Don't use swap partition for hibernation! (was Re: laptop experiences...)

[david.gardner@philips.com]

>I see. Thanks for explaining, John. I think it would be possible, in
>most cases, to use the swap partition to save the state for hibernate,
>though. What do you think? Does the APM code in the kernel rely on the
>APM BIOS to do the actual writing of the RAM image? If it does,
>wouldn't it be possible to fool it and make do without any FAT
>partitions?

Whoa!  You won't be able to use the swap partition for this purpose!
Consider: You are running lots of apps (and hence using lots of memory
and swap-space), you tell the machine to hibernate and it saves it's
memory contents _over_the_top_of_the_swap_space_!  This, as you can
imagine, will cause all sorts of bad things to happen when you resume...

As for the requiring of FAT partitions, I can't comment as for Thinkpads,
but I have a Libretto 50 and a Vaio C1XE, and both of these have no need
for a FAT partition for the hibernation features: The Libretto `hides'
part of the disk at the end, so reserving it for the hibernation, and 
the Vaio needs you to run a utility called `phdisk.exe' (dos app, there
is a Linux one which someone has written called `lphdisk' - check 
Freshmeat.net) to create _either_ a hibernation file on a FAT partition,
or a hibernation partition.  Both of these work under Linux after having
been created, but the partition is usually a better option for Linux as
it avoids the need to have a nasty FAT partition laying around (why 
would you want one anyway?  ;-)

Also, in both these cases (Lib, Vaio) the BIOS does the writing of the 
RAM image.  The Lib even hibernates automagically when the battery gets
_really_ low, whereas the Vaio just carries on until it turns off... 
*grrr*


Hope this helps a bit

	Dave