Around 17 o'clock on May 8, Alan Cox wrote: > FB already has mode switch context waits, so does DRM. And you don't > want to break direct access to registers for most things because its > a performance consideration of high importance.
Direct access to device registers from user mode really isn't necessary; placing a sequence of commands in a ring buffer shared between user and kernel space will yield essentially identical performance, allow for interrupt-managed waits and provide kernel-mediated device context switching. Existing "direct" rendering systems (like DRI or the nVidia GL driver) demonstrate how this kind of mechanism can work effectively. -keith
pgp00000.pgp
Description: PGP signature
