Date: Wed, 11 Feb 2015 10:49:31 +0800 From: Sepherosa Ziehau <sepherosa@gmail.com> To: Sean Bruno <sbruno@freebsd.org> Cc: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org> Subject: Re: Intel 82574L (em) Message-ID: <CAMOc5cwG%2BmbcArAt5xSjKYrbZeA40ETZUPgi4y51r_GYYsitQw@mail.gmail.com> In-Reply-To: <54CBF396.3090903@ignoranthack.me> References: <54CBF396.3090903@ignoranthack.me>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Jan 31, 2015 at 5:11 AM, Sean Bruno <sbruno@ignoranthack.me> wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA512 > > http://www.intel.com/content/dam/doc/datasheet/82574l-gbe-controller-datasheet.pdf > > According to 7.1.11, this device does indeed have 2 queues for stuff and > or things. So, basic RSS would be possible in something like an Atom box. > > I note that the em(4) driver intentionally disables this on > initialization. I'm up for some science on my new shiny, soon to be > router box. Any reason not to default to 1 queue and allow loader.conf > to raise it to 2? You could actually enable 2 RX rings w/o MSI-X on 82574; you still get the benefit of hardware calculated RSS hash at least. And as far as I have tested, 2 RX rings work for 82574L, but 2 TX rings don't work (gave me TX watchdog timeout). And you could also use 2 RX rings on 82571/82572/82573 and i217/i218; 2 TX rings work on 82571 at least (you need to setup TX context for each TX descriptor though). Best Regards, sephe
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAMOc5cwG%2BmbcArAt5xSjKYrbZeA40ETZUPgi4y51r_GYYsitQw>