[ltp] radeonfb sleep patch for kernel 2.6.15

[Guarded Identity]

> Laptops with everything working/sleeping as it should do 4-6 days
> suspend-to-ram on a full batt. You get batts ~45Wh with the R-Models, so
> about 0.46 Watt for 4 days.
>
> I think the 1-Watt-limit is OK.

I agree that the limit isn't so bad... it's just that my capacity isn't very 
close to the 45Wh I originally had, and also, my battery is often only 
partially charged.

Here's two follow up questions I have:

1) I've read that APM might offer significant savings in power lost during 
than ACPI S3.  Is this something that might have been true in the past, but 
not so much these days?  Can APM do better than the 0.50 Watts I burn with 
S3?  I realize that a physical trial on my part is the best answer to this.  
I'll try to get around to tinkering with the kernel sometime.  In the 
meantime, maybe someone has an opinion.

2) Since it seems I'm sometimes in a position to "over-sleep" my laptop.  it 
would be nice to convert my sleep into a Suspend2 hibernation.  I was talking 
to someone in #suspend2 who mentioned that his Acer laptop automatically 
wakes up from an ACPI S3 when the battery reaches a certain limit.  He says 
he couldn't find anywhere in the BIOS that might control this behavior; maybe 
it's some weird Acer-specific behavior.  But this does allow a script to be 
run that can hibernate upon sensing the battery is too low.  Is it possible 
to get Thinkpads to similarly wake up from an ACPI S3 sleep?  If not, then 
can APM be set up to suspend-to-disk if a sleeping computer's battery gets 
dangerously low?

- Sukant