I checked in fixes for a number of problems this morning:

 - sti driver works again (on B180 at least)
 - spurious SIGILLs from xmpyu instructions are gone
 - set correct iisq* when delivering signals

with all these applied I was able to run a native "build.sh -j 8"
for more than 12 hours, which is way more than I've ever been able
to do before.

I tried booting a 715/100 and I got the same problem that some posted about
a few weeks back:  HPMC trap after the "mem0" line of autoconf.
this is due to the sti driver, the "devtype = bus_space_read_1(...)" line
in sti_sgc_probe() is where the trap happens.  this card revision also
appears to use the pagezero info for the ROM.  there must be a better way
to determine this, but for now I just added this revision to this exception
list.

the boot-from-disk problem (init gets infinite SIGSEGVs) is still there,
and the hack from before (adding padding to struct scsipi_xfer) didn't help.
I'll look into this next.

-Chuck