Incomplete (potentially broken) Raven Chipset support for PREP machines.

[Gabriel Paubert]

On Tue, 8 Dec 1998, Cort Dougan wrote:

> 
> We'd then have to have one of those per board type.  That would be pretty
> much what we have now.  Perhaps I don't follow you.
> 
> }Why dont you have a table with two entries : one being the device func (device + function
> }byte) and the other the desired interrupt or directly (dev, func, intr)
> 
> I'm happy to take a look at any patches you have for that section of code.
> I don't have one of your machines so can't test for it.  The pci/irq code
> is a place where I was thinking we need a good cleanup in 2.2.

Indeed, I have attached the files which allow the Raven top be used on the
MVME2600, I can't make a diff, because my access to the CVS tree is too
slow. Beware that  the interrupt handling is ugly... 

But the interrupt mapping from residual data is probably clean enough to
be included directly.

BTW: I also use another BAT to map at 0xf0000000 for PCI interrupt
acknowledges, that's a huge overkill and it should be done in a different
way, it's only used to read one byte and can be done with ioremap, when it 
works on PreP. 

	Gabriel.

P.S: why is this list so slow, I've just got this morning tens of messages
from Novembre 23rd--25th.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/octet-stream
Size: 28012 bytes
Desc: Some support for Raven bridges
URL: <http://lists.ozlabs.org/pipermail/linuxppc-dev/attachments/19981209/df945855/attachment.obj>