Date: Sun, 30 Jul 2006 08:23:41 -0400 From: Chuck Swiger <cswiger@mac.com> To: Peter Jeremy <peterjeremy@optushome.com.au> Cc: freebsd-current@freebsd.org Subject: Re: vmstat's entries type Message-ID: <44CCA4CD.1050309@mac.com> In-Reply-To: <20060730115427.GD1310@turion.vk2pj.dyndns.org> References: <20060728210154.GC748@turion.vk2pj.dyndns.org> <1154189612.1565.10.camel@genius.i.cz> <20060729205655.GE748@turion.vk2pj.dyndns.org> <20060729211530.GA50342@uk.tiscali.com> <1154212340.3609.18.camel@genius.i.cz> <20060729230214.GI12597@groat.ugcs.caltech.edu> <1154216319.23616.23.camel@genius.i.cz> <1154217036.23616.28.camel@genius.i.cz> <20060730005342.GJ12597@groat.ugcs.caltech.edu> <1154250451.23616.37.camel@genius.i.cz> <20060730115427.GD1310@turion.vk2pj.dyndns.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Peter Jeremy wrote: [ ... ] > This was based on an examination of /usr/src/sys/*/include/atomic.h. > It looks like I was overly quick because the RISC architectures are > using compare-and-swap, rather than test-and-set primitives - sorry > about the error. The details are: [ ... ] > - ppc doesn't support 64-bit atomic operations lwarx/stwcx...? http://publibn.boulder.ibm.com/doc_link/en_US/a_doc_lib/aixassem/alangref/lwarx.htm http://www-128.ibm.com/developerworks/library/pa-atom/ > - sparc64 uses cas in a loop. > > Whilst true spinlocks are not needed, amd64 is the only architecture > FreeBSD supports that does not require some sort of loop to perform > a 64-bit atomic operation. You can use membar instructions on the SPARCv9 to enforce various memory models (RMO, PSO, TSO, if not full sequential consistency). -- -Chuck
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?44CCA4CD.1050309>