From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759044Ab0GAVOH (ORCPT ); Thu, 1 Jul 2010 17:14:07 -0400 Received: from kroah.org ([198.145.64.141]:33775 "EHLO coco.kroah.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758805Ab0GAVNA (ORCPT ); Thu, 1 Jul 2010 17:13:00 -0400 X-Mailbox-Line: From gregkh@clark.site Thu Jul 1 10:42:48 2010 Message-Id: <20100701174248.299881798@clark.site> User-Agent: quilt/0.48-10.1 Date: Thu, 01 Jul 2010 10:41:52 -0700 From: Greg KH To: linux-kernel@vger.kernel.org, stable@kernel.org Cc: stable-review@kernel.org, torvalds@linux-foundation.org, akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk, =?ISO-8859-15?q?Linus=20L=C3=BCssing?= , Sven Eckelmann , Andrew Lunn Subject: [022/200] Staging: batman-adv: Fixing wrap-around bug in vis In-Reply-To: <20100701175201.GA2149@kroah.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 2.6.34-stable review patch. If anyone has any objections, please let me know. ------------------ From: =?UTF-8?q?Linus=20L=C3=BCssing?= commit ea4ceb18b525fd7016c10995c0f1313a729c7e2b upstream. When the seqno for a vis packet had a wrap around from i.e. 255 to 0, add_packet() would falsely claim the older packet with the seqno 255 as newer as the one with the seqno of 0 and would therefore ignore the new packet. This happens with all following vis packets until the old vis packet expires after 180 seconds timeout. This patch fixes this issue and gets rid of these highly undesired 3min. breaks for the vis-server. Signed-off-by: Linus Lüssing Signed-off-by: Sven Eckelmann Signed-off-by: Andrew Lunn Signed-off-by: Greg Kroah-Hartman --- drivers/staging/batman-adv/vis.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) --- a/drivers/staging/batman-adv/vis.c +++ b/drivers/staging/batman-adv/vis.c @@ -27,6 +27,22 @@ #include "hard-interface.h" #include "hash.h" +/* Returns the smallest signed integer in two's complement with the sizeof x */ +#define smallest_signed_int(x) (1u << (7u + 8u * (sizeof(x) - 1u))) + +/* Checks if a sequence number x is a predecessor/successor of y. + they handle overflows/underflows and can correctly check for a + predecessor/successor unless the variable sequence number has grown by + more then 2**(bitwidth(x)-1)-1. + This means that for a uint8_t with the maximum value 255, it would think: + * when adding nothing - it is neither a predecessor nor a successor + * before adding more than 127 to the starting value - it is a predecessor, + * when adding 128 - it is neither a predecessor nor a successor, + * after adding more than 127 to the starting value - it is a successor */ +#define seq_before(x, y) ({typeof(x) _dummy = (x - y); \ + _dummy > smallest_signed_int(_dummy); }) +#define seq_after(x, y) seq_before(y, x) + struct hashtable_t *vis_hash; DEFINE_SPINLOCK(vis_hash_lock); static DEFINE_SPINLOCK(recv_list_lock); @@ -212,7 +228,7 @@ static struct vis_info *add_packet(struc old_info = hash_find(vis_hash, &search_elem); if (old_info != NULL) { - if (vis_packet->seqno - old_info->packet.seqno <= 0) { + if (!seq_after(vis_packet->seqno, old_info->packet.seqno)) { if (old_info->packet.seqno == vis_packet->seqno) { recv_list_add(&old_info->recv_list, vis_packet->sender_orig);