QMI IPv6 crashes

[Dan Williams]

On Tue, 2014-04-22 at 23:00 +0200, Bjørn Mork wrote:
> Dan Williams <dcbw at redhat.com> writes:
> > On Tue, 2014-04-22 at 22:22 +0200, Bjørn Mork wrote:
> >> Bjørn Mork <bjorn at mork.no> writes:
> >> 
> >> > But I guess I shouldn't let the kernel manage these things, but instead
> >> > test with NM?  It will deal with the RS and neighbour discovery on its
> >> > own, is that right?
> >> 
> >> Struggling a bit with this.. How do I configure NM to start an IPv6 only
> >> connection?  I don't see any gsm.ip-type or similar attribute here:
> >
> > [ipv4]
> > method=disabled
> >
> > [ipv6]
> > method=auto
> >
> > will do the trick,
> 
> That makes sense, and actually was one of the things I tried.
> 
> > assuming you're using the dcbw/wwan-ipv6 branch for
> > NM?
> 
> Right.  Didn't think about looking at the NM branches.  Will rebuild
> now.
> 
> Hmm, that failed in unexpected ways.  Is this me screwing up something:
> 
> make[3]: Entering directory `/usr/local/src/git/networkmanager/src'
> make[3]: *** No rule to make target `platform/wifi/wifi-utils-nl80211.h', needed by `nm-enum-types.h.stamp'.  Stop.

make distclean (or find and remove all .deps directories)
re-autogen
rebuild

due to some movement of the wifi bits in the source tree, the .deps/
directories have stale information, so they need to be removed.

> >  After editing, make sure you "nmcli con reload" since NM by default
> > doesn't watch config files edited directly.
> 
> Thanks for the tip.  Now I don't have to restart NM every time I modify
> a connection :-)

Oh hah, you've never needed to do that!  Previously NM would watch the
files, but when we switched that off we added "nmcli con reload" at the
same time.  Perhaps not publicized as well?

> > To test dual-stack, set the ipv4 method to 'auto' too.
> 
> I'm currently unable to test dual-stack. Don't have access to any APNs
> allowing it.

Ah, ok.  At least confirmation that MM dcbw/icera-ipv6 + NM
dcbw/wwan-ipv6 works in IPv6-only mode would be good.

Dan