public inbox for gentoo-dev@lists.gentoo.org
 help / color / mirror / Atom feed
* Re: [gentoo-dev] Start enhancement
       [not found] <004001c17bce$99f982f0$6400a8c0@server>
@ 2001-12-03 13:54 ` Joshua Pierre
  2001-12-03 15:12   ` Vitaly Kushneriuk
  2001-12-03 16:25 ` jano
  1 sibling, 1 reply; 4+ messages in thread
From: Joshua Pierre @ 2001-12-03 13:54 UTC (permalink / raw
  To: gentoo-dev

On Mon, Dec 03, 2001 at 08:46:21AM +0100 or thereabouts, Sebastian Werner wrote:
> I have a good dream last night. Some days ago I find the tool from
> Microsoft to optimize the boot time of their new system (winxp). The
> idea is to run some services to system-init parallel. Or in other words:
> Why we must wait for samba to start xdm. Couldn't we use the
> multitaskting experience of linux in our init-scripts. So we can start
> Linux in near the half time I think. Good or bad idea?

Not sure if it is possible but maybe a manual way to set the order in which the init scripts start.

However, a few of the init scripts might have whacky depends making things wait uncessarily.

Who knows but a way to manually change the boot order would be great but as I said above I am not sure if it is possible or not.

Cheers,

-- 
Joshua Pierre
Developer & Release Technician
Themes.Org -- Open Source Interface Enhancement


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [gentoo-dev] Start enhancement
  2001-12-03 13:54 ` [gentoo-dev] Start enhancement Joshua Pierre
@ 2001-12-03 15:12   ` Vitaly Kushneriuk
  0 siblings, 0 replies; 4+ messages in thread
From: Vitaly Kushneriuk @ 2001-12-03 15:12 UTC (permalink / raw
  To: gentoo-dev

It is definitly possible.
Might require writing a little c proggy: 
   waitpids pid1 pid2 ....
which returns when ANY of the processes terminates. 
The rest can be implemented in sh. 
Or the whole process can be implemented in Python ( thought I don't like
the idea of boot process in python)

The following is a simplified python/sh algo :-). All marking of who is
started + dependencies etc. is done, like in present version, through
files and dirs in /dev/shm
-----------------------------------------------------------

function can_start():
    if depend($1) == ''
       return TRUE;# no dependencies
    for service in depend($1)
       if not_started(service):
          return FALSE;
    return TRUE;




while not all services started:
    for every service that not started:
       if can_start(service):
	   service&
	   PIDS="$PIDS $!"
    PIDS = `waitpids $PIDS` # will wait for at least one process to
                            # terminate, so we MAY be able to start more
                            # services
-----------------------------------------------------------

It is more complicated than that, but not much.
exit satatus, error handling, service start output, circular dependancy
and timeouts are some of the issues that need to be solved.

I might implement it when I'll have some more time then right now (busy
working)



On Mon, 2001-12-03 at 15:54, Joshua Pierre wrote:
> On Mon, Dec 03, 2001 at 08:46:21AM +0100 or thereabouts, Sebastian Werner wrote:
> > I have a good dream last night. Some days ago I find the tool from
> > Microsoft to optimize the boot time of their new system (winxp). The
> > idea is to run some services to system-init parallel. Or in other words:
> > Why we must wait for samba to start xdm. Couldn't we use the
> > multitaskting experience of linux in our init-scripts. So we can start
> > Linux in near the half time I think. Good or bad idea?
> 
> Not sure if it is possible but maybe a manual way to set the order in which the init scripts start.
> 
> However, a few of the init scripts might have whacky depends making things wait uncessarily.
> 
> Who knows but a way to manually change the boot order would be great but as I said above I am not sure if it is possible or not.
> 
> Cheers,
> 
> -- 
> Joshua Pierre
> Developer & Release Technician
> Themes.Org -- Open Source Interface Enhancement
> _______________________________________________
> gentoo-dev mailing list
> gentoo-dev@gentoo.org
> http://lists.gentoo.org/mailman/listinfo/gentoo-dev



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [gentoo-dev] Start enhancement
       [not found] <004001c17bce$99f982f0$6400a8c0@server>
  2001-12-03 13:54 ` [gentoo-dev] Start enhancement Joshua Pierre
@ 2001-12-03 16:25 ` jano
  2001-12-03 16:33   ` Daniel Robbins
  1 sibling, 1 reply; 4+ messages in thread
From: jano @ 2001-12-03 16:25 UTC (permalink / raw
  To: gentoo-dev

I concur.  I wonder what sort of drawbacks there would be to doing this?

> Hey,
>
> I have a good dream last night. Some days ago I find the tool from
> Microsoft to optimize the boot time of their new system (winxp). The
> idea is to run some services to system-init parallel. Or in other
> words: Why we must wait for samba to start xdm. Couldn't we use the
> multitaskting experience of linux in our init-scripts. So we can start
> Linux in near the half time I think. Good or bad idea?
>
> Thanks in advance.
>
> Sebastian Werner
>





^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [gentoo-dev] Start enhancement
  2001-12-03 16:25 ` jano
@ 2001-12-03 16:33   ` Daniel Robbins
  0 siblings, 0 replies; 4+ messages in thread
From: Daniel Robbins @ 2001-12-03 16:33 UTC (permalink / raw
  To: gentoo-dev

On Mon, Dec 03, 2001 at 08:25:48AM -0800, jano wrote:
> > I have a good dream last night. Some days ago I find the tool from
> > Microsoft to optimize the boot time of their new system (winxp). The
> > idea is to run some services to system-init parallel. Or in other
> > words: Why we must wait for samba to start xdm. Couldn't we use the
> > multitaskting experience of linux in our init-scripts. So we can start
> > Linux in near the half time I think. Good or bad idea?
> 
> I concur.  I wonder what sort of drawbacks there would be to doing this?

No real drawbacks and should be possible to add due to our dep-based 
initscripts.  Right now, we are focusing on more fundamental areas, though.
In the future, we'll add support for this.  The disadvantage would be
that without a higher-level script output mechanism, the output from
the scripts would get interleaved line by line.  Also, let's take a step
back and think about a system that boots 1 second faster.  Big whoop.
Is this where we want to be focusing our precious development time?

Best Regards,

-- 
Daniel Robbins                                  <drobbins@gentoo.org>
Chief Architect/President                       http://www.gentoo.org 
Gentoo Technologies, Inc.


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2001-12-03 16:33 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <004001c17bce$99f982f0$6400a8c0@server>
2001-12-03 13:54 ` [gentoo-dev] Start enhancement Joshua Pierre
2001-12-03 15:12   ` Vitaly Kushneriuk
2001-12-03 16:25 ` jano
2001-12-03 16:33   ` Daniel Robbins

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox