netdev - Re: Fw: Linux tcp_xmit_retransmit_queue use after free on 4.8-rc1 / master]

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]

Date:	Tue, 16 Aug 2016 11:39:08 -0700
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Greg KH <gregkh@...uxfoundation.org>
Cc:	netdev@...r.kernel.org, Marco Grassi <marco.gra@...il.com>,
	Yuchung Cheng <ycheng@...gle.com>,
	Neal Cardwell <ncardwell@...gle.com>
Subject: Re: Fw: Linux tcp_xmit_retransmit_queue use after free on 4.8-rc1 /
 master]

On Tue, 2016-08-16 at 08:39 -0700, Eric Dumazet wrote:
> On Tue, 2016-08-16 at 12:45 +0200, Greg KH wrote:
> > For some reason Marco's emails can't make it to netdev, so I'm
> > forwarding this on.  Please cc: him on responses.
> 
> Thanks for the report Greg and Marco.
> 
> My first guess is this is caused by 
> 
> d41a69f1d390 tcp: make tcp_sendmsg() aware of socket backlog
> 
> And a combination of funky sendmsg() flags (like FastOpen)
> 
> I will look at this problem today.
> 

No, above commit was innocent ;)

It looks like the bug is very old, and following patch would fix it.
I will submit it formally after few tests.


diff --git a/include/net/tcp.h b/include/net/tcp.h
index c00e7d51bb18..7717302cab91 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -1523,6 +1523,8 @@ static inline void tcp_check_send_head(struct sock *sk, struct sk_buff *skb_unli
 {
 	if (sk->sk_send_head == skb_unlinked)
 		sk->sk_send_head = NULL;
+	if (tcp_sk(sk)->highest_sack == skb_unlinked)
+		tcp_sk(sk)->highest_sack = NULL;
 }
 
 static inline void tcp_init_send_head(struct sock *sk)