Date: Fri, 08 Mar 2002 13:00:28 -0500 (EST) From: John Baldwin <jhb@FreeBSD.org> To: Poul-Henning Kamp <phk@critter.freebsd.dk> Cc: Matthew Jacob <mjacob@feral.com>, Jeff Roberson <jeff@FreeBSD.org>, cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org Subject: Re: cvs commit: src/sys/sys smp.h src/sys/kern subr_smp.c src/sy Message-ID: <XFMail.020308130028.jhb@FreeBSD.org> In-Reply-To: <6023.1015543411@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On 07-Mar-02 Poul-Henning Kamp wrote: > In message <XFMail.020307181140.jhb@FreeBSD.org>, John Baldwin writes: >> >>On 07-Mar-02 Poul-Henning Kamp wrote: >>> In message <XFMail.020307171639.jhb@FreeBSD.org>, John Baldwin writes: >>> >>>>Does that make sense? I'm not say we need to support some wildly sparse >>>>range, >>>>but we shouldn't assume 0 and 1 for any dual CPU system. >>> >>> What is the problem with putting a logical CPU id in a word in the >>> per-cpu area ? As far as I know, that would even be faster to read >>> than the APIC-id ? >> >>Nothing. We actully do this now. We just base the logical ID on the >>physical >>ID now in a 1:1 fashion. > > So if we change this not to, JeffR and others needing a per-cpu array > index will be happy. > > Going first ? > > Going second ? > > Going ? It doesn't _matter_. For consumers of the code, it is a sparse logical ID in the range 0...MAXCPU. It is an _implementation_ detail that on the alpha arch, the logical ID just happens to be the physical ID. I just wanted to make _sure_ that people were aware that the ID's were sparse and was simply bringing up the alpha as an example of _how_ it could be sparse. -- John Baldwin <jhb@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.020308130028.jhb>