Date: Wed, 18 May 2011 12:52:43 -0400 From: Attilio Rao <attilio@freebsd.org> To: Garrett Cooper <yanegomi@gmail.com> Cc: freebsd-current@freebsd.org, Andriy Gapon <avg@freebsd.org>, freebsd-arch@freebsd.org Subject: Re: [rfc] remove hlt_cpus et al sysctls and related code Message-ID: <BANLkTimBwVb4TO497C3NFMsMq11jDY-AKQ@mail.gmail.com> In-Reply-To: <BANLkTinVGrLoAOS_ZQ1YVB_Fw1cvf5kHyA@mail.gmail.com> References: <4DD3F662.9040603@FreeBSD.org> <BANLkTikOTe9ut3GFx0bhOernKandRGLhPg@mail.gmail.com> <BANLkTinVGrLoAOS_ZQ1YVB_Fw1cvf5kHyA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
2011/5/18 Attilio Rao <attilio@freebsd.org>: > 2011/5/18 Garrett Cooper <yanegomi@gmail.com>: >> On Wed, May 18, 2011 at 9:40 AM, Andriy Gapon <avg@freebsd.org> wrote: >>> >>> I think that it is a well known fact that currently we do not have any = support for >>> dynamically offlining processors. =C2=A0Yet, we have some code that loo= ks like it does >>> provide that support and even provides a user interface to supposedly d= o that. >>> >>> What we don't currently do specifically: >>> - rebinding interrupts away from an offlined processor >>> - updating relevant cpu sets and masks >>> - protecting the above for concurrent access >>> - moving threads away from an offlined processor >>> - notifying potentially interested parties >>> - maybe more... >>> >>> The code has been in this shape for a long while and I would dare to sa= y that it >>> never really worked, not in "production ready" sense anyway. >>> An example of troubles caused by using that code can be found e.g. in t= he >>> followups to the following PR: >>> http://www.freebsd.org/cgi/query-pr.cgi?pr=3D145385 >>> And also discussed here: >>> http://thread.gmane.org/gmane.os.freebsd.stable/74462/focus=3D74510 >>> >>> I think that there already have been a proposal to remove the systcls a= nd the >>> code. =C2=A0I would like to re-submit that proposal. >>> Removing that code would: >>> 1) prevent users from hurting themselves by executing broken code >>> 2) potentially make things easier for largeSMP project >>> >>> Once we grow correct code for offlining CPUs, then we could re-introduc= e the >>> sysctls without any problems. >>> While the offlining code doesn't seem terribly hard to develop, it's a = big piece >>> of work and requires time and effort. >> >> =C2=A0 =C2=A0What would be nice too (even though it might not be possibl= e) is >> to make this more MI than it is today (i.e. sysctls that work for >> amd64, sparc64, etc), but that might be a pipe dream. >> Thanks! >> -Garrett > > That is actually the purpose. =C2=A0We should have a real online/offline > system for hotplugging CPUs, not only tied to x86 hyperthreading. > The htt specific parts are mostly hacks that don't take into account > all the necessary handover for it. For instance, I always promised to implement them and I never did, mostly because it is quite a bit of work in my idea and we don't really have a big pressure for it and there are really bigger things on my plate. If someone wants to step up and implementing offline/online CPUs I would be glad to discuss approaches I have in mind and help formalizing a plan for it, thus offering guidance for the implementation. Attilio --=20 Peace can only be achieved by understanding - A. Einstein
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?BANLkTimBwVb4TO497C3NFMsMq11jDY-AKQ>