[Mesa-users] Need help getting start with EGL/DRM/KMS/GLES2

[Albert Freeman]

My mesa config: I am using the git development tree for both libdrm
and mesa. Linux 4.1.6-1 (the 1 is an arch linux maintainer patch).

./autogen.sh --prefix=/usr \
    --sysconfdir=/etc \
    --with-dri-driverdir=/usr/lib/xorg/modules/dri \
    --with-gallium-drivers= \
    --with-dri-drivers=i965 \
    --with-egl-platforms=x11,drm,wayland \
    --enable-egl \
    --enable-gbm \
    --enable-shared-glapi \
    --enable-glx \
    --enable-glx-tls \
    --enable-dri \
    --enable-gles1 \
    --enable-gles2 \
    --enable-texture-float \
    --enable-xa \
    --enable-vdpau

On 2 September 2015 at 00:52, Albert Freeman <albertwdfreeman at gmail.com> wrote:
> My mesa config: I am using the git development tree for both libdrm
> and mesa. Linux 4.1.6-1 (the 1 is an arch linux maintainer patch).
>
> ./autogen.sh --prefix=/usr \
>     --sysconfdir=/etc \
>     --with-dri-driverdir=/usr/lib/xorg/modules/dri \
>     --with-gallium-drivers= \
>     --with-dri-drivers=i965 \
>     --with-egl-platforms=x11,drm,wayland \
>     --enable-egl \
>     --enable-gbm \
>     --enable-shared-glapi \
>     --enable-glx \
>     --enable-glx-tls \
>     --enable-dri \
>     --enable-gles1 \
>     --enable-gles2 \
>     --enable-texture-float \
>     --enable-xa \
>     --enable-vdpau
>
> On 1 September 2015 at 17:16, Jean-François Doyon <jfdoyon at gmail.com> wrote:
>> Albert,
>>
>> Thanks for taking the time to investigate ... I'm afraid that made no
>> difference on my end.  I guess there's something about my software stack or
>> environment that's causing this then, now that I know I wasn't crazy and my
>> code was indeed supposed to work!
>>
>> Can I ask what your environment is like? Were you using the vmwgfx driver?
>> What version of Mesa, libdrm, kernel, etc?
>>
>> I'll try to get gdb into my image to see if I can glean some more details
>> ...
>>
>> I did try to use buildroot's included libdrm/mesa configs (instead of my own
>> "recipes"), which are also slightly older versions, but that changed nothing
>> either ...
>>
>> Thanks,
>> J.F.
>>
>> On Tue, Sep 1, 2015 at 12:10 AM, Albert Freeman <albertwdfreeman at gmail.com>
>> wrote:
>>>
>>> Well with these quick and dirty changes I made, it works fine on my
>>> system, no segfault and the screen displays red.
>>>
>>> --- gleskmsorg.c    2015-09-01 14:04:32.853672410 +0000
>>> +++ gleskms.c    2015-09-01 14:03:57.797005474 +0000
>>> @@ -24,8 +24,6 @@ EGL_CONFORMANT, EGL_OPENGL_ES2_BIT,
>>>  EGL_NONE };
>>>
>>>  static const char device_name[] = "/dev/dri/card0";
>>> -static const uint32_t drm_mode_connector_id = 18;
>>> -static const uint32_t drm_mode_crtc_id = 21;
>>>
>>>  int main(void) {
>>>
>>> @@ -86,10 +84,12 @@ int main(void) {
>>>          exit(EXIT_FAILURE);
>>>      }
>>>
>>> +    drmModeRes *res;
>>>      drmModeConnector *conn;
>>>      uint32_t conn_id, width, height;
>>>      drmModeModeInfo modeinfo;
>>> -    conn = drmModeGetConnector(fd, drm_mode_connector_id);
>>> +    res = drmModeGetResources(fd);
>>> +    conn = drmModeGetConnector(fd, *res->connectors);
>>>      conn_id = conn->connector_id;
>>>      width = conn->modes[0].hdisplay;
>>>      height = conn->modes[0].vdisplay;
>>> @@ -129,7 +129,7 @@ int main(void) {
>>>          exit(EXIT_FAILURE);
>>>      }
>>>
>>> -    ret = drmModeSetCrtc(fd, drm_mode_crtc_id, drm_fb_id, 0, 0, &conn_id,
>>> 1,
>>> +    ret = drmModeSetCrtc(fd, *res->crtcs, drm_fb_id, 0, 0, &conn_id, 1,
>>>              &modeinfo);
>>>      if (ret) {
>>>          printf("failed to set mode: %m\n");
>>> @@ -143,8 +143,8 @@ int main(void) {
>>>      gbm_surface_destroy(gs);
>>>      eglMakeCurrent(dpy, EGL_NO_SURFACE, EGL_NO_SURFACE, EGL_NO_CONTEXT);
>>>      eglDestroyContext(dpy, ctx);
>>> -//    eglDestroySurface(dpy, surface);
>>> -//    eglTerminate(dpy);
>>> +    eglDestroySurface(dpy, surface);
>>> +    eglTerminate(dpy);
>>>      gbm_device_destroy(gbm);
>>>      close(fd);
>>>      return 0;
>>
>>