Date: Mon, 23 Mar 2015 16:42:14 -0700 From: hiren panchasara <hiren@strugglingcoder.info> To: freebsd-net@freebsd.org, jfv@freebsd.org, erj@freebsd.org Cc: Adrian Chadd <adrian@freebsd.org>, Jason Wolfe <nitroboost@gmail.com> Subject: Full 32bit flowid from igb(4) Message-ID: <20150323234214.GU53237@strugglingcoder.info> In-Reply-To: <20150323233908.GT53237@strugglingcoder.info> References: <20150323233908.GT53237@strugglingcoder.info>
next in thread | previous in thread | raw e-mail | index | archive | help
--R7Dyui215VKdTDYA Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Correcting Eric's email and subject line. On 03/23/15 at 04:39P, hiren panchasara wrote: > scottl@freebsd.org > Bcc:=20 > Subject: Full 32bit flowid from igb(4) [was: Re: Unbalanced LACP link] > Reply-To:=20 > In-Reply-To: <20150319175145.GH53237@strugglingcoder.info> >=20 > On 03/19/15 at 10:51P, hiren panchasara wrote: > > On 03/17/15 at 12:34P, Adrian Chadd wrote: > > > On 17 March 2015 at 11:33, Jason Wolfe <nitroboost@gmail.com> wrote: > > > > On Mon, Mar 16, 2015 at 2:43 AM, Hans Petter Selasky <hps@selasky.o= rg> wrote: > > > >> On 03/16/15 10:37, Vitalii Duk wrote: > > > >>> > > > >>> I've changed use_flowid to 0 and it helped! But isn't it setting > > > >>> significant? In a description it says "Shift flowid bits to preve= nt > > > >>> multiqueue collisions". > > > >> > > > >> > > > >> Hi, > > > >> > > > >> Maybe your ethernet hardware is not properly setting the m_flowid = =2E.. > > > >> > > > >> --HPS > > > >> > > > > > > > > Flip use_flowid back to 1 and try setting > > > > net.link.lagg.default_flowid_shift / net.link.lagg.X.flowid_shift t= o 0 > > > > as Hiren suggested. r260179 added this shift, which has caused us > > > > balancing issues with the i350/igb. > > > > > > > > https://svnweb.freebsd.org/base?view=3Drevision&revision=3D260179 > > > > > > > > Based on Adrian's comment about igb/ixgbe not setting the 'full > > > > flowid' under normal conditions, does that mean this shift should b= e 0 > > > > by default to ensure we don't break balancing for devices that only > > > > set the CPU/MSIX queue? > > >=20 > > > Or we can just see if there's anything wrong with putting the full 32 > > > bit RSS flowid in received packets that have them. > >=20 > > It'd be nice to have but for now I am proposing following to fix a known > > broken case because of an optimization: > > https://reviews.freebsd.org/D2098 >=20 > Turns out, setting flowid_shift to 0 is not the correct solution. Please > look at the review above for more details.=20 >=20 > Looking at the code, we have a way to get full flowid but it's hidden > under "ifdef RSS": > rxr->fmp->m_pkthdr.flowid =3D le32toh(cur->wb.lower.hi_dword.rss); >=20 > Using just this line gives me good hash on I350 igb(4) chipset that we > have. >=20 > Is it possible to expose this outside RSS? Would this break other > things/chips? >=20 > Cheers, > Hiren --R7Dyui215VKdTDYA Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (FreeBSD) iQF8BAEBCgBmBQJVEKTVXxSAAAAAAC4AKGlzc3Vlci1mcHJAbm90YXRpb25zLm9w ZW5wZ3AuZmlmdGhob3JzZW1hbi5uZXRBNEUyMEZBMUQ4Nzg4RjNGMTdFNjZGMDI4 QjkyNTBFMTU2M0VERkU1AAoJEIuSUOFWPt/lomgIAKIvM6TwFTT+PaFvN8TPRdNc LJ1iY+8iAeGkwz1YJZ1oOVYONScc/zl4rYQdxzKvYwMges0eHTdJFEltwbSGGyzG BkYZR/BzvMzpLzGe/aDqATbotxOCBzbMdoWNuY3mwhmNfNLphoOq/YLb9y/202SY d9642VJInhtURcrD2t1isOzXg9bNzQISpN4Ti1I/+x+p2Cptzx4UwjOPoiJVd59s HxolAbawUckOW+3Ze4SOyP+nM470rlmWuRx+ifmh6uJgl0lsjD/3AtqY/dHaE88a rh+tYU9cHwo4PRuNUE7wiQekCXaVwEYwf4A5jb7cTE+8CbB/O5tJna9lgZz4HkA= =F6OG -----END PGP SIGNATURE----- --R7Dyui215VKdTDYA--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150323234214.GU53237>