Here are the results of this weeks tests!

Is seems that the original problem - that knoppix chooses an incorrect driver for my nVidia card - might not be the real issue here. It still might need fixing, I just need to figure out where to send the lspci -v output to, and it'll get fixed eventually. In my case it probably won't help, however... so this discussion should probably be moved to another forum, as it's not knoppix-specific. Maybe I'll try debian-user next.

I'll post these last results, though, in case anyone here is a PCI-interrupt guru...

The reply from "Loper" opened a real can of worms for me, as I struggled to see how the BIOS assigns interrupts to PCI cards. As he guessed, I had a card in slot "1" - my SCSI controller card. In this position it is assigned IRQ 9, and everything works fine with my ati card in place.

I thought the best way to approach this problem would be to try to find a new home for the SCSI controller, that works with the ati card, then to see if I can insert the nVidia card and have it work OK.

The detailed results are in the updated version of http://www.netg.se/~hugge/archives/k...ia_problem.bz2 but basically, none of the other slots allow the scsi card to work usefully, as they all cause the BIOS to assign IRQ 11 to it. Interestingly, this clashes with my soundcard and stops the mouse working in exactly the same way as my original attempt with the nVidia card - the startup/shutdown sounds got chopped up in the same way, and the mouse was frozen. It is particularly interesting, as the nVidia card seems to like IRQ 11 in the one case that I have tried (with the SCSI controller in slot "1").

If anyone is interested, or would like to confirm my analysis, the contents of the archive mentioned above look like this:

Code:
knoppix_nvidia_problem/
knoppix_nvidia_problem/ati/
knoppix_nvidia_problem/ati/scsi_in_any_slot/
knoppix_nvidia_problem/ati/scsi_in_any_slot/XF86Config-4
knoppix_nvidia_problem/ati/scsi_in_slot1_working/
knoppix_nvidia_problem/ati/scsi_in_slot1_working/dmesg.txt
knoppix_nvidia_problem/ati/scsi_in_slot1_working/lspcin.txt
knoppix_nvidia_problem/ati/scsi_in_slot1_working/lspciv.txt
knoppix_nvidia_problem/ati/scsi_in_slot2/
knoppix_nvidia_problem/ati/scsi_in_slot2/dmesg.txt
knoppix_nvidia_problem/ati/scsi_in_slot2/lspcin.txt
knoppix_nvidia_problem/ati/scsi_in_slot2/lspciv.txt
knoppix_nvidia_problem/ati/scsi_in_slot3/
knoppix_nvidia_problem/ati/scsi_in_slot3/dmesg.txt
knoppix_nvidia_problem/ati/scsi_in_slot3/lspcin.txt
knoppix_nvidia_problem/ati/scsi_in_slot3/lspciv.txt
knoppix_nvidia_problem/ati/scsi_in_slot4/
knoppix_nvidia_problem/ati/scsi_in_slot4/dmesg.txt
knoppix_nvidia_problem/ati/scsi_in_slot4/lspcin.txt
knoppix_nvidia_problem/ati/scsi_in_slot4/lspciv.txt
knoppix_nvidia_problem/ati_and_nv/
knoppix_nvidia_problem/ati_and_nv/fstab
knoppix_nvidia_problem/ati_and_nv/lsmod.txt
knoppix_nvidia_problem/ati_and_nv/modules.conf
knoppix_nvidia_problem/ati_and_nv/XF86Config
knoppix_nvidia_problem/ati_and_nv/XF86Config-4.in
knoppix_nvidia_problem/ati_and_nv/XF86Config.in
knoppix_nvidia_problem/nv/
knoppix_nvidia_problem/nv/scsi_in_slot1/
knoppix_nvidia_problem/nv/scsi_in_slot1/dmesg.txt
knoppix_nvidia_problem/nv/scsi_in_slot1/lspciv.txt
knoppix_nvidia_problem/nv/scsi_in_slot1/XF86Config-4
knoppix_nvidia_problem/README
Next opportunity I get, and as long as there is still any point in doing so, I shall update this archive with test results of the nVidia card, moving the scsi controller to the other three slots.

The thing that worries me is that the nVidia card reports IRQ 11 from lspci -v, whereas the ati card doesn't report any IRQ at all. If the nVidia is so particular about which interrupt it wants to use, there may be no way to use it in my machine .

--
Cheers!

.~.
/V\
// \\
/( )\
^`~´^
< hugge >