[Intel-gfx] [PATCH i-g-t] i915/gem_eio: Trim kms workload

[Martin Peres]

On 2020-02-20 18:53, Chris Wilson wrote:
> We don't need to try reset-stress on every engine with the display, just
> once is enough to stress any interlinkage.

If you say so :)

> 
> This should reduce the runtime to 10s.

That would be appreciated!

> 
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Martin Peres <martin.peres at linux.intel.com>

The only difference between the previous behaviour and the new one
introduced by the patch is that reset_stress() is called for only the
default exec engine as opposed to all of them previously. So, the patch is:

Reviewed-by: Martin Peres <martin.peres at linux.intel.com>

> ---
>  tests/i915/gem_eio.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/tests/i915/gem_eio.c b/tests/i915/gem_eio.c
> index 0fe51efeb..1ec609410 100644
> --- a/tests/i915/gem_eio.c
> +++ b/tests/i915/gem_eio.c
> @@ -898,8 +898,14 @@ static void test_kms(int i915, igt_display_t *dpy)
>  
>  	test_inflight(i915, 0);
>  	if (gem_has_contexts(i915)) {
> -		test_reset_stress(i915, 0);
> -		test_reset_stress(i915, TEST_WEDGE);
> +		uint32_t ctx = context_create_safe(i915);
> +
> +		reset_stress(i915, ctx,
> +			     "default", I915_EXEC_DEFAULT, 0);
> +		reset_stress(i915, ctx,
> +			     "default", I915_EXEC_DEFAULT, TEST_WEDGE);
> +
> +		gem_context_destroy(i915, ctx);
>  	}
>  
>  	*shared = 1;
>