Messages in this thread | | | Subject | TAHI IPv6 testing of kernel 2.6.20 | From | William Seppeler <> | Date | Thu, 15 May 2008 15:21:14 -0400 |
| |
I noticed on the TAHI website that kernel.org had the 2.6.20 kernel certified for IPv6 Ready Logo Phase2 as a router device.
The application was submitted on 20070827 (Application ID: US-2-C-20070827-000142) and certified on 20070926.
I'm currently testing a 2.6.20 kernel using Self_Test_1.4.14 (performing Phase2 host testing - ipv6ready_p2_host). It passes all tests except for one. I'm failing test:
v6LC_2_3_6_G - Redirected to Alternate Router: Invalid (Target Address is Multicast)
Based on the kernel.org application date, I assume the application was submitted using TAHI Self_Test_1.4.8. I see from the Self_Test CHANGELOG that test v6LC_2_3_6_G was modified after the application date:
2007/09/26 Yukiyo Akisada nd.p2/v6LC_2_3_6_G.def - correct packet definition
Q: Has the kernel.org group retested the 2.6.20 kernel using any of the recent Self_Test releases? In particular, Self_Test_1.4.14 or greater? And were there any failures?
As for details into the v6LC_2_3_6_G test failure:
The log shows the test fails due to processing an invalid redirect packet (ie: it's a negative test). The redirect packet contains a multicast address as a target address. RFC2461 (p.26) doesn't explicitly state this is invalid:
Target Address An IP address that is a better first hop to use for the ICMP Destination Address. When the target is the actual endpoint of communication, i.e., the destination is a neighbor, the Target Address field MUST contain the same value as the ICMP Destination Address field. Otherwise the target is a better first-hop router and the Target Address MUST be the router's link-local address so that hosts can uniquely identify routers.
But the test does appear to make sense based on the last sentence. The linux kernel I'm testing shows an echo response trying to use the multicast address from the redirect packet as it's first hop. This is what causes the test to fail. The expectation is that the linux kernel would ignore the redirect packet entirely.
As stated, this is the ONLY test in the whole Phase2 test suite that fails. I understand the failure, but I don't understand how the 2.6.20 kernel is Phase2 certified. At least not with the current Self_Test test suite.
Q: Has anyone else had issues in getting the 2.6.20 kernel to pass all TAHI IPv6 Ready Logo Phase2 tests?
Please CC answers to wrs@pt.com
Thank You,
======================================== William Seppeler Test Engineer Performance Technologies Inc. 205 Indigo Creek Dr. Rochester, NY 14626
Phone: (585) 256-0200, Fax: (585) 256-0791 Email: wrs@pt.com, Web: www.pt.com ========================================
| |