Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 May 2009 15:19:18 GMT
From:      Randall Stewart <rrs@FreeBSD.org>
To:        freebsd-gnats-submit@FreeBSD.org
Subject:   kern/134369: IPV6 in Head broken for routing table updates
Message-ID:  <200905081519.n48FJIKY069413@www.freebsd.org>
Resent-Message-ID: <200905081520.n48FK1AR034769@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help

>Number:         134369
>Category:       kern
>Synopsis:       IPV6 in Head broken for routing table updates
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri May 08 15:20:01 UTC 2009
>Closed-Date:
>Last-Modified:
>Originator:     Randall Stewart
>Release:        HEAD - 8.0
>Organization:
FreeBSD Consultant LLC
>Environment:
uname -a
FreeBSD bsd4 8.0-CURRENT FreeBSD 8.0-CURRENT #3 r191882:191891M: Fri May  8 07:35:56 EDT 2009     randall@bsd4:/usr/src/sys/amd64/compile/bsd4  amd64


>Description:
The routing socket no longer gets routing messages
for IPv6.  Also when an IPv6 address is removed
the route is NOT removed. This seriously breaks
any IPv6 routing protocol which may use a
routing socket. It also breaks SCTP as well.


>How-To-Repeat:
Simply add a new in6 address to an interface. If you have
a routing socket open you will see that NO information is
given.

Then delete the in6 address. When its deleted aka ifconfig xxx inet6 addr -alias
then do a netstat -r -n inet6 and you will still find the address.

I can also supply a small program that reads the routing socket
if you want to run it before the add of the inet6 address.
>Fix:
Well there are two possibilities.

Qing Li can go in can add back the rt_newaddrmsg() calls
that he took out in the right places in the refactor'd code he
committed around Dec/Jan.

Or

We can back out these changes before 8 goes out. We need
to get this fixed BEFORE the 8 code slush IMO. So if Qing Li
does not have time to fix this.. then  backing out his changes
are the right answer.




>Release-Note:
>Audit-Trail:
>Unformatted:



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200905081519.n48FJIKY069413>