netdev - Problem facing with ipv6 router advertisement in kernel-3.0.26

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>] [day] [month] [year] [list]

Message-ID: <4A71D24947E78D43BC584A7CD4391A41017DCF73@SIXPRD0410MB359.apcprd04.prod.outlook.com>
Date:	Wed, 18 Jul 2012 13:02:02 +0000
From:	BALAKUMARAN KANNAN <balakumarank@...aelxsi.co.in>
To:	"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Problem facing with ipv6 router advertisement in kernel-3.0.26

       As per RFC, curhoplimit of router advertisement(ra) should be kept as hoplimit for the routing table entry. It expects that the hoplimit of a routing path should not affected if a ra comes for the path with curhoplimit is 0. But in the case of kernel-3.0.26, the value is altered to 255. But it happens only in the router cache(#ip -6 route show cached) not in the main routing table (#ip -6 route show). But if I disable 'multiple routing table' option while building the kernel, it also affects the main routing table.

	Timer expiry for a routing table entry. I am facing this scenario. I have the gc_interval value 30 seconds. I am receinving a ra with preferred lifetime 20 seconds. So the routing table entry expirs and removed from the routing table in 20 seconds. But the cache remains with the expired routing entry till next flush.

	As per my knowledge MTU value is not stored in the routing table. I can see the the MTU value in routing cache but not in the main routing table. So in my case, once the routing cahce is flushed, the kernel not properly fragments the packet according to the MTU value even the timer doesn't expire. I don't know am I right or wrong. I will tell what is happening. I have a router connected with a target board. The router sends a ra with mtu 1200 and sends a ICMP_REQUEST with length 1500. And it expects the target board to send ICMP_REPLY as fragmented into two packets (1200 and 300). But the target board does correctly if the cache is present. But once the cache is flushed, it fails.

Note: I am using Kernel-3.0.26 for arm architecture.

Thank you,

--Regards,
K.Balakumaran

--
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