Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

CVS commit: syssrc

1 view
Skip to first unread message

Soren S. Jorvang

unread,
Jul 18, 2000, 3:00:00 AM7/18/00
to
On Tue, Jul 18, 2000 at 03:11:50PM -0400, John Hawkinson wrote:
> Well, GENERIC does not have cardbus support right now. Perhaps
> enabling it in GENERIC but not in CARDBUS?

I think for 1.5 we should leave cbb/pnpbios/pcibios disabled in
the default kernels, but also provide a "laptop" set of install
and generic kernels with those options enabled.


--
Soren

Frank van der Linden

unread,
Jul 18, 2000, 3:00:00 AM7/18/00
to
On Tue, Jul 18, 2000 at 09:14:40PM +0200, Soren S. Jorvang wrote:
> I think for 1.5 we should leave cbb/pnpbios/pcibios disabled in
> the default kernels, but also provide a "laptop" set of install
> and generic kernels with those options enabled.

Yep. I'm overhauling the install floppies anyway, and want to add
a "laptop" install floppy plus kernels. To be able to test that,
I need things to work right on my own laptop though :-)

Fortunately, Soda-san just checked in a fix that makes the PCIBIOS
options not panic on my laptop (and I'm pretty sure it'll solve
problems for others too). Now I just need to fix the memory mapping
problems on the O2 Micro cardbus bridge. Somehow it looks like it
will only allow mappings in the normal, legacy ISA memhole
for PCMCIA cards (i.e. 0xd0000). I'm trying to see what's going
on there.

- Frank

David Brownlee

unread,
Jul 18, 2000, 3:00:00 AM7/18/00
to
Does that leave any known cases where the PCIBIOS fixup stuff
makes things worse than not having it? If not, maybe we could
at least enable it in -current?

David/absolute
-- www.netbsd.org: A pmap for every occasion --

Bernd Sieker

unread,
Jul 19, 2000, 3:00:00 AM7/19/00
to
On 18.07.00, 21:14:40, Soren S. Jorvang wrote:
> On Tue, Jul 18, 2000 at 03:11:50PM -0400, John Hawkinson wrote:
> > Well, GENERIC does not have cardbus support right now. Perhaps
> > enabling it in GENERIC but not in CARDBUS?
>
> I think for 1.5 we should leave cbb/pnpbios/pcibios disabled in
> the default kernels, but also provide a "laptop" set of install
> and generic kernels with those options enabled.

I have found that the PnP-BIOS is utterly broken in some laptop BIOSs
(namely the "SystemSoft MobilePRO" BIOS), and causes the computer to
reset immediately after the kernel has found a pnpbios.

I have seen this on several different models, all with the "MobilePRO"
BIOS.

So I don't think this should be enabled by default.

I was also having big trouble getting PCMCIA NE2000-cards to work if I
have CardBus enabled (with "pcmcia* at cardslot?"); with a "pcmcia* at
pcic?"

I have also seen several models of notebooks that would not configure
the CardBus Bridge due to "unconfigured interrupt", unless
PCIBIOS_INTR_FIXUP was enabled.

It seems to me that in general PC-Cards of any kind are still bound to
cause problems.

>
>
> --
> Soren
>

--
Bernd Sieker

NetBSD: Linux without the hype.
-- Hubert Feyrer

John Hawkinson

unread,
Jul 19, 2000, 3:00:00 AM7/19/00
to
| > I think for 1.5 we should leave cbb/pnpbios/pcibios disabled in
| > the default kernels, but also provide a "laptop" set of install
| > and generic kernels with those options enabled.
|
| I have found that the PnP-BIOS is utterly broken in some laptop BIOSs
| (namely the "SystemSoft MobilePRO" BIOS), and causes the computer to
| reset immediately after the kernel has found a pnpbios.
| I have seen this on several different models, all with the "MobilePRO"
| BIOS.
|
| So I don't think this should be enabled by default.

Please submit open PR ("send-pr").

Including the output of the kernel messages to the serial console
for a kernel built with PNPBIOSVERBOSE and PNPBIOSDEBUG would be
quite helpful, if you can manage it.

If not, please boot the kernel debugger and set a breakpoint
in the pnpbios code and keep continuing-through it until you
determine the device which fails. If you need more fine-grained
instructions on how to do this, please say-so.

Thank you.

--jhawk

Frank van der Linden

unread,
Jul 19, 2000, 3:00:00 AM7/19/00
to
On Wed, Jul 19, 2000 at 01:35:13PM +0200, Bernd Sieker wrote:
> I have found that the PnP-BIOS is utterly broken in some laptop BIOSs
> (namely the "SystemSoft MobilePRO" BIOS), and causes the computer to
> reset immediately after the kernel has found a pnpbios.

Yep. I have one of those, a "SystemSoft" BIOS, and I'm seeing that too.
But, Windows does work, and I'm pretty sure that it uses the PnP BIOS
too, so there must be some way to get it to work (or detect that it's
broken and not use it, at least).

> I was also having big trouble getting PCMCIA NE2000-cards to work if I
> have CardBus enabled (with "pcmcia* at cardslot?"); with a "pcmcia* at
> pcic?"

On what kind of cardbus bridge was that? Not the "O2 Micro Inc 6832" by
any chance? I'm having some problems getting that to work as well.
I had to tell it to map memory in the legacy ISA range to make it
work with a 3c589D card. Now I'm trying a 3c575 (cardbus card), but
it's hanging somewhere.

- Frank

Bernd Sieker

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
On 19.07.00, 12:36:02, John Hawkinson wrote:
> | > I think for 1.5 we should leave cbb/pnpbios/pcibios disabled in
> | > the default kernels, but also provide a "laptop" set of install
> | > and generic kernels with those options enabled.
> |
> | I have found that the PnP-BIOS is utterly broken in some laptop BIOSs
> | (namely the "SystemSoft MobilePRO" BIOS), and causes the computer to
> | reset immediately after the kernel has found a pnpbios.
> | I have seen this on several different models, all with the "MobilePRO"
> | BIOS.
> |
> | So I don't think this should be enabled by default.
>
> Please submit open PR ("send-pr").
>
> Including the output of the kernel messages to the serial console
> for a kernel built with PNPBIOSVERBOSE and PNPBIOSDEBUG would be
> quite helpful, if you can manage it.
>
> If not, please boot the kernel debugger and set a breakpoint
> in the pnpbios code and keep continuing-through it until you
> determine the device which fails. If you need more fine-grained
> instructions on how to do this, please say-so.

I would have sent a PR, but unfortunately (well, rather fortunately
for me) I do not have access to those offending notebooks any longer,
but instead got one with a Phoenix BIOS, which seems to work fine.

I know someone who has a MobilePRO BIOS notebook, but I cannot say
when I will be able to lay hands on his machine again, let alone do
some debugging with the NetBSD kernel.

Sorry for that, I'd liked to have helped.

>
> Thank you.
>
> --jhawk
>

--
Bernd Sieker

Bernd Sieker

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to

As I said in another message, I got rid of that "bad" notebook, and
the one I have now suits my needs. Sorry again for not being able to
provide any more help.

>
> - Frank
>

--
Bernd Sieker

0 new messages