Re: udevsynthesize as a udevstart+coldplug replacement

[Christian Zoz <zoz@suse.de>]

On Thu, Aug 25, Marco d'Itri wrote:
> On Aug 25, Christian Zoz <zoz@suse.de> wrote:
> 
> > > multiple init scripts would depend on the udev-done (or something like
> > > this) init script, which will not terminale until udevcontrol will.
> > > This way other init scripts would be able to continue working without
> > > needing to wait for udev.
> > > Or maybe it's not worth the effort, because they would need to use
> > > /some/ device node anyway.
> > Why not make the init scripts wait individually for whatever they need?
> How?

That is a bigger change in init scripts. Each init script needs a list
of requirements, similar to the required services now. The scripts
could be called like it is done up to now and then each script has to
check its requirements and fail after reasonable timeout. Or there is
an instance that checks requirements of different services and calls
them as soon as all requirements are fulfilled.

There is a third way: We currently initialize devices and interfaces
(class devices, block devices) via events. Why not extend this and
trigger events when a certain state of system intialisation is
reached.

This all are still very raw ideas. But i don't like the current way of
system initialisation, that tries to serialize everything, not so
much. It makes to much trouble with hotplug.

-- 

ciao, christian

睡眠不足はいい仕事の敵だ。


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel