Getting X Off the Hardware

The X window system is generally imple- mented by directly inserting hardware manip- ulation code into the X server. Mode selection and 2D acceleration code are often executed in user mode and directly communicate with the hardware. The current architecture provides for separate 2D and 3D acceleration code, with the 2D code executed within the X server and the 3D code directly executed by the applica- tion, partially in user space and partially in the kernel. Video mode selection remains within the X server, creating an artificial dependency for 3D graphics on the correct operation of the window system. This paper lays out an alterna- tive structure for X within the Linux environ- ment where the responsibility for acceleration lies entirely within the existing 3D user/kernel library, the mode selection is delegated to an external library and the X server becomes a simple application layered on top of both of these. Various technical issues related to this architecture along with a discussion of input device handling will be discussed.