public inbox for igt-dev@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Martin Peres <martin.peres@linux.intel.com>
To: Ville Syrjala <ville.syrjala@linux.intel.com>,
	igt-dev@lists.freedesktop.org
Subject: Re: [igt-dev] [PATCH i-g-t] tests/kms_rotation_crc: Speed up by not testing every format
Date: Fri, 21 Feb 2020 09:16:00 +0200	[thread overview]
Message-ID: <23cbeaa2-70fd-58b1-5899-dbb91d8bca21@linux.intel.com> (raw)
In-Reply-To: <20200214165502.21990-1-ville.syrjala@linux.intel.com>

On 2020-02-14 18:55, Ville Syrjala wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> As we did with kms_plane_scaling let's not test every pixel
> format for rotation fails either. Instead we test each "class"
> of formats as defined by igt_reduce_format(). The hope being
> that there are no specific bugs in the hw for different
> swizzles of the same underlying base format.
> 
> On KBL:
> $ time ./build/tests/kms_rotation_crc --r sprite-rotation-90
> - real	0m39,851s
> - user	0m1,037s
> - sys	0m12,895s
> + real	0m16,773s
> + user	0m0,357s
> + sys	0m5,475s
> 
> As per the usual recipe we add an --extended cmdline knob to
> force the full coverage.
> 
> Only cuts ~30% of the total kms_rotation_crc runtime though.
> There are other subtests which take over 80s on this particular
> machine, so further work is clearly required to get the testing
> time to sane levels.
> 
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

This cuts the execution time of the test from 00:15:07 -> 00:06:53 on
TGL, so this is definitely welcomed!

As for the patch itself, I agree with the approach and I like that you
added the --extended flag just to cover our basis. One day, we should
work on running these extended tests randomly, but today is not the day.

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

Thanks!

> ---
>  tests/kms_rotation_crc.c | 58 +++++++++++++++++++++++++++++++++++++---
>  1 file changed, 55 insertions(+), 3 deletions(-)
> 
> diff --git a/tests/kms_rotation_crc.c b/tests/kms_rotation_crc.c
> index 2a7b10e990e2..fc4c13389953 100644
> --- a/tests/kms_rotation_crc.c
> +++ b/tests/kms_rotation_crc.c
> @@ -23,6 +23,7 @@
>   */
>  
>  #include "igt.h"
> +#include "igt_vec.h"
>  #include <math.h>
>  
>  #define MAX_FENCES 32
> @@ -65,6 +66,7 @@ typedef struct {
>  
>  	struct p_struct *multiplaneoldview;
>  	struct p_point planepos[MAXMULTIPLANESAMOUNT];
> +	bool extended;
>  } data_t;
>  
>  typedef struct {
> @@ -371,6 +373,28 @@ static void test_single_case(data_t *data, enum pipe pipe,
>  	}
>  }
>  
> +static bool test_format(data_t *data,
> +			struct igt_vec *tested_formats,
> +			uint32_t format)
> +{
> +	if (!igt_fb_supported_format(format))
> +		return false;
> +
> +	if (!is_i915_device(data->gfx_fd) ||
> +	    data->extended)
> +		return true;
> +
> +	format = igt_reduce_format(format);
> +
> +	/* only test each format "class" once */
> +	if (igt_vec_index(tested_formats, &format) >= 0)
> +		return false;
> +
> +	igt_vec_push(tested_formats, &format);
> +
> +	return true;
> +}
> +
>  static void test_plane_rotation(data_t *data, int plane_type, bool test_bad_format)
>  {
>  	igt_display_t *display = &data->display;
> @@ -408,15 +432,21 @@ static void test_plane_rotation(data_t *data, int plane_type, bool test_bad_form
>  				continue;
>  
>  			if (!data->override_fmt) {
> +				struct igt_vec tested_formats;
> +
> +				igt_vec_init(&tested_formats, sizeof(uint32_t));
> +
>  				for (j = 0; j < plane->drm_plane->count_formats; j++) {
>  					uint32_t format = plane->drm_plane->formats[j];
>  
> -					if (!igt_fb_supported_format(format))
> +					if (!test_format(data, &tested_formats, format))
>  						continue;
>  
>  					test_single_case(data, pipe, output, plane, i,
>  							 format, test_bad_format);
>  				}
> +
> +				igt_vec_fini(&tested_formats);
>  			} else {
>  				test_single_case(data, pipe, output, plane, i,
>  						 data->override_fmt, test_bad_format);
> @@ -738,7 +768,30 @@ static const char *tiling_test_str(uint64_t tiling)
>  	}
>  }
>  
> -igt_main
> +static int opt_handler(int opt, int opt_index, void *_data)
> +{
> +	data_t *data = _data;
> +
> +	switch (opt) {
> +	case 'e':
> +		data->extended = true;
> +		break;
> +	}
> +
> +	return IGT_OPT_HANDLER_SUCCESS;
> +}
> +
> +static const struct option long_opts[] = {
> +	{ .name = "extended", .has_arg = false, .val = 'e', },
> +	{}
> +};
> +
> +static const char help_str[] =
> +	"  --extended\t\tRun the extended tests\n";
> +
> +static data_t data;
> +
> +igt_main_args("", long_opts, help_str, opt_handler, &data)
>  {
>  	struct rot_subtest {
>  		unsigned plane;
> @@ -771,7 +824,6 @@ igt_main
>  		{ 0, 0 }
>  	};
>  
> -	data_t data = {};
>  	int gen = 0;
>  
>  	igt_fixture {
> 
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev

      parent reply	other threads:[~2020-02-21  7:16 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-14 16:55 [igt-dev] [PATCH i-g-t] tests/kms_rotation_crc: Speed up by not testing every format Ville Syrjala
2020-02-14 18:48 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
2020-02-15 21:02 ` [igt-dev] [PATCH i-g-t] " Chris Wilson
2020-02-17 22:43 ` [igt-dev] ✗ Fi.CI.IGT: failure for " Patchwork
2020-02-21  7:16 ` Martin Peres [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=23cbeaa2-70fd-58b1-5899-dbb91d8bca21@linux.intel.com \
    --to=martin.peres@linux.intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=ville.syrjala@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox