Date: Wed, 22 Aug 2018 09:01:02 +0200 From: Sebastian Huber <sebastian.huber@embedded-brains.de> To: Matthew Macy <mmacy@freebsd.org> Cc: freebsd-hackers@freebsd.org Subject: Re: epoch(9) background information? Message-ID: <26445c95-17c5-1a05-d290-0741d91b7721@embedded-brains.de> In-Reply-To: <CAPrugNpZ5CihCW6hz3ztXAZrNn1qJNRsE=yGCvw1rzqNPQYRvg@mail.gmail.com> References: <db397431-2c4c-64de-634a-20f38ce6a60e@embedded-brains.de> <CALX0vxBAN6nckuAnYR3_mOfwbCjJCjHGuuOFh9njpxO%2BGUzo3w@mail.gmail.com> <fc088eb4-f306-674c-7404-ebe17a60a5f8@embedded-brains.de> <15e3f080-2f82-a243-80e9-f0a916445828@embedded-brains.de> <CAPrugNpZ5CihCW6hz3ztXAZrNn1qJNRsE=yGCvw1rzqNPQYRvg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 22/08/18 08:49, Matthew Macy wrote: > > > On Tue, Aug 21, 2018 at 11:42 PM Sebastian Huber=20 > <sebastian.huber@embedded-brains.de=20 > <mailto:sebastian.huber@embedded-brains.de>> wrote: > > On 22/08/18 08:34, Sebastian Huber wrote: > > On 21/08/18 15:38, Jacques Fourie wrote: > >> > >> > >> On Tue, Aug 21, 2018 at 8:33 AM, Sebastian Huber > >> <sebastian.huber@embedded-brains.de > <mailto:sebastian.huber@embedded-brains.de> > >> <mailto:sebastian.huber@embedded-brains.de > <mailto:sebastian.huber@embedded-brains.de>>> wrote: > >> > >> =C2=A0=C2=A0=C2=A0 Hello, > >> > >> =C2=A0=C2=A0=C2=A0 I update currently a port of the FreeBSD netw= ork stack, etc. to > >> =C2=A0=C2=A0=C2=A0 the real-time operating system RTEMS from the= head version at > >> =C2=A0=C2=A0=C2=A0 2017-04-04 to the head version of today. I no= ticed that some > >> =C2=A0=C2=A0=C2=A0 read-write locks are replaced by a relatively= new stuff called > >> =C2=A0=C2=A0=C2=A0 EPOCH(9). Is there some background informatio= n available > for this? > >> =C2=A0=C2=A0=C2=A0 The man page is a bit vague and searching for= something named > >> =C2=A0=C2=A0=C2=A0 epoch on the internet is not really great. Fo= r example, what is > >> =C2=A0=C2=A0=C2=A0 the motivation for this change? How is this r= elated to > >> =C2=A0=C2=A0=C2=A0 read-copy-update (RCU)? > >> > >> =C2=A0=C2=A0=C2=A0 -- =C2=A0=C2=A0=C2=A0 Sebastian Huber, embedd= ed brains GmbH > >> > >> =C2=A0=C2=A0=C2=A0 Address : Dornierstr. 4, D-82178 Puchheim, Ge= rmany > >> > <https://maps.google.com/?q=3DDornierstr.+4,+D-82178+Puchheim,+Germ= any&entry=3Dgmail&source=3Dg> > >> =C2=A0=C2=A0=C2=A0 Phone=C2=A0 =C2=A0: +49 89 189 47 41-16 > >> =C2=A0=C2=A0=C2=A0 Fax=C2=A0 =C2=A0 =C2=A0: +49 89 189 47 41-09 > >> =C2=A0=C2=A0=C2=A0 E-Mail=C2=A0 : sebastian.huber@embedded-brain= s.de > <mailto:sebastian.huber@embedded-brains.de> > >> =C2=A0=C2=A0=C2=A0 <mailto:sebastian.huber@embedded-brains.de > <mailto:sebastian.huber@embedded-brains.de>> > >> =C2=A0=C2=A0=C2=A0 PGP=C2=A0 =C2=A0 =C2=A0: Public key available= on request. > >> > >> =C2=A0=C2=A0=C2=A0 Diese Nachricht ist keine gesch=C3=A4ftliche = Mitteilung im Sinne > des > >> EHUG. > >> > >> =C2=A0=C2=A0=C2=A0 _____________________________________________= __ > >> freebsd-hackers@freebsd.org > <mailto:freebsd-hackers@freebsd.org> > <mailto:freebsd-hackers@freebsd.org > <mailto:freebsd-hackers@freebsd.org>> > >> =C2=A0=C2=A0=C2=A0 mailing list > >> https://lists.freebsd.org/mailman/listinfo/freebsd-hackers > >> <https://lists.freebsd.org/mailman/listinfo/freebsd-hackers>; > >> =C2=A0=C2=A0=C2=A0 To unsubscribe, send any mail to > >> =C2=A0=C2=A0=C2=A0 "freebsd-hackers-unsubscribe@freebsd.org > <mailto:freebsd-hackers-unsubscribe@freebsd.org> > >> =C2=A0=C2=A0=C2=A0 <mailto:freebsd-hackers-unsubscribe@freebsd.o= rg > <mailto:freebsd-hackers-unsubscribe@freebsd.org>>" > >> > >> > >> Additional information is available here : > >> http://concurrencykit.org/presentations/ebr.pdf > >> <http://concurrencykit.org/presentations/ebr.pdf>. The way I > >> understand it is that it is mostly used in place of read locks t= o > >> provide liveness guarantees without using atomics. Additional > detail > >> is available in the commit messages. As an example see r333813 f= or > >> some performance data. > >> > > > > Thanks, for the reference. The "epoch reclamation" are good > keywords > > to find more information. > > > > What is the right mailing list to ask questions about the epoch > > implementation of the FreeBSD kernel? > > > > To support this machinery in RTEMS is a bit difficult (in > particular > > EPOCH_LOCKED). Since RTEMS is supposed to be a real-time operatin= g > > system it supports only fixed-priority and job-level fixed priori= ty > > (EDF) schedulers. To allow some scaling to larger SMP systems it > > supports clustered scheduling together with the mutual exclusion > > locking protocols MrsP > > (http://www-users.cs.york.ac.uk/~burns/MRSPpaper.pdf > <http://www-users.cs.york.ac.uk/%7Eburns/MRSPpaper.pdf>) and OMIP > > (http://www.mpi-sws.org/~bbb/papers/pdf/ecrts13b.pdf > <http://www.mpi-sws.org/%7Ebbb/papers/pdf/ecrts13b.pdf>). This > makes the > > thread pinning hard to implement (which is very easy to support i= n > > FreeBSD). The locking protocols may temporarily move a thread whi= ch > > owns a mutex to a foreign scheduler instance, e.g. a thread which > > wants to obtain the mutex helps the owner to make progress if it > was > > pre-empted in its home scheduler instance. Due to a timeout of th= e > > helper the owner may loose the right to execute in the foreign > > scheduler instance. This would make it impossible to fulfil the > > processor pinning constraint (e.g. the thread priority in the > foreign > > scheduler instance is undefined). > > > > It would save me a lot of trouble if I could assume that > EPOCH_LOCKED > > is an exotic feature which is unlikely to get used in FreeBSD. > > > > Another question, is it a common use case to call > epoch_enter_preempt() > and epoch_exit_preempt() while owning a mutex? > > > Yes. Very. It is generally not permitted to hold a mutex across=20 > epoch_wait() that's why there's the special flag EPOCH_LOCKED. If you=20 > have a very limited number of threads, you might want to have each=20 > thread have its own record registered with the epoch. Then you=20 > wouldn't need the CPU pinning. The pinning is just away of providing a=20 > limited number of records to an unbounded number of threads. Thanks for the prompt answer. Do I need a record per thread and per epoch? Could I use only one (maybe=20 dependent on the nest level?) record per thread? --=20 Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone : +49 89 189 47 41-16 Fax : +49 89 189 47 41-09 E-Mail : sebastian.huber@embedded-brains.de PGP : Public key available on request. Diese Nachricht ist keine gesch=C3=A4ftliche Mitteilung im Sinne des EHUG= .
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?26445c95-17c5-1a05-d290-0741d91b7721>