linux-kernel - Re: [PATCH] xfrm: Fix xfrm_state refcnt leak in xfrm

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]

Date:   Thu, 23 Apr 2020 08:01:39 +0200
From:   Steffen Klassert <steffen.klassert@...unet.com>
To:     Xiyu Yang <xiyuyang19@...an.edu.cn>
CC:     Herbert Xu <herbert@...dor.apana.org.au>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>, <netdev@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>, <yuanxzhang@...an.edu.cn>,
        <kjlu@....edu>, Xin Tan <tanxin.ctf@...il.com>
Subject: Re: [PATCH] xfrm: Fix xfrm_state refcnt leak in xfrm_input()

On Thu, Apr 23, 2020 at 01:19:20PM +0800, Xiyu Yang wrote:
> xfrm_input() invokes xfrm_state_lookup(), which returns a reference of
> the specified xfrm_state object to "x" with increased refcnt and then
> "x" is escaped to "sp->xvec[]".
> 
> When xfrm_input() encounters error, it calls kfree_skb() to free the
> "skb" memory. Since "sp" comes from one of "skb" fields, this "free"
> behavior causes "sp" becomes invalid, so the refcount for its field
> should be decreased to keep refcount balanced before kfree_skb() calls.
> 
> The reference counting issue happens in several exception handling paths
> of xfrm_input(). When those error scenarios occur such as skb_dst()
> fails, the function forgets to decrease the refcnt increased by
> xfrm_state_lookup() and directly calls kfree_skb(), causing a refcnt
> leak.

kfree_skb() drops these refcounts already, why should we do that here
too?