[Intel-gfx] [PATCH v2 1/1] drm/i915/guc/slpc: Add selftest for slpc tile-tile interaction

Tvrtko Ursulin tvrtko.ursulin at linux.intel.com
Fri Oct 28 10:23:34 UTC 2022


Hi,

I peeked inside from curiosity and was pleasantly surprise to see 
kthread_work is used! Some comments below.

On 28/10/2022 10:50, Riana Tauro wrote:
> Run a workload on tiles simultaneously by requesting for RP0 frequency.
> Pcode can however limit the frequency being granted due to throttling
> reasons. This test fails if there is any throttling
> 
> v2: Fix build error
> 
> Signed-off-by: Riana Tauro <riana.tauro at intel.com>
> ---
>   drivers/gpu/drm/i915/gt/selftest_slpc.c | 63 +++++++++++++++++++++++++
>   1 file changed, 63 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/gt/selftest_slpc.c b/drivers/gpu/drm/i915/gt/selftest_slpc.c
> index 82ec95a299f6..d19486772f5a 100644
> --- a/drivers/gpu/drm/i915/gt/selftest_slpc.c
> +++ b/drivers/gpu/drm/i915/gt/selftest_slpc.c
> @@ -13,6 +13,14 @@ enum test_type {
>   	VARY_MAX,
>   	MAX_GRANTED,
>   	SLPC_POWER,
> +	TILE_INTERACTION,
> +};
> +
> +struct slpc_thread {
> +	struct kthread_worker *worker;
> +	struct kthread_work work;
> +	struct intel_gt *gt;
> +	int result;
>   };
>   
>   static int slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 freq)
> @@ -310,6 +318,7 @@ static int run_test(struct intel_gt *gt, int test_type)
>   			break;
>   
>   		case MAX_GRANTED:
> +		case TILE_INTERACTION:
>   			/* Media engines have a different RP0 */
>   			if (engine->class == VIDEO_DECODE_CLASS ||
>   			    engine->class == VIDEO_ENHANCEMENT_CLASS) {
> @@ -325,6 +334,7 @@ static int run_test(struct intel_gt *gt, int test_type)
>   		case SLPC_POWER:
>   			err = slpc_power(gt, engine);
>   			break;
> +

Noise.

>   		}
>   
>   		if (test_type != SLPC_POWER) {
> @@ -426,6 +436,58 @@ static int live_slpc_power(void *arg)
>   	return ret;
>   }
>   
> +static void slpc_spinner_thread(struct kthread_work *work)
> +{
> +	struct slpc_thread *thread = container_of(work, typeof(*thread), work);
> +
> +	thread->result = run_test(thread->gt, TILE_INTERACTION);
> +}
> +
> +static int live_slpc_tile_interaction(void *arg)
> +{
> +	struct drm_i915_private *i915 = arg;
> +	struct intel_gt *gt;
> +	struct slpc_thread *threads;
> +	int i = 0, ret = 0;
> +
> +	threads = kcalloc(I915_MAX_GT, sizeof(*threads), GFP_KERNEL);
> +	if (!threads)
> +		return -ENOMEM;
> +
> +	for_each_gt(gt, i915, i) {
> +		pr_info("Running on GT: %d\n", gt->info.id);

Not sure logging these makes sense since test runs on all tiles in 
parallel anyway.

> +		threads[i].worker = kthread_create_worker(0, "igt/slpc_parallel:%d", gt->info.id);
> +
> +		if (IS_ERR(threads[i].worker)) {
> +			ret = PTR_ERR(threads[i].worker);
> +			break;
> +		}
> +
> +		threads[i].gt = gt;
> +		threads[i].result = 0;

No need to zero result due kcalloc and runs do not repeat.

> +		kthread_init_work(&threads[i].work, slpc_spinner_thread);
> +		kthread_queue_work(threads[i].worker, &threads[i].work);
> +	}
> +
> +	for_each_gt(gt, i915, i) {
> +		int status;
> +
> +		if (!threads[i].worker)
> +			continue;

Could be ERR_PTR by the look of it so it would crash below. Either gate 
on threads[i].gt or use IS_ERR_OR_NULL.

Regards,

Tvrtko

> +
> +		kthread_flush_work(&threads[i].work);
> +		status = READ_ONCE(threads[i].result);
> +		if (status && !ret) {
> +			pr_err("%s GT %d failed ", __func__, gt->info.id);
> +			ret = status;
> +		}
> +		kthread_destroy_worker(threads[i].worker);
> +	}
> +
> +	kfree(threads);
> +	return ret;
> +}
> +
>   int intel_slpc_live_selftests(struct drm_i915_private *i915)
>   {
>   	static const struct i915_subtest tests[] = {
> @@ -433,6 +495,7 @@ int intel_slpc_live_selftests(struct drm_i915_private *i915)
>   		SUBTEST(live_slpc_vary_min),
>   		SUBTEST(live_slpc_max_granted),
>   		SUBTEST(live_slpc_power),
> +		SUBTEST(live_slpc_tile_interaction),
>   	};
>   
>   	struct intel_gt *gt;


More information about the Intel-gfx mailing list