netdev - Re: [PATCH 1/1] netfilter: nft_fib: Fix for rpath check with VRF devices

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 30 Sep 2022 16:47:52 +0200
From:   Florian Westphal <fw@...len.de>
To:     Guillaume Nault <gnault@...hat.com>
Cc:     Florian Westphal <fw@...len.de>, netdev@...r.kernel.org,
        Jakub Kicinski <kuba@...nel.org>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Paolo Abeni <pabeni@...hat.com>,
        netfilter-devel@...r.kernel.org, Phil Sutter <phil@....cc>
Subject: Re: [PATCH 1/1] netfilter: nft_fib: Fix for rpath check with VRF
 devices

Guillaume Nault <gnault@...hat.com> wrote:
> On Wed, Sep 28, 2022 at 01:39:08PM +0200, Florian Westphal wrote:
> > diff --git a/net/ipv6/netfilter/nft_fib_ipv6.c b/net/ipv6/netfilter/nft_fib_ipv6.c
> > index 8970d0b4faeb..1d7e520d9966 100644
> > --- a/net/ipv6/netfilter/nft_fib_ipv6.c
> > +++ b/net/ipv6/netfilter/nft_fib_ipv6.c
> > @@ -41,6 +41,9 @@ static int nft_fib6_flowi_init(struct flowi6 *fl6, const struct nft_fib *priv,
> >  	if (ipv6_addr_type(&fl6->daddr) & IPV6_ADDR_LINKLOCAL) {
> >  		lookup_flags |= RT6_LOOKUP_F_IFACE;
> >  		fl6->flowi6_oif = get_ifindex(dev ? dev : pkt->skb->dev);
> > +	} else if ((priv->flags & NFTA_FIB_F_IIF) &&
> > +		   (netif_is_l3_master(dev) || netif_is_l3_slave(dev))) {
> > +		fl6->flowi6_oif = dev->ifindex;
> >  	}
> 
> I'm not very familiar with nft code, but it seems dev can be NULL here,
> so netif_is_l3_master() can dereference a NULL pointer.

No, this should never be NULL, NFTA_FIB_F_IIF is restricted to
input/prerouting chains.

> Shouldn't we test dev in the 'else if' condition?

We could do that in case it makes review easier.