Anyone very familiar with EGL/GLESv2 who can give me a hand getting the Broadcom-specifics of the Raspberry Pi working with the EGL surface/context code in Chrome? I've added the dispmanx stuff, and here's where I am at the moment:
./chrome --use-gl=egl -log-level=0 and --enable-logging=stderr --vmodule=gl*=3
[1597:1597:1219673032:INFO:gpu_main.cc(85)] gpu_info_collector::CollectGraphicsInfo failed
[1597:1597:1219831736:VERBOSE1:gles2_cmd_decoder.cc(2139)] GL_OES_packed_depth_stencil not supported.
[1597:1597:1219842298:ERROR:gles2_cmd_decoder.cc(3031)] GLES2DecoderImpl::ResizeOffscreenFrameBuffer failed because offscreen FBO was incomplete.
[1597:1597:1219844701:ERROR:gles2_cmd_decoder.cc(2211)] Could not allocate offscreen buffer storage.
[1573:1573:1219850261:ERROR:command_buffer_proxy_impl.cc(134)] Failed to initialize command buffer service.
[1573:1573:1219854068:FATAL:image_transport_factory.cc(377)] Failed to make compositor shared context current.
Any pointers in the right direction would be hugely appreciated. Maybe the entire idea is bypassing X for graphics isn't sane? If you'd like to build for the Raspberry Pi, you can find a CL for my board overlay here:
https://gerrit.chromium.org/gerrit/#change,22836
Thanks,
Liam McLoughlin