From mboxrd@z Thu Jan 1 00:00:00 1970 From: Martin KaFai Lau Subject: Re: NULL pointer dereference in rt6_get_cookie() Date: Tue, 13 Oct 2015 12:30:16 -0700 Message-ID: <20151013193016.GC68583@kafai-mba.local> References: <20151010132437.GB25926@orbit.nwl.cc> <20151013181443.GB68583@kafai-mba.local> <20151013190941.5BF3C213B6@mail.nwl.cc> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: , Hannes Frederic Sowa , Steffen Klassert , Julian Anastasov Return-path: Received: from mx0b-00082601.pphosted.com ([67.231.153.30]:62347 "EHLO mx0b-00082601.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751594AbbJMTa3 (ORCPT ); Tue, 13 Oct 2015 15:30:29 -0400 Content-Disposition: inline In-Reply-To: <20151013190941.5BF3C213B6@mail.nwl.cc> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, Oct 13, 2015 at 09:10:39PM +0200, Phil Sutter wrote: > Hi Martin, > > On Tue, Oct 13, 2015 at 11:14:43AM -0700, Martin KaFai Lau wrote: > > On Sat, Oct 10, 2015 at 03:24:37PM +0200, Phil Sutter wrote: > > > The conditional at the start of the function evaluates true, since > > > 'rt->rt6i_flags & RTF_PCPU' is non-zero. > > Hi Phil, can you try the following patch and capture the dmesg output > > to confirm the value of rt->rt6i_flags and the rt->dst.flags. > > > > Thanks, > > Martin > > > > --- a/include/net/ip6_fib.h > > +++ b/include/net/ip6_fib.h > > @@ -167,8 +167,15 @@ static inline void rt6_update_expires(struct rt6_info *rt0, int timeout) > > > > static inline u32 rt6_get_cookie(const struct rt6_info *rt) > > { > > - if (rt->rt6i_flags & RTF_PCPU || unlikely(rt->dst.flags & DST_NOCACHE)) > > + if (rt->rt6i_flags & RTF_PCPU || unlikely(rt->dst.flags & DST_NOCACHE)) { > > rt = (struct rt6_info *)(rt->dst.from); > > + if (!rt) > > + pr_err("rt6i_dst:%pI6c/%d rt6i_gateway:%pI6c " > > + "rt6i_flags:%08X dst.flags:%08X\n", > > + &rt->rt6i_dst.addr, rt->rt6i_dst.plen, > > + &rt->rt6i_gateway, rt->rt6i_flags, > > + rt->dst.flags); > > + } > > > > return rt->rt6i_node ? rt->rt6i_node->fn_sernum : 0; > > } > > This code is not sane. Your pr_err() statement tries to dereference the > NULL pointer in question. Are you interested in the originally passed > rt6_info? Good catch. sorry about that.