netdev - Re: [PATCH] net: Do not include padding in TCP GRO checksum

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] [day] [month] [year] [list]

Message-ID: <1384807463.29151.88.camel@bwh-desktop.uk.level5networks.com>
Date:	Mon, 18 Nov 2013 20:44:23 +0000
From:	Ben Hutchings <bhutchings@...arflare.com>
To:	Alexander Duyck <alexander.h.duyck@...el.com>
CC:	<davem@...emloft.net>, <netdev@...r.kernel.org>,
	<edumazet@...gle.com>, <herbert@...dor.apana.org>
Subject: Re: [PATCH] net: Do not include padding in TCP GRO checksum

On Thu, 2013-11-14 at 17:18 -0800, Alexander Duyck wrote:
> In some recent tests where I was generating invalid frames I found that
> the checksum was being treated as valid for certain frames that computed
> the checksum with padding included.  On closer inspection I found the
> issue was that GRO was using the skb->len instead of the length recorded in
> the IP/IPv6 header to determine the number of bytes to checksum.  As such
> padded frames that actually had invalid checksums generated by adding the
> padding to the checksum were being incorrectly tagged as valid.
> 
> This change corrects that by using the tot_len from IPv4 headers and the
> payload_len from IPv6 headers to compute the total number of bytes to be
> included in the checksum.
> 
> Signed-off-by: Alexander Duyck <alexander.h.duyck@...el.com>
> ---
> 
> I haven't had a chance to test this much yet and I am not that familiar
> some of this GRO code so any review of this would be greatly appreciated.
> I will try to get this tested by end-of-day tomorrow to verify it resolves
> the issues I saw with invalid padded frames being marked as valid and
> doesn't introduce any new issues.
> 
>  net/ipv4/tcp_offload.c   |   13 ++++++++-----
>  net/ipv6/tcpv6_offload.c |   11 +++++++----
>  2 files changed, 15 insertions(+), 9 deletions(-)
> 
> diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
> index a2b68a1..3dabb76 100644
> --- a/net/ipv4/tcp_offload.c
> +++ b/net/ipv4/tcp_offload.c
[...]
> @@ -288,11 +291,11 @@ flush:
>  		return NULL;
>  
>  	case CHECKSUM_NONE:
> -		wsum = csum_tcpudp_nofold(iph->saddr, iph->daddr,
> -					  skb_gro_len(skb), IPPROTO_TCP, 0);
> +		wsum = csum_tcpudp_nofold(iph->saddr, iph->daddr, length,
> +					  IPPROTO_TCP, 0);
>  		sum = csum_fold(skb_checksum(skb,
>  					     skb_gro_offset(skb),
> -					     skb_gro_len(skb),
> +					     length,

length may also be > skb_gro_len() which makes this unsafe.

>  					     wsum));
>  		if (sum)
>  			goto flush;
> diff --git a/net/ipv6/tcpv6_offload.c b/net/ipv6/tcpv6_offload.c
> index c1097c7..53fc71d 100644
> --- a/net/ipv6/tcpv6_offload.c
> +++ b/net/ipv6/tcpv6_offload.c
[...]
> @@ -52,11 +56,10 @@ flush:
>  
>  	case CHECKSUM_NONE:
>  		wsum = ~csum_unfold(csum_ipv6_magic(&iph->saddr, &iph->daddr,
> -						    skb_gro_len(skb),
> -						    IPPROTO_TCP, 0));
> +						    length, IPPROTO_TCP, 0));
>  		sum = csum_fold(skb_checksum(skb,
>  					     skb_gro_offset(skb),
> -					     skb_gro_len(skb),
> +					     length,
>  					     wsum));
>  		if (sum)
>  			goto flush;

Same problem here.

inet_gro_receive() and ipv6_gro_receive() already check whether the IP
length field is consistent with skb_gro_len(), and set
NAPI_GRO_CB(skb)->flush if it is not.  I wonder whether it would make
sense to skip calling the transport-layer gro_receive function in that
case.

Ben.

-- 
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html