Date: Thu, 19 Apr 2001 16:42:00 -0700 (PDT) From: "Peter C. Wallace" <pcw@mesanet.com> To: <freebsd-small@FreeBSD.ORG> Subject: Re: The ultimate board! Message-ID: <Pine.BSF.4.31.0104191631220.29802-100000@freeby.mesanet.com> In-Reply-To: <200104192310.f3JNAa807989@harmony.village.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 19 Apr 2001, Warner Losh wrote: > In message <200104192226.PAA21832@mina.soco.agilent.com> Darryl Okahata writes: > : ... but, you can't assume even wear. Don't directory entries, > : etc., get written in the same place, over and over? What about > : superblocks? > > You can assume even wear becahse wear averaging is in the CF hardware. > There's a pool of extra blocks that are used to average the wear. At > least that's what the CF propiganda says from SandDisk. Thats what the data sheet for the CF controller chips (Toshiba) that we use says also. The extra blocks are not for wear leveling, thats taken care of by having enough free space. The free blocks are for remapping newly developed bad blocks: so called "grown" defects. The Toshiba controller chip implements wear leveling, bad block forwarding and ECC. So basically if you leave enough free space on the drive for your frequently written files to "rotate" through. You can have _very_ long life. The internal flash format the Toshiba uses is published (its the same as the SmartMedia format) > > Warner > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-small" in the body of the message > Peter Wallace To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-small" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.31.0104191631220.29802-100000>