DragonFly bugs List (threaded) for 2007-06
Re: [issue647] rough-draft VKERNEL host-initiated shutdown patch
:In any case,
:hoping to clean up / resubmit this weekend, more than likely with a
:second signal handler to take advantage of Matt's new reboot code.
Definitely! Having a properly working shutdown signal would be good.
In fact, it might be easiest to just set it up as a mailbox signal and
monitor it in the clock interrupt, then create a kernel thread to initiate
the actual shutdown when the signal is detected. Or something like that.
I noticed another thing, which is when you run the vkernel from
a shell script, and you interrupt it with ^\ + ^C (that sequence
will enter ddb, then kill the vkernel), it currently causes the vkernel
to do its shutdown/reboot thing, but also kills the shell script and
disconnects the terminal so you wind up getting a shell prompt with the
vkernel still running in the background. Its very odd and annoying
I think there is some sort of termios / tty session magic we have
to do to make the tty interrupt signals only go to the vkernel,
but I'm not sure what it is.