Patchwork [hardy,lucid,oneiric,precise,quantal,CVE,1/1] net: fix divide by zero in tcp algorithm illinois

login
register
mail settings
Submitter Luis Henriques
Date Nov. 12, 2012, 12:04 p.m.
Message ID <1352721866-16193-1-git-send-email-luis.henriques@canonical.com>
Download mbox | patch
Permalink /patch/198378/
State New
Headers show

Comments

Luis Henriques - Nov. 12, 2012, 12:04 p.m.
From: Jesper Dangaard Brouer <brouer@redhat.com>

BugLink: http://bugs.launchpad.net/bugs/1077091

Reading TCP stats when using TCP Illinois congestion control algorithm
can cause a divide by zero kernel oops.

The division by zero occur in tcp_illinois_info() at:
 do_div(t, ca->cnt_rtt);
where ca->cnt_rtt can become zero (when rtt_reset is called)

Steps to Reproduce:
 1. Register tcp_illinois:
     # sysctl -w net.ipv4.tcp_congestion_control=illinois
 2. Monitor internal TCP information via command "ss -i"
     # watch -d ss -i
 3. Establish new TCP conn to machine

Either it fails at the initial conn, or else it needs to wait
for a loss or a reset.

This is only related to reading stats.  The function avg_delay() also
performs the same divide, but is guarded with a (ca->cnt_rtt > 0) at its
calling point in update_params().  Thus, simply fix tcp_illinois_info().

Function tcp_illinois_info() / get_info() is called without
socket lock.  Thus, eliminate any race condition on ca->cnt_rtt
by using a local stack variable.  Simply reuse info.tcpv_rttcnt,
as its already set to ca->cnt_rtt.
Function avg_delay() is not affected by this race condition, as
its called with the socket lock.

Cc: Petr Matousek <pmatouse@redhat.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

(cherry picked from commit 8f363b77ee4fbf7c3bbcf5ec2c5ca482d396d664)
CVE-2012-4565
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
---
 net/ipv4/tcp_illinois.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
Andy Whitcroft - Nov. 12, 2012, 12:25 p.m.
On Mon, Nov 12, 2012 at 12:04:26PM +0000, Luis Henriques wrote:
> From: Jesper Dangaard Brouer <brouer@redhat.com>
> 
> BugLink: http://bugs.launchpad.net/bugs/1077091
> 
> Reading TCP stats when using TCP Illinois congestion control algorithm
> can cause a divide by zero kernel oops.
> 
> The division by zero occur in tcp_illinois_info() at:
>  do_div(t, ca->cnt_rtt);
> where ca->cnt_rtt can become zero (when rtt_reset is called)
> 
> Steps to Reproduce:
>  1. Register tcp_illinois:
>      # sysctl -w net.ipv4.tcp_congestion_control=illinois
>  2. Monitor internal TCP information via command "ss -i"
>      # watch -d ss -i
>  3. Establish new TCP conn to machine
> 
> Either it fails at the initial conn, or else it needs to wait
> for a loss or a reset.
> 
> This is only related to reading stats.  The function avg_delay() also
> performs the same divide, but is guarded with a (ca->cnt_rtt > 0) at its
> calling point in update_params().  Thus, simply fix tcp_illinois_info().
> 
> Function tcp_illinois_info() / get_info() is called without
> socket lock.  Thus, eliminate any race condition on ca->cnt_rtt
> by using a local stack variable.  Simply reuse info.tcpv_rttcnt,
> as its already set to ca->cnt_rtt.
> Function avg_delay() is not affected by this race condition, as
> its called with the socket lock.
> 
> Cc: Petr Matousek <pmatouse@redhat.com>
> Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
> Acked-by: Eric Dumazet <edumazet@google.com>
> Acked-by: Stephen Hemminger <shemminger@vyatta.com>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> 
> (cherry picked from commit 8f363b77ee4fbf7c3bbcf5ec2c5ca482d396d664)
> CVE-2012-4565
> Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
> ---
>  net/ipv4/tcp_illinois.c |    8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/net/ipv4/tcp_illinois.c b/net/ipv4/tcp_illinois.c
> index 5aa5f54..f2c2ff0 100644
> --- a/net/ipv4/tcp_illinois.c
> +++ b/net/ipv4/tcp_illinois.c
> @@ -314,11 +314,13 @@ static void tcp_illinois_info(struct sock *sk, u32 ext,
>  			.tcpv_rttcnt = ca->cnt_rtt,
>  			.tcpv_minrtt = ca->base_rtt,
>  		};
> -		u64 t = ca->sum_rtt;
>  
> -		do_div(t, ca->cnt_rtt);
> -		info.tcpv_rtt = t;
> +		if (info.tcpv_rttcnt > 0) {
> +			u64 t = ca->sum_rtt;
>  
> +			do_div(t, info.tcpv_rttcnt);
> +			info.tcpv_rtt = t;
> +		}
>  		nla_put(skb, INET_DIAG_VEGASINFO, sizeof(info), &info);
>  	}
>  }

That is a little subtle, using the copy already made in 'info' as a
parameter to nla_put is a little unexpected.  Otherwise it seems to do
what is expected, to be race tollerant and, matches the upstream commit.

Acked-by: Andy Whitcroft <apw@canonical.com>

-apw
Herton Ronaldo Krzesinski - Nov. 12, 2012, 12:43 p.m.

Andy Whitcroft - Nov. 12, 2012, 2:16 p.m.
Applied to hardy, lucid, oneiric, precise, and quantal.

-apw

Patch

diff --git a/net/ipv4/tcp_illinois.c b/net/ipv4/tcp_illinois.c
index 5aa5f54..f2c2ff0 100644
--- a/net/ipv4/tcp_illinois.c
+++ b/net/ipv4/tcp_illinois.c
@@ -314,11 +314,13 @@  static void tcp_illinois_info(struct sock *sk, u32 ext,
 			.tcpv_rttcnt = ca->cnt_rtt,
 			.tcpv_minrtt = ca->base_rtt,
 		};
-		u64 t = ca->sum_rtt;
 
-		do_div(t, ca->cnt_rtt);
-		info.tcpv_rtt = t;
+		if (info.tcpv_rttcnt > 0) {
+			u64 t = ca->sum_rtt;
 
+			do_div(t, info.tcpv_rttcnt);
+			info.tcpv_rtt = t;
+		}
 		nla_put(skb, INET_DIAG_VEGASINFO, sizeof(info), &info);
 	}
 }