> Alan Cox wrote:
> > Its simple, elegant but does need optimising. If you have little ram left
> > but DMA ram your page allocation rate is noticably slower. On a 20Mb box
> > thats not ideal.
>
> An obvious solution is: put DMA pages at one of the list and non-DMA at
> the other end. RMQUEUE_DMA searches forward, RMQUEUE searches backward.
> Each type stays contiguous. Optimal, better than Ben's counting
> strategy, d'ya see?
You misread the patch. The counting is purely optional, and I'll remove
it if people don't like it. The really useful part of the patch is that
dmaable pages go on a separate list from non-dma. It's a simple, obvious
extension of the existing mechanism that's easy to get right -- whereas
putting pages on the back of the list is going to be a lot more subtle.
-ben
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/