Some of the devices I write linux drivers for deliver images to the host
at pretty high speed. I have allocated kernel memory with get_free_page
to serve this purpose. It is generally better if the board in question
can support DMA chaining so that you can allocate lots of pages.
I have found that allocating a bunch of pages and allowing the user
mode process to mmap that memory works well enough. Think of it as a
virtual frame buffer for a frame grabber. The problem with this is of course
that you need the physical memory available. (Then again, the hardware
really requires that to be the case anyhow.
-- Steve Williams steve@icarus.com steve@picturel.com"The woods are lovely, dark and deep. But I have promises to keep, And lines to code before I sleep, And lines to code before I sleep."