Fri, 29 Sep 2000 16:51:53 -0600
Module name: linux
Changes by: dhd 00/09/29 16:51:53
Duh! Duh duh duh. I was wondering how emulation would work given
that delayed traps happen on the instruction after the faulting one.
Now it makes perfect sense, and I'm no longer as angry at HP as I was
half an hour ago :)
Obviously we do *not* want to advance the IAOQ, because the idea is to
emulate the faulted insn then rerun the ftest (or whatever)
instruction that caught the trap with the updated status word to get
the proper results.
Now awk passes its tests, hooray.