[ltp] T41: swsusp vs. swsusp2 (+)

[Andrey Kislyuk]

>>> Hard one to call.  Swsusp (1) is ancient and pretty much unmaintained
>>> as far as I can see.
It looks to be quite maintained, otherwise I suspect it would break 
pretty fast as things got changed around. I did see a few times when it 
didn't wake up though; that never happened in s3 sleep.

>> It's not entirely clear that suspend2 (in its current form) will be in 
>> the kernel in any sort of sensible timeframe. There's still 
>> disagreement over whether certain aspects of it are the correct approach.
> Ok, so does the in-kernel swsusp work?  I have a TP G41 with a HT P4 
> processor.  I gather that I cannot suspend to RAM while running in HT. 
> Will swsusp work?  Does it suspend to swap?  How much swap do  need? 
> Where can the answers to this and a host of other questions be found?

http://www.suspend2.net/ maybe? You know, the project's homepage.

It suspends to swap, you don't need much swap (it compresses the image). 
I don't know if it works with HT though.

I got frustrated at the fact that swsusp2, the ck patches, and reiser4 
could not be had together in one package in gentoo portage, so I 
installed a custom gentoo patchset (acid-sources) yesterday. Previously 
I used swsusp but mostly s3 sleep. With the handy hibernate script, the 
switchover from my old acpi scripts to this was absolutely smooth. On 
this T40, swsusp2 is at least 3 times as fast as swsusp. swsusp2 with s3 
sleep (image-backed sleep) is several times slower than just s3 sleep, 
with the obvious benefit of having an image in case the laptop doesn't 
wake up for some reason.

The new kernel is also much snappier, probably due to the ck patchset, 
and once my 7k100 drive with Reiser4 is up it should get even better :)

-ak