re: mostly working SMP again

[matthew green <mrg%eterna.com.au@localhost>]

      > i'd like folks to make the body of sparc/sparc/intr.c:strayintr()
      > be empty and boot test kernel,
      
      It looks this is required only on using zs console.
      It works without stray interrupts if console is cgsix and keyboard,
      so some more locking is required around zscngetc()/zscnputc() functions?
      (though my tree has some local changes around zs)
   
   
   this looks reasonable to test out.  i'll see how i go.


this doesn't seem to help.

i've noticed that there's a comment in zscnputc()'s guts about always
using delay(2) to avoid stry intrs.  i wonder if this is happening..

i think i've notice alpha port sends console output via IPI's to one
cpu... or something.



.mrg.