Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!spool.mu.edu!uwm.edu!cs.utexas.edu!howland.erols.net!newsfeed.internetmci.com!news-in2.uu.net!news.artisoft.com!usenet
From: Terry Lambert <terry@lambert.org>
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: What is a zombie?
Date: Mon, 30 Sep 1996 12:21:16 -0700
Organization: Me
Lines: 61
Message-ID: <32501DAC.2BEEACF9@lambert.org>
References: <52a5dt$qef@vixen.cso.uiuc.edu> <3249D8F7.2EA39DDA@lambert.org>
<52jtk1$8s@uriah.heep.sax.de> <324F080E.4BFCA0F@lambert.org> <52nr0k$baa@uriah.heep.sax.de>
NNTP-Posting-Host: hecate.artisoft.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 2.01 (X11; I; Linux 1.1.76 i486)
J Wunsch wrote:
] Terry Lambert <terry@lambert.org> wrote:
]
] > ] Implementing SA_NOCLDWAIT for sa_flags in sigaction(2) would
] > ] be the way to go, but nobody has done this so far in FreeBSD.
] > ] So your only chance is to actually install the d*mn signal
] > ] handler, and have it
] > ] fetch the exit status of the child.
] >
] > Then FreeBSD is broken. Someone with commit priveledges to the
] > tree should commit the (trivial) fix.
]
] It's not `broken' per se, since the API allows for various options.
] They are called options since they are optional. Would you call
] SVR4's implementation `broken' since it doesn't implement SunOS'
] SA_USERTRAMP option? Btw., SA_SIGINFO isn't implemented in SVR4
] either. Btw^2, it has already been discussed that an implementation
] where setting SIGCHLD to SIG_IGN has the same effect like SA_NOCLDWAIT
] must be considered broken: SIGCHLD is already ignored, ignoring it
] twice should not change anything in the behaviour.
I would call any implementation that leaves zombies with no one
to reap tham, "in error".
If the default is "discard signal", then there should be no reap
required.
Typically, most other implementations draw a distinction between
SIG_DFL and SIG_IGN, and so with SIG_DFL create children that need
to be reaped, and with SIG_IGN create children which don't need
to be reaped.
] Anyway, please use send-pr to submit your patches. ;-)
The "send-pr" system is ineffective.
Using "send-pr" to submit patches is the FreeBSD equivalent
of getting in line at the Motor Vehicle Department: the line
is too long, and there are too few people servicing the people
in line. The end result is that there are send-pr'ed patches
which have been (verifiably, via www.freebsd.org) sitting "in
the queue" for a year or more waiting for someone with commit
priviledges to service them. Just like by the time you get
out of the DMV after renewing your license, you have aged and
lost all your hair and have to get back into line to get a new
picture taken so you will match.
Besides, the fix is so trivial it is beneath my level of interest
to pound it through the constipated send-pr pipe. I would rather
beat my head against the brick wall of trying to get my platform
portability and FS architecture changes rolled in, instead. My
head is only so sturdy, so I must pick my battles.
Regards,
Terry Lambert
terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.