[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180706145548.kirmwesxx5nviwyj@lakrids.cambridge.arm.com>
Date: Fri, 6 Jul 2018 15:55:49 +0100
From: Mark Rutland <mark.rutland@....com>
To: David Ahern <dsahern@...il.com>
Cc: linux-kernel@...r.kernel.org,
Alexey Kuznetsov <kuznet@....inr.ac.ru>,
"David S . Miller" <davem@...emloft.net>,
Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
netdev@...r.kernel.org
Subject: Re: [PATCH] ipv4: fib: avoid NULL dereference
On Fri, Jul 06, 2018 at 08:49:14AM -0600, David Ahern wrote:
> On 7/6/18 8:28 AM, Mark Rutland wrote:
> > diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
> > index 5bc0c89e81e4..8d98c8162554 100644
> > --- a/net/ipv4/fib_trie.c
> > +++ b/net/ipv4/fib_trie.c
> > @@ -497,11 +497,11 @@ static void tnode_free(struct key_vector *tn)
> > struct callback_head *head = &tn_info(tn)->rcu;
> >
> > while (head) {
> > + tn = container_of(head, struct tnode, rcu)->kv;
> > +
> > head = head->next;
> > tnode_free_size += TNODE_SIZE(1ul << tn->bits);
> > node_free(tn);
>
> you are changing tn before the above 2 calls are made on the tn passed in
As tn_info(tn) is container_of(kv, struct tnode, tn), the end result is
the same tn value, so this is not a problem.
Thanks,
Mark.
Powered by blists - more mailing lists