[igt-dev] [PATCH i-g-t 2/2] tests/i915/i915_pm_dc: Fix DC9 test

[Bhatt, Jigar]

-----Original Message-----
From: Gupta, Anshuman <anshuman.gupta at intel.com> 
Sent: Tuesday, September 21, 2021 4:55 PM
To: Bhatt, Jigar <jigar.bhatt at intel.com>; igt-dev at lists.freedesktop.org
Cc: Shankar, Uma <uma.shankar at intel.com>; Varide, Nischal <nischal.varide at intel.com>; Sharma, Swati2 <swati2.sharma at intel.com>
Subject: RE: [PATCH i-g-t 2/2] tests/i915/i915_pm_dc: Fix DC9 test



> -----Original Message-----
> From: Bhatt, Jigar <jigar.bhatt at intel.com>
> Sent: Tuesday, September 21, 2021 3:25 PM
> To: igt-dev at lists.freedesktop.org
> Cc: Bhatt, Jigar <jigar.bhatt at intel.com>; Shankar, Uma 
> <uma.shankar at intel.com>; Varide, Nischal <nischal.varide at intel.com>; 
> Sharma, Swati2 <swati2.sharma at intel.com>; Gupta, Anshuman 
> <anshuman.gupta at intel.com>
> Subject: [PATCH i-g-t 2/2] tests/i915/i915_pm_dc: Fix DC9 test
> 
> V1:
> Currently, check_dc9 is being called with reference
> (previous) counter being read after dpms_off call.
> At this time, already the counter is 0.
> We need to read the counter before dpms_off is called so that it holds 
> the previous value which had incremented while testing shallow states (DC5 or DC6).
> 
> V2:
> Bit cleaner code.[Imre]
> dump the i915_pm_runtime_status debugfs file in case of test 
> failure.[Anshuman]
> 
> V3:
> Disabling the polling during DC9 test “echo 0 > 
> /sys/module/drm_kms_helper/parameters/poll”.[Anshuman]
> 
> Signed-off-by: Jigar Bhatt <jigar.bhatt at intel.com>
> ---
>  tests/i915/i915_pm_dc.c | 39 +++++++++++++++++++++++++--------------
>  1 file changed, 25 insertions(+), 14 deletions(-)
> 
> diff --git a/tests/i915/i915_pm_dc.c b/tests/i915/i915_pm_dc.c index
> 04acf839..d4fa96f3 100644
> --- a/tests/i915/i915_pm_dc.c
> +++ b/tests/i915/i915_pm_dc.c
> @@ -40,6 +40,7 @@
>  #define CHECK_DC3CO(1 << 2)
> 
>  #define PWR_DOMAIN_INFO "i915_power_domain_info"
> +#define RPM_STATUS "i915_runtime_pm_status"
> 
>  typedef struct {
>  double r, g, b;
> @@ -393,23 +394,31 @@ static bool support_dc6(int debugfs_fd)  return 
> strstr(buf, "DC5 -> DC6 count");  }
> 
> -static bool check_dc9(uint32_t debugfs_fd, int prev_dc, bool 
> dc6_supported, int seconds)
> +static bool dc9_wait_entry(uint32_t debugfs_fd, int dc_target, int 
> +prev_dc, int seconds)
>  {
>  /*
>   * since we do not have DC9 counter,
>   * so we rely on dc5/dc6 counter reset to check if display engine was 
> in DC9.
>   */
> -return igt_wait(dc6_supported ? read_dc_counter(debugfs_fd,
> CHECK_DC6) <
> -prev_dc : read_dc_counter(debugfs_fd, CHECK_DC5) <
> +return igt_wait(read_dc_counter(debugfs_fd, dc_target) >
>  prev_dc, seconds, 100);
>  }
> 
> -static void setup_dc9_dpms(data_t *data, int prev_dc, bool 
> dc6_supported)
> +static void check_dc9(data_t *data, int dc_target, int prev_dc)
>  {
> +igt_assert_f(dc9_wait_entry(data->debugfs_fd, dc_target, prev_dc,
> 3000),
> +"DC9 state is not achieved\n%s:\n%s\n", RPM_STATUS,
> +data->debugfs_dump = igt_sysfs_get(data-
> >debugfs_fd, RPM_STATUS)); }
> +
> +static void setup_dc9_dpms(data_t *data, int dc_target) { int 
> +prev_dc;
> +
> +prev_dc = read_dc_counter(data->debugfs_fd, dc_target);
>  setup_dc_dpms(data);
>  dpms_off(data);
> -igt_skip_on_f(!(igt_wait(dc6_supported ? read_dc_counter(data-
> >debugfs_fd, CHECK_DC6) >
> -prev_dc : read_dc_counter(data->debugfs_fd,
> CHECK_DC5) >
> +igt_skip_on_f(!(igt_wait(read_dc_counter(data->debugfs_fd, dc_target)
> +>
>  prev_dc, 3000, 100)), "Unable to enters shallow DC states\n");  
> dpms_on(data);  cleanup_dc_dpms(data); @@ -417,18 +426,20 @@ static 
> void setup_dc9_dpms(data_t *data, int prev_dc, bool dc6_supported)
> 
>  static void test_dc9_dpms(data_t *data)  { -bool dc6_supported;
> +int prev_dc, dc_target;
> +bool buf;
> +const char *kms_poll =
> "/sys/module/drm_kms_helper/parameters/poll";
> 
> +buf = igt_sysfs_get_boolean(data->debugfs_fd, kms_poll); 
> +igt_sysfs_set_boolean(data->debugfs_fd, kms_poll, 0);
Passing debugfs_fd for module parameter is completely wrong!
I am not sure whether this will work or not.
Also please move this restore and setup part of polling to setup_dc9_dpms().
Thanks,
Anshuman


Hi Anshuman,

 igt_sysfs_get_boolean(data->debugfs_fd, kms_poll);  what else we can use ?
 if we move restore part to setup_dc9_dpms() again it may be resume when check_dc9 is checking counter reset.

Thanks and regards,
Jigar Bhatt

>  require_dc_counter(data->debugfs_fd, CHECK_DC5); -dc6_supported = 
> support_dc6(data->debugfs_fd); -setup_dc9_dpms(data, dc6_supported ? 
> read_dc_counter(data-
> >debugfs_fd, CHECK_DC6) :
> -read_dc_counter(data->debugfs_fd, CHECK_DC5), dc6_supported);
> +dc_target = support_dc6(data->debugfs_fd) ? CHECK_DC6 :
> CHECK_DC5;
> +setup_dc9_dpms(data, dc_target);
> +prev_dc = read_dc_counter(data->debugfs_fd, dc_target);
>  dpms_off(data);
> -igt_assert_f(check_dc9(data->debugfs_fd, dc6_supported ?
> -read_dc_counter(data->debugfs_fd,
> CHECK_DC6) :
> -read_dc_counter(data->debugfs_fd,
> CHECK_DC5),
> -dc6_supported, 3000), "Not in DC9\n");
> +check_dc9(data, dc_target, prev_dc);
>  dpms_on(data);
> +igt_sysfs_set_boolean(data->debugfs_fd, kms_poll, buf);
>  }
> 
>  IGT_TEST_DESCRIPTION("These tests validate Display Power DC states");
> --
> 2.25.1