Date: Sat, 17 Jun 1995 03:34:17 +1000 From: Bruce Evans <bde@zeta.org.au> To: aflundi@sandia.gov, freebsd-hackers@freebsd.org Subject: Re: HD Geometry dirty trick Message-ID: <199506161734.DAA29552@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>> I.e., 64/32 geometry works if it was the correct (BIOS) geometry all >> along. Otherwise, it is unlikely to work. >What do you mean by "correct"? I have a Fujitsu M2624 with I mean that it is identical with the BIOS geometry. >an Adaptec 1542B in my home machine and a Micropolis 2217 >with a Buslogic Bt445S in my machine at work, and neither >have real geometries even in the ball park of xxxx/64/32, >yet xxxx/64/32 is what DOS fdisk gave and is the only geometry >of the several "reality based geometries" I tried that "worked" >with 2.0.5R. The BIOS can invent any geometry that it wants (subject to the constraints 1 <= nsectors <= 63, 1 <= nheads <= 256, 1 <= ncyls <= 1024). >I also have a Maxtor 7120a IDE drive that did not suffer >from the need for this trick. I presume this is because >the CMOS contains the IDE geometry, but not the SCSI geometries. No, FreeBSD doesn't look in the CMOS for disk geometries. It asks the drive. Small IDE drives report the same geometry as the BIOS (unless the BIOS has been configured weirdly). FreeBSD should have no problems in this case. "Small" means that a suitable geometry can be invented subject to the above constraints and also subject to the "physical" constraint nheads <= 16, i.e., <= 504MB. Above that various extensions are used and various things break. E.g., for more than 16 heads, old drivers (such as FreeBSD's) break. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199506161734.DAA29552>