Date: Wed, 06 Jan 2010 13:12:43 +0200 From: Alexander Motin <mav@FreeBSD.org> To: =?UTF-8?B?RGFnLUVybGluZyBTbcO4cmdyYXY=?= <des@des.no> Cc: Miroslav Lachman <000.fbsd@quip.cz>, Thomas Backman <serenity@exscape.org>, Poul-Henning Kamp <phk@phk.freebsd.dk>, freebsd-current@freebsd.org, freebsd-arch@freebsd.org, Pieter de Goeje <pieter@degoeje.nl> Subject: Re: File system blocks alignment Message-ID: <4B44702B.9090503@FreeBSD.org> In-Reply-To: <86637fjypj.fsf@ds4.des.no> References: <65036.1262386032@critter.freebsd.dk> <201001020101.16450.pieter@degoeje.nl> <86637fjypj.fsf@ds4.des.no>
next in thread | previous in thread | raw e-mail | index | archive | help
Dag-Erling Smørgrav wrote: > Pieter de Goeje <pieter@degoeje.nl> writes: >> Poul-Henning Kamp <phk@phk.freebsd.dk> writes: >>> Ideally, newfs/UFS should do a quick test to look for any obvious >>> boundaries, and DTRT, a nice little task for somebody :-) >> A search for the offset for which newfs (or a simpler test) runs >> fastest? Interesting idea :-) > > Or just default to 4 kB or even 8 kB, which has zero performance impact > (either way) for disks with 512 B sectors. AFAIK it is what latest Windows'es do, just with bigger value. While it is usually good, I am not sure it is a perfect solution. These WDC HDDs with installed "XP compatibility" jumper artificially shift addresses, adding some offset between logical and physical sectors to fit legacy 63 sectors partition offset. Such case should be detected and handled. The only problem is that these drives doesn't report this info now. :( -- Alexander Motin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4B44702B.9090503>