ipv6 activity causing system hang in kernel 2.4.4

[Tim Haynes <kernel@stirfried.vegetable.org.uk>]

    [I can't see evidence of this being reported before; apols if I'm
     wrong. Please also Cc: me as I only read l-k intermittently but would
     like to help out more.]

Hi,

I've been making very tentative forays into IPv6. However, in my simple
experiments thus far I appear to have located a bug:

1/ configure 2 machines with site-local IP#s - I'm using 
        ifconfig eth0 inet6 add fec0:1234:5:6::n 
2/ flood-ping from one to the other
3/ after about 15s, watch one box hang, needing magic-sysreq or hard reset

This is only with kernel 2.4.4; 2.4.2, 2.4.3 and NetBSD boxes are not
affected. It is independent of platform; I've reproduced it at will on a
lowly p75, an athlon, a p3-800 and on a powerbook/PPC.

All kernels are compiled to have ipv6 modular, netfilter modular...
everything with which I'm playing, modular.

Compiler versions:
| zsh, 12:06AM % gcc -v
| Reading specs from /usr/lib/gcc-lib/powerpc-linux/2.95.4/specs
| gcc version 2.95.4 20010319 (Debian prerelease)
| zsh, 12:06AM % gcc -v
| Reading specs from /usr/lib/gcc-lib/i386-linux/2.95.4/specs
| gcc version 2.95.4 20010319 (Debian prerelease)
(I'm tracking Debian/Unstable here.)

I have tcpdump logs (<http://spodzone.org.uk/~tim/ipv6/> - they're 570K
apiece); the `victim' machine receives nothing but ping-requests and sends
nothing but ping-replies until the file is truncated; the surviving box
sends nothing but requests and receives nothing but replies until it
becomes requests-only. (IOW there is no evidence of ARP, fragmentation
traffic, only the pings.)

The Changelog lists an `IPv6 packet re-assembly fix' in -pre2; my
suspicions lie in this area or with my compiler.

If there's anything else I can provide by way of diagnostics, please let me
know.

Cheers,

~Tim
-- 
<http://spodzone.org.uk/>