netdev - Re: [Patch net] gre: refetch erspan header from skb->data after pskb_may

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]

Message-ID: <20191206104212.GE27144@netronome.com>
Date:   Fri, 6 Dec 2019 11:42:13 +0100
From:   Simon Horman <simon.horman@...ronome.com>
To:     Cong Wang <xiyou.wangcong@...il.com>
Cc:     netdev@...r.kernel.org,
        Lorenzo Bianconi <lorenzo.bianconi@...hat.com>
Subject: Re: [Patch net] gre: refetch erspan header from skb->data after
 pskb_may_pull()

Hi Cong,

On Thu, Dec 05, 2019 at 07:39:02PM -0800, Cong Wang wrote:
> After pskb_may_pull() we should always refetch the header
> pointers from the skb->data in case it got reallocated.
> 
> In gre_parse_header(), the erspan header is still fetched
> from the 'options' pointer which is fetched before
> pskb_may_pull().
> 
> Found this during code review of a KMSAN bug report.
> 
> Fixes: cb73ee40b1b3 ("net: ip_gre: use erspan key field for tunnel lookup")
> Cc: Lorenzo Bianconi <lorenzo.bianconi@...hat.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@...il.com>
> ---
>  net/ipv4/gre_demux.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/net/ipv4/gre_demux.c b/net/ipv4/gre_demux.c
> index 44bfeecac33e..5fd6e8ed02b5 100644
> --- a/net/ipv4/gre_demux.c
> +++ b/net/ipv4/gre_demux.c
> @@ -127,7 +127,7 @@ int gre_parse_header(struct sk_buff *skb, struct tnl_ptk_info *tpi,
>  		if (!pskb_may_pull(skb, nhs + hdr_len + sizeof(*ershdr)))
>  			return -EINVAL;
>  
> -		ershdr = (struct erspan_base_hdr *)options;
> +		ershdr = (struct erspan_base_hdr *)(skb->data + nhs + hdr_len);

It seems to me that in the case of WCCPv2 hdr_len will be 4 bytes longer
than where options would be advanced to. Is that a problem here?

>  		tpi->key = cpu_to_be32(get_session_id(ershdr));
>  	}
>  
> -- 
> 2.21.0
>