Date: Fri, 12 Feb 2016 21:07:10 +0100 From: Kristof Provost <kp@FreeBSD.org> To: Larry Rosenman <ler@lerctr.org> Cc: Freebsd current <freebsd-current@freebsd.org>, owner-freebsd-current@freebsd.org Subject: Re: IPV6 TCP6 Slow Connect Message-ID: <ED881300-6C2D-4CD0-9F46-B8BD1C354B01@FreeBSD.org> In-Reply-To: <7d399dddcde79d950e0817acddcbcfc1@thebighonker.lerctr.org> References: <8ca0068954b5cb25891c3324b2e8c863@thebighonker.lerctr.org> <20160211101449.GI19084@vega.codepro.be> <b152cabdaafbbfea7b17c7ebb4a0fb35@thebighonker.lerctr.org> <88197BB9-F03F-48BB-97DC-9BAAA2FA299F@FreeBSD.org> <125b65d56d310c0436ceb8047d1d0417@thebighonker.lerctr.org> <563b81912fc2de4670be51a832467ba9@thebighonker.lerctr.org> <06d72e0d3ae9f20bea04af39dba74847@thebighonker.lerctr.org> <D9EE0405-3818-400C-B52C-50BC5EB88CF4@FreeBSD.org> <33469a0ef87e3299819248f153d842d2@thebighonker.lerctr.org> <3df20fcac20d532ef77bf63ea2e042cb@thebighonker.lerctr.org> <A504FBD9-53B8-4AC0-876A-A9C37A139880@FreeBSD.org> <7d399dddcde79d950e0817acddcbcfc1@thebighonker.lerctr.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 12 Feb 2016, at 15:33, Larry Rosenman <ler@lerctr.org> wrote: >=20 > On 2016-02-12 08:31, Kristof Provost wrote: >>> On 12 Feb 2016, at 15:29, Larry Rosenman <ler@lerctr.org> wrote: >>> On 2016-02-12 08:13, Larry Rosenman wrote: >>>> sysctl net.inet.tcp.rfc1323=3D0 >>>> makes it work >>> Shouldn't the stack do the right thing here? For the record, the = other side >>> is also FreeBSD (10.2-STABLE). >> Yes, but it=E2=80=99s possible that there=E2=80=99s a problem with = the pf scrubbing of >> the window scaling or timestamp options. >> I have a vague recollection of having looked at that in the past. >> Bug 172648 also claims there is/was an issue with checksums in that >> case, but I=E2=80=99ve never been able to reproduce it. >> Regards, >> Kristof > Ok. Since I can reproduce this at will, and the 2 firewalls are = pfSense, how can I help? I=E2=80=99ll still need to reproduce it locally to fix it, but it might = be interesting to know if the packet is dropped by the router, or sent = out again with an incorrect checksum. Can you take a capture on the WAN interface and see if the TCP SYN makes = it out (if it does, I=E2=80=99d expect the checksum to be wrong) or not? Regards, Kristof=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?ED881300-6C2D-4CD0-9F46-B8BD1C354B01>