Well, we just got a spanking brand new Procom Tech.
PR-IDE drive for the HP Vectra here running 1.2.3.
The drive specs out at 826cyl 48hds 63s/t for about
1280meg. This is a plain Slackware 2.1 install.
The system took it fine, and fdisk /dev/hdb sees it nicely.
Disk /dev/hdb: 48 heads, 63 sectors, 826 cylinders
Units = cylinders of 3024 * 512 bytes
Device Boot Begin Start End Blocks Id System
/dev/hdb1 1 1 34 51376+ 83 Linux native
/dev/hdb2 35 35 826 1197504 5 Extended
/dev/hdb5 35 35 67 49864+ 82 Linux swap
/dev/hdb6 68 68 420 533704+ 83 Linux native
/dev/hdb7 421 421 515 143608+ 83 Linux native
/dev/hdb8 516 516 826 470200+ 83 Linux native
Originally, the system saw the drive as having 2497cyls, 16hds 63s/t
and fdisk seemed to like it OK. I was a little worried about the
number of cyls, so in the BIOS setup I changed it to what's above,
which is what the drive really has according to the data sheet that
came with it. It was close to 2497 at least, I don't remember exactly.
This will eventually be the only drive in the system, hence
the various filesystems, / /usr /var /var/spool/news in order
in this list. The mke2fs seems to go OK on each one, a couple
of them giving this msg :
Checking for bad blocks (read-only test): Bad block 0 out of range; ignored.
The filesystems seem to work fine though. The main problem is in
the "mkswap -c /dev/hdb5 49864". With the mkswap from SW 2.1, it
gives this :
jedi:~# mkswap /dev/hdb5 49864
Setting up swapspace, size = 51056640 bytes
mkswap: unable to write signature page
I grabbed the latest mkswap from the a2/util.tgz in the sunsite
Slackware 2.2 distribution, but it gives the same thing ...
What's going on here ?