Date: Thu, 23 Jan 2020 18:18:15 -1000 (HST) From: Jeff Roberson <jroberson@jroberson.net> To: Gleb Smirnoff <glebius@freebsd.org> Cc: Ryan Stone <rysto32@gmail.com>, src-committers <src-committers@freebsd.org>, svn-src-all <svn-src-all@freebsd.org>, svn-src-head <svn-src-head@freebsd.org> Subject: Re: svn commit: r357051 - head/sys/dev/bge Message-ID: <alpine.BSF.2.21.9999.2001231808020.1198@desktop> In-Reply-To: <20200124033243.GL1268@FreeBSD.org> References: <202001231636.00NGawrr080128@repo.freebsd.org> <CAFMmRNyP5nKPZ62c9yRktH4t23GAgnpc2u8kHhUevUCu3doAQQ@mail.gmail.com> <20200123230546.GG1268@FreeBSD.org> <CAFMmRNwhcV-%2BSsq_a1RZJYXC8wvnUDLDxdFeHD7PgL5wuierkg@mail.gmail.com> <20200124012458.GI1268@FreeBSD.org> <CAFMmRNz3TVPfP7_fhBMa=gdN%2BM4OdCco3z1qO=mPOm57-QOsng@mail.gmail.com> <20200124024356.GK1268@FreeBSD.org> <alpine.BSF.2.21.9999.2001231657400.1198@desktop> <20200124033243.GL1268@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 23 Jan 2020, Gleb Smirnoff wrote: > On Thu, Jan 23, 2020 at 05:09:14PM -1000, Jeff Roberson wrote: > J> While we don't have a policy strictly requiring reviews it is the norm to > J> have substantial changes socialized and reviewed. I appreciate the work > J> that you are doing but it likely should've been discussed somewhere > J> more publicly. I apologized if I missed it but I don't see reference to > J> anything. > > That was https://reviews.freebsd.org/D23242 Ok thank you. Can you tag commits so people can see the discussion? Was it in one I missed? When I'm committing a long patch series I include the link in all of them. Ryan, are you subscribed to @networking? > > J> Architecturally I am more concerned with the coarseness of net_epoch and > J> the duration of hold becoming a resource utilization problem in high > J> turn-over workloads. Like short connection tcp. Has anyone done > J> substantial testing here? epoch as it is today will hold every free > J> callback for a minimum of several clock ticks and a maximum of 2x the > J> duration of the longest epoch section time. With preemption, etc. this > J> could be 100s of ms of PCBs held. > > We also are concerned about that theoretically. Haven't yet seen effect > in practice, but our sessions are mostly longer living. First we have the > tunable to limit batching. Second, there are some ideas on how to improve > the garbage collector performance if it becomes an issue. I am often surprised at how much cpu time I see on linux spent in RCU free processing. Many of these things are written in such a way that everything is cache cold by the time you swing back around. So the callout walk is a giant cold linked list. Thanks, Jeff > > -- > Gleb Smirnoff >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.BSF.2.21.9999.2001231808020.1198>