On Thu, 2019-08-29 at 13:52 -0700, Matt Roper wrote: > When choosing a low resolution mode to switch to, we search the > connector's mode list for a mode that's small enough that a plane > positioned in the bottom left corner of the screen in the original mode > will be fully offscreen in the new, smaller mode. However if no > suitable modes are found in the modelist, we blindly fall back to a 10x7 > mode without actually checking that the canary plane will now be > offscreen. If, for example, the original mode was 13x7, then switching > to 10x7 will leave the plane visible onscreen, causing the test to fail. > > Before falling back to a 10x7 mode, check whether that's actually small > enough. If it isn't just skip this subtest. > > v2: Restructure the skip condition to make it more obvious why this size > limitation is important. (Simon) > > Cc: Simon Ser <simon.ser@xxxxxxxxx> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105581 > Signed-off-by: Matt Roper <matthew.d.roper@xxxxxxxxx> Thanks! Reviewed-by: Simon Ser <simon.ser@xxxxxxxxx> > --- > tests/kms_plane_lowres.c | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/tests/kms_plane_lowres.c b/tests/kms_plane_lowres.c > index 68b85025..1a316c84 100644 > --- a/tests/kms_plane_lowres.c > +++ b/tests/kms_plane_lowres.c > @@ -59,8 +59,12 @@ get_lowres_mode(int drmfd, igt_output_t *output, drmModeModeInfo *mode_default) > } > } > > - if (!found) > + if (!found) { > + igt_require_f(mode_default->vdisplay - SIZE > 768, > + "Current mode not tall enough; plane would still be onscreen after switch to 10x7.\n"); > + > return *igt_std_1024_mode_get(); > + } > > return mode; > } _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
