netdev - Re: [PATCH net] udp: fix a skb extensions leak

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 for Android: free password hash cracker in your pocket

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

Date:   Mon, 30 Mar 2020 15:45:31 +0200
From:   Steffen Klassert <steffen.klassert@...unet.com>
To:     Florian Westphal <fw@...len.de>
CC:     Xin Long <lucien.xin@...il.com>,
        network dev <netdev@...r.kernel.org>, <davem@...emloft.net>,
        Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net] udp: fix a skb extensions leak

On Mon, Mar 30, 2020 at 03:27:59PM +0200, Florian Westphal wrote:
> Xin Long <lucien.xin@...il.com> wrote:
> > On udp rx path udp_rcv_segment() may do segment where the frag skbs
> > will get the header copied from the head skb in skb_segment_list()
> > by calling __copy_skb_header(), which could overwrite the frag skbs'
> > extensions by __skb_ext_copy() and cause a leak.
> > 
> > This issue was found after loading esp_offload where a sec path ext
> > is set in the skb.
> > 
> > On udp tx gso path, it works well as the frag skbs' extensions are
> > not set. So this issue should be fixed on udp's rx path only and
> > release the frag skbs' extensions before going to do segment.
> > Reported-by: Xiumei Mu <xmu@...hat.com>
> > Fixes: cf329aa42b66 ("udp: cope with UDP GRO packet misdirection")
> 
> Hmm, I suspect this bug came in via
> 3a1296a38d0cf62bffb9a03c585cbd5dbf15d596 , net: Support GRO/GSO fraglist chaining.

I overlooked the explicit mentioning of skb_segment_list()
in the commit message. Fraglist GRO is disabled by default,
this means that it was enabled explicitely. So yes, the bug
came via that commit.

> 
> I suspect correct fix is:
> 
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index 621b4479fee1..7e29590482ce 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -3668,6 +3668,7 @@ struct sk_buff *skb_segment_list(struct sk_buff *skb,
> 
>                 skb_push(nskb, -skb_network_offset(nskb) + offset);
> 
> +               skb_release_head_state(nskb);
>                  __copy_skb_header(nskb, skb);
> 
>                 skb_headers_offset_update(nskb, skb_headroom(nskb) - skb_headroom(skb));
> 
> AFAICS we not only leak reference of extensions, but also skb->dst and skb->_nfct.

Would be nice if we would not need to drop the resources
just to add them back again in the next line. But it is ok
as a quick fix for the bug.