On Fri, Apr 14, 2006 at 04:18:17PM -0400, Jeff Garzik wrote:
> Jack Carroll wrote:
> > Device server? How about a motherboard that breaks the 16-interrupt
> >limit of the classic PC? It could use multi-level vectored interrupts,
> >like
>
> This limit is already long gone. Please read up on PCI Message
> Signalled Interrupts (MSI) and PCI MSI-X.
>
>
> >the 68000 on VME bus, which gives practially unlimited interrupts. Then it
> >could have a whole lot of PCI slots, and the resources to actually use them
> >in an I/O-heavy application.
> > Of course, MS wouldn't support all those I/O slots, but do we care?
>
> The standard code in Linux, and very likely Windows, already supports
> this sort of thing.
>
> Jeff
????
How do I tell Linux 2.6.15 to assign three PCI Ethernet boards to
interrupts 16, 17, and 18, leaving 5, 9, 10, 11, and 12 free for the ISA
boards?
According to "Upgrading & Repairing PCs", 14th edition, page 329,
"For the PCI bus to function in a PC, the PCI interrupts must be mapped to
ISA interrupts." And on page 327, it shows that there are only 16 ISA
interrupts -- and 10 of them are hardwired to fixed functions. With 2.6.15
grabbing IRQ 9 through 11 for internal functions, there aren't enough left
for more than a couple of actual PCI and ISA card sockets.
Is there some way around that?
Every physical socket in a maxed-out motherboard should be
guaranteed non-conflicting I/O resources, regardless of what else is present
in the computer.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
