From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: bridge/netfilter: regression in 2.6.39.1 Date: Mon, 06 Jun 2011 15:09:22 +0200 Message-ID: <1307365762.3098.11.camel@edumazet-laptop> References: <4DE93422.3070000@ahsoftware.de> <20110606111507.GA1000@hmsreliant.think-freely.org> <4DECBEA3.6070408@ahsoftware.de> <1307362358.3098.6.camel@edumazet-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Neil Horman , linux-kernel@vger.kernel.org, David Miller , Herbert Xu , netdev@vger.kernel.org To: Alexander Holler Return-path: In-Reply-To: <1307362358.3098.6.camel@edumazet-laptop> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Le lundi 06 juin 2011 =C3=A0 14:12 +0200, Eric Dumazet a =C3=A9crit : > Le lundi 06 juin 2011 =C3=A0 13:48 +0200, Alexander Holler a =C3=A9cr= it : > > Am 06.06.2011 13:15, schrieb Neil Horman: > > > On Fri, Jun 03, 2011 at 09:21:06PM +0200, Alexander Holler wrote: > > >> Hello, > > >> > > >> I'm getting a oops in the bridge code in br_change_mtu() with > > >> 2.6.39.1. The patch below seems to fix that. > > >> > > >> I'm not sure about the usage of dst_cow_metrics_generic() in > > >> fake_dst_ops, but after having a quick look at it seems to be ok= to > > >> use that here. > > >> > > >> Regards, > > >> > > >> Alexander > > >> > > > How did the flags of the dst entry on which we're callnig dst_ent= ry_write_ptr > > > wind up getting the READ_ONLY flag set on it? I don't see how w= e'er falling > > > into that clause in which we call cow_metrics when we call dst_me= tric_set. It > > > seems like that flag is set erroneously. perhaps we should just = update > > > fake_rtable.dst to have the correct flags? > > > Neil > >=20 > > It is set by that change: > >=20 > > -------- > > @@ -124,7 +128,7 @@ void br_netfilter_rtable_init(struct net_bridge= *br) > > atomic_set(&rt->dst.__refcnt, 1); > > rt->dst.dev =3D br->dev; > > rt->dst.path =3D &rt->dst; > > - dst_metric_set(&rt->dst, RTAX_MTU, 1500); > > + dst_init_metrics(&rt->dst, br_dst_default_metrics, true); > > rt->dst.flags =3D DST_NOXFRM; > > rt->dst.ops =3D &fake_dst_ops; > > } > > -------- > >=20 > > The true in dst_init_metrics() is responsible for that flag. > >=20 >=20 > You are aware this change fixed an oops ? >=20 > read_only in this context means : In case this must be written, we ma= ke > a COW first > (allocate a piece of memory, copy the source in it before applying an= y > change) >=20 > It would be nice you send us the stack trace, so that we can have a c= lue > of whats going on. >=20 Alexander, you should take a look at : git show 0972ddb2 To get an idea of how to deal with this problem=20 (See how Held Bernhard included a backtrace to help us make a diagnostic) We dont want to even allocate a piece of memory to copy br_dst_default_metric for a fake dst.