mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Eric Dumazet <edumazet@google.com>,
	Neal Cardwell <ncardwell@google.com>,
	Yuchung Cheng <ycheng@google.com>,
	Stephen Hemminger <stephen@networkplumber.org>,
	"David S. Miller" <davem@davemloft.net>
Subject: [ 04/19] tcp: cubic: fix overflow error in bictcp_update()
Date: Thu, 12 Sep 2013 10:26:13 -0700	[thread overview]
Message-ID: <20130912172524.646274753@linuxfoundation.org> (raw)
In-Reply-To: <20130912172524.135620246@linuxfoundation.org>

3.0-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <edumazet@google.com>

[ Upstream commit 2ed0edf9090bf4afa2c6fc4f38575a85a80d4b20 ]

commit 17a6e9f1aa9 ("tcp_cubic: fix clock dependency") added an
overflow error in bictcp_update() in following code :

/* change the unit from HZ to bictcp_HZ */
t = ((tcp_time_stamp + msecs_to_jiffies(ca->delay_min>>3) -
      ca->epoch_start) << BICTCP_HZ) / HZ;

Because msecs_to_jiffies() being unsigned long, compiler does
implicit type promotion.

We really want to constrain (tcp_time_stamp - ca->epoch_start)
to a signed 32bit value, or else 't' has unexpected high values.

This bugs triggers an increase of retransmit rates ~24 days after
boot [1], as the high order bit of tcp_time_stamp flips.

[1] for hosts with HZ=1000

Big thanks to Van Jacobson for spotting this problem.

Diagnosed-by: Van Jacobson <vanj@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 net/ipv4/tcp_cubic.c |   10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

--- a/net/ipv4/tcp_cubic.c
+++ b/net/ipv4/tcp_cubic.c
@@ -204,8 +204,8 @@ static u32 cubic_root(u64 a)
  */
 static inline void bictcp_update(struct bictcp *ca, u32 cwnd)
 {
-	u64 offs;
-	u32 delta, t, bic_target, max_cnt;
+	u32 delta, bic_target, max_cnt;
+	u64 offs, t;
 
 	ca->ack_cnt++;	/* count the number of ACKs */
 
@@ -248,9 +248,11 @@ static inline void bictcp_update(struct
 	 * if the cwnd < 1 million packets !!!
 	 */
 
+	t = (s32)(tcp_time_stamp - ca->epoch_start);
+	t += msecs_to_jiffies(ca->delay_min >> 3);
 	/* change the unit from HZ to bictcp_HZ */
-	t = ((tcp_time_stamp + msecs_to_jiffies(ca->delay_min>>3)
-	      - ca->epoch_start) << BICTCP_HZ) / HZ;
+	t <<= BICTCP_HZ;
+	do_div(t, HZ);
 
 	if (t < ca->bic_K)		/* t - K */
 		offs = ca->bic_K - t;



  parent reply	other threads:[~2013-09-12 17:26 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-09-12 17:26 [ 00/19] 3.0.96-stable review Greg Kroah-Hartman
2013-09-12 17:26 ` [ 01/19] htb: fix sign extension bug Greg Kroah-Hartman
2013-09-12 17:26 ` [ 02/19] net: check net.core.somaxconn sysctl values Greg Kroah-Hartman
2013-09-12 17:26 ` [ 03/19] fib_trie: remove potential out of bound access Greg Kroah-Hartman
2013-09-12 17:26 ` Greg Kroah-Hartman [this message]
2013-09-12 17:26 ` [ 05/19] tcp: cubic: fix bug in bictcp_acked() Greg Kroah-Hartman
2013-09-12 17:26 ` [ 06/19] ipv6: dont stop backtracking in fib6_lookup_1 if subtree does not match Greg Kroah-Hartman
2013-09-12 17:26 ` [ 07/19] tun: signedness bug in tun_get_user() Greg Kroah-Hartman
2013-09-12 17:26 ` [ 08/19] ipv6: remove max_addresses check from ipv6_create_tempaddr Greg Kroah-Hartman
2013-09-12 17:26 ` [ 09/19] ipv6: drop packets with multiple fragmentation headers Greg Kroah-Hartman
2013-09-12 17:26 ` [ 10/19] ipv6: Dont depend on per socket memory for neighbour discovery messages Greg Kroah-Hartman
2013-09-12 17:26 ` [ 11/19] net: bridge: convert MLDv2 Query MRC into msecs_to_jiffies for max_delay Greg Kroah-Hartman
2013-09-12 17:26 ` [ 12/19] ICMPv6: treat dest unreachable codes 5 and 6 as EACCES, not EPROTO Greg Kroah-Hartman
2013-09-12 17:26 ` [ 13/19] tipc: fix lockdep warning during bearer initialization Greg Kroah-Hartman
2013-09-12 17:26 ` [ 14/19] PARISC: include <linux/prefetch.h> in drivers/parisc/iommu-helpers.h Greg Kroah-Hartman
2013-09-12 17:26 ` [ 15/19] pci: frv architecture needs generic setup-bus infrastructure Greg Kroah-Hartman
2013-09-12 17:26 ` [ 16/19] m32r: consistently use "suffix-$(...)" Greg Kroah-Hartman
2013-09-12 17:26 ` [ 17/19] m32r: add memcpy() for CONFIG_KERNEL_GZIP=y Greg Kroah-Hartman
2013-09-12 17:26 ` [ 18/19] m32r: make memset() global for CONFIG_KERNEL_BZIP2=y Greg Kroah-Hartman
2013-09-12 17:26 ` [ 19/19] KVM: s390: move kvm_guest_enter,exit closer to sie Greg Kroah-Hartman
2013-09-12 22:29 ` [ 00/19] 3.0.96-stable review Guenter Roeck
2013-09-12 23:07   ` Greg Kroah-Hartman
2013-09-13 23:04 ` Shuah Khan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20130912172524.646274753@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=ncardwell@google.com \
    --cc=stable@vger.kernel.org \
    --cc=stephen@networkplumber.org \
    --cc=ycheng@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome