Re: Floating point in the kernel

[Adrian M. <ad...@ne...>]

On Mon, 2007-10-08 at 05:39 +0900, Paul Mundt wrote:

> At this point you're stuck with doing all of your extensive optimizations
> in something like X or DirectFB. It's still worth doing deferred I/O for
> bulk page write-out via the DMAC for the framebuffer, which should at
> least give you some more bandwidth for things like X and mplayer which
> mmap() the device.

The problem is that the DMA doesn't work for the framebuffer as such but
for the TA. Direct DMA is texture DMA and, afaics - I am still hacking
at this when I get the time, which has been in short supply in the last
few weeks - that is still about loading in a lot of TA parameters, just
not asking the TA to process them.

As you suggest, what I am looking at is replacing the generic PIO
bitblit etc. Just not convinced it is possible with FP