tech-userlevel archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Making powerd=YES default

On 01.08.2011 14:36, Martin Husemann wrote:
> On Mon, Aug 01, 2011 at 12:54:24PM +0200, Jean-Yves Migeon wrote:
>> What do you want to achieve?
> I want a way to not run powerd (even on i386) and still make postinstall
> quiet.

The sentence contradicts itself: postinstall cannot guess what the user
wants, it can only suggest defaults. Adding powerd checks to the
disabled items list nullifies the chance to fix the requirement on upgrade.

BTW, there's no way for a user to specify that a service shall remain
disabled within rc(8) upon update. It's not rc(8) concern anyway.

There are multiple ways to fix my requirement. Sadly, this needs to add
MD hooks somewhere: rc.d/powerd, etcupdate/postinstall, rc.local,
whatever. Or even a pure-kernel solution, thereby circumventing all the
powerd/sysmon code (if you want to go the Linux way).

While rc.conf(5) is parsed as a shell script, it's somewhat
controversial to add conditions within (remember the discussion about
Turing complete configuration files).

I don't really care what solution gets accepted in the end, I need one
that can command kernel to suspend, and currently, powerd is the only
one, as Xen power events are managed via sysmon. Note that suspend isn't
idempotent under Xen, and can only be submitted once. If powerd(8) isn't
there to capture the event, it's lost forever.

So now, I suggest you propose an alternative, so that we can make some
progress. Thanks.

Jean-Yves Migeon

Home | Main Index | Thread Index | Old Index