On Mon, 4 Oct 1999, Andreas Beck wrote:
> But I'll contact Stephen about it. Looking at the Linux-Kernel reviews,
> we'll probably get along well :-).
I'm in a interesting debate with him. Well as you know I'm seeking a
solution to the accel and framebuffer access issue. He is totally sold on
the DRI method of coorperative locking from userland. I think this method
sucks. What I have done is started to rewrite the schedular function to
sync process access to these devices. If a process is scheduled and the
mmio region is already in use then the next process is considered instead.
It ensures only one process can access a MMIO resource at anytime. Marcus
knows whats the debate is all about.
