[LMH] Lisp timing issue solved

Daniel Seagraves dseagrav@lunar-tokyo.net
Mon, 5 Dec 2005 04:38:08 -0600 (CST)


I solved the Lisp timing bug. Fixing this enables lisp to keep proper
time, boot with a network card, and causes the mouse to work.

The clue was that setting the retrace-interval timer to it's correct value
caused a crash. On further investigation, lisp was writing the Interval 0
count while the interval timer was in latched-read mode, THEN crashing. In
my code that wasn't anticipated and ended up being a no-op. Lisp is trying
to count the refresh rate. I made writing in latched-read mode operate
like writing in MSB/LSB mode, and lisp works now. Lisp is trying to check
the refresh rate.

The rate of the 100MS/10MS/100NS RTC clocks are checked also, failing the
checks doesn't cause lisp to crash but causes it to disable most of the SIB,
which creates the odd behaviour we were seeing.

I still haven't solved the keyboard repeating bug. Timing doesn't seem to
affect it.