Re: Kernel 2.6, NVIDIA and 460GX
Hello David and others,
> If you want to test this, I appended the NV patch that I'm using. The
> patch is for kernel v2.6.10-rc3 and if you're using an older kernel,
> you may have to undo the s/pci_find_class/pci_get_class/ changes. The
> important change would be the one around line 403 where a new call to
> pci_map_single() appears. There are a couple of places like that
> later on. If you compare the patch with what you're using, it should
> be pretty obvious whether or not this will make a difference.
>
> --david
Nope, the patch didn't help (BTW, is it complete? It seems it ends
abruptly in the Debian IA-64 list). Here's what I've recorded when
starting a graphical session and then invoking glxinfo:
nvidia: module license 'NVIDIA' taints kernel.
ACPI: PCI interrupt 0000:03:00.0[A] -> GSI 55 (level, low) -> IRQ 55
0: nvidia: loading NVIDIA Linux ia64 NVIDIA Kernel Module 1.0-5347 Wed
Mar 10
14:25:56 PST 2004
agpgart: Found an AGP 0.2 compliant device at 0000:04:14.1.
agpgart: Putting AGP V2 device at 0000:03:00.0 into 0x mode
agpgart: Putting AGP V2 device at 0000:04:14.1 into 0x mode
Badness in pci_find_subsys at drivers/pci/search.c:167
Call Trace:
[<a00000010001acc0>] show_stack+0x80/0xa0
sp=e00000000477f910 bsp=e0000000047795e0
[<a00000010022ff40>] pci_find_subsys+0x2e0/0x300
sp=e00000000477fae0 bsp=e000000004779588
[<a00000010022f990>] pci_find_slot+0x30/0xc0
sp=e00000000477fae0 bsp=e000000004779560
[<a0000002011c8420>] os_pci_init_handle+0x40/0xc0 [nvidia]
sp=e00000000477fae0 bsp=e000000004779520
[<a000000200d35950>] _nv000904rm+0x50/0x80 [nvidia]
sp=e00000000477fae0 bsp=e0000000047794d8
[<a000000201146910>] _nv000711rm+0xb70/0xd00 [nvidia]
sp=e00000000477fae0 bsp=e000000004779488
[<a0000002011455e0>] _nv000704rm+0x80/0xa0 [nvidia]
sp=e00000000477fb20 bsp=e000000004779468
[<a000000200f5c0d0>] _nv002984rm+0x3d0/0x420 [nvidia]
sp=e00000000477fb20 bsp=e0000000047793d0
[<a000000200f57260>] _nv002963rm+0x1580/0x21c0 [nvidia]
sp=e00000000477fb50 bsp=e000000004779340
[<a0000002010b5480>] _nv001134rm+0x6a0/0xc80 [nvidia]
sp=e00000000477fba0 bsp=e000000004779230
[<a000000200d3e1f0>] _nv005317rm+0x150/0x1c0 [nvidia]
sp=e00000000477fbc0 bsp=e0000000047791e8
[<a000000200d43300>] rm_isr_bh+0x20/0x40 [nvidia]
sp=e00000000477fbc0 bsp=e0000000047791c8
[<a0000001000a5680>] tasklet_action+0x200/0x340
sp=e00000000477fbc0 bsp=e000000004779178
[<a0000001000a4d60>] __do_softirq+0x1c0/0x1e0
sp=e00000000477fbc0 bsp=e0000000047790f8
[<a0000001000a4e00>] do_softirq+0x80/0xe0
sp=e00000000477fbc0 bsp=e000000004779098
[<a0000001000198a0>] ia64_handle_irq+0x1c0/0x200
sp=e00000000477fbc0 bsp=e000000004779050
[<a000000100012360>] ia64_leave_kernel+0x0/0x270
sp=e00000000477fbc0 bsp=e000000004779050
[<a00000010001a030>] ia64_pal_call_static+0x90/0xc0
sp=e00000000477fd90 bsp=e000000004779000
[<a00000010001b740>] default_idle+0xe0/0x180
sp=e00000000477fd90 bsp=e000000004778fa8
[<a00000010001b940>] cpu_idle+0x160/0x420
sp=e00000000477fe30 bsp=e000000004778f18
[<a00000010005aca0>] start_secondary+0x80/0xa0
sp=e00000000477fe30 bsp=e000000004778f00
[<a000000100008580>] _start+0x260/0x290
sp=e00000000477fe30 bsp=e000000004778f00
The call trace is repeated several time with various memory adresses and
register values, but the symbols are the same.
nVIDIA kernel 1.0-5336 exhibits the same callstack. FYI, I don't think
the "Putting device... in 0x mode" is the problem. I get the same
messages with a 2.6.6 kernel and the nVIDIA drivers run fine.
Reply to: