Re: RTNL: assertion failed at net/ipv6/addrconf.c (1699)

[Hannes Frederic Sowa <hannes@stressinduktion.org>]

Hi Cong,

On Tue, Sep 2, 2014, at 18:50, Cong Wang wrote:
> On Fri, Aug 29, 2014 at 6:51 PM, Hannes Frederic Sowa
> <hannes@stressinduktion.org> wrote:
> >
> > Also rtnl_lock and rcu_read_lock compose in that order, so we don't need
> > to change dev_get_by_flags, but as this is the only user it sure is
> > possible. RCU locked version is just easier composeable, so I wouldn't
> > touch that if needed in future, just also take rcu lock as before.
> 
> There is no point to keep RCU read lock if we have rtnl lock,
> I don't know why you don't want to change dev_get_by_flags(),
> it is pretty easy to do since it only has one caller.

I definitely don't have a problem cleaning this up in net-next. I wanted
a minimal patch for stable because I didn't check history where and when
additional users of dev_get_by_flags_rcu were removed.

> Even if you really need RCU in future, you are always welcome
> to bring it back when you do, sorry we should never be blocked by
> code NOT merged yet.
> 
> >
> > Also we should move ASSERT_RTNL checks from addrconf_join_solict to
> > ipv6_dev_mc_inc/dec.
> >
> 
> Make it another patch.

It is just one logical change, moving ASSERT_RTNLs to places where they
better catch invalid callstacks.

Bye,
Hannes