Date: Thu, 15 Jun 2006 17:48:15 +0800 From: Xin LI <delphij@delphij.net> To: Robert Watson <rwatson@FreeBSD.org> Cc: cvs-src@FreeBSD.org, src-committers@FreeBSD.org, David Xu <davidxu@FreeBSD.org>, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/sys sched.h src/sys/posix4 ksched.c src/sys/kern sched_4bsd.c sched_core.c sched_ule.c Message-ID: <1150364895.2108.13.camel@spirit> In-Reply-To: <20060615102900.C9158@fledge.watson.org> References: <200606150637.k5F6bdax000157@repoman.freebsd.org> <1150363262.937.44.camel@spirit> <20060615102900.C9158@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] 在 2006-06-15四的 10:29 +0100，Robert Watson写道： > On Thu, 15 Jun 2006, Xin LI wrote: > [...]> A question that might be related: We have sched_sizeof_proc(), etc. in > > every scheduler, which returns a constant. Is it worthy to convert them to > > constants that just expand inline during the kernel compile? > > It should make no difference -- sched_sizeof_proc() is evaluated exactly once, > when the process zone is created. After that, the size is cached in the > process zone and never re-evaluated. The current model allows the size to be > calculated based on data structures private to sched_foo.c, which has some > nice implementation hiding advantages :-). Oh... Yes that's quite reasonable. Thanks for the explanation! Cheers, -- Xin LI <delphij delphij net> http://www.delphij.net/ [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (FreeBSD) iD8DBQBEkSzfhcUczkLqiksRAortAJ9gxwVg55Xm6hW3wVIkGevTFYPDcgCgitqA Og6d3B6TF/w547MPuN1sHpw= =UrlE -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1150364895.2108.13.camel>