Created attachment 11474 [details]
xorg log

The bios is not detected, this is the initial problem that needs solving.

Created attachment 11523 [details]
Xorg.0.log

Today, a freshly compiled nouveau_drv.so gives me a new crash.

I uploaded a new log from the latest nouveau_drv.so. The new crash is:

/usr/lib/xorg/modules/drivers//nouveau_drv.so(nv30UpdateArbitrationSettings+0x23) [0x2ac7262f1823]
4: /usr/lib/xorg/modules/drivers//nouveau_drv.so(NVCalcStateExt+0x41c) [0x2ac7262f277c]
5: /usr/lib/xorg/modules/drivers//nouveau_drv.so(NVDACInit+0x44f) [0x2ac7262ea60f]
6: /usr/lib/xorg/modules/drivers//nouveau_drv.so [0x2ac7262ee5d5]
7: /usr/lib/xorg/modules/drivers//nouveau_drv.so [0x2ac7262f002c]
8: /usr/lib/xorg/modules/drivers//nouveau_drv.so [0x2ac7262f050a]
9: /usr/bin/X11/X(AddScreen+0x222) [0x439522]
10: /usr/bin/X11/X(InitOutput+0x268) [0x468a48]
11: /usr/bin/X11/X(main+0x275) [0x439d25]
12: /lib/libc.so.6(__libc_start_main+0xf4) [0x2ac72413bb44]
13: /usr/bin/X11/X(FontFileCompleteXLFD+0x229) [0x439209]

it seems to involve the nvGetClocks call in line 694 in nv_hw.c.

Digging into nvGetClocks, the crash happens at line 191 where *MClk is computed:

      *MClk = ((N * NB * pNv->CrystalFreqKHz) / (M * MB)) >> P;

In my case, pll had the value 0x0b01, causing MB to evaluate to zero, thus causing a divide by zero error.

Created attachment 11535 [details] [review]
Fix BIOS location on g80s

This patch, mostly lifted from nv fixes the missing bios problem on my card

Created attachment 11536 [details]
xorglog after patch applied

Latest nouveau updates from nv makes this bug obsolete (I now get X up on the screen, although somwhat garbled)

I had no valid modes by default on my Thinkpad R61 (neither 1680x1050 internal or standard external 1920x1080p was working by default). The only way to start X up was to feed it manually generated modelines.