[igt-dev] [PATCH i-g-t] tests/kms_flip_scaled_crc: add flip to downscaled subtests

Shankar, Uma uma.shankar at intel.com
Tue Sep 15 12:08:51 UTC 2020



> -----Original Message-----
> From: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
> Sent: Monday, September 14, 2020 5:29 PM
> To: Shankar, Uma <uma.shankar at intel.com>; igt-dev at lists.freedesktop.org
> Subject: Re: [igt-dev] [PATCH i-g-t] tests/kms_flip_scaled_crc: add flip to
> downscaled subtests
> 
> On 4.9.2020 17.31, Shankar, Uma wrote:
> >
> >
> >> -----Original Message-----
> >> From: igt-dev <igt-dev-bounces at lists.freedesktop.org> On Behalf Of
> >> Juha-Pekka Heikkila
> >> Sent: Wednesday, September 2, 2020 2:42 PM
> >> To: igt-dev at lists.freedesktop.org
> >> Subject: [igt-dev] [PATCH i-g-t] tests/kms_flip_scaled_crc: add flip
> >> to downscaled subtests
> >
> > I think this would sound better "Add Subtests for Flip with fb downscaling"
> >
> >>
> >> attempt to cause cdclk changes and stress scalers with flipping to
> >> different size fb with different modifiers. Verify results with crc.
> >
> > This is failing on some platforms like in GLK:
> > https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_4945/shard-glk9/igt@kms
> > _flip_scaled_crc at flip-32bpp-ytile-to-32bpp-ytileccs.html
> >
> > Can you check why it could be happening, I can see you have the modifier
> check. But still its rejecting the combination in driver.
> > May be something to fix.
> 
> This I suspect is something problematic on i915 side, what is returned there is -
> EINVAL from flip. Those GLK dmesgs generally look somehow really messy, it is
> for some reason all the time trying to do something with pipe b even when it's
> supposing to be disabled.

Ok, we may have to ensure that this doesn't fail else enabling it in CI will crib.
May be create a separate Bug for this and get that tracked separately.

> >
> > Overall this is really a good plan to have this combination tested, something
> we have lacked.
> 
> There was regression on drm-tip on these which was just recently fixed
> hence this test. If for example would have run some videoplayer which
> uses hw downscaling result would've been just black screen ;)
> >
> >> Signed-off-by: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
> >> ---
> >>   tests/Makefile.sources      |   1 +
> >>   tests/kms_flip_scaled_crc.c | 255 ++++++++++++++++++++++++++++++++++++
> >>   tests/meson.build           |   1 +
> >>   3 files changed, 257 insertions(+)
> >>   create mode 100644 tests/kms_flip_scaled_crc.c
> >>
> >> diff --git a/tests/Makefile.sources b/tests/Makefile.sources index
> >> 269b506d..fe9fb7ed 100644
> >> --- a/tests/Makefile.sources
> >> +++ b/tests/Makefile.sources
> >> @@ -53,6 +53,7 @@ TESTS_progs = \
> >>   	kms_fence_pin_leak \
> >>   	kms_flip \
> >>   	kms_flip_event_leak \
> >> +	kms_flip_scaled_crc \
> >>   	kms_flip_tiling \
> >>   	kms_force_connector_basic \
> >>   	kms_frontbuffer_tracking \
> >> diff --git a/tests/kms_flip_scaled_crc.c b/tests/kms_flip_scaled_crc.c new file
> >> mode 100644 index 00000000..9daa2f92
> >> --- /dev/null
> >> +++ b/tests/kms_flip_scaled_crc.c
> >> @@ -0,0 +1,255 @@
> >> +/*
> >> + * Copyright © 2020 Intel Corporation
> >> + *
> >> + * Permission is hereby granted, free of charge, to any person
> >> +obtaining a
> >> + * copy of this software and associated documentation files (the
> >> +"Software"),
> >> + * to deal in the Software without restriction, including without
> >> +limitation
> >> + * the rights to use, copy, modify, merge, publish, distribute,
> >> +sublicense,
> >> + * and/or sell copies of the Software, and to permit persons to whom
> >> +the
> >> + * Software is furnished to do so, subject to the following conditions:
> >> + *
> >> + * The above copyright notice and this permission notice (including the
> >> +next
> >> + * paragraph) shall be included in all copies or substantial portions
> >> +of the
> >> + * Software.
> >> + *
> >> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
> >> +EXPRESS OR
> >> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
> >> +MERCHANTABILITY,
> >> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO
> EVENT
> >> +SHALL
> >> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
> DAMAGES
> >> OR
> >> +OTHER
> >> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
> >> +ARISING
> >> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
> >> OTHER
> >> +DEALINGS
> >> + * IN THE SOFTWARE.
> >> + *
> >> + */
> >> +
> >> +#include "igt.h"
> >> +
> >> +IGT_TEST_DESCRIPTION("Test flipping between scaled/nonscaled
> >> +framebuffers");
> >> +
> >> +typedef struct {
> >> +	int drm_fd;
> >> +	igt_display_t display;
> >> +	igt_output_t *output;
> >> +	enum pipe pipe;
> >> +	uint32_t gen;
> >> +	struct igt_fb small_fb;
> >> +	struct igt_fb big_fb;
> >> +	igt_pipe_crc_t *pipe_crc;
> >> +} data_t;
> >> +
> >> +const struct {
> >> +	const char * const name;
> >> +	const char * const describe;
> >> +	const uint64_t firstmodifier;
> >> +	const uint32_t firstformat;
> >> +	const uint64_t secondmodifier;
> >> +	const uint32_t secondformat;
> >> +} flip_scenario_test[] = {
> >> +	{
> >> +		"flip-32bpp-ytile-to-64bpp-ytile",
> >> +		"Try to flip from 32bpp non scaled fb to 64bpp downscaled fb in
> >> attempt to stress cdclk reprogramming.",
> >
> > We can trim this description, Like: "Flip from 32bpp non scaled fb to 64bpp
> downscaled fb to stress CD clock programming"
> > Trim this for all combinations.
> >
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED, DRM_FORMAT_XRGB8888,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED,
> >> DRM_FORMAT_XRGB16161616F
> >> +	},
> >> +	{
> >> +		"flip-64bpp-ytile-to-32bpp-ytile",
> >> +		"Try to flip from 64bpp non scaled fb to 32bpp downscaled fb.",
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED,
> >> DRM_FORMAT_XRGB16161616F,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED, DRM_FORMAT_XRGB8888
> >> +	},
> >> +	{
> >> +		"flip-64bpp-ytile-to-16bpp-ytile",
> >> +		"Try to flip from 64bpp non scaled fb to 16bpp downscaled fb.",
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED,
> >> DRM_FORMAT_XRGB16161616F,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED, DRM_FORMAT_RGB565
> >> +	},
> >> +	{
> >> +		"flip-32bpp-ytileccs-to-64bpp-ytile",
> >> +		"Try to flip from 32bpp ccs non scaled fb to 64bpp downscaled fb
> >> ytile.",
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED_CCS,
> >> DRM_FORMAT_XRGB8888,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED,
> >> DRM_FORMAT_XRGB16161616F
> >> +	},
> >> +	{
> >> +		"flip-32bpp-ytile-to-32bpp-ytilegen12rcccs",
> >> +		"Try to flip from 32bpp non scaled fb to 32bpp downscaled gen12
> >> rc ccs fb in attempt to stress cdclk reprogramming.",
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED, DRM_FORMAT_XRGB8888,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS,
> >> DRM_FORMAT_XRGB8888
> >> +	},
> >> +	{
> >> +		"flip-32bpp-ytile-to-32bpp-ytileccs",
> >> +		"Try to flip from 32bpp non scaled ytiled fb to 32bpp downscaled
> >> ytiled ccs fb.",
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED, DRM_FORMAT_XRGB8888,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED_CCS,
> >> DRM_FORMAT_XRGB8888
> >> +	},
> >> +	{
> >> +		"flip-64bpp-ytile-to-32bpp-ytilercccs",
> >> +		"Try to flip from 64bpp non scaled ytiled fb to 32bpp downscaled
> >> gen12 rc ccs fb.",
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED,
> >> DRM_FORMAT_XRGB16161616F,
> >> +		LOCAL_I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS,
> >> DRM_FORMAT_XRGB8888
> >> +	},
> >> +};
> >> +
> >> +static void setup_fb(data_t *data, struct igt_fb *newfb, uint32_t width,
> >> +		     uint32_t height, uint64_t format, uint64_t modifier) {
> >> +	struct drm_mode_fb_cmd2 f = {0};
> >> +	cairo_t *cr;
> >> +
> >> +	igt_require(igt_display_has_format_mod(&data->display, format,
> >> +					       modifier));
> >> +
> >> +	igt_create_bo_for_fb(data->drm_fd, width, height, format, modifier,
> >> +			     newfb);
> >> +	igt_assert(newfb->gem_handle > 0);
> >> +
> >> +	f.width = newfb->width;
> >> +	f.height = newfb->height;
> >> +	f.pixel_format = newfb->drm_format;
> >> +	f.flags = LOCAL_DRM_MODE_FB_MODIFIERS;
> >> +
> >> +	for (int n = 0; n < newfb->num_planes; n++) {
> >> +		f.handles[n] = newfb->gem_handle;
> >> +		f.modifier[n] = newfb->modifier;
> >> +		f.pitches[n] = newfb->strides[n];
> >> +		f.offsets[n] = newfb->offsets[n];
> >> +	}
> >> +
> >> +	cr = igt_get_cairo_ctx(data->drm_fd, newfb);
> >> +	igt_paint_color(cr, 0, 0, newfb->width, newfb->height, 0, 1, 0);
> >> +	igt_put_cairo_ctx(cr);
> >> +
> >> +	igt_assert(drmIoctl(data->drm_fd, LOCAL_DRM_IOCTL_MODE_ADDFB2,
> >> &f) == 0);
> >> +	newfb->fb_id = f.fb_id;
> >> +}
> >> +
> >> +static void free_fbs(data_t *data)
> >> +{
> >> +	igt_remove_fb(data->drm_fd, &data->small_fb);
> >> +	igt_remove_fb(data->drm_fd, &data->big_fb); }
> >> +
> >> +static void test_flip_to_scaled(data_t *data, uint32_t index, enum pipe pipe,
> >> +				igt_output_t *output)
> >> +{
> >> +	igt_plane_t *primary;
> >> +	igt_crc_t small_crc, big_crc;
> >> +
> >> +	drmModeModeInfo *mode;
> >> +	struct drm_event_vblank ev;
> >> +	int ret;
> >> +
> >> +	igt_display_reset(&data->display);
> >> +
> >> +	igt_debug("running on output %s pipe %s\n", output->name,
> >> +		  kmstest_pipe_name(pipe));
> >> +	if (data->big_fb.fb_id == 0) {
> >> +		mode = igt_output_get_mode(output);
> >> +
> >> +		// big fb will be 4k unless running on older than ICL
> >> +		if (data->gen < 11) {
> >> +			setup_fb(data, &data->small_fb,
> >> +				 min(mode->hdisplay, 640),
> >> +				 min(mode->vdisplay, 480),
> >> +				 flip_scenario_test[index].firstformat,
> >> +				 flip_scenario_test[index].firstmodifier);
> >> +
> >> +			setup_fb(data, &data->big_fb,
> >> +				 1280, 960,
> >
> > Not sure, but just something to check. Whether this downscaling will trigger a
> cd clock change for all platforms.
> > We should somehow check that as well if cd clock change really happened.  CD
> clock steps are not uniform across platforms.
> > Also if driver hardcodes a higher cd clock this test may pass without its
> intended objective.
> 
> I cannot enforce cdclk changes but what happen here with downscaling is
> it always go 4:1 I figure there's minimum four times more samples needed
> with filtering. As much as I verified those dmesgs from CI there was
> always display clock reprogrammed on all platforms but it is just what
> happen on today's drm-tip. If would run this on i915 from year ago there
> would be no such change thus on description it just says 'attempt to
> stress cdclk'.
> 
> As a flipping test I think final intended objective is flipping between
> framebuffers succeeded and it is verified with crc below.

Yeah agree, as long as this combination works we should be fine.

> 
> >
> >> +				 flip_scenario_test[index].secondformat,
> >> +				 flip_scenario_test[index].secondmodifier);
> >> +			igt_debug("small fb %dx%d\n", data->small_fb.width,
> >> +			          data->small_fb.height);
> >> +			igt_debug("big fb %dx%d\n", data->big_fb.width,
> >> +			          data->big_fb.height);
> >> +		} else {
> >> +			setup_fb(data, &data->small_fb,
> >> +				 min(mode->hdisplay, 1920),
> >> +				 min(mode->vdisplay, 1080),
> >> +				 flip_scenario_test[index].firstformat,
> >> +				 flip_scenario_test[index].firstmodifier);
> >> +
> >> +			setup_fb(data, &data->big_fb,
> >> +				 3840, 2160,
> >> +				 flip_scenario_test[index].secondformat,
> >> +				 flip_scenario_test[index].secondmodifier);
> >> +			igt_debug("small fb %dx%d\n", data->small_fb.width,
> >> +			          data->small_fb.height);
> >> +			igt_debug("big fb %dx%d\n", data->big_fb.width,
> >> +			          data->big_fb.height);
> >> +		}
> >> +	}
> >> +
> >> +	igt_output_set_pipe(output, pipe);
> >> +
> >> +	primary = igt_output_get_plane_type(output,
> >> DRM_PLANE_TYPE_PRIMARY);
> >> +	igt_display_commit_atomic(&data->display,
> >> DRM_MODE_ATOMIC_ALLOW_MODESET,
> >> +				  NULL);
> >> +	data->pipe_crc = igt_pipe_crc_new(data->drm_fd, pipe,
> >> +					  INTEL_PIPE_CRC_SOURCE_AUTO);
> >> +	igt_pipe_crc_start(data->pipe_crc);
> >> +
> >> +	igt_plane_set_position(primary, 0, 0);
> >> +	igt_plane_set_fb(primary, &data->small_fb);
> >> +
> >> +	igt_display_commit_atomic(&data->display,
> >> +				  DRM_MODE_ATOMIC_ALLOW_MODESET, NULL);
> >> +	igt_pipe_crc_get_current(data->drm_fd, data->pipe_crc, &small_crc);
> >> +
> >> +	igt_plane_set_fb(primary, &data->big_fb);
> >> +	igt_plane_set_size(primary, data->small_fb.width,
> >> +			   data->small_fb.height);
> >> +	ret = igt_display_try_commit_atomic(&data->display,
> >> +
> >> DRM_MODE_ATOMIC_ALLOW_MODESET  |
> >> +					    DRM_MODE_PAGE_FLIP_EVENT, NULL);
> >> +
> >> +	igt_require_f(ret != -ERANGE,
> >> +		      "Platform scaling limits exceeded, skipping.");
> >> +	igt_assert_eq(ret, 0);
> >> +
> >> +	igt_assert(read(data->drm_fd, &ev, sizeof(ev)) == sizeof(ev));
> >> +
> >> +	igt_pipe_crc_get_current(data->drm_fd, data->pipe_crc, &big_crc);
> >> +	igt_assert(igt_check_crc_equal(&small_crc, &big_crc));
> >
> > With this downscaling, are we able to match crc. Scalar may do some rounding,
> so not sure if this will
> > match in all cases. Good to check this.
> 
> We cannot really do crc checking with scalers but I have idea here
> maximum solid green would stay maximum solid green regardless of the
> scaler, all components for scaler are maximum solid green. With these
> starting values if scaler couldn't produce maximum solid green I'd call
> scaler broken.

Idea looks good, should be fine.

> >
> >> +
> >> +	igt_pipe_crc_stop(data->pipe_crc);
> >> +	igt_pipe_crc_free(data->pipe_crc);
> >> +	data->pipe_crc = NULL;
> >> +}
> >> +
> >> +igt_main
> >> +{
> >> +	enum pipe pipe;
> >> +	data_t data = {};
> >> +	igt_output_t *output;
> >> +
> >> +	igt_fixture {
> >> +		data.drm_fd = drm_open_driver_master(DRIVER_INTEL);
> >> +		data.gen = intel_gen(intel_get_drm_devid(data.drm_fd));
> >> +		igt_require(data.gen >= 9);
> >> +		igt_display_require(&data.display, data.drm_fd);
> >> +		igt_require(data.display.is_atomic);
> >> +		igt_require_pipe_crc(data.drm_fd);
> >> +		kmstest_set_vt_graphics_mode();
> >> +	}
> >> +
> >> +	for (int index = 0; index < ARRAY_SIZE(flip_scenario_test); index++) {
> >> +		igt_describe(flip_scenario_test[index].describe);
> >> +		igt_subtest(flip_scenario_test[index].name) {
> >> +			for_each_pipe_with_single_output(&data.display, pipe,
> >> +							 output)
> >> +				test_flip_to_scaled(&data, index, pipe, output);
> >> +
> >> +			free_fbs(&data);
> >> +		}
> >> +	}
> >> +	igt_fixture {
> >> +		free_fbs(&data);
> >> +		if (data.pipe_crc) {
> >> +			igt_pipe_crc_stop(data.pipe_crc);
> >> +			igt_pipe_crc_free(data.pipe_crc);
> >> +			data.pipe_crc = NULL;
> >> +		}
> >> +		kmstest_set_vt_text_mode();
> >> +		igt_display_fini(&data.display);
> >> +	}
> >> +}
> >> diff --git a/tests/meson.build b/tests/meson.build index 684de043..8e17a6ae
> >> 100644
> >> --- a/tests/meson.build
> >> +++ b/tests/meson.build
> >> @@ -37,6 +37,7 @@ test_progs = [
> >>   	'kms_fence_pin_leak',
> >>   	'kms_flip',
> >>   	'kms_flip_event_leak',
> >> +	'kms_flip_scaled_crc',
> >>   	'kms_flip_tiling',
> >>   	'kms_force_connector_basic',
> >>   	'kms_frontbuffer_tracking',
> >> --
> >> 2.26.0
> >>
> >> _______________________________________________
> >> igt-dev mailing list
> >> igt-dev at lists.freedesktop.org
> >> https://lists.freedesktop.org/mailman/listinfo/igt-dev



More information about the igt-dev mailing list