[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160215195930.GA9001@salvia>
Date: Mon, 15 Feb 2016 20:59:30 +0100
From: Pablo Neira Ayuso <pablo@...filter.org>
To: Zhouyi Zhou <zhouzhouyi@...il.com>
Cc: eric.dumazet@...il.com, kaber@...sh.net, kadlec@...ckhole.kfki.hu,
davem@...emloft.net, netfilter-devel@...r.kernel.org,
coreteam@...filter.org, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, fw@...len.de,
gnomes@...rguk.ukuu.org.uk, sergei.shtylyov@...entembedded.com,
Zhouyi Zhou <yizhouzhou@....ac.cn>
Subject: Re: [PATCH V6] netfilter: h323: avoid potential attack
On Tue, Feb 02, 2016 at 09:40:04PM +0800, Zhouyi Zhou wrote:
> diff --git a/net/netfilter/nf_conntrack_h323_main.c b/net/netfilter/nf_conntrack_h323_main.c
> index 9511af0..8d24c4b 100644
> --- a/net/netfilter/nf_conntrack_h323_main.c
> +++ b/net/netfilter/nf_conntrack_h323_main.c
> @@ -110,6 +110,21 @@ int (*nat_q931_hook) (struct sk_buff *skb,
>
> static DEFINE_SPINLOCK(nf_h323_lock);
> static char *h323_buffer;
> +static int h323_buffer_valid_bytes;
> +
> +static bool h323_buffer_ref_valid(void *p, int len)
I'd rather see you pass a parameter to this function with the
remaining size in the buffer, so we don't need this global variable
h323_buffer_valid_bytes.
You can probably add an initial patch to add a structure to store the
state information so we reduce the function parameter footprint.
struct h323_ct_state {
...
int buflen;
};
So you pass up the h323_ct_state structure to function calls,
something like that.
Thanks.
> +{
> + if ((unsigned long)len > h323_buffer_valid_bytes)
> + return false;
> +
> + if (p + len > (void *)h323_buffer + h323_buffer_valid_bytes)
> + return false;
> +
> + if (p < (void *)h323_buffer)
> + return false;
> +
> + return true;
> +}
>
> static struct nf_conntrack_helper nf_conntrack_helper_h245;
> static struct nf_conntrack_helper nf_conntrack_helper_q931[];
Powered by blists - more mailing lists