Date: Fri, 28 Apr 1995 16:49:37 +1000 From: Bruce Evans <bde@zeta.org.au> To: bde@zeta.org.au, wpaul@skynet.ctr.columbia.edu Cc: hackers@FreeBSD.org Subject: Re: 0412 SNAP install hangs after 'changing root device ...' Message-ID: <199504280649.QAA22098@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>> >> 3. During the install of 0412 newfs produces the following warning >> >> message 'Calculated sectors per cylinder 4096 disagrees with disklabel >> >> (2046). No other warnings/errors are produced during the install. >> ... >Y'know, I see this same message whenever I mount my swap partition >as an mfs filesystem on /tmp, and I saw it today when trying to newfs >a floppy. At least in the case of the floppy, this caused me all >sorts of grief until I forced it to use 36 sectors per cylinder with >a command line option (4096 sectors per cylinder on a 1.44 meg floppy? >yer kidding, right?). Could somebody please explain to me what's so >normal about this value? newfs used to default to the geometry in the label. It was br^H^Hchanged to default to a fictitious geometry with each cylinder having one track of 4096 sectors. This is supposed to give a better filesystem layout for ZBR drives (a fictitious geometry specialized for running newfs is likely to be better than a fictitions geometry specialized for avoiding BIOS braindamage) and to avoid some newfs bugs for certain geometries. It works poorly for 1.xMB floppies. The most amusing brokenness is that it gives more inodes than there are sectors on a floppy. This is interesting because it shows that the cpg and inospercg calculations don't work in boundary cases. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504280649.QAA22098>