Date: Fri, 6 Nov 2015 13:15:50 -0800 From: Adrian Chadd <adrian.chadd@gmail.com> To: Ian Lepore <ian@freebsd.org> Cc: Hans Petter Selasky <hps@selasky.org>, Luigi Rizzo <rizzo@iet.unipi.it>, Rasool Al-Saadi <ralsaadi@swin.edu.au>, "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, Alexander Motin <mav@freebsd.org> Subject: Re: Timing issue with Dummynet on high kernel timer interrupt Message-ID: <CAJ-Vmom2KQO5P94UJNUAcpcKR36nO=RTKdAS1xn4ojypyUMF_g@mail.gmail.com> In-Reply-To: <1446843506.91534.443.camel@freebsd.org> References: <6545444AE21C2749939E637E56594CEA3C0DCCC4@gsp-ex02.ds.swin.edu.au> <5638B7B5.3030802@selasky.org> <6545444AE21C2749939E637E56594CEA3C0DE7FF@gsp-ex02.ds.swin.edu.au> <563B2703.5080402@selasky.org> <6545444AE21C2749939E637E56594CEA3C0E0BD9@gsp-ex02.ds.swin.edu.au> <563C6864.2090907@selasky.org> <CA%2BhQ2%2Bhm2z0MkB-8w5xJM7%2Biz13r_ZjwmpZBnb30_D_48gaf-w@mail.gmail.com> <563C786C.1050305@selasky.org> <CA%2BhQ2%2Bj0WiGgzV119M1ZQiXP5Z7fq7deVxDPkOhvTc7hpTETKw@mail.gmail.com> <563CC186.9000807@selasky.org> <563CD533.2000909@selasky.org> <1446828229.91534.417.camel@freebsd.org> <563CDA8F.5010901@selasky.org> <563CDBF9.3090800@selasky.org> <1446830585.91534.435.camel@freebsd.org> <CAJ-Vmo=1ie6XqMCFZ1FcuD4TmD7PQGMXjODUo1eGPdaQ8wc4Tg@mail.gmail.com> <1446843506.91534.443.camel@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 6 November 2015 at 12:58, Ian Lepore <ian@freebsd.org> wrote: > On Fri, 2015-11-06 at 11:15 -0800, Adrian Chadd wrote: >> Ideally there'd be both behaviours: >> >> * You'd specify whether a timer/sleep needs to be exact or can >> withstand some jitter (which is what linux provides); and >> * You can communicate to the kernel its aggressiveness for coalescing >> wakeups. >> > > We already implement exactly both of these things (the former in args > to scheduling callouts, the latter in the form of the sysctl I > referenced earlier which lets you set the "agressivness for > coalescing"). > > The problem with the former is very little code actually uses the > flavor of callout scheduling that lets you specify precision. The > problem with the latter is that its default value is to allow enough > deviation that people think the system is misbehaving. Right. What about sleep, though? (eg userland calling sleep, usleep, etc.) Would the original poster problem be fixed just by setting the exact precision? -adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmom2KQO5P94UJNUAcpcKR36nO=RTKdAS1xn4ojypyUMF_g>