/etc/disktab entry seems weird

[John Martin]

The "/etc/disktab" entry for my DECstation 3100 running Ultrix 4.0 with
an RZ55 looks like this:

------
rz55|RZ55|DEC RZ55 Winchester:\
	:ty=winchester:ns#36:nt#15:nc#1224:\
			  ^^    ^^    ^^^^
	:pa#32768:ba#8192:fa#1024:\
	:pb#131072:bb#8192:fb#1024:\
	:pc#649040:bc#8192:fc#1024:\
	    ^^^^^^
	:pd#152446:bd#8192:fd#1024:\
	:pe#152446:be#8192:fe#1024:\
	:pf#180308:bf#8192:ff#1024:\
	:pg#485200:bg#8192:fg#1024:\
	:ph#0:bh#8192:fh#1024:
------

Looking at the highlighted numbers here, though, something just doesn't
seem to work out.  The value for "pc", by convention, should be the
number of available sectors on the disk, but a naive calculation of
sectors*tracks*cylinders would yield 36*15*1224=660960 sectors, not
649040 sectors as indicated by the "pc" entry.

There seem to be several disks listed in "/etc/disktab" that have similar
sorts of discrepancies:

------
rd31|RD31|DEC RD31 Winchester: ns=17, nt= 4, nc= 615, pc=  41560, calc=  41820
rd32|RD32|DEC RD32 Winchester: ns=17, nt= 6, nc= 820, pc=  83236, calc=  83640
rd33|RD33|DEC RD33 Winchester: ns=17, nt= 7, nc=1170, pc= 138565, calc= 139230
rd53|RD53|DEC RD53 Winchester: ns=17, nt= 8, nc=1024, pc= 138672, calc= 139264
ra80|RA80|DEC RA80 Winchester: ns=31, nt=14, nc= 546, pc= 237212, calc= 236964
rd54|RD54|DEC RD54 Winchester: ns=17, nt=15, nc=1221, pc= 311200, calc= 311355
rz55|RZ55|DEC RZ55 Winchester: ns=36, nt=15, nc=1224, pc= 649040, calc= 660960
rz56|RZ56|DEC RZ56 Winchester: ns=54, nt=15, nc=1632, pc=1299174, calc=1321920
rz57|RZ57|DEC RZ57 Winchester: ns=71, nt=15, nc=1925, pc=2025788, calc=2050125
ra92|RA92|DEC RA92 Winchester: ns=69, nt=13, nc=3279, pc=2940951, calc=2941263
------

What am I missing?

* Where does the magic 649040 value come from?

* Is the "pc" entry wrong?  Are the "ns", "nt" and "nc" values wrong?

* Perhaps the remaining 11920 sectors on the RZ55 are preallocated to
  hold bad sector remappings?

  * If so, how/when do they get preallocated?

  * How's the 11920 calculated?

  * Which entries in TFM describe the process, or what other sources
    talk about it?

* Is this really worth worrying about?

Please email any ideas you have, I'll summarize in about two weeks.

Regards,
--
John Martin, Applied Dynamics International, Ann Arbor, MI, (313)973-1300
jcm at adi.com || jcm%amara.uucp at mailgw.cc.umich.edu || ...uunet!sharkey!amara!jcm
"How many times over the past six months did you, or people working for you,
 write some program fragment for table searching?" -- Bertrand Meyer