Date: Tue, 24 Mar 2015 13:51:37 -0700 From: Adrian Chadd <adrian@freebsd.org> To: hiren panchasara <hiren@strugglingcoder.info> Cc: Jack F Vogel <jfv@freebsd.org>, FreeBSD Net <freebsd-net@freebsd.org>, erj@freebsd.org, Jack Vogel <jfvogel@gmail.com>, Jason Wolfe <nitroboost@gmail.com> Subject: Re: Full 32bit flowid from igb(4) Message-ID: <CAJ-Vmo=ajfWokQ3bdu1Ww8kV3AdVwwsDpoS9y_%2Bywk%2BgcrbBvQ@mail.gmail.com> In-Reply-To: <20150324154931.GC53237@strugglingcoder.info> References: <20150323233908.GT53237@strugglingcoder.info> <20150323234214.GU53237@strugglingcoder.info> <CAFOYbc=f3OT85z50uqYEnBUkEej=8FRnRP8TpQuTStgqUEoDTA@mail.gmail.com> <20150324154931.GC53237@strugglingcoder.info>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, The main reason I didn't add it outside of RSS is that I didn't want to impact the behaviour that was there before. Before, it wasn't using the flowid - only the msix/queue id. I read the intel datasheets about that particular field - I'm pretty sure that by default we'll only see RSS hashed packets for IPv4/IPv6, however non v4/v6 packets won't have a flowid. There are also cases of the flow director or some hardware checksum config using the same field as the flowid. The /full/ solution would very carefully check the return status and ensure what's in the flowid field is a flowid. The "sometimes it may have a flowid, sometimes it won't" problem isn't so bad with kernel RSS enabled - it'll just software hash it. -adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmo=ajfWokQ3bdu1Ww8kV3AdVwwsDpoS9y_%2Bywk%2BgcrbBvQ>