On Mon, 11 Jun 2001, Greg A. Woods wrote:

> I'm not entirely opposed to having stand-alone kernel-only reboot
> trigger support in the console, just so long as it asks me nicely
> whether I really mean it or not and will go back to normal operation if
> I say "no" (i.e. just like BSD/OS does).  I just don't think it's really
> that important an option given that 99.999% of the times it's needed it
> can be provided by DDB, and I think it would have to be a kernel
> compile-time _option_ anyway, even on i386.

Using DDB seems overkill here. While it provides the required
functionality, I'm not entirely convinced that providing access to site
staff or whoever to reboot the machine warrants allowing them to read and
write kernel memory at will. Or that compiling in all of DDB just for this
feature is worth the size penalty. OK, I'm not sure how big DDB is, but I
imagine it's bigger than a console reboot hook.

While personally I find that there seems something wrong about the idea of
having a key sequence to reboot the machine, I have to admit that I've
found it useful in FreeBSD on occasions, and I can see it being useful on
one or two of my NetBSD machines.

As we don't seem to be able to reach consesus here, it seems like the
obvious thing to suggest would be:

* An option (off by default) to enable a console trigger for kernel reboot
* An option (on by default) to ask for confirmation a la BSD/OS
* An option (not sure about the default here) to halt instead of reboot

That way nobody loses. I get to keep most of my NetBSD machines free of
the Ctrl-Alt-Delete evilness, while the ones that would benefit from it
will be rebootable without confirmation. Those that want to protect
against accidental rebooting get to do so, by default.

I think the only person who loses is the one pressing Ctrl-Alt-Del to log
on at an xdm screen only to find that the computer stops responding as the
reboot question is printed to the console and so doesn't appear because X
is running.

-- 
Gillette - the best a man can forget