[Intel-gfx] [PATCH 1/2] i915/pmu: Add support for total context runtime for GuC back-end

[Dixit, Ashutosh]

On Thu, 27 Apr 2023 15:47:04 -0700, Umesh Nerlige Ramappa wrote:
>
> GPU accumulates the context runtime in a 32 bit counter - CTX_TIMESTAMP
> in the context image. This value is saved/restored on context switches.
> KMD accumulates these values into a 64 bit counter taking care of any
> overflows as needed. This count provides the basis for client specific
> busyness in the fdinfo interface.
>
> KMD accumulation happens just before the context is unpinned and when
> context switches out. This works for execlist back-end since execlist
> scheduling has visibility into context switches. With GuC mode, KMD does
> not have visibility into context switches and this counter is
> accumulated only when context is unpinned. Context is unpinned once the
> context scheduling is successfully disabled. Disabling context
> scheduling is an asynchronous operation. Also if a context is servicing
> frequent requests, scheduling may never be disabled on it.
>
> For GuC mode, since updates to the context runtime may be delayed, add
> hooks to update the context runtime in a worker thread as well as when
> a user queries for it.
>
> Limitation:
> - If a context is never switched out or runs for a long period of time,
>   the runtime value of CTX_TIMESTAMP may never be updated, so the
>   counter value may be unreliable. This patch does not support such
>   cases. Such support must be available from the GuC FW and it is WIP.
>
> This patch is an extract from previous work authored by John/Umesh here -
> https://patchwork.freedesktop.org/patch/496441/?series=105085&rev=4
>
> v2: (Ashutosh)
> - Drop COPS_RUNTIME_ACTIVE_TOTAL
> - s/guc_context_update_clks/__guc_context_update_stats
> - Pin context before accessing in guc_timestamp_ping
> - In guc_context_unpin, use spinlock to serialize access to runtime stats

LGTM now:

Reviewed-by: Ashutosh Dixit <ashutosh.dixit at intel.com>

I have made a few notes for myself below.

>
> Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa at intel.com>
> Co-developed-by: John Harrison <John.C.Harrison at Intel.com>
> Signed-off-by: John Harrison <John.C.Harrison at Intel.com>
> ---
>  drivers/gpu/drm/i915/gt/intel_context.c       |  5 ++-
>  drivers/gpu/drm/i915/gt/intel_context.h       |  2 +-
>  drivers/gpu/drm/i915/gt/intel_context_types.h |  2 ++
>  .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 31 +++++++++++++++++++
>  4 files changed, 38 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
> index 2aa63ec521b8..a53b26178f0a 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context.c
> +++ b/drivers/gpu/drm/i915/gt/intel_context.c
> @@ -578,10 +578,13 @@ void intel_context_bind_parent_child(struct intel_context *parent,
>	child->parallel.parent = parent;
>  }
>
> -u64 intel_context_get_total_runtime_ns(const struct intel_context *ce)
> +u64 intel_context_get_total_runtime_ns(struct intel_context *ce)
>  {
>	u64 total, active;
>
> +	if (ce->ops->update_stats)
> +		ce->ops->update_stats(ce);

This is indeed called from atomic context due to rcu_read_lock in
show_client_class, so we can't sleep in update_stats (and we are not due to
use of intel_context_pin_if_active).

Because we are already updating stats from guc_timestamp_ping, we could
have skipped updating them again from here without losing too much accuracy
but it's fine.

> +
>	total = ce->stats.runtime.total;
>	if (ce->ops->flags & COPS_RUNTIME_CYCLES)
>		total *= ce->engine->gt->clock_period_ns;
> diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
> index 48f888c3da08..43ed92f8dc83 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context.h
> +++ b/drivers/gpu/drm/i915/gt/intel_context.h
> @@ -375,7 +375,7 @@ intel_context_clear_nopreempt(struct intel_context *ce)
>	clear_bit(CONTEXT_NOPREEMPT, &ce->flags);
>  }
>
> -u64 intel_context_get_total_runtime_ns(const struct intel_context *ce);
> +u64 intel_context_get_total_runtime_ns(struct intel_context *ce);
>  u64 intel_context_get_avg_runtime_ns(struct intel_context *ce);
>
>  static inline u64 intel_context_clock(void)
> diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
> index e36670f2e626..aceaac28a33e 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context_types.h
> +++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
> @@ -58,6 +58,8 @@ struct intel_context_ops {
>
>	void (*sched_disable)(struct intel_context *ce);
>
> +	void (*update_stats)(struct intel_context *ce);
> +
>	void (*reset)(struct intel_context *ce);
>	void (*destroy)(struct kref *kref);
>
> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> index ee3e8352637f..c869ddc73e69 100644
> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> @@ -1402,13 +1402,34 @@ static void __update_guc_busyness_stats(struct intel_guc *guc)
>	spin_unlock_irqrestore(&guc->timestamp.lock, flags);
>  }
>
> +static void __guc_context_update_stats(struct intel_context *ce)
> +{
> +	struct intel_guc *guc = ce_to_guc(ce);
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&guc->timestamp.lock, flags);
> +	lrc_update_runtime(ce);
> +	spin_unlock_irqrestore(&guc->timestamp.lock, flags);
> +}
> +
> +static void guc_context_update_stats(struct intel_context *ce)
> +{
> +	if (!intel_context_pin_if_active(ce))

Correct, all we need here is intel_context_pin_if_active. If the context is
not pinned we updated stats when we unpinned.

> +		return;
> +
> +	__guc_context_update_stats(ce);
> +	intel_context_unpin(ce);
> +}
> +
>  static void guc_timestamp_ping(struct work_struct *wrk)
>  {
>	struct intel_guc *guc = container_of(wrk, typeof(*guc),
>					     timestamp.work.work);
>	struct intel_uc *uc = container_of(guc, typeof(*uc), guc);
>	struct intel_gt *gt = guc_to_gt(guc);
> +	struct intel_context *ce;
>	intel_wakeref_t wakeref;
> +	unsigned long index;
>	int srcu, ret;
>
>	/*
> @@ -1424,6 +1445,10 @@ static void guc_timestamp_ping(struct work_struct *wrk)
>	with_intel_runtime_pm(&gt->i915->runtime_pm, wakeref)
>		__update_guc_busyness_stats(guc);
>
> +	/* adjust context stats for overflow */
> +	xa_for_each(&guc->context_lookup, index, ce)
> +		guc_context_update_stats(ce);

This helps prevent overflow if the context switched out (which updated the
runtime in the context image) but was not not unpinned.

> +
>	intel_gt_reset_unlock(gt, srcu);
>
>	guc_enable_busyness_worker(guc);
> @@ -2774,6 +2799,7 @@ static void guc_context_unpin(struct intel_context *ce)
>  {
>	struct intel_guc *guc = ce_to_guc(ce);
>
> +	__guc_context_update_stats(ce);

Correct, cannot call intel_context_unpin (from guc_context_update_stats)
since this function is called from intel_context_unpin.

>	unpin_guc_id(guc, ce);
>	lrc_unpin(ce);
>
> @@ -3455,6 +3481,7 @@ static void remove_from_context(struct i915_request *rq)
>  }
>
>  static const struct intel_context_ops guc_context_ops = {
> +	.flags = COPS_RUNTIME_CYCLES,
>	.alloc = guc_context_alloc,
>
>	.close = guc_context_close,
> @@ -3473,6 +3500,8 @@ static const struct intel_context_ops guc_context_ops = {
>
>	.sched_disable = guc_context_sched_disable,
>
> +	.update_stats = guc_context_update_stats,
> +
>	.reset = lrc_reset,
>	.destroy = guc_context_destroy,
>
> @@ -3728,6 +3757,7 @@ static int guc_virtual_context_alloc(struct intel_context *ce)
>  }
>
>  static const struct intel_context_ops virtual_guc_context_ops = {
> +	.flags = COPS_RUNTIME_CYCLES,
>	.alloc = guc_virtual_context_alloc,
>
>	.close = guc_context_close,
> @@ -3745,6 +3775,7 @@ static const struct intel_context_ops virtual_guc_context_ops = {
>	.exit = guc_virtual_context_exit,
>
>	.sched_disable = guc_context_sched_disable,
> +	.update_stats = guc_context_update_stats,

The approach should also work for virtual engines.

>
>	.destroy = guc_context_destroy,
>
> --
> 2.36.1
>

Thanks.
--
Ashutosh