Date: Tue, 06 Feb 2007 23:46:30 -0600 From: Alan Cox <alc@cs.rice.edu> To: Robert Watson <rwatson@FreeBSD.org> 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_contig.c vm_page.c vm_pageq.c vm_zeroidle.c Message-ID: <45C967B6.5020304@cs.rice.edu> In-Reply-To: <20070206115809.K32369@fledge.watson.org> References: <200702050602.l1562tnI034504@repoman.freebsd.org> <20070206115809.K32369@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Robert Watson wrote: > On Mon, 5 Feb 2007, Alan Cox wrote: > >> alc 2007-02-05 06:02:55 UTC >> >> FreeBSD src repository >> >> Modified files: >> sys/vm vm_contig.c vm_page.c vm_pageq.c >> vm_zeroidle.c >> Log: >> Change the free page queue lock from a spin mutex to a default >> (blocking) >> mutex. With the demise of Alpha support, there is no longer a >> reason for >> it to be a spin mutex. > > > Is there any measurable performance change from this? In the past, > I've observed significantly higher cost for acquiring spin mutexes vs. > sleep mutexes in micro-benchmarking. The change was most striking on > Intel Xeon P4 hardware with an SMP kernel, and quite a bit less so on > other platforms. I didn't notice any difference for buildworld on a 4-way Opteron. Alan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?45C967B6.5020304>