netdev - Re: Bug in net/ipv6/ip6_fib.c:fib6_dump

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:	Sat, 23 Jun 2012 07:37:31 +0200
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	David Miller <davem@...emloft.net>
Cc:	johunt@...mai.com, kaber@...sh.net, dbavatar@...il.com,
	netdev@...r.kernel.org, yoshfuji@...ux-ipv6.org, jmorris@...ei.org,
	pekkas@...core.fi, kuznet@....inr.ac.ru,
	linux-kernel@...r.kernel.org, Ben Greear <greearb@...delatech.com>
Subject: Re: Bug in net/ipv6/ip6_fib.c:fib6_dump_table()

From: Eric Dumazet <edumazet@...gle.com>

> 1) Patrick McHardy has been inactive for a while, so do not expect
>    any insight from him.
> 
> 2) Ben Greear isn't even on the CC: list of this discussion yet he
>    appears to be the person who reproduced the crash way back then
>    and is listed in the Tested-by tag of the commit.
> 
>    As a result we aren't likely to get any insight from the one person
>    who actually could hit the crash.
> 
> I'm inclined to just revert simply because we have people active who
> can reproduce regressions introduced by this change and nobody can
> understand why the change is even necessary.

Well, except that :

I spent 3 hours trying to understand Alexey code and failed.

All other /proc/net files don't have a such sophisticated walkers aware
mechanism (easily DOSable by the way, if some guy opens 10.000 handles
and suspend in the middle the dumps).

cat /proc/net/tcp for example can display same socket twice or miss a
socket, because a 'suspend/restart' remembers offsets/counts in a hash
chain, not a pointer to 'next socket'

The fix I submitted is a real one, based on my analysis and tests.

Patrick patch was restarting the dump at the root of the tree, and
setting skip = count was doing nothing at all, since all entries were
dumped again.

This is more a stable candidate fix.

If someones smarter than me can find the real bug, then we certainly can
revert Patrick patch ?

[PATCH] ipv6: fib: fix fib dump restart

Commit 2bec5a369ee79576a3 (ipv6: fib: fix crash when changing large fib
while dumping it) introduced ability to restart the dump at tree root,
but failed to skip correctly a count of already dumped entries. Code
didn't match Patrick intent.

We must skip exactly the number of already dumped entries.

Note that like other /proc/net files or netlink producers, we could
still dump some duplicates entries.

Reported-by: Debabrata Banerjee <dbavatar@...il.com>
Reported-by: Josh Hunt <johunt@...mai.com>
Signed-off-by: Eric Dumazet <edumazet@...gle.com>
Cc: Patrick McHardy <kaber@...sh.net>
Cc: Ben Greear <greearb@...delatech.com>
Cc: Alexey Kuznetsov <kuznet@....inr.ac.ru>
---
 net/ipv6/ip6_fib.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index 74c21b9..6083276 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -1349,8 +1349,8 @@ static int fib6_walk_continue(struct fib6_walker_t *w)
 			if (w->leaf && fn->fn_flags & RTN_RTINFO) {
 				int err;

-				if (w->count < w->skip) {
-					w->count++;
+				if (w->skip) {
+					w->skip--;
 					continue;
 				}

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html