[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Hibarnation / lphdisk / Destroyed MBR (solved)



On Thu, 2003-02-13 at 00:50, mi wrote:
> Hi Nils Anders, Ceasr and whoever's interested !

Thanks a lot for the information.  It is pretty much what I expected.  I
think I won't bother with hibernation yet --not with BIOS hibernation
anyway.  If my machine behaves as yours, I can reboot faster than a
suspend/resume cycle :-/

> Amazing was the comparison to the new-installed Win2000 (with  an empty word 
> doc ):  10 | 10.
> This is definiteley another mechanism, though.
> It doesn't show up the BIOS hibernation screen at all.
> Waking up by pressing the power button starts up grub as uusual !
> Then, after choosing Win2000 from the bootmenu, 
> it reinitializes the session very quick (else needs about 3 minutes to come 
> up). 
> btw. In Win2000 there's is a menu chooser for hibernation which shows up the 
> amount of free spacce on C: compared to the needed space for hib.

I'm almost sure that W2K implements hibernation in its kernel, bypassing
the BIOS.  There is a project for doing the same on Linux:

  http://falcon.sch.bme.hu/~seasons/linux/swsusp.html

There are a number of advantages with this approach:

 - You couldn't care less if the BIOS supports hibernation, or does so
   in a broken, stupid fashion (my laptop's case).  Actually, you should
   not even need APM/ACPI, except for powering down without user
   intervention.

 - You do not need to waste disk on a hibernation partition/file.  It
   uses the swap space for hibernation.

 - Suspending/resuming is much faster: you do not need to write to disk
   pages already swapped out, nor restore them on resume; and you do not
   need to dump memory that is not in use.

The page includes patches for kernels 2.4.10 and 2.5.1 (among other,
older, versions).  I don't know if it would work on a debianish 2.4.18
or 2.4.20, but I think I'll give it a go when I can spare some time. 
That could mean a long wait, though; and it may not work at all, or
break other stuff.  It still looks quite experimental.

Anyway, if someone tries it before (hint, hint, Michael :-), I would
*greatly* appreciate a full account of the adventure.

Best regards,

 -CR




Reply to: