[igt-dev] [Intel-gfx] [PATCH i-g-t] igt/pm_rpm: Ignore modesets for basic tests with no KMS

[Souza, Jose]

On Fri, 2018-10-05 at 09:05 +0100, Chris Wilson wrote:
> If KMS is not available, we cannot simply turn on an output and
> expect
> that to wake the device up. As such we have to ignore that part of
> the
> basic subtest and simply proclaim victory if the device is able to
> sleep!

We could replace the 'turn on a screen' to submit some job batch to one
of the engines, this way we could test the waking path too when KMS is
disabled but for now it is better than skip the tests at all as you
said.

Reviewed-by: José Roberto de Souza <jose.souza at intel.com>

> 
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> ---
>  tests/pm_rpm.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/tests/pm_rpm.c b/tests/pm_rpm.c
> index 7488efd7..731c9cf6 100644
> --- a/tests/pm_rpm.c
> +++ b/tests/pm_rpm.c
> @@ -777,7 +777,8 @@ static void basic_subtest(void)
>  {
>  	disable_all_screens_and_wait(&ms_data);
>  
> -	enable_one_screen_and_wait(&ms_data);
> +	if (ms_data.res)
> +		enable_one_screen_and_wait(&ms_data);
>  }
>  
>  static void pc8_residency_subtest(void)
> @@ -1405,8 +1406,10 @@ static void pci_d3_state_subtest(void)
>  	disable_all_screens_and_wait(&ms_data);
>  	igt_assert(igt_wait(device_in_pci_d3(), 2000, 100));
>  
> -	enable_one_screen_and_wait(&ms_data);
> -	igt_assert(!device_in_pci_d3());
> +	if (ms_data.res) {
> +		enable_one_screen_and_wait(&ms_data);
> +		igt_assert(!device_in_pci_d3());
> +	}
>  }
>  
>  static void __attribute__((noreturn)) stay_subtest(void)