcheck active, fires in %dms\n",
> + jiffies_to_msecs(dev_priv->gpu_error.hangcheck_timer.expires - jiffies));
> + } else
> + seq_printf(m, "Hangcheck inactive\n");
> +
More information about the Intel-gfx
mailing list