[TUHS] Partition table rationales?

Random832 random832 at fastmail.com
Sat Dec 16 13:43:58 AEST 2017

Is there any information on the rationales behind the sizes and
positions of the hardcoded partitions in the rp/hp drivers in V6? There
are odd gaps of unused space between ones that appear to be intended to
be used together (I think I have some pieces of the puzzle worked out
below, but not all).

For the rp driver, only rp[01] uses the whole disk - rp2 and rp3 are
9200 blocks. The hp partitions looked more complicated, but it looks
like the manpage is wrong about where partition 2 starts - 84018 vs
48018. Partitions 0123 are 161584 blocks, 4567 are 162400 blocks. 4567
start on 100-cylinder boundaries, but are only a bit over 97 cylinders
long - in fact, they are the same length as rp[01]: 203 RP03 cylinders.

rp6 and rp7 are not mentioned in the manpage, and are 15600 blocks,
making rp[65] and 47 reasonably non-wasteful configurations.

Is there something special about 9200 blocks? rp2 and rp3 are that size,
and hp1 is that size rounded to the next cylinder.
V7 is much simpler: rp0 is the whole disk, 123 is a non-overlapping set
with no wasted space. hp[0145], 016, and 017 are non-overlapping sets.

More information about the TUHS mailing list