Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 25 Jun 2021 13:33:04 +0300
From:   Dan Carpenter <dan.carpenter@oracle.com>
To:     Pablo Neira Ayuso <pablo@netfilter.org>
Cc:     Colin King <colin.king@canonical.com>,
        Jozsef Kadlecsik <kadlec@netfilter.org>,
        Florian Westphal <fw@strlen.de>,
        "David S . Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>,
        netfilter-devel@vger.kernel.org, coreteam@netfilter.org,
        netdev@vger.kernel.org, kernel-janitors@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH][next] netfilter: nf_tables: Fix dereference of null
 pointer flow
Message-ID: <20210625103304.GI2040@kadam>
References: <20210624195718.170796-1-colin.king@canonical.com>
 <20210625095901.GH2040@kadam>
 <20210625102021.GA32352@salvia>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20210625102021.GA32352@salvia>
User-Agent: Mutt/1.9.4 (2018-02-28)
Precedence: bulk

On Fri, Jun 25, 2021 at 12:20:21PM +0200, Pablo Neira Ayuso wrote:
> Hi,
> 
> On Fri, Jun 25, 2021 at 12:59:01PM +0300, Dan Carpenter wrote:
> > Btw, why is there no clean up if nft_table_validate() fails?
> 
> See below.
> 
> > net/netfilter/nf_tables_api.c
> >   3432                                  list_add_tail_rcu(&rule->list, &old_rule->list);
> >   3433                          else
> >   3434                                  list_add_rcu(&rule->list, &chain->rules);
> >   3435                  }
> >   3436          }
> >   3437          kvfree(expr_info);
> >   3438          chain->use++;
> >   3439  
> >   3440          if (flow)
> >   3441                  nft_trans_flow_rule(trans) = flow;
> >   3442  
> >   3443          if (nft_net->validate_state == NFT_VALIDATE_DO)
> >   3444                  return nft_table_validate(net, table);
> >                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > The cleanup for this would be quite involved unfortunately...  Not
> > necessarily something to attempt without being able to test the code.
> 
> At this stage, the transaction has been already registered in the
> list, and the nf_tables_abort() path takes care of undoing what has
> been updated in the preparation phase.
> 

Ah...  Thanks.

regards,
dan carpenter