Date: Tue, 26 May 2020 18:52:21 -0700 From: Adrian Chadd <adrian.chadd@gmail.com> To: Matthew Macy <mmacy@freebsd.org> Cc: freebsd-current <freebsd-current@freebsd.org> Subject: Re: iflib and options RSS is a no go for igbX Message-ID: <CAJ-Vmom8AcNFzTNDAqsKpsqMXW0uQn=NzgqwUyUZ35fKcEMjuA@mail.gmail.com> In-Reply-To: <CAPrugNpgaceaNFPJn7M%2B0g6pjJJvYUXgxTC7A6x4MijWcG3ycw@mail.gmail.com> References: <mailman.1179.1590537564.4503.freebsd-current@freebsd.org> <CAPrugNpgaceaNFPJn7M%2B0g6pjJJvYUXgxTC7A6x4MijWcG3ycw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 26 May 2020 at 17:01, Matthew Macy <mmacy@freebsd.org> wrote: > > > > > I can go setup my second PC this week (I have PTO!) and go see if I > > have any PCIe igb hardware here. I /think/ I do but it's fibre. > > The RSS compile time option used to be disabled in part because of > this. The drivers had working RSS that relied on packet receipt before > setting flowid. And it should also work fine if you don't compile in the RSS option. The reason it's annoying is when you do RSS + PCBGROUPS; as I said in the ena review the challenge was getting traffic into the right pcbgroup inpcb list and that isn't doable if you're waiting for hardware flowid first. Honestly I think we should rename it to something else besides RSS, cause it's more like PCBGROUPS + microsoft RSS. It's not very well named right now and I think it's confusing tf out a lot of people. -adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmom8AcNFzTNDAqsKpsqMXW0uQn=NzgqwUyUZ35fKcEMjuA>