It would seem impossible and unreasonable for uugetty to crash the
computer cold like that. It also seems that the serial drivers should
be well tested and debugged by now. I am also running a Comtrol
RocketPort and it seems to be flawless. It uses no IRQs and is at
its default IO address. It is a 4 port, RJ-45 connector, ISA type.
This doesn't happen every time someone logs in and out from the modem.
Only sometimes and the sometimes is still unpredictable. Before upgrading
to a 2.0 kernel I occasionally experienced the hung uugetty, but HUP'ing
it never crashed the system. I am running CND-1.0 -> Picasso -> updates
-> 2.0-updates -> 2.0.10 kernel.
uugetty is from Red Hat's getty_ps-2.0.7h-1 rpm. I don't see how to get
the version out of the uugetty binary.
A clip from /var/log/messages is below. ...If you look closely at
the reported date on the messages you will see that my bios clock
is off a bit, but I don't expect that is related and has been fixed.
Also, where is the kernel looking for the map file? System.map
is in / with vmlinuz and has a matching date/time. What is it
looking for and what is the effect of it not finding it? /proc/ksyms
seems to be complete.
Aug 19 21:02:52 falcon sudo: tyson : TTY=ttyp7 ; PWD=/home/tyson ; USER=root ; COMMAND=/bin/kill -HUP 945
Aug 19 21:02:52 falcon uugetty: exiting on HANGUP signal
Aug 19 20:53:20 falcon syslogd 1.3-0: restart.
Aug 19 20:53:20 falcon kernel: klogd 1.3-0, log source = /proc/kmsg started.
Aug 19 20:53:21 falcon kernel: Cannot find map file.