Date: Tue, 21 Aug 2007 18:17:48 +0200 From: Jacek Zapala <jacek@ipv6.jacek.it.pl> To: Daniel Hartmeier <daniel@benzedrine.cx> Cc: freebsd-net@freebsd.org Subject: Re: kern/115413: [ipv6] ipv6 pmtu not working Message-ID: <1187713068.3973.6.camel@localhost.localdomain> In-Reply-To: <20070821145047.GC32421@insomnia.benzedrine.cx> References: <200708211010.l7LAA6V7082258@freefall.freebsd.org> <20070821121118.GF27160@insomnia.benzedrine.cx> <1187703472.22531.4.camel@localhost.localdomain> <20070821135048.GA32421@insomnia.benzedrine.cx> <1187705811.30269.5.camel@localhost.localdomain> <20070821143125.GB32421@insomnia.benzedrine.cx> <1187707117.846.3.camel@localhost.localdomain> <20070821145047.GC32421@insomnia.benzedrine.cx>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 2007-08-21 at 16:50 +0200, Daniel Hartmeier wrote: > Could you possibly try the patch below, and see if it fixes the > problem > (with rfc1323 enabled again)? > > We're accessing th_flags from the TCP header embedded in the ICMPv6 > packet, even though we only pulled up the first 8 bytes of the mbuf, > because the sender doesn't have to provide more of the header. > > There could be random garbage there, and the bit corresponding to > TH_SYN might be set, so the window scale factor is not applied. Not > sure > if that would be reproducable so reliably, but it sure is a bug ;) > I have applied the patch and it looks like it has helped. But I'm not sure if I understood well - you suspect that only 8 bytes of tcp header are copied from the original tcp packet to the icmp message by the router? TCP packet in the icmp message looks like almost complete (of course it is shorter than the original due to 1280 bytes icmp packet limit). I have captured that traffic so I can send you full packets if you would like to look at them. Jacek
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1187713068.3973.6.camel>