Subject: Re: Experimental new cosc(4)
To: Leo Smiers <l.smiers@chello.nl>
From: Richard Earnshaw <rearnsha@arm.com>
List: port-arm32
Date: 07/06/2001 16:49:00
> In message <E15IVAm-00083I-00@chiark.greenend.org.uk> you wrote:
> 
> > 
> > Incidentally, I suspect the problem is that the driver's not turning on
> > interrupts.  The new cosc.c I got mentioned COSC_INTERRUPTS_ON, but nothing 
> > ever defined it, so the test version I produced doesn't use it.  If anyone  
> > knows how to turn on interrupts on the cosc, that might help.
> > 
> I might be wrong, but I a long tim eago someone/somewere told me that the
> cosc card has no interrupt line, so interrupts don't work. That's way it
> was still a polling driver. 

How will DMA be made to work if there are no interrupts for you to tell 
when a DMA transfer has completed?  Or would you rely on DMA interrupts 
from IOMD?

> In november I wanted to do DMA but as there was
> no solid DMA interface yet for the RPC I stopped and hoped that someone
> was going to add it. In fact I had a message from Richard Earnshaw who asked
> for a couple of evenings to sort it out. I am still waiting..., well I
> stopped wating and stopped at all.

It's a fair cop.  I've got plenty of excuses, but few that you would want 
to hear ;-)  I do still intend to look at it, but at this point I can make 
no promises as to when.  What I must do first, however, is get those 
changes I've made to my local tree either thrown away or committed to the 
master tree.  Then I can start looking at new things...

R.