Date: Tue, 02 Jun 2009 20:53:01 +0200 From: Michael <freebsdusb@bindone.de> To: Jack Vogel <jfvogel@gmail.com>, freebsd-bugs@freebsd.org Subject: Re: Bug in igb driver Message-ID: <4A25750D.2030809@bindone.de> In-Reply-To: <2a41acea0906021052m1aa394bfq2d30c0588e41443c@mail.gmail.com> References: <4A25185D.2060304@bindone.de> <4A2558EB.4060604@bindone.de> <2a41acea0906021052m1aa394bfq2d30c0588e41443c@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Thanks for your respsonse. That code is obviously not in CURRENT yet, if you can provide me a patch against CURRENT of today I can give it a try. In the meantime this should really get into the 7.2 release errata - similar to those regarding TSO on fxp (in the end these devices are pretty common and tracking this down was an annoying experience I would like to spare others). If that's beyond your scope maybe you can tell me how to trigger this. cheers Michael Jack Vogel wrote: > Thanks for noting this, code that is coming has the driver disable LRO > when the stack calls for it, I believe it will solve this problem. > > Regards, > > Jack > > > On Tue, Jun 2, 2009 at 9:52 AM, Michael <freebsdusb@bindone.de> wrote: > >> Michael wrote: >>> Hello, >>> >>> we encountered problems using igb driver on a Supermicro dual quad-core >>> system (X7DWU). >>> It features an Intel Gigabit dual port controller (82575). >>> >>> The IP forwarding performance was extremely bad (5 kb/s up to 50 kb/s) >>> until we disabled >>> Large Receive Offloading in /etc/sysctl.conf: >>> dev.igb.0.enable_lro=0 >>> dev.igb.1.enable_lro=0 >>> >>> (changing it using ifconfig -lro and down/up the interfaces didn't help). >>> >>> With the systctl values set as shown above, we get the expected >>> performance ob 50-60 MB/s. >>> We tested this on 7.1-RELEASE, 7.2-RELEASE and 7-STABLE. Right now we're >>> installing >>> 8-CURRENT on the machine to see if it makes a difference. >>> >>> It think it would be worthwhile to list this issue in the 7.2 release >>> errata section (that would've >>> saved us a significant amount of time :). >>> >>> Cheers >>> Michael >>> >>> >>> >>> >>> _______________________________________________ >>> freebsd-bugs@freebsd.org mailing list >>> http://lists.freebsd.org/mailman/listinfo/freebsd-bugs >>> To unsubscribe, send any mail to "freebsd-bugs-unsubscribe@freebsd.org" >>> >> Turns out that 8/CURRENT of today has the same issues. >> Is this okay as a bug report, or should I file a PR seperately? >> >> _______________________________________________ >> freebsd-bugs@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-bugs >> To unsubscribe, send any mail to "freebsd-bugs-unsubscribe@freebsd.org" >> > _______________________________________________ > freebsd-bugs@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-bugs > To unsubscribe, send any mail to "freebsd-bugs-unsubscribe@freebsd.org"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4A25750D.2030809>