Here's a trace ('t' in the debugger) of the startup crash with GENERIC-27
and the Quadra/LC/Performa 630:

_Debugger(0,167c9c,e2952,e2668,1) + 6
_panic(e2668,1,6bbc940,167e22,0) + 40
_trap(0,105,7fe01c)+1e6
_addrerr(6bbc940,f020e,167d5c)+e8
_obio_search(6bbc940,f020e,0,f016a,f020e) + 40
_mapply(167da4,f020e) + 3c
_config_search(b87c8,6bbc940,0,b8764,0) + 56
_obio_attach(6bb980,6bbc940,0,6bbc980,0) + 24
_config_attach(6bbc980,f019e,0,0,6bbc980) + 28c
_mainbus_search(6bbc980,f019e,0,f0176,f019e) + 2e
_mapply(167e7c,f019e) + 3c
_config_search(dfb3c,6bbc980,0,dfb0e,0) + 56
_mainbus_attach(0,6bbc980,0,dfb0e,0) + <oops, forgot to write this part down>
_config_attach(0,f0182,da9bd,0,124) + 28c
_config_rootfound(da9bd,da9bd) + 3a
_configure(dbaf6,cc,cc000,dbae5,1047000) + 24
_cpu_startup(f0f9c,4147,16340,800b8a,f9001000) + 4a6
_main() + 4a
_main() + 4a

Hope this helps.

Also: Allen Briggs' netbsd-20 test kernel (reports itself as GENERIC-20,
also has DAFB and SONIC patches) doesn't work either. Same crash exactly.

Note: GENERIC-26 *does* work on the 630. GENERIC-28 also fails to work, but
this is because of the hardware Cuda driver's immaturity?

Hunch: Maybe changes to support B&W X and dt on DAFB messed up something
needed on the 630's Valkyrie video? (btw, if the ite0 minimal console is
b&w video, why does X and dt's grf device need to be a separate driver?)

- Rolf Braun ... rbraun@geocities.com ... TidalWave on IRC
- Sassy Software: cool stuff for Macintosh & Apple II
- http://www.geocities.com/SiliconValley/Heights/3110/