Date: Mon, 11 Oct 2004 13:09:07 -0700 From: Sean McNeil <sean@mcneil.com> To: Robert Watson <rwatson@freebsd.org> Cc: freebsd-current@freebsd.org Subject: Re: 20% packet loss with re0 in gigE mode vs. 0% in 100BT Message-ID: <1097525346.1123.27.camel@server> In-Reply-To: <Pine.NEB.3.96L.1041011152957.31040Y-100000@fledge.watson.org> References: <Pine.NEB.3.96L.1041011152957.31040Y-100000@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--=-wGBpgMQ6iNVtOIUk5xNY Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Mon, 2004-10-11 at 12:31, Robert Watson wrote: > On Mon, 11 Oct 2004, Sean McNeil wrote: >=20 > > > Assuming you're not filling the send buffer, it would definitely sugg= est a > > > driver, configuration, or hardware bug. There have recently been a n= umber > > > of changes to the if_re driver to fix support for jumbo frames, etc. = It > > > would be interesting to know whether backing out to earlier revisions= of > > > the if_re driver affect the problem you're seeing. In particular, > > > ifre.c:1.35 was the jumbo frame change, so 1.34 would be interesting,= and > > > 1.31 is before some other related changes. Likewise, you could try > > > backing out to before locking was introduced by setting debug.mpsafen= et=3D0 > > > in loader.conf and then backing out to if_re.c:1.29. I might be gene= rally > > > useful to try setting debug.mpsafenet=3D0 with the current driver to > > > eliminate that as a possible concern. > >=20 > > These are good suggestions as well, but I have heard from another user > > that has seen this kind of thing over all of these versions. It is les= s > > likely then that the jumbo or locking changes caused the issue.=20 >=20 > Have you tried 4.x to see if the same problem occurs there? >=20 > Have you tried the normal juggling of ACPI, APIC, and other configuration > variables? It sounds like things work fine at minimal bandwidth rates, s= o > that probably rules out interrupt problems and the like, but I figured it > was worth asking. >=20 > I assume you've probably looked at substituting the switch, forcing > auto-negotiation to 1gb, etc? Again, I have failed to provide information that was sent in previous emails under a different topic. I haven't tried 4.x, no, but I have tried with various configurations: FreeBSD/re0/amd64/gigE fails FreeBSD/re0/amd64/100BT passes WinXP/re0/gigE passes WinXP/dc0/100BT passes ... Everything I have tested works just fine. It is only when the re0 under FreeBSD is using PHY gigE that I have any packet loss. There is a small chance that there is bursting that is causing the loss but I doubt it as the output rate is 15Mbps. This is a pretty low rate and not very likely to cause serious bursting. I now have a source stream and a saved stream from a receiver. I will attempt to examine the differences in them to see if it is just a matter of some packets being dropped or if it is more nefarious. Sean --=-wGBpgMQ6iNVtOIUk5xNY Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (FreeBSD) iD8DBQBBauhiyQsGN30uGE4RArogAJwOEoGgYV2/RBdjiQVD05+y1me6FwCgpfBM ArfSly+NeFO15eH1Qj7cTmY= =dCkm -----END PGP SIGNATURE----- --=-wGBpgMQ6iNVtOIUk5xNY--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1097525346.1123.27.camel>