Date: Fri, 06 Jun 2008 23:27:47 +0000 From: "Poul-Henning Kamp" <phk@phk.freebsd.dk> To: Alan Cox <alc@cs.rice.edu> Cc: Alan Cox <alc@FreeBSD.org>, cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/vm vm_page.c Message-ID: <3171.1212794867@critter.freebsd.dk> In-Reply-To: Your message of "Fri, 06 Jun 2008 18:05:52 EST." <4849C2D0.9060704@cs.rice.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <4849C2D0.9060704@cs.rice.edu>, Alan Cox writes: >Poul-Henning Kamp wrote: >>> Essentially, neither madvise(..., MADV_DONTNEED) nor madvise(..., MADV_FREE) >>> work. (Moreover, I don't believe that they have ever worked as intended.) >> >> I remember they did have a measurable but minor impact on performance >> in systems which paged a lot, back when they were introduced, sort >> of 1995-ish. But with the subsequent plunge in RAM prices they >> never saw much use. > >After more thorough digging through CVS history, I now believe that >MADV_DONTNEED worked properly between revision 1.70 (May 1996) and 1.76 >(June 1996) of vm/vm_object.c. :-) It's more difficult to say for >MADV_FREE. It may have been effective until 1999. Sounds quite probable. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3171.1212794867>