Possible future problems?: c partition

[Chris Torek]

> From: eichelbe at nadc.ARPA
>> From: niel <@CSNET-RELAY.ARPA, at ucsc.csnet (Niel Warren):niel at ucsc.csnet>
>>
>> About that c partition... c accesses the whole disk including 
>> the space reserved bad blocks. 

Like all blanket statements, this is false.  :-)  It depends on
how you've set up your partition tables.  The distributed 4.1 and
4.2 code does indeed include the bad block areas in the C partition.

> Wizards:  Is this really true?  Am I playing with fire by having
> a c partition on a CDC 9766 (like an RM05) controlled by an SI 9400
> controller?

Not necessarily.  Not only does it depend on your tables, there is
nothing that says you must use the entire available area for a file
system or swap/paging region.

As I recall, unmodified 4.1 demanded that all paging areas be the
same size; this size was wired into a variable in the various swap
configuration files.  Under 4.2 (and modified 4.1) there is a third
entry in ``struct swdevt'' which indicates the size of the area.
If this third value is zero, the block device's ``dsize'' routine
is called during swap configuration to find out how much space is
available.  This routine typically returns the value from the
appropriate partition table.

In any case, if you have set your swap size to 600 blocks, then
you have 300K of swap space, even though you can access 600000
blocks (300M) via the C partition.  (600 blocks is a silly size;
this is just an example.)

One can also make a file system that is smaller than the available
space.  This is usually pointless (but consider Sun's network disk
for a counterexample).

All in all, you are probably better off if you avoid the C partition,
unless you have adjusted your tables; if you do not use it, you cannot
get into trouble.  However, there is nothing inherently wrong with
using the area.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris at umcp-cs		ARPA:	chris at maryland