summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2017-02-21split gbm outRob Clark4-32/+98
2017-02-21make eglGetPlatformDisplayEXT optionalRob Clark1-5/+11
Pull in get_proc() macro from yuv and atomic-fence branches. Once those parts get merged into a unified kmscube we'll want to look up more extension fxns.
2017-02-21update gitignoreRob Clark1-0/+2
2017-02-21add cmdline arg to specify kms deviceRob Clark1-4/+34
We'll want to add more args in the future, otherwise getopt might be a bit overkill.
2017-01-12Don't check for every kind of DRM driver. Just use the CARD !Myy1-23/+1
The DRM checks are WAY too fragile. Testing every potential DRM driver available goes against what device drivers are for. They're supposed to provide transparent abstractions. Currently the patch uses only card0. Some people *might* have multiple graphics cards. In this case, feel free to add options to choose the right DRM device entry (/dev/dri/cardX). Signed-off-by: Myy <myy@miouyouyou.fr>
2017-01-11Used eglGetPlatformDisplayEXT in order to use this sample with MaliMyy1-1/+9
For some reason, ARM Mali Wayland/DRM drivers does not seem to support eglGetDisplay(gbm.dev) and prefers instead eglGetPlatformDisplayEXT(EGL_PLATFORM_GBM_KHR, gbm.dev, NULL). Used the Weston DRM and GL renderers as reference. Signed-off-by: Myy <myy@miouyouyou.fr>
2016-09-20add vc4Eric Anholt1-1/+10
2016-09-07Make search for a crtc more robustNicholas Bishop1-4/+51
Running the kmscube test on qemu with "-vga virtio" fails for me with a "no encoder!" error. Fixed by searching all crtcs for one that is supported by one of the encoders in the connector. This code is only used if the existing `encoder->crtc_id` check is zero, so there should be no regression for systems where it already works. I think this should match the advice given here: https://dvdhrm.wordpress.com/2012/09/13/linux-drm-mode-setting-api/
2016-08-30also print GL extensionsRob Clark1-0/+1
2016-08-30misc cleanupRob Clark1-27/+33
2016-08-30fix typoRob Clark1-4/+4
2016-08-14add virtio_gpuGustavo Padovan1-1/+1
2016-08-14add tegraChristoph Haag1-1/+1
2014-11-04Prefer mode of type "DRM_MODE_TYPE_PREFERED"Gabriel Laskar1-1/+6
On screen with native resolution of 1440x900, there is sometime a compatibility mode for 1280x1024, with shrinked output and vertical bands to maintain ratio. If we select the resolution with the higher area, we select the compatibility mode and not the native resolution. When using mode_type == DRM_MODE_TYPE_PREFERED, we avoid that issue. Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
2013-11-13add msmRob Clark1-1/+1
The msm drm/kms + freedreno gallium driver can run kmscube too. Signed-off-by: Rob Clark <robdclark@gmail.com>
2012-10-24Makefile.am: remove need for m4 folderNicolas Dechesne1-2/+0
The current Makefile.am requires the 'm4' folder, however there is no need for it for the moment. that causes troubles when building packages with Debian CDBS 'autoreconf' which does not use the local autogen.sh script. Signed-off-by: Nicolas Dechesne <n-dechesne@ti.com>
2012-10-24Put vertex attribute data into a VBO.Jesse Barker1-109/+118
This makes the division between set up (init_gl()), and draw (draw()) code cleaner, and should be more efficient in general. Signed-off-by: Jesse Barker <jesse.barker@linaro.org>
2012-09-04update copyrightRob Clark1-0/+3
Found the original author of the gl cube code, so update the copyright notice accordingly.
2012-09-03initial commitRob Clark7-0/+1355