Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 28 Sep 2009 22:42:57 +0200 (CEST)
From:      Alexander Best <alexbestms@math.uni-muenster.de>
To:        Kevin Oberman <oberman@es.net>
Cc:        Dag-Erling =?iso-8859-1?Q?Sm=F8rgrav?= <des@des.no>, freebsd-current@FreeBSD.org, Andriy Gapon <avg@icyb.net.ua>
Subject:   Re: ipv6 related warnings
Message-ID:  <permail-200909282042571e86ffa800001c4f-a_best01@message-id.uni-muenster.de>
In-Reply-To: <20090928164051.BBDC51CC13@ptavv.es.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Kevin Oberman schrieb am 2009-09-28:
> > Date: Mon, 28 Sep 2009 15:24:02 +0200 (CEST)
> > From: Alexander Best <alexbestms@math.uni-muenster.de>
> > Sender: owner-freebsd-current@freebsd.org

> > Dag-Erling Sm=F8rgrav schrieb am 2009-09-28:
> > > Andriy Gapon <avg@icyb.net.ua> writes:
> > > > Dag-Erling Sm=F8rgrav <des@des.no> writes:
> > > > > No, you can build a *kernel* without IPv6, just not userland.
> > > > > Parts
> > > > > of userland will respect WITHOUT_INET6, others won't, and
> > > > > others
> > > > > yet
> > > > > will build but not run correctly.
> > > > So shouldn't those disrespecting parts of userland be fixed?

> > > Why?  In most (userland) cases, it takes more work to write code
> > > that
> > > supports IPv4 but not IPv6 than to write code that supports both.

> > > > Besides, returning to the original post, those messages weren't
> > > > there
> > > > just a few days ago, this is a new addition.

> > > ...if by "just a few days" you mean "two weeks".  If this really
> > > is
> > > "a
> > > new addition", it dates back to one of hrs@'s commits on
> > > 2009-09-13.
> > > However, the most likely cause of the OP's problem is incorrect
> > > use
> > > of
> > > mergemaster.  If he did indeed run "mergemaster -p" and
> > > "mergemaster
> > > -i", but nothing more, his /etc is completely out of sync.

> > mergemaster -p and mergemaster -i aren't sufficient to sync /etc
> > with
> > /usr/src/etc? but that's what UPDATING recommends.

> mergemaster -i (or, for sanity, -iUF) is adequate as long as the
> proper
> things are done when mergemaster notes a difference to be merged. But
> there is nothing "wrong" with using just '-i'. The real question is
> whether you install, delete, or merge any of the changes mergemaster
> reports.

> If you always install, you will probably break localizations. If you
> always skip or delete, you will end up with a /etc that is not going
> to
> work right.

> mergemaster is a wonderful tool as anyone who ran FreeBSD 2.2 or 3
> can
> tell you, but it is not a panacea. You still need to take enough time
> to
> properly evaluate each merge and DTRT.

> (And, to be clear, I have no reason to assume that you did not DTRT.)

just to clear things up. i just did `mergemaster -i` to see which files fro=
m
/etc and /usr/src/etc i'm getting asked about (which means they differ).
here's the list:

--- /boot/device.hints  2009-08-03 18:52:42.000000000 +0200
+++ ./boot/device.hints 2009-09-28 22:21:46.000000000 +0200

--- /etc/bluetooth/hcsecd.conf  2009-08-12 15:19:10.000000000 +0200
+++ ./etc/bluetooth/hcsecd.conf 2009-09-28 22:21:45.000000000 +0200

--- /etc/bluetooth/hosts        2009-05-03 22:03:57.000000000 +0200
+++ ./etc/bluetooth/hosts       2009-09-28 22:21:45.000000000 +0200

--- /etc/mail/freebsd.mc        2009-09-12 11:17:23.000000000 +0200
+++ ./etc/mail/freebsd.mc       2009-09-28 22:21:46.000000000 +0200

--- /etc/mail/freebsd.cf        2009-09-12 11:17:28.000000000 +0200
+++ ./etc/mail/freebsd.cf       2009-09-28 22:21:46.000000000 +0200

--- /etc/mail/freebsd.submit.cf 2009-09-12 11:08:55.000000000 +0200
+++ ./etc/mail/freebsd.submit.cf        2009-09-28 22:21:46.000000000 +0200

--- /etc/mail/sendmail.cf       2009-09-27 16:35:24.000000000 +0200
+++ ./etc/mail/sendmail.cf      2009-09-28 22:21:46.000000000 +0200

--- /etc/mail/submit.cf 2009-09-27 16:35:24.000000000 +0200
+++ ./etc/mail/submit.cf        2009-09-28 22:21:46.000000000 +0200

--- /etc/crontab        2009-09-24 18:13:24.000000000 +0200
+++ ./etc/crontab       2009-09-28 22:21:45.000000000 +0200

--- /etc/devd.conf      2009-09-12 02:38:27.000000000 +0200
+++ ./etc/devd.conf     2009-09-28 22:21:45.000000000 +0200

--- /etc/ftpusers       2009-06-13 22:35:46.000000000 +0200
+++ ./etc/ftpusers      2009-09-28 22:21:45.000000000 +0200

--- /etc/group  2009-09-08 11:23:58.000000000 +0200
+++ ./etc/group 2009-09-28 22:21:45.000000000 +0200

--- /etc/hosts  2009-08-16 16:59:41.000000000 +0200
+++ ./etc/hosts 2009-09-28 22:21:45.000000000 +0200

--- /etc/login.conf     2009-07-01 12:49:55.000000000 +0200
+++ ./etc/login.conf    2009-09-28 22:21:45.000000000 +0200

--- /etc/motd   2009-09-27 16:39:30.000000000 +0200
+++ ./etc/motd  2009-09-28 22:21:45.000000000 +0200

--- /etc/shells 2009-09-27 17:19:31.000000000 +0200
+++ ./etc/shells        2009-09-28 22:21:45.000000000 +0200

--- /etc/sysctl.conf    2009-08-16 13:45:47.000000000 +0200
+++ ./etc/sysctl.conf   2009-09-28 22:21:45.000000000 +0200

--- /etc/ttys   2009-06-16 15:14:43.000000000 +0200
+++ ./etc/ttys  2009-09-28 22:21:45.000000000 +0200

--- /etc/master.passwd  2009-09-08 11:23:58.000000000 +0200
+++ ./etc/master.passwd 2009-09-28 22:21:45.000000000 +0200

so to me this means that apart from these files /etc and /usr/src/etc are i=
n
sync and i'm quite sure none of these files have anything to do with ipv6
apart from sysctl.conf, but as you can see:

--- /etc/sysctl.conf    2009-08-16 13:45:47.000000000 +0200
+++ /usr/src/etc/sysctl.conf    2009-09-23 21:55:11.000000000 +0200
@@ -7,12 +7,3 @@
 # Uncomment this to prevent users from seeing information about processes
 that
 # are being run under another UID.
 #security.bsd.see_other_uids=3D0
-#
-kern.ipc.shmmax=3D67108864
-kern.ipc.shmall=3D32768
-vfs.usermount=3D1
-net.inet.tcp.blackhole=3D1
-net.inet.udp.blackhole=3D1
-hw.snd.vpc_autoreset=3D0
-kern.ipc.shm_allow_removed=3D1
-compat.linux.osrelease=3D2.6.16

i have nothing ipv6 related in that file.

as another reference i'm posting all the options i have in my /etc/src.conf=
:

WITHOUT_ACCT=3Dtrue
WITHOUT_AMD=3Dtrue
WITHOUT_APM=3Dtrue
WITHOUT_ATM=3Dtrue
WITHOUT_AUDIT=3Dtrue
WITHOUT_BIND=3Dtrue
WITHOUT_BSNMP=3Dtrue
WITHOUT_CDDL=3Dtrue
WITHOUT_CTM=3Dtrue
WITHOUT_FLOPPY=3Dtrue
WITHOUT_FREEBSD_UPDATE=3Dtrue
WITHOUT_GAMES=3Dtrue
WITHOUT_GPIB=3Dtrue
WITH_IDEA=3Dtrue
WITHOUT_INET6=3Dtrue
WITHOUT_IPFILTER=3Dtrue
WITHOUT_IPFW=3Dtrue
WITHOUT_IPX=3Dtrue
WITHOUT_JAIL=3Dtrue
WITHOUT_LPR=3Dtrue
WITHOUT_MAILWRAPPER=3Dtrue
WITHOUT_NCP=3Dtrue
WITHOUT_NDIS=3Dtrue
WITHOUT_NETCAT=3Dtrue
WITHOUT_NIS=3Dtrue
WITHOUT_NLS=3Dtrue
WITHOUT_NLS_CATALOGS=3Dtrue
WITHOUT_NS_CACHING=3Dtrue
WITHOUT_PAM=3Dtrue
WITHOUT_PF=3Dtrue
WITHOUT_PORTSNAP=3Dtrue
WITHOUT_PPP=3Dtrue
WITHOUT_QUOTAS=3Dtrue
WITHOUT_RCMDS=3Dtrue
WITHOUT_RCS=3Dtrue
WITHOUT_ROUTED=3Dtrue
WITHOUT_SLIP=3Dtrue
WITHOUT_TCSH=3Dtrue
WITHOUT_TELNET=3Dtrue
WITHOUT_WPA_SUPPLICANT_EAPOL=3Dtrue
WITHOUT_CDDL=3Dtrue

so...as far as i can see there's no way of building an ipv6-unaware world a=
ny
longer. it seems like the WITHOUT_INET6 option has become obsolete.

if this is the case the option should completely be removed from the whole
source and defining it should no longer trigger any ifdef's. also it should=
 be
removed from `man src.conf` and a note should be added to UPDATING telling
users that WITHOUT_INET6 won't work any longer.

if however WITHOUT_INET6 as an option should remain supported and infact le=
t
the user build an ipv6-unaware world this should be fixed imo.

cheers.
alex



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?permail-200909282042571e86ffa800001c4f-a_best01>