All of lore.kernel.org
 help / color / mirror / Atom feed
From: Harish Chegondi <harish.chegondi@intel.com>
To: "Dixit, Ashutosh" <ashutosh.dixit@intel.com>
Cc: <igt-dev@lists.freedesktop.org>,
	Lukasz Laguna <lukasz.laguna@intel.com>,
	 Adam Miszczak <adam.miszczak@linux.intel.com>,
	Jakub Kolakowski <jakub1.kolakowski@intel.com>,
	Kamil Konieczny <kamil.konieczny@linux.intel.com>,
	Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>
Subject: Re: [PATCH i-g-t 3/3] tests/intel/xe_eu_stall: Use query IOCTL to check for EU stall support
Date: Wed, 11 Jun 2025 09:37:12 -0700	[thread overview]
Message-ID: <aEmwuNhRbhPwR7X8@intel.com> (raw)
In-Reply-To: <875xhbuoev.wl-ashutosh.dixit@intel.com>

On Wed, Jun 04, 2025 at 03:34:16PM -0700, Dixit, Ashutosh wrote:
> On Tue, 03 Jun 2025 16:57:36 -0700, Harish Chegondi wrote:
> >
> > Use EU stall query IOCTL to check if EU stall monitoring feature is
> > available on the platform and if it is supported in the driver.
> > This would replace the checks in the test with those in the driver.
> >
> > Cc: Lukasz Laguna <lukasz.laguna@intel.com>
> > Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
> > Cc: Adam Miszczak <adam.miszczak@linux.intel.com>
> > Cc: Jakub Kolakowski <jakub1.kolakowski@intel.com>
> > Cc: Kamil Konieczny <kamil.konieczny@linux.intel.com>
> > Cc: Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>
> > Signed-off-by: Harish Chegondi <harish.chegondi@intel.com>
> > Resolves: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/5131
> > ---
> >  tests/intel/xe_eu_stall.c | 55 +++++++++++++++++++++------------------
> >  1 file changed, 30 insertions(+), 25 deletions(-)
> >
> > diff --git a/tests/intel/xe_eu_stall.c b/tests/intel/xe_eu_stall.c
> > index b3a0ee742..4e76018b6 100644
> > --- a/tests/intel/xe_eu_stall.c
> > +++ b/tests/intel/xe_eu_stall.c
> > @@ -74,6 +74,8 @@ static int stream_fd = -1;
> >
> >  static volatile bool child_is_running = true;
> >
> > +static struct drm_xe_query_eu_stall *query_eu_stall_data;
> > +
> >  /*
> >   * EU stall data format for PVC
> >   */
> > @@ -506,33 +508,12 @@ static void test_eustall(int drm_fd, uint32_t devid, bool blocking_read, int ite
> >		.properties_ptr = to_user_pointer(properties),
> >	};
> >
> > -	struct drm_xe_query_eu_stall *query_eu_stall_data;
> > -	struct drm_xe_device_query query = {
> > -		.extensions = 0,
> > -		.query = DRM_XE_DEVICE_QUERY_EU_STALL,
> > -		.size = 0,
> > -		.data = 0,
> > -	};
> > -
> >	igt_info("User buffer size: %u\n", p_user);
> >	if (p_args[0])
> >		igt_info("Workload: %s\n", p_args[0]);
> >	else
> >		igt_info("Workload: GPGPU fill\n");
> > -
> > -	igt_assert_eq(igt_ioctl(drm_fd, DRM_IOCTL_XE_DEVICE_QUERY, &query), 0);
> > -	igt_assert_neq(query.size, 0);
> > -
> > -	query_eu_stall_data = malloc(query.size);
> > -	igt_assert(query_eu_stall_data);
> > -
> > -	query.data = to_user_pointer(query_eu_stall_data);
> > -	igt_assert_eq(igt_ioctl(drm_fd, DRM_IOCTL_XE_DEVICE_QUERY, &query), 0);
> > -
> > -	igt_assert(query_eu_stall_data->num_sampling_rates > 0);
> > -	if (p_rate == 0)
> > -		properties[3] = query_eu_stall_data->sampling_rates[0];
> > -	igt_info("Sampling Rate: %lu\n", properties[3]);
> > +	igt_info("Sampling Rate: %lu\n", p_rate);
> >
> >	stream_fd = eu_stall_open(drm_fd, &props);
> >
> > @@ -660,18 +641,42 @@ static struct option long_options[] = {
> >
> >  igt_main_args("e:g:o:r:u:w:", long_options, help_str, opt_handler, NULL)
> >  {
> > -	int drm_fd;
> > +	bool blocking_read = true;
> > +	int drm_fd, ret;
> >	uint32_t devid;
> >	struct stat sb;
> > -	bool blocking_read = true;
> > +	struct drm_xe_device_query query = {
> > +		.extensions = 0,
> > +		.query = DRM_XE_DEVICE_QUERY_EU_STALL,
> > +		.size = 0,
> > +		.data = 0,
> > +	};
> >
> >	igt_fixture {
> >		drm_fd = drm_open_driver(DRIVER_XE);
> >		igt_require_fd(drm_fd);
> >		devid = intel_get_drm_devid(drm_fd);
> > -		igt_require(IS_PONTEVECCHIO(devid) || intel_graphics_ver(devid) >= IP_VER(20, 0));
> > +
> >		igt_require_f(igt_get_gpgpu_fillfunc(devid), "no gpgpu-fill function\n");
> >		igt_require_f(!stat(OBSERVATION_PARANOID, &sb), "no observation_paranoid file\n");
> > +
> > +		ret = igt_ioctl(drm_fd, DRM_IOCTL_XE_DEVICE_QUERY, &query);
> > +		igt_skip_on_f((ret == -1 && errno == ENODEV),
> > +			      "EU stall monitoring is not available on this platform\n");
> > +		igt_skip_on_f((ret == -1 && errno == EINVAL),
> 
> inner brackets not needed...
Will remove.
> 
> > +			      "EU stall monitoring is not supported in the driver\n");
> > +		igt_assert_neq(query.size, 0);
> > +
> > +		query_eu_stall_data = malloc(query.size);
> > +		igt_assert(query_eu_stall_data);
> > +
> > +		query.data = to_user_pointer(query_eu_stall_data);
> > +		igt_assert_eq(igt_ioctl(drm_fd, DRM_IOCTL_XE_DEVICE_QUERY, &query), 0);
> > +
> > +		igt_assert(query_eu_stall_data->num_sampling_rates > 0);
> > +		if (p_rate == 0)
> > +			p_rate = query_eu_stall_data->sampling_rates[0];
> > +
> 
> Maybe cleaner to put all this query stuff into a function and calling the
> function from the fixture? Also, queries are generally cached in 'struct
> xe_device' so we could do the same here? See xe_device_get() and also
> xe_query_oa_units_new(). Since that's in the lib, make sure IGT still works
> even if the eu_stall query fails (see 851e62faf971).
I am moving the query IOCTL into the lib. Probably those changes should
go into a separate patch?
> 
> Also, another nit, use igt_require or igt_require_f instead of
> igt_skip_on_f? Though they are mostly NOT of each other so they are
> equivalent, so either is ok.
I initially used igt_require_f, but later changed it to igt_skip_on_f
as the condition logic is simpler with igt_skip_on_f.
> 
> >		if (output_file) {
> >			output = fopen(output_file, "w");
> >			igt_require(output);
> > --
> > 2.48.1
> >
Thank You
Harish.

  parent reply	other threads:[~2025-06-11 16:37 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-03 23:57 [PATCH i-g-t 1/3] tests/intel/xe_eu_stall: Close any open EU stall fd before opening a new one Harish Chegondi
2025-06-03 23:57 ` [PATCH i-g-t 2/3] tests/intel/xe_eu_stall: Do not check for presence of data on simulation Harish Chegondi
2025-06-05  0:10   ` Dixit, Ashutosh
2025-06-03 23:57 ` [PATCH i-g-t 3/3] tests/intel/xe_eu_stall: Use query IOCTL to check for EU stall support Harish Chegondi
2025-06-04 13:35   ` Kolakowski, Jakub1
2025-06-04 22:21     ` Dixit, Ashutosh
2025-06-04 22:34   ` Dixit, Ashutosh
2025-06-05  2:29     ` Dixit, Ashutosh
2025-06-11 16:37     ` Harish Chegondi [this message]
2025-06-04  0:22 ` ✓ Xe.CI.BAT: success for series starting with [i-g-t,1/3] tests/intel/xe_eu_stall: Close any open EU stall fd before opening a new one Patchwork
2025-06-04  0:42 ` ✓ i915.CI.BAT: " Patchwork
2025-06-04  2:36 ` ✗ i915.CI.Full: failure " Patchwork
2025-06-04 23:53 ` [PATCH i-g-t 1/3] " Dixit, Ashutosh
2025-06-11 16:38   ` Harish Chegondi
2025-06-05  0:22 ` ✗ Xe.CI.Full: failure for series starting with [i-g-t,1/3] " Patchwork

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=aEmwuNhRbhPwR7X8@intel.com \
    --to=harish.chegondi@intel.com \
    --cc=adam.miszczak@linux.intel.com \
    --cc=ashutosh.dixit@intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=jakub1.kolakowski@intel.com \
    --cc=kamil.konieczny@linux.intel.com \
    --cc=lukasz.laguna@intel.com \
    --cc=marcin.bernatowicz@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.