Help - Search - Members - Calendar
Full Version: Agp With Via Pt800
Linuxhelp > Support > Hardware Support
onety-three
Hello!
I've been trying to get 3D acceleration to work without success for several weeks now... Guess it's time to ask someone more experienced than me :)

My main problem seems to be the AGP support as fglrx (ATI display drivers) prints the well-known
(EE) fglrx(0): [agp] unable to acquire AGP, error "xf86_ENODEV"
and agpgart loads without printing any log lines except for
Linux agpgart interface v0.100 © Dave Jones

Having googled a lot, I know there should be some information about the memory architecture or at least a warning about my chipset not being supported. I compiled a 2.6.8-rc2-kernel with any debug option only remotely related to the problem enabled and a current agpgart-patch by Dave Jones applied (this one) but it didn't help. At least I got an additional log line related to agp:
bus pci: add driver agpgart-via

I tried all the agp_try_unsupported-stuff, too, although I didn't get the log line instructing me to do it, but of course it didn't help either.

Well, in case you need some technical information: The mb is an MSI MS-7042 which uses the VIA PT800 chipset (which has the VT8237 southbridge). Actually, this constant defined in pci_ids.h seems to be related to it:
#define PCI_DEVICE_ID_VIA_8237 0x3227
So, is my chipset known, but not yet supported?

This is what lspci -vvv prints: click me

If you could help me, I'd finally be able to trash my windowz... Thanks in advance!
christian
I've had this problem too, trying to get 3-D acceleration running for a MS-7042 board.
I am using SuSE 9.1, with the SuSE-adjusted ATI fglrx driver

The key point of my solution was to replace the line
#define PCI_DEVICE_ID_VIA_PT880 0x3258
in pci_ids.h
with the line
#define PCI_DEVICE_ID_VIA_PT880 0x0258

With this adjustment, I got the driver working properly.

The line
#define PCI_DEVICE_ID_VIA_8237 0x3227
refers only to the south bridge (sound etc.), not to the PT800 chip.
onety-three
It works! I can't believe it, it works :D
Thank you very, very much.

It seems to me this should be corrected in the official kernel. Has anybody got an idea how to contact the right people? I'm quite confused about all those different instructions on kernel.org, and the guy who's listed as maintainer for pci_ids.h writes on his webpage that he _was_ maintaining the file which seems to imply he isn't anymore. Hm.
Edwardos
I really want to say: THANK YOU. this small change made it work.

I surfed the net for some time. Could not find a solution till now!
PasiC
QUOTE (christian @ Aug 12 2004, 09:12 AM)
The key point of my solution was to replace the line
#define PCI_DEVICE_ID_VIA_PT880              0x3258
in pci_ids.h
with the line
#define PCI_DEVICE_ID_VIA_PT880              0x0258

With this adjustment, I got the driver working properly.

The line
#define PCI_DEVICE_ID_VIA_8237 0x3227
refers only to the south bridge (sound etc.), not to the PT800 chip.

hi,

I've got a lamer's question:P
after these modifs. whole kernel rebuild required?

bests: PasiC

UHU-Linux 1.2 beta (Hungary)
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2017 Invision Power Services, Inc.