Re: [Intel-gfx] [PATCH v5 3/4] drm/i915: Make the heartbeat play nice with long pre-emption timeouts

[Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>]

On 06/10/2022 22:38, John.C.Harrison@Intel.com wrote:
> From: John Harrison <John.C.Harrison@Intel.com>
> 
> Compute workloads are inherently not pre-emptible for long periods on
> current hardware. As a workaround for this, the pre-emption timeout
> for compute capable engines was disabled. This is undesirable with GuC
> submission as it prevents per engine reset of hung contexts. Hence the
> next patch will re-enable the timeout but bumped up by an order of
> magnitude.
> 
> However, the heartbeat might not respect that. Depending upon current
> activity, a pre-emption to the heartbeat pulse might not even be
> attempted until the last heartbeat period. Which means that only one
> period is granted for the pre-emption to occur. With the aforesaid
> bump, the pre-emption timeout could be significantly larger than this
> heartbeat period.
> 
> So adjust the heartbeat code to take the pre-emption timeout into
> account. When it reaches the final (high priority) period, it now
> ensures the delay before hitting reset is bigger than the pre-emption
> timeout.
> 
> v2: Fix for selftests which adjust the heartbeat period manually.
> v3: Add FIXME comment about selftests. Add extra FIXME comment and
> drm_notices when setting heartbeat to a non-default value (review
> feedback from Tvrtko)
> 
> Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
> ---
>   .../gpu/drm/i915/gt/intel_engine_heartbeat.c  | 39 +++++++++++++++++++
>   1 file changed, 39 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
> index a3698f611f457..9a527e1f5be65 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
> +++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
> @@ -22,9 +22,37 @@
>   
>   static bool next_heartbeat(struct intel_engine_cs *engine)
>   {
> +	struct i915_request *rq;
>   	long delay;
>   
>   	delay = READ_ONCE(engine->props.heartbeat_interval_ms);
> +
> +	rq = engine->heartbeat.systole;
> +
> +	/*
> +	 * FIXME: The final period extension is disabled if the period has been
> +	 * modified from the default. This is to prevent issues with certain
> +	 * selftests which override the value and expect specific behaviour.
> +	 * Once the selftests have been updated to either cope with variable
> +	 * heartbeat periods (or to override the pre-emption timeout as well,
> +	 * or just to add a selftest specific override of the extension), the
> +	 * generic override can be removed.
> +	 */
> +	if (rq && rq->sched.attr.priority >= I915_PRIORITY_BARRIER &&
> +	    delay == engine->defaults.heartbeat_interval_ms) {
> +		long longer;
> +
> +		/*
> +		 * The final try is at the highest priority possible. Up until now
> +		 * a pre-emption might not even have been attempted. So make sure
> +		 * this last attempt allows enough time for a pre-emption to occur.
> +		 */
> +		longer = READ_ONCE(engine->props.preempt_timeout_ms) * 2;
> +		longer = intel_clamp_heartbeat_interval_ms(engine, longer);
> +		if (longer > delay)
> +			delay = longer;
> +	}
> +
>   	if (!delay)
>   		return false;
>   
> @@ -288,6 +316,17 @@ int intel_engine_set_heartbeat(struct intel_engine_cs *engine,
>   	if (!delay && !intel_engine_has_preempt_reset(engine))
>   		return -ENODEV;
>   
> +	/* FIXME: Remove together with equally marked hack in next_heartbeat. */
> +	if (delay != engine->defaults.heartbeat_interval_ms &&
> +	    delay < 2 * engine->props.preempt_timeout_ms) {
> +		if (intel_engine_uses_guc(engine))
> +			drm_notice(&engine->i915->drm, "%s heartbeat interval adjusted to a non-default value which may downgrade individual engine resets to full GPU resets!\n",
> +				   engine->name);
> +		else
> +			drm_notice(&engine->i915->drm, "%s heartbeat interval adjusted to a non-default value which may cause engine resets to target innocent contexts!\n",
> +				   engine->name);
> +	}
> +
>   	intel_engine_pm_get(engine);
>   
>   	err = mutex_lock_interruptible(&ce->timeline->mutex);

LGTM - hope it is agreeable to you too.

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

Regards,

Tvrtko