backends/drm: use explicit sync where possible
Instead of calling glFinish, which blocks until it's done and has high CPU usage on NVidia, use EGL_ANDROID_native_fence_fd to get an explicit sync fd, which the commit thread automatically waits on before committing the buffer to KMS.
Note that this still doesn't entirely fix bug 452219 (at least not on its own), but it at least reduces the CPU usage a bit for me and should reduce it for setups with NVidia as primary as well.