public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 08/25] drm/i915/pmu: Always sample an active ringbuffer
Date: Fri, 22 Feb 2019 14:10:33 +0200	[thread overview]
Message-ID: <878sy8vxae.fsf@gaia.fi.intel.com> (raw)
In-Reply-To: <20190219122215.8941-8-chris@chris-wilson.co.uk>

Chris Wilson <chris@chris-wilson.co.uk> writes:

> As we no longer have a precise indication of requests queued to an
> engine, make no presumptions and just sample the ring registers to see
> if the engine is busy.
>
> v2: Report busy while the ring is idling on a semaphore/event.
> v3: Give the struct a name!
> v4: Always 0 outside the powerwell; trusting the powerwell is
> accurate enough for our sampling pmu.
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> ---
>  drivers/gpu/drm/i915/i915_pmu.c         | 60 ++++++++++---------------
>  drivers/gpu/drm/i915/intel_ringbuffer.h |  2 +-
>  2 files changed, 24 insertions(+), 38 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c
> index 13d70b90dd0f..21adad72bd86 100644
> --- a/drivers/gpu/drm/i915/i915_pmu.c
> +++ b/drivers/gpu/drm/i915/i915_pmu.c
> @@ -148,14 +148,6 @@ void i915_pmu_gt_unparked(struct drm_i915_private *i915)
>  	spin_unlock_irq(&i915->pmu.lock);
>  }
>  
> -static bool grab_forcewake(struct drm_i915_private *i915, bool fw)
> -{
> -	if (!fw)
> -		intel_uncore_forcewake_get(i915, FORCEWAKE_ALL);
> -
> -	return true;
> -}
> -
>  static void
>  add_sample(struct i915_pmu_sample *sample, u32 val)
>  {
> @@ -168,49 +160,43 @@ engines_sample(struct drm_i915_private *dev_priv, unsigned int period_ns)
>  	struct intel_engine_cs *engine;
>  	enum intel_engine_id id;
>  	intel_wakeref_t wakeref;
> -	bool fw = false;
>  
>  	if ((dev_priv->pmu.enable & ENGINE_SAMPLE_MASK) == 0)
>  		return;
>  
> -	if (!dev_priv->gt.awake)
> -		return;
> -
> -	wakeref = intel_runtime_pm_get_if_in_use(dev_priv);
> +	wakeref = 0;
> +	if (READ_ONCE(dev_priv->gt.awake))

Is this gt.awake check just to be more lightweight on sampling?

> +		wakeref = intel_runtime_pm_get_if_in_use(dev_priv);
>  	if (!wakeref)
>  		return;
>  
>  	for_each_engine(engine, dev_priv, id) {
> -		u32 current_seqno = intel_engine_get_seqno(engine);
> -		u32 last_seqno = intel_engine_last_submit(engine);
> +		struct intel_engine_pmu *pmu = &engine->pmu;
> +		bool busy;
>  		u32 val;
>  
> -		val = !i915_seqno_passed(current_seqno, last_seqno);
> -
> -		if (val)
> -			add_sample(&engine->pmu.sample[I915_SAMPLE_BUSY],
> -				   period_ns);
> -
> -		if (val && (engine->pmu.enable &
> -		    (BIT(I915_SAMPLE_WAIT) | BIT(I915_SAMPLE_SEMA)))) {
> -			fw = grab_forcewake(dev_priv, fw);
> -
> -			val = I915_READ_FW(RING_CTL(engine->mmio_base));
> -		} else {
> -			val = 0;
> -		}
> +		val = I915_READ_FW(RING_CTL(engine->mmio_base));
> +		if (val == 0) /* powerwell off => engine idle */
> +			continue;
>  
>  		if (val & RING_WAIT)
> -			add_sample(&engine->pmu.sample[I915_SAMPLE_WAIT],
> -				   period_ns);
> -
> +			add_sample(&pmu->sample[I915_SAMPLE_WAIT], period_ns);
>  		if (val & RING_WAIT_SEMAPHORE)
> -			add_sample(&engine->pmu.sample[I915_SAMPLE_SEMA],
> -				   period_ns);
> -	}
> +			add_sample(&pmu->sample[I915_SAMPLE_SEMA], period_ns);
>  
> -	if (fw)
> -		intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
> +		/*
> +		 * MI_MODE reports IDLE if the ring is waiting, but we regard
> +		 * this as being busy instead, as the engine is busy with the
> +		 * user request.
> +		 */
> +		busy = val & (RING_WAIT_SEMAPHORE | RING_WAIT);
> +		if (!busy) {
> +			val = I915_READ_FW(RING_MI_MODE(engine->mmio_base));
> +			busy = !(val & MODE_IDLE);

The comment makes sense if you do
busy = val & MODE_IDLE;

-Mika

> +		}
> +		if (busy)
> +			add_sample(&pmu->sample[I915_SAMPLE_BUSY], period_ns);
> +	}
>  
>  	intel_runtime_pm_put(dev_priv, wakeref);
>  }
> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
> index 710ffb221775..5d45ad4ecca9 100644
> --- a/drivers/gpu/drm/i915/intel_ringbuffer.h
> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
> @@ -392,7 +392,7 @@ struct intel_engine_cs {
>  		bool irq_armed;
>  	} breadcrumbs;
>  
> -	struct {
> +	struct intel_engine_pmu {
>  		/**
>  		 * @enable: Bitmask of enable sample events on this engine.
>  		 *
> -- 
> 2.20.1
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2019-02-22 12:10 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-19 12:21 [PATCH 01/25] drm/i915: Move verify_wm_state() to heap Chris Wilson
2019-02-19 12:21 ` [PATCH 02/25] drm/i915: Use time based guilty context banning Chris Wilson
2019-02-19 12:21 ` [PATCH 03/25] drm/i915: Prevent user context creation while wedged Chris Wilson
2019-02-19 13:07   ` Mika Kuoppala
2019-02-19 13:11     ` Chris Wilson
2019-02-19 13:31       ` Mika Kuoppala
2019-02-19 13:20     ` Chris Wilson
2019-02-19 12:21 ` [PATCH 04/25] drm/i915: Avoid reset lock in writing fence registers Chris Wilson
2019-02-20 14:55   ` Mika Kuoppala
2019-02-19 12:21 ` [PATCH 05/25] drm/i915: Reorder struct_mutex-vs-reset_lock in i915_gem_fault() Chris Wilson
2019-02-19 12:21 ` [PATCH 06/25] drm/i915: Trim i915_do_reset() to minimum delays Chris Wilson
2019-02-19 12:21 ` [PATCH 07/25] drm/i915: Trim delays for wedging Chris Wilson
2019-02-19 12:48   ` Mika Kuoppala
2019-02-19 12:21 ` [PATCH 08/25] drm/i915/pmu: Always sample an active ringbuffer Chris Wilson
2019-02-22 12:10   ` Mika Kuoppala [this message]
2019-02-22 12:17     ` Chris Wilson
2019-02-22 12:31       ` Mika Kuoppala
2019-02-19 12:21 ` [PATCH 09/25] drm/i915: Replace global_seqno with a hangcheck heartbeat seqno Chris Wilson
2019-02-19 12:22 ` [PATCH 10/25] drm/i915: Remove access to global seqno in the HWSP Chris Wilson
2019-02-19 12:22 ` [PATCH 11/25] drm/i915: Remove i915_request.global_seqno Chris Wilson
2019-02-19 12:22 ` [PATCH 12/25] drm/i915/selftests: Exercise resetting during non-user payloads Chris Wilson
2019-02-19 12:22 ` [PATCH 13/25] drm/i915: Reduce the RPS shock Chris Wilson
2019-02-19 21:00   ` Lyude Paul
2019-02-20 12:05     ` Chris Wilson
2019-02-20 15:14   ` Mika Kuoppala
2019-02-19 12:22 ` [PATCH 14/25] drm/i915/pmu: Use GT parked for estimating RC6 while asleep Chris Wilson
2019-02-19 12:22 ` [PATCH 15/25] drm/i915: Skip scanning for signalers if we are already inflight Chris Wilson
2019-02-19 12:22 ` [PATCH 16/25] drm/i915/execlists: Suppress mere WAIT preemption Chris Wilson
2019-02-19 12:22 ` [PATCH 17/25] drm/i915/execlists: Suppress redundant preemption Chris Wilson
2019-02-19 12:22 ` [PATCH 18/25] drm/i915: Make request allocation caches global Chris Wilson
2019-02-19 12:22 ` [PATCH 19/25] drm/i915: Introduce i915_timeline.mutex Chris Wilson
2019-02-19 12:22 ` [PATCH 20/25] drm/i915: Keep timeline HWSP allocated until idle across the system Chris Wilson
2019-02-19 12:22 ` [PATCH 21/25] drm/i915: Compute the global scheduler caps Chris Wilson
2019-02-19 12:22 ` [PATCH 22/25] drm/i915: Use HW semaphores for inter-engine synchronisation on gen8+ Chris Wilson
2019-02-19 12:22 ` [PATCH 23/25] drm/i915: Prioritise non-busywait semaphore workloads Chris Wilson
2019-02-19 12:22 ` [PATCH 24/25] drm/i915/execlists: Skip direct submission if only lite-restore Chris Wilson
2019-02-19 12:22 ` [PATCH 25/25] drm/i915: Use __ffs() in for_each_priolist for more compact code Chris Wilson
2019-02-19 12:56 ` [PATCH 01/25] drm/i915: Move verify_wm_state() to heap Ville Syrjälä
2019-02-19 13:05 ` ✗ Fi.CI.CHECKPATCH: warning for series starting with [01/25] " Patchwork
2019-02-19 13:15 ` ✗ Fi.CI.SPARSE: " Patchwork
2019-02-19 13:31 ` ✓ Fi.CI.BAT: success " Patchwork
2019-02-19 17:14 ` ✗ Fi.CI.IGT: failure " Patchwork
2019-02-19 17:16   ` Chris Wilson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=878sy8vxae.fsf@gaia.fi.intel.com \
    --to=mika.kuoppala@linux.intel.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox