----- Forwarded message from Grant Grundler <grundler@parisc-linux.org> -----
From: Grant Grundler <grundler@parisc-linux.org>
Date: Sun, 9 Oct 2005 15:14:42 -0600
To: Joey Hess <joeyh@debian.org>
Cc: Debian Bug Tracking System <submit@bugs.debian.org>,
debian-hppa@lists.debian.org
Subject: Re: tulip module no longer works on a500
User-Agent: Mutt/1.5.9i
On Sun, Oct 09, 2005 at 04:18:16PM -0400, Joey Hess wrote:
> I run daily installation tests with d-i on a hppa a500 machine.
Awesome!
> /proc/pci does show that the ethernet card is there:
>
> Bus 0, device 0, function 0:
> Ethernet controller: Digital Equipment Corporation DECchip 21142/43 (rev 65).
> IRQ 65.
> Master Capable. Latency=255. Min Gnt=20.Max Lat=40.
> I/O at 0x0 [0x7f].
> Non-prefetchable 32 bit memory at 0xfffffffff8020000 [0xfffffffff80203ff].
>
> p2-mate points out that I/O at 0x0 is quite strange.
It's definitely wrong. It implies the resource never got assigned
or someone/something clobbered it.
PAT PDC machines (e.g. A500) need to assign resources for all PCI
devices except boot device and console. And the kernel definitely
does it for all the PAT boxes I've seen (e.g. 2.6.14-rc2):
ios:/usr/src# cat /proc/pci
PCI devices found:
Bus 0, device 0, function 0:
Class 0200: PCI device 1011:0019 (rev 65).
IRQ 18.
Master Capable. Latency=128. Min Gnt=20.Max Lat=40.
I/O at 0x80 [0xff].
Non-prefetchable 32 bit memory at 0xfffffffff8006000 [0xfffffffff80063ff]....
Can you post the entire console boot log someplace?
> Note the port 0x80... The other devices in pspci (scsi, etc) seem to be
> at the same ports with either kernel.
ok - good to know.
> With 2.6.12 I do see another device showing up at the 0x80 port:
>
> Bus 0, device 4, function 0:
> Serial controller: Hewlett-Packard Company Diva Serial [GSP] Multiport UART (rev 2).
This is the serial console. Sounds like a change in the resource
assignment is breakng your box.
> Note that this machine has two other (external) ethernet cards that, AFAIK, I
> cannot use with the linux kernel.
Eh?!!
"AceNIC Gigabit Ethernet" are most certainly supported by acenic driver.
I've submitted necessary driver fixes years ago and tested them.
Can you try "modprobe acenic"?
> Until this bug is fixed, I will be unable to provide daily installation tests
> for the hppa architecture.
We should be able to narrow this down pretty quickly.
I'm pretty sure we'll need more help from you or remote access
to the machine.
thanks,
grant
----- End forwarded message -----
--
see shy jo
Attachment:
signature.asc
Description: Digital signature