I decided to give generic-27 a shot, to see if it would help my PB145's
SCSI from hell.  I was a little bit surprised by the results.

adb bus subsystem
unknown OS trap 0x109f, no trap address available
panic: mrg_aline_super()
Stopped at _Debugger+0x6: unlk a6

I take it the 100-series ADB hardware support isn't in the source tree
yet, right?  (Or if it is, why was it doing anything with MRG?)  I'm
putting it through the stress test via serial console now (multiple large
makes and make cleans in a near-useless old directory of outdated versions
of my talker sources) and it appears to be behaving itself _much_ better.
(multiple fscks, 0 errors!)  If there's a kernel in the works by anybody
with those ncrscsi (yeah, NCRSCSI... sbc explodes on my PB and leaves an
awful mess behind, soo) patches and with the PB145 hardware support,
please let me know.  Finally, the PB145 isn't misbehaving!  :-)


Later,

David A. Gatwood
davagatw@mars.utm.edu
dgatwood@globegate.utm.edu
marsmail@globegate.utm.edu (temporary)