[buug] Using swap on an SSD...or not.

[Rick Moen]

Quoting Jon McClintock (jammer at weak.org):

> Apparently there's also a dependency on a swap partition to enable 
> hibernation.

Good point.  Actually, of the two hibernation implementations, I believe
only uswsusp needs an active swap partition to write out RAM contents
to.  The other implementation, suspend2, can save to a file on the HD,
instead, which you can programmatically delete when not being needed 
for hibernation, and create programmatically as needed.

And, just now found, here's one suggested way to do the same thing with
the default uswusp implementation:
http://superuser.com/questions/21020/can-i-hibernate-linux-without-a-swap-partition

(Untested, as I don't have a system where I've had to solve this
problem.)  

A completely different approach might be to allocate a regular swap
partition, but set the kernel's swappiness parameters such that it
doesn't get used for VM despite being known to the swap manager.