lkml.org 
[lkml]   [2000]   [Mar]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: forcing PCI device IRQ numbers ?
"Richard B. Johnson" wrote:

> On Thu, 23 Mar 2000, Paul Barton-Davis wrote:
>
> > is there any way to force the assignment of PCI device IRQ's ?
> >
> > i have a situation in one of my machines where it ends up with 2
> > digital audio interfaces plus the SCSI adapter sitting on the same
> > IRQ. Not a disaster, but not good either. There are, of course, free
> > IRQs, otherwise I wouldn't be asking.
> >
>
> Swapping boards around on the PCI bus helps on some motherboards. Also
> turn off PNP in the BIOS. This makes things cleaner when Linux gets
> control. The ASUS P2B-DS is one that has the problem. At one time I
> found that everything was on IRQ12. Turning OFF PNP made some boards
> look like they were on IRQ9 (phony, it's the cascade for the old
> controller). Moving a few boards around got them all onto separate
> IRQs.
>

Actually, I usually have better luck enabling plug and play OS in the BIOS.
Both with Linux and Windows 98SE. Enabling plug and play OS will cause most
BIOSes not to initialize ISA cards. There are also usually settings in the
BIOS to reserve IRQs for "Legacy Device". All that means is that it won't
use it when it goes to do PnP init. If PnPOS is turned on, then that means
it won't assign a PCI card to that IRQ. When Linux starts, that IRQ will be
available for use with isapnp. In Windows, it will use that IRQ for the
ISA board.

The downside is that PCI devices will be sharing IRQs. But it is my
understanding that PCI devices sharing IRQs is not necessarily a bad thing
unless the drivers for them are poorly written. One card that comes to mind
is the STB TVPCI under Windows, it hates sharing an IRQ. Played with it
under Linux a long time ago, and got a picture from the composite input, but
the tuner didn't work right (and I haven't had the thing working in Windows
for so long, I have no idea if it works or not).

Really, the Plug and Play OS setting can be your friend if you beat it into
submission ;).

-Dave


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:57    [W:0.035 / U:1.304 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site