Suspend-to-disk in ubuntu main?

[Matthew Garrett]

On Sat, 2004-12-04 at 17:50 -0500, thully at umich.edu wrote:
> I'm aware that suspend-to-disk support for Hoary is being tested in an
> experimental kernel (look for PMTesting in wiki).  However, I wondered when and
> if this will be in Ubuntu main.

We're moving forward with kernel development again now - I'm planning on
working on the initrd support for swsusp with upstream in order to come
up with a solution everyone is happy with, rather than merging the
current (slightly hacky) solution.

> Also, I wondered why swsusp2 is considered so unstable by everyone at Ubuntu? 
> I've used it on another Debian distribution and it seems to work more smoothly
> than swsusp (and as a bonus it has a script to suspend the system which
> terminates conflicting process, a nice status inticator, and it doesn't depend
> on power management - so it still works on systems w/broken ACPI).

Because:

a) I think the code is on crack
b) I suspect the upstream author is on crack
c) The kernel developers appear to unanimously think that it's crack
d) swsusp doesn't require working ACPI anyway
e) terminating processes during suspend is an invitation to really,
really irritate people

The two main advantages swsusp 2 has over the in-kernel code are speed
and the fact that it has an utterly insane shell script that draws
pretty ascii-art all over your screen while trying to remove modules.
The first is being sorted (this is probably 2.6.11 timescale), and the
second we can deal with anyway.  

> I really hope this makes Hoary, as my laptop doesn't do ACPI suspend-to-RAM
> correctly - I have battery usage of 10% per hour of suspend (I can use APM OK -
> but then I lose an hour of battery life).

This is a T-series Thinkpad, right? I've a couple of suspicions - I'll
try to get my hands on some test hardware to see what we can do about
it.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org