netfilter-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jan Engelhardt <jengelh@medozas.de>
To: kaber@trash.net
Cc: netfilter-devel@vger.kernel.org
Subject: [PATCH 8/9] xtables: fix the broken detection/removal of redundant addresses
Date: Sun, 27 Feb 2011 02:31:18 +0100	[thread overview]
Message-ID: <1298770280-7652-9-git-send-email-jengelh@medozas.de> (raw)
In-Reply-To: <1298770280-7652-1-git-send-email-jengelh@medozas.de>

From: Wes Campaigne <westacular@gmail.com>

	[To observe this issue, populate a hostname (DNS or local db)
	with multiple adresses across multiple subnets (cf. prefixlen
	below)

	# e.g. /etc/hosts
	127.0.0.2       lo-x
	127.0.0.3       lo-x
	127.0.1.4       lo-x
	127.0.1.5       lo-x
	127.0.2.6       lo-x

	Then invoke xtables_ipparse_any by e.g. `-m conntrack
	--ctorigsrc lo-x/24`. -j.eng]

This same block of code, apparently to detect if addresses are
identical after applying the mask, and to skip the duplicates and the
ones made redundant by the mask, has been present and unchanged from
as far back as I could find (circa iptables 1.2).

By inspection, it was wrong, and always has been: once the code finds
a duplicate, it will drop the rest of the array one by one as it
re-detects the same duplicate over and over. When the addresses came
from a single hostname lookup, and their order was random, then this
created unpredictable behaviour by iptables, which seem to ignore some
of those addresses at random times.

I suspect the original idea also involved a swap between the duplicate
and the address from the (current) end of the array, but a line of
code to do that seems to have never existed. I have finally added it.
(Well, as much as is needed: there does not need to be a full swap,
because we are just going to ignore the duplicate, pretend the array
is one shorter, and never look at the contents of the end again. So,
we can get away with just copying from the end.)

[Reword comment about shuffle: replace by mentioning tail copy to
replace dup. -j.eng]

Signed-off-by: Wes Campaigne <westacular@gmail.com>
---
 xtables.c |   18 ++++++++++++++----
 1 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/xtables.c b/xtables.c
index b45bf92..660fbd0 100644
--- a/xtables.c
+++ b/xtables.c
@@ -1311,8 +1311,13 @@ void xtables_ipparse_any(const char *name, struct in_addr **addrpp,
 		addrp[j++].s_addr &= maskp->s_addr;
 		for (k = 0; k < j - 1; ++k)
 			if (addrp[k].s_addr == addrp[j-1].s_addr) {
-				--*naddrs;
-				--j;
+				/*
+				 * Nuke the dup by copying an address from the
+				 * tail here, and check the current position
+				 * again (--j).
+				 */
+				memcpy(&addrp[--j], &addrp[--*naddrs],
+				       sizeof(struct in_addr));
 				break;
 			}
 	}
@@ -1620,8 +1625,13 @@ void xtables_ip6parse_any(const char *name, struct in6_addr **addrpp,
 		++j;
 		for (k = 0; k < j - 1; ++k)
 			if (IN6_ARE_ADDR_EQUAL(&addrp[k], &addrp[j - 1])) {
-				--*naddrs;
-				--j;
+				/*
+				 * Nuke the dup by copying an address from the
+				 * tail here, and check the current position
+				 * again (--j).
+				 */
+				memcpy(&addrp[--j], &addrp[--*naddrs],
+				       sizeof(struct in_addr));
 				break;
 			}
 	}
-- 
1.7.1


  parent reply	other threads:[~2011-02-27  1:31 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-02-27  1:31 iptables: docs & address parsing Jan Engelhardt
2011-02-27  1:31 ` [PATCH 1/9] doc: mention other possible nf_loggers for TRACE Jan Engelhardt
2011-02-27  1:31 ` [PATCH 2/9] doc: fix odd partial sentence in libipt_TTL Jan Engelhardt
2011-02-27  1:31 ` [PATCH 3/9] libxt_quota: require --quota to be specified Jan Engelhardt
2011-02-27  1:31 ` [PATCH 4/9] doc: rateest options can be optional Jan Engelhardt
2011-02-27  1:31 ` [PATCH 5/9] libxtables: fix memory scribble beyond end of array Jan Engelhardt
2011-02-27  1:31 ` [PATCH 6/9] libxtables: avoid confusing use of ai_protocol=IPPROTO_IPV6 Jan Engelhardt
2011-02-27  1:31 ` [PATCH 7/9] xtables: fix excessive memory allocation in host_to_ipaddr Jan Engelhardt
2011-02-27  1:31 ` Jan Engelhardt [this message]
2011-02-27  1:31 ` [PATCH 9/9] xtables: use all IPv6 addresses resolved from a hostname Jan Engelhardt
2011-02-27 15:19 ` iptables: docs & address parsing Patrick McHardy

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=1298770280-7652-9-git-send-email-jengelh@medozas.de \
    --to=jengelh@medozas.de \
    --cc=kaber@trash.net \
    --cc=netfilter-devel@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).