2001-04-22 01:32:32

by Manuel McLure

[permalink] [raw]
Subject: 2.4.3-ac10/ac11 crash at boot in PDC20265 check

It looks like the check for the PDC20265 added in ac10 causes some
problems. I have an MSI K7T Turbo R which has a Promise FastTrak built in -
when I try to boot ac10 or ac11 on it I get an Oops just after the "ide:
Found promise 20265 in RAID mode." message. I diffed ide-pci.c between ac5
(which worked) and ac10 (which fails) and found that the only change was
the check for the PDC20265 - I commented this out and the kernel boots fine
now.

On another topic, does anyone know how to disable the onboard FastTrak? I
don't use it, and it makes my boot time longer...

Thanks,
--
Manuel A. McLure KE6TAW | ...for in Ulthar, according to an ancient
<[email protected]> | and significant law, no man may kill a cat.
<http://www.mclure.org> | -- H.P. Lovecraft


2001-04-22 02:01:22

by Alan

[permalink] [raw]
Subject: Re: 2.4.3-ac10/ac11 crash at boot in PDC20265 check

> Found promise 20265 in RAID mode." message. I diffed ide-pci.c between ac5
> (which worked) and ac10 (which fails) and found that the only change was
> the check for the PDC20265 - I commented this out and the kernel boots fine
> now.

Can you send me the oops data. I'll take a look and figure out what is up.
My first guess would be its a NULL pointer error ?

2001-04-22 02:20:24

by Manuel McLure

[permalink] [raw]
Subject: Re: 2.4.3-ac10/ac11 crash at boot in PDC20265 check

On 2001.04.21 19:02 Alan Cox wrote:
> > Found promise 20265 in RAID mode." message. I diffed ide-pci.c between
> ac5
> > (which worked) and ac10 (which fails) and found that the only change
> was
> > the check for the PDC20265 - I commented this out and the kernel boots
> fine
> > now.
>
> Can you send me the oops data. I'll take a look and figure out what is
> up.
> My first guess would be its a NULL pointer error ?

Yes, it's a NULL pointer dereference. I copied down part of the Oops - I'll
have to put the code back before I can copy down the rest. Here's what I
have until now:

Unable to handle kernel NULL pointer dereference at virtual address
00000024 printing eip:
c02a71de

pgd entry c0101000: 0000000000000000
pmd entry c0101000: 0000000000000000
.. pmd not present!
Oops: 0000
CPU: 0
EIP: 0010:[<c02a71de>]
EFLAGS: 00010286

According to System.map, c02a71de is in the middle of ide_setup_pci_device
which makes sense. ide_setup_pci_device starts at c02a7020.

--
Manuel A. McLure KE6TAW | ...for in Ulthar, according to an ancient
<[email protected]> | and significant law, no man may kill a cat.
<http://www.mclure.org> | -- H.P. Lovecraft