netdev - Re: [PATCH] ipv4: tcp: security_sk_alloc() needed for unicast

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: <1344547743.31104.582.camel@edumazet-glaptop>
Date:	Thu, 09 Aug 2012 23:29:03 +0200
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Eric Paris <eparis@...isplace.org>
Cc:	Paul Moore <paul@...l-moore.com>,
	David Miller <davem@...emloft.net>,
	Casey Schaufler <casey@...aufler-ca.com>,
	John Stultz <johnstul@...ibm.com>,
	"Serge E. Hallyn" <serge@...lyn.com>,
	lkml <linux-kernel@...r.kernel.org>,
	James Morris <james.l.morris@...cle.com>,
	selinux@...ho.nsa.gov, john.johansen@...onical.com,
	LSM <linux-security-module@...r.kernel.org>,
	netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH] ipv4: tcp: security_sk_alloc() needed for unicast_sock

On Thu, 2012-08-09 at 16:06 -0400, Eric Paris wrote:
> NAK.
> 
> I personally think commit be9f4a44e7d41cee should be reverted until it
> is fixed.  Let me explain what all I believe it broke and how.
> 

Suggesting to revert this commit while we have known working fixes is a
bit of strange reaction.

I understand you are upset, but I believe we tried to fix it.

> Old callchain of the creation of the 'equivalent' socket previous to
> the patch in question just for reference:
> 
>     inet_ctl_sock_create
>       sock_create_kern
>         __sock_create
>           pf->create (inet_create)
>             sk_alloc
>               sk_prot_alloc
>                 security_sk_alloc()
> 
> 
> This WAS working properly.  All of it. 

Nobody denies it. But acknowledge my patch uncovered a fundamental
issue.

What kind of 'security module' can decide to let RST packets being sent
or not, on a global scale ? (one socket for the whole machine)

smack_sk_alloc_security() uses smk_of_current() : What can be the
meaning of smk_of_current() in the context of 'kernel' sockets...

Your patch tries to maintain this status quo.

In fact I suggest the following one liner patch, unless you can really
demonstrate what can be the meaning of providing a fake socket for these
packets.

This mess only happened because ip_append_data()/ip_push_pending_frames()
are so complex and use an underlying socket.

But this socket should not be ever used outside of its scope.

diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 76dde25..ec410e0 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -1536,6 +1536,7 @@ void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr,
 			  arg->csumoffset) = csum_fold(csum_add(nskb->csum,
 								arg->csum));
 		nskb->ip_summed = CHECKSUM_NONE;
+		skb_orphan(nskb);
 		skb_set_queue_mapping(nskb, skb_get_queue_mapping(skb));
 		ip_push_pending_frames(sk, &fl4);
 	}

--
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