Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 28 Mar 2001 06:41:15 +1000
From:      Peter Jeremy <peter.jeremy@alcatel.com.au>
To:        Robert Watson <rwatson@FreeBSD.ORG>
Cc:        John Baldwin <jhb@FreeBSD.ORG>, cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG
Subject:   Re: cvs commit: src/sbin/reboot reboot.c
Message-ID:  <20010328064115.Q26138@gsmx07.alcatel.com.au>
In-Reply-To: <Pine.NEB.3.96L.1010326204002.81313C-100000@fledge.watson.org>; from rwatson@FreeBSD.ORG on Mon, Mar 26, 2001 at 08:42:04PM -0500
References:  <XFMail.010326151018.jhb@FreeBSD.org> <Pine.NEB.3.96L.1010326204002.81313C-100000@fledge.watson.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 2001-Mar-26 20:42:04 -0500, Robert Watson <rwatson@FreeBSD.ORG> wrote:
>
>On Mon, 26 Mar 2001, John Baldwin wrote:
>
>> Since reboot will still need the code for the fallback case, this
>> doesn't avoid any of the code duplication.  This would just be more work
>> than what we have now w/o any gain that I can see. :-P If you want to
>> share the code, stick it in a library. 
>
>Well, my feeling was that part of the potential to "stick" init comes from
>the complex shutdown procedure.  Maybe we need to just assert more
>strongly that neither "reboot" nor "halt" provides a "gentle" shutdown,
>and stick with the current signalling technique (not attempt to run any
>more complex shutdown code), perhaps spacing the time periods a little
>more.  I.e., the status quo is right, modulo a bit of tweaking.

If init gets stuck as a result of the process killing approach[1],
then halt/reboot are likely to fall into the same knot.  The logical
extension is that one command does the gentle shutdown approach
and the other command just calls reboot(2) without attempting to
kill processes.  I dislike situations where two different processes
implement basically the same functionality in different ways.

[1] I have seen commercial Unices where processes can get into a
    state where attempting to kill them will wedge the process
    issuing the kill(2).

Peter

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010328064115.Q26138>