Date: Thu, 03 Jul 1997 16:00:02 +1000 From: Stephen McKay <syssgm@dtir.qld.gov.au> To: freebsd-current@freebsd.org Subject: Re: ppp & HUP. Message-ID: <199707030600.QAA18716@ogre.dtir.qld.gov.au> In-Reply-To: <33BB1473.33590565@whistle.com> from Julian Elischer at "Thu, 03 Jul 1997 02:54:43 %2B0000" References: <Pine.BSF.3.96.970703060227.268B-100000@nagual.pp.ru> <33BB1473.33590565@whistle.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[Somehow I missed the start of this. Strange. Am I losing mail?] On Thursday, 3rd July 1997, Julian Elischer wrote: >=?KOI8-R?B?4c7E0sXKIP7F0s7P1w==?= wrote: >> >> On Wed, 2 Jul 1997, Brian Somers wrote: >> >> > > > reboot(8) will send SIGTERMs to the processes, but not SIGHUPs. >> > >> > death() in init.c :( >> >> I start thinking that sending HUP from init is a bug because: >> 1) It increase disk activity just before shutdown since all daemons >> re-read their configs. >> 2) It cause redials for redial-able software as uucp (and ppp some time >> ago). >> >> If nobody will explain why this HUP is neded (hanging shells perfectly >> killed by SIGKILL), I'll remove HUP sending from init. >> >I have processes that close down assuming they have > 5 or 6 seconds to do so when they receive HUP. >it's traditional. I don't like changing such traditions.. Well, back in the good old days (4.2 BSD), it was SIGTERM (from shutdown), then SIGKILL (from halt or reboot or init). If init ever wanted you dead, it just used SIGKILL. Your tradition must be a young tradition. :-) >From my earliest UNIX days, it has been SIGTERM == "Please exit soon" and SIGKILL == "Die now!". SIGHUP was for "Oops, you have no terminal now", and usually there was no point in carrying on. But that was before the days of POSIX, and their version of job control. If they've stuffed that up, someone should set me straight. It seems to me that the real problem here is that a lot of daemons now abuse all sorts of signals for a bunch of naff things. If anything, we should be removing all that crap. There are other, better IPC mechanisms. On the other hand, I can't see any reason for init to send SIGHUP, so I won't object if that is removed. Stephen.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199707030600.QAA18716>