Date: Tue, 5 Jan 2010 14:28:59 +0100 From: Ivan Voras <ivoras@freebsd.org> To: Alexander Motin <mav@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r201545 - head/sys/geom/raid3 Message-ID: <9bbcef731001050528k5b46b8ect2f9576f7b47e563a@mail.gmail.com> In-Reply-To: <201001050206.o05265Ln053910@svn.freebsd.org> References: <201001050206.o05265Ln053910@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
2010/1/5 Alexander Motin <mav@freebsd.org>: > Author: mav > Date: Tue Jan =C2=A05 02:06:05 2010 > New Revision: 201545 > URL: http://svn.freebsd.org/changeset/base/201545 > > Log: > =C2=A0Slightly optimize XOR calculation. Linux has an interesting approach to RAID XOR operations - they calculate at boot time which method is faster, from a set of i386, MMX, SSE and the like and then use it for what appears to be every context requiring fast XORs. E.g. http://fxr.watson.org/fxr/source/arch/x86/include/asm/xor_64.h?v=3Dlinux-2.= 6 (I know SIMD instructions would require a FP-enabled kernel, I'm just pointing out an interesting idea)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9bbcef731001050528k5b46b8ect2f9576f7b47e563a>