diff mbox series

[i-g-t] tests/kms_panel_fitting: Fix plane scaling avoidance on gen7/gen8

Message ID 20190830223023.16396-1-matthew.d.roper@intel.com (mailing list archive)
State New, archived
Headers show
Series [i-g-t] tests/kms_panel_fitting: Fix plane scaling avoidance on gen7/gen8 | expand

Commit Message

Matt Roper Aug. 30, 2019, 10:30 p.m. UTC 
Most gen7 and gen8 platforms can't do plane scaling, so we need to
ensure the test doesn't try to do plane scaling on those platforms.  The
legacy non-atomic subtest bakes these platform characteristics into the
test itself since legacy modesetting interfaces don't provide a way to
probe platform capabilities like atomic does.

Maarten previously tried to address this with commit 24c5e0778
("tests/kms_panel_fitting: Do not use scaling on gen7 and gen8, v2."),
but he augmented an existing test for gen9+pipeC which comes too late
since we've already turned on the plane at that point.  We can fix this
by moving the test up higher; not only before we enable panel fitting,
but also before we turn on the sprite.

Note that this still isn't a great subtest since it's very
Intel-specific, despite being part of a generic KMS test that's intended
for all vendors.  A future enhancement might be to try to probe the
platform capabilities with a TEST_ONLY atomic operation before using the
legacy interfaces to actually program them.

While touching the code, I also added some slight restructuring, added
additional comments, and broke up >80 char lines to add clarity to
what's going on.

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
---
 tests/kms_panel_fitting.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

Comments

Ville Syrjälä Sept. 2, 2019, 4:51 p.m. UTC | #1 
On Fri, Aug 30, 2019 at 03:30:23PM -0700, Matt Roper wrote:
> Most gen7 and gen8 platforms can't do plane scaling, so we need to
> ensure the test doesn't try to do plane scaling on those platforms.  The
> legacy non-atomic subtest bakes these platform characteristics into the
> test itself since legacy modesetting interfaces don't provide a way to
> probe platform capabilities like atomic does.
> 
> Maarten previously tried to address this with commit 24c5e0778
> ("tests/kms_panel_fitting: Do not use scaling on gen7 and gen8, v2."),
> but he augmented an existing test for gen9+pipeC which comes too late
> since we've already turned on the plane at that point.  We can fix this
> by moving the test up higher; not only before we enable panel fitting,
> but also before we turn on the sprite.
> 
> Note that this still isn't a great subtest since it's very
> Intel-specific, despite being part of a generic KMS test that's intended
> for all vendors.  A future enhancement might be to try to probe the
> platform capabilities with a TEST_ONLY atomic operation before using the
> legacy interfaces to actually program them.
> 
> While touching the code, I also added some slight restructuring, added
> additional comments, and broke up >80 char lines to add clarity to
> what's going on.
> 
> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

> ---
>  tests/kms_panel_fitting.c | 24 +++++++++++++++++++-----
>  1 file changed, 19 insertions(+), 5 deletions(-)
> 
> diff --git a/tests/kms_panel_fitting.c b/tests/kms_panel_fitting.c
> index 491e429f..a942294b 100644
> --- a/tests/kms_panel_fitting.c
> +++ b/tests/kms_panel_fitting.c
> @@ -131,11 +131,10 @@ static void test_panel_fitting(data_t *d)
>  		igt_fb_set_position(&d->fb2, d->plane2, 100, 100);
>  		igt_fb_set_size(&d->fb2, d->plane2, d->fb2.width-200, d->fb2.height-200);
>  		igt_plane_set_position(d->plane2, 100, 100);
> -		igt_plane_set_size(d->plane2, mode->hdisplay-200, mode->vdisplay-200);
> -		igt_display_commit2(display, COMMIT_UNIVERSAL);
>  
>  		/*
> -		 * most of gen7 and all of gen8 doesn't support scaling at all.
> +		 * Most of gen7 and all of gen8 doesn't support plane scaling
> +		 * at all.
>  		 *
>  		 * gen9 pipe C has only 1 scaler shared with the crtc, which
>  		 * means pipe scaling can't work simultaneously with panel
> @@ -144,9 +143,24 @@ static void test_panel_fitting(data_t *d)
>  		 * Since this is the legacy path, userspace has to know about
>  		 * the HW limitations, whereas atomic can ask.
>  		 */
> -		if (IS_GEN8(devid) || (IS_GEN7(devid) && !IS_IVYBRIDGE(devid)) ||
> +		if (IS_GEN8(devid) ||
> +		    (IS_GEN7(devid) && !IS_IVYBRIDGE(devid)) ||
>  		    (IS_GEN9(devid) && pipe == PIPE_C))
> -			igt_plane_set_size(d->plane2, d->fb2.width-200, d->fb2.height-200);
> +			/* same as visible area of fb => no scaling */
> +			igt_plane_set_size(d->plane2,
> +					   d->fb2.width-200,
> +					   d->fb2.height-200);
> +		else
> +			/*
> +			 * different than visible area of fb => plane scaling
> +			 * active
> +			 */
> +			igt_plane_set_size(d->plane2,
> +					   mode->hdisplay-200,
> +					   mode->vdisplay-200);
> +
> +		/* Plane scaling active (if possible), pfit off */
> +		igt_display_commit2(display, COMMIT_UNIVERSAL);
>  
>  		/* enable panel fitting along with sprite scaling */
>  		mode->hdisplay = 1024;
> -- 
> 2.20.1
> 
> _______________________________________________
> igt-dev mailing list
> igt-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/igt-dev
Matt Roper Sept. 2, 2019, 9:10 p.m. UTC | #2 
On Mon, Sep 02, 2019 at 07:51:41PM +0300, Ville Syrjälä wrote:
> On Fri, Aug 30, 2019 at 03:30:23PM -0700, Matt Roper wrote:
> > Most gen7 and gen8 platforms can't do plane scaling, so we need to
> > ensure the test doesn't try to do plane scaling on those platforms.  The
> > legacy non-atomic subtest bakes these platform characteristics into the
> > test itself since legacy modesetting interfaces don't provide a way to
> > probe platform capabilities like atomic does.
> > 
> > Maarten previously tried to address this with commit 24c5e0778
> > ("tests/kms_panel_fitting: Do not use scaling on gen7 and gen8, v2."),
> > but he augmented an existing test for gen9+pipeC which comes too late
> > since we've already turned on the plane at that point.  We can fix this
> > by moving the test up higher; not only before we enable panel fitting,
> > but also before we turn on the sprite.
> > 
> > Note that this still isn't a great subtest since it's very
> > Intel-specific, despite being part of a generic KMS test that's intended
> > for all vendors.  A future enhancement might be to try to probe the
> > platform capabilities with a TEST_ONLY atomic operation before using the
> > legacy interfaces to actually program them.
> > 
> > While touching the code, I also added some slight restructuring, added
> > additional comments, and broke up >80 char lines to add clarity to
> > what's going on.
> > 
> > Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> > Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> 
> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Pushed, thanks for the review.


Matt

> 
> > ---
> >  tests/kms_panel_fitting.c | 24 +++++++++++++++++++-----
> >  1 file changed, 19 insertions(+), 5 deletions(-)
> > 
> > diff --git a/tests/kms_panel_fitting.c b/tests/kms_panel_fitting.c
> > index 491e429f..a942294b 100644
> > --- a/tests/kms_panel_fitting.c
> > +++ b/tests/kms_panel_fitting.c
> > @@ -131,11 +131,10 @@ static void test_panel_fitting(data_t *d)
> >  		igt_fb_set_position(&d->fb2, d->plane2, 100, 100);
> >  		igt_fb_set_size(&d->fb2, d->plane2, d->fb2.width-200, d->fb2.height-200);
> >  		igt_plane_set_position(d->plane2, 100, 100);
> > -		igt_plane_set_size(d->plane2, mode->hdisplay-200, mode->vdisplay-200);
> > -		igt_display_commit2(display, COMMIT_UNIVERSAL);
> >  
> >  		/*
> > -		 * most of gen7 and all of gen8 doesn't support scaling at all.
> > +		 * Most of gen7 and all of gen8 doesn't support plane scaling
> > +		 * at all.
> >  		 *
> >  		 * gen9 pipe C has only 1 scaler shared with the crtc, which
> >  		 * means pipe scaling can't work simultaneously with panel
> > @@ -144,9 +143,24 @@ static void test_panel_fitting(data_t *d)
> >  		 * Since this is the legacy path, userspace has to know about
> >  		 * the HW limitations, whereas atomic can ask.
> >  		 */
> > -		if (IS_GEN8(devid) || (IS_GEN7(devid) && !IS_IVYBRIDGE(devid)) ||
> > +		if (IS_GEN8(devid) ||
> > +		    (IS_GEN7(devid) && !IS_IVYBRIDGE(devid)) ||
> >  		    (IS_GEN9(devid) && pipe == PIPE_C))
> > -			igt_plane_set_size(d->plane2, d->fb2.width-200, d->fb2.height-200);
> > +			/* same as visible area of fb => no scaling */
> > +			igt_plane_set_size(d->plane2,
> > +					   d->fb2.width-200,
> > +					   d->fb2.height-200);
> > +		else
> > +			/*
> > +			 * different than visible area of fb => plane scaling
> > +			 * active
> > +			 */
> > +			igt_plane_set_size(d->plane2,
> > +					   mode->hdisplay-200,
> > +					   mode->vdisplay-200);
> > +
> > +		/* Plane scaling active (if possible), pfit off */
> > +		igt_display_commit2(display, COMMIT_UNIVERSAL);
> >  
> >  		/* enable panel fitting along with sprite scaling */
> >  		mode->hdisplay = 1024;
> > -- 
> > 2.20.1
> > 
> > _______________________________________________
> > igt-dev mailing list
> > igt-dev@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/igt-dev
> 
> -- 
> Ville Syrjälä
> Intel
diff mbox series

Patch

diff --git a/tests/kms_panel_fitting.c b/tests/kms_panel_fitting.c
index 491e429f..a942294b 100644
--- a/tests/kms_panel_fitting.c
+++ b/tests/kms_panel_fitting.c
@@ -131,11 +131,10 @@  static void test_panel_fitting(data_t *d)
 		igt_fb_set_position(&d->fb2, d->plane2, 100, 100);
 		igt_fb_set_size(&d->fb2, d->plane2, d->fb2.width-200, d->fb2.height-200);
 		igt_plane_set_position(d->plane2, 100, 100);
-		igt_plane_set_size(d->plane2, mode->hdisplay-200, mode->vdisplay-200);
-		igt_display_commit2(display, COMMIT_UNIVERSAL);
 
 		/*
-		 * most of gen7 and all of gen8 doesn't support scaling at all.
+		 * Most of gen7 and all of gen8 doesn't support plane scaling
+		 * at all.
 		 *
 		 * gen9 pipe C has only 1 scaler shared with the crtc, which
 		 * means pipe scaling can't work simultaneously with panel
@@ -144,9 +143,24 @@  static void test_panel_fitting(data_t *d)
 		 * Since this is the legacy path, userspace has to know about
 		 * the HW limitations, whereas atomic can ask.
 		 */
-		if (IS_GEN8(devid) || (IS_GEN7(devid) && !IS_IVYBRIDGE(devid)) ||
+		if (IS_GEN8(devid) ||
+		    (IS_GEN7(devid) && !IS_IVYBRIDGE(devid)) ||
 		    (IS_GEN9(devid) && pipe == PIPE_C))
-			igt_plane_set_size(d->plane2, d->fb2.width-200, d->fb2.height-200);
+			/* same as visible area of fb => no scaling */
+			igt_plane_set_size(d->plane2,
+					   d->fb2.width-200,
+					   d->fb2.height-200);
+		else
+			/*
+			 * different than visible area of fb => plane scaling
+			 * active
+			 */
+			igt_plane_set_size(d->plane2,
+					   mode->hdisplay-200,
+					   mode->vdisplay-200);
+
+		/* Plane scaling active (if possible), pfit off */
+		igt_display_commit2(display, COMMIT_UNIVERSAL);
 
 		/* enable panel fitting along with sprite scaling */
 		mode->hdisplay = 1024;