nvidia and severn2 -- an alternative

Pekka Pietikainen pp at ee.oulu.fi
Sat Oct 4 21:56:19 UTC 2003


On Sat, Oct 04, 2003 at 04:03:35PM -0400, Robert P. J. Day wrote:
>   if you're running the 2.6.0-xxx kernel, go to www.minion.de, and grab
> the single diff file NVIDIA_kernel-1.0-4496-2.6.diff. (note: while i'm
> grateful for that site, i find the explanation there virtually
> incomprehensible.)
>   i use this constantly as i like to keep up with the 2.6.0 bk patches.
> 
Which reminds me, anyone else having problems with the minion.de stuff?

It used to work with 2.6 without tweaking, but broke some time
ago. It looks AGP related, if I actually disable AGP support (or enable
NVAGP which doesn't get enabled due to agpgart being statically in the
kernel) it works with a performance hit. I think it actually used
NVAGP when it did work with 2.6...

With agpgart I get a messed up text screen (fixable by setfont) and X
consuming 99% CPU with

Kernel Module  1.0-4496  Wed Jul 16 19:03:09 PDT 2003

Oct 1 21:27:24 connecting kernel: agpgart: Found an AGP 3.5 compliant device
at 0000:00:00.0.
Oct  1 21:27:24 connecting kernel: agpgart: Putting AGP V3 device at
0000:00:00.
0 into 8x mode
Oct  1 21:27:24 connecting kernel: agpgart: Putting AGP V3 device at
0000:01:00.
0 into 8x mode
Oct  1 21:27:25 connecting kernel: SysRq : SAK
Oct  1 21:27:32 connecting kernel: Badness in pci_find_subsys at
drivers/pci/sea
rch.c:132
Oct  1 21:27:32 connecting kernel: Call Trace:
Oct  1 21:27:32 connecting kernel:  [<c01bcbf7>] pci_find_subsys+0xc7/0xd0
Oct  1 21:27:32 connecting kernel:  [<c01bcc2f>] pci_find_device+0x2f/0x40
Oct  1 21:27:32 connecting kernel:  [<c01bcb08>] pci_find_slot+0x28/0x50
Oct  1 21:27:32 connecting kernel:  [<e0d6f699>]
os_pci_init_handle+0x39/0x70 [nvidia]
Oct  1 21:27:32 connecting kernel:  [<e0c442af>] __nvsym00057+0x1f/0x24
[nvidia]

This is on a KT400, Ti4200 with AGP 8x support. 
-- 
Pekka Pietikainen





More information about the fedora-test-list mailing list