lkml.org 
[lkml]   [1999]   [Mar]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: PlugNPlay Bios expert wanted ;-)
On Tue, Mar 16, 1999 at 09:55:11AM +0100, Ulf Jaenicke-Roessler wrote:

> > Basically, the Realtek people are right. The W9x simply removes the value
> > for the interrupt number of the ethernet card from the PCI config space
> > register of that card.
>
> Who knows why Windows does this :( I don't see any sense in it.

Because it simply tries to reinitialize the whole PCI subsystem its
way while starting (and doing PnP stuff), and as it does that, the values
get lost (and even should they stay, they'd be bogus), and it doesn't care
about writing the correct value back, because IT knows them and thinks
it's enough.

> > The number was written there by the BIOS, while initializing the PCI
> > subsystem to tell the operating system which IRQ was assigned to the
> > card.
>
> Well, as I wrote in my original message, if I use BIOS access for the PCI
> card, it can be seen, that the I/O information is destroyed too.

When BIOS is queried about the irq number, it just looks in the place where
it has written it during boot and blindly assumes it's still there. If it
is not, it'll report the value which is there.

> There was one way, I could keep the IRQ information. I needed to disable
> 'interrupt steering' in Windows. However, it didn't restore the I/O
> information and there were certain problems in Windows itself (although I
> can't remember exactly - it's related to the PnP support of Windows, but
> not only), so I decided to stick with the default behaviour.

Hmm, but, with the different i/o assigned, but the interrupt working, the
realtek card could work, doesn't it?

> I hoped, that there might be a way to trigger the BIOS to rewrite the
> information about I/O and IRQ.

There isn't.

> > This is a problem of W9x and, as far as I know could only be fixed by
> > re-assigning all the interrupts to PCI cards upon Linux boot, which doesn't
> > happen now (though is possible to happen in the future).
>
> This would be great.

I'm working on that with a friend here. A new bus architecture for Linux 2.3.

> > So, the only solution for you now is to mark the linux batchfile as needing
> > a different configuration in W98, so that W98 reboots prior to running it.
>
> Yes, but actually I tried to circumvent rebooting, because it's faster.

It should be possible to make a program that'd save the information, and then
put it back ...

Goodd luck,
Vojtech

-
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 13:50    [W:0.046 / U:0.672 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site