>> swap_pager_clean: clean of page 22cd000 failed
>> swap_pager_clean: clean of page 231c000 failed
>> swap_pager_clean: clean of page 1348000 failed
>> swap_pager_clean: clean of page 21e3000 failed

> This has happened to me amid SCSI problems, albeit on different
> ports.

Well, if it's caused by SCSI problems here, they're software - the very
same machine works just fine if I boot a late-May kernel.

Incidentally, I have a bit more info on this.  I left the "make build"
running overnight, and when I got in this morning, it was scrolling
swap_pager_clean messages as fast as the console could scroll.  I
nerve-pinched it and brought it back up with the old kernel, and
checked the logfile from the "make build".  It had gotten as far as
running ctags in /usr/src/lib/libc, as part of the "cd lib; make depend
&& make && make install" stage.  Normally "make build" would have
completely finished in that much time.

I'm going to compare code between the latest-known-working kernel
source I have and the current source, see if I can see what could be
wrong.  _Something_ certainly is.

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu