net.ipv6.conf.eth1.autoconf = 0
net.ipv6.conf.eth0.autoconf = 0
net.ipv6.conf.default.autoconf = 0
net.ipv6.conf.all.autoconf = 0
net.ipv6.conf.lo.autoconf = 0

in /etc/sysctl.conf is normally enough to disable ipv6

but apparently it doesn't work... What's so difficult in making ipv6 as module. I'm usually compiling my own kernel anyway but the stock kernel should be as flexible as possible...

#> netstat -pant
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 ::ffff:127.0.0.1:6880 :::* LISTEN 1700/java
tcp 0 0 :::6881 :::* LISTEN 1700/java
tcp 0 0 ::ffff:127.0.0.1:45100 :::* LISTEN 1700/java
tcp 0 0 :::80 :::* LISTEN 1604/httpd

>> I think (judging from reading the forums)

Could you be more precise?

Does this mean we should make ipv4 a module too then?

Judd,

http://bbs.archlinux.org/viewtopic.php?t=17093&highlight=disable+ipv6
http://bbs.archlinux.org/viewtopic.php?t=17022&highlight=disable+ipv6
http://bbs.archlinux.org/viewtopic.php?t=13001&highlight=disable+ipv6

It seems there is also problem with the speed of Firefox and ipv6, more threads are about this.

Dale, don't be ironic.. how many people really use ipv6? ipv4 is essential on the other hand....
If you want to know the opinion of the community on this make a poll, im sure most would like to see ipv6 as a module.

Funny that we have to fix a userspace problem with a kernel based solution. Maybe you ought to bark at the firefox developers?

Firefox is ONLY one of the minor problems (and it's fixable by turning off ipv6 in about:config anyway), but the main reason is that most people don't use. I opened a thread asking people what they think about it and most are sharing my views.

http://bbs.archlinux.org/viewtopic.php?t=20269

Please don't do this. I posted in pixel's forum post about why I think it's good that IPv6 is NOT optional, and I'd think a bit less of the devs if they made the choice to move a large group of people (Archers) away from supporting IPv6.

It's NOT about supporting ipv6, it's about flexibility of the kernel. How many people REALLY need ipv6 nowadays? I'm not saying to erase ipv6 completely from the kernl, just make it modular so it would be real easy to disable it if not needed. that's all

and about 80% of forums members who casted their votes share my view.

Calm down.

Besides, it's a broken concept. Fixing a userspace (Firefox) issue with a kernel change.

To me, that's stupid. File a bug at the firefox bug tracker and tell them to improve how they handle ipv6.

iphitus, it has NOTHING to do with Firefox really. why you keep coming up with that? NOONE really use ipv6 so why force people to use it?

It being there, does not force anybody to use it.

The reason for people complaining about it, is because Firefox has had problems with it. If that werent the case, the flammable topic would never have been raised. It does go back to firefox.

It's somewhat ironic that you say it has nothing to do with firefox, when one of your arguments, is exactly that -- Firefox. Take a look at your own comment, fifth down, three links, firefox.

no, it does not. I said only one of the minor reasons is apparently connected to Firefox but that's it, and i also explicitly written you can disable it in config:about. How many times i need to repeat it before you get it? :)

"It being there, does not force anybody to use it."

You can't disable it, so that's forcing people to use it. I ask, what's so hard in making ipv6 modular? it's one switch in config. It still can be enabled by deafult if devs really want it, but this time it would be easy to disable it... think about it. I'm not asking to erase it completely from the kernel, what im askin for is only to make it more flexible, because right now you CANT disable it...

"NOONE really use ipv6 so why force people to use it?"

You are wrong, for example I _AM_ using IPv6. You can use it too, just register to some tunelbroker (SixXS.net, etc.).
IPv6 is the future and it should be promoted, not given away from kernel as module. IPv6 is standard and should be always in kernel.

And if there is bug in Firefox, then you should talk to Firefox developers, not Arch Linux developers.

I have to side with iphitus here. First off, X tries to default ti ip6 sockets, and spits out some errors if not found. IP6 is supported and used by alot of ISPs, though not around here. Saying "no one uses IP6" sounds like an easilly falsifiable claim. In fact, I've used a IP6 from hexago myself, just messing with it.

Here's the break down:
* Why disable IP6?
"It causes problems with some programs" - see iphitus' arguement - those programs are broken, not IP6.
* Make it modular, everything else is.
If this were a valid argument, it would also follow that ip4 becomes modular. And any other networking layer that *can* be made modular. Want to use a loopback device? Whoops, gotta modprobe for that. Domain sockets? Sorry, modprobe again.
* But if it's modular, you can still use it!
Again, this goes back to the previous 2 arguments. Regardless of actual usage, *why* does it need to be modular? Because of userspace issues? Fix those apps.

phrakture, ok blame me for putting out this information about firefox issues. i repeat ONCE AGAIN tho - ipv6 should be modularized because its usage is very rare.. to say at least, not because of some userspace applications. Make it a module and make all parties happy :)

Just a short one, the discussion has really become stupid in some parts.
All pro-ipv6 folks here said that there is no reason for making it modular. I don't see a reason for not making it modular:

When I have ipv6 as a module, it autoloads everytime I boot unless I explicitly prevent that. I don't know what application did this (I suspect apache or openssh).

There are some people who want to use ipv6, they can load the module if even necessary. There are people who do not care about ipv6, and there are people who do not want to have ipv6 in their kernels for whatever reason (yes, I really don't care about the reason right now).
Why only make two of those groups happy if you can make all of them happy? I really do not see the point in having something in the kernel that only a small percentage of the Arch users use. It is like saying "let's compile driver XYZ in, some people have that device".

So, please, tell me a good reason for not having it as a module? And please don't tell me that this really bothers anyone at all:

_XSERVTransSocketOpenCOTSServer: Unable to open socket for inet6
_XSERVTransOpen: transport open failed for inet6/architect:0
_XSERVTransMakeAllCOTSServerListeners: failed to open listener for inet6

And if it does, why don't you fix it instead of "solving a userspace problem with a kernel solution"?

added as module in .2 kernel