I just found this thread lying around in my inbox, sorry I didn't answer
Quoted from Tim Waugh on Fri, Aug 23, 2002 at 09:33:56AM +0100.
> On Thu, Aug 22, 2002 at 06:42:20AM -0400, Blaise Gassend wrote:
> > I suspect that the absence in DMA read support probably explains why
> > this would be so CPU intensive.
> > I have a DMA read routine working for my cpia camera, but am having
> > trouble adapting it to the general parport driver. I sent a message
> > about my problems a few months ago, but got no replies. If any parport
> > gurus were willing to give me a hand it should be possible to get the
> > routine working with very little work.
> The trouble is that the DMA support in the parport driver is known to
> be unreliable. We never traced it down; it could be due to hardware
> for all I know.
Which part exactly is unreliable. As far as I know, the standard 4.x
kernel contains working functions that write to the parallel port using
hardware assisted PIO or DMA. There is a hardware assisted PIO read
function that is broken, and there is no DMA read.
Which part exactly were you refering to in your post when you said that
DMA support was unreliable? The DMA write?
I have written a DMA read function that has been working great for me
for months at a time (my camera has a faulty cable and I think that that
is what is causing problems every few months). I would be interested in
integrating it into the parport_pc driver, but have a few issues that I
would like advice from a parport guru for.
-- To unsubscribe, send mail to: firstname.lastname@example.org --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2b29 : Tue Oct 01 2002 - 14:20:01 EDT