Location of partition table on disk

[Chris Torek]

In article <7195 at utzoo.UUCP> henry at utzoo.UUCP (Henry Spencer) writes:
>I too thought it would be a good idea to have the partition table
>on the disk it partitioned, until I read a paper on the Cray-2 Unix
>port. ... the size and disk location of each partition is recorded
>in the /dev inode for it. ... It seems to me that this is a cleaner
>solution all around.

I remain unconvinced.  It is certainly an intriguing idea, and not
entirely without precedent (we have a local kernel hack that already
puts things into the disk block fields of /dev inodes).  It has some
definite advantages:

	- the number of partitions per drive is not fixed;
	- one can record striping information cleanly;
	- it takes very little driver code.

It also has some definite disadvantages:

	- the number of partitions per drive is not fixed;
	- there is a bootstrap problem (avoided by compiling
	  in the root partition info);
	- it does little for removable packs (indeed, it is
	  quite possibly harmful in such cases);
	- it discourages sharing the information among operating
	  systems (VMS to read Unix /dev directories anyone?).

I still believe that the disk layout is a property of the disk
pack, not of the file systems mounted thereon.  There is certainly
some crossover, though.  Perhaps the information should be in
several places, and cross-verified---though that brings up the
question of what to do if the verification fails. . . .
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 1516)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris at umcp-cs		ARPA:	chris at mimsy.umd.edu