On Mon, 2005-11-07 at 08:35 +0000, Ted.Harding@nessie.mcc.ac.uk wrote:
I can imagine, for instance, that when someone puts Windows on a machine for the first time, with APM set up, the Windows installation simply makes sure that the end of its formatted partition space stops short of the physucal end of the disk so that there's a kind of anonymous Limbo into which Windows simply squirts a binary image of the system state on hibernation (sort of 'dd' equivalent).
It depends on the version of Windows we are talking about, Windows XP creates a file c:\hiberfil.sys which is where the memory contents are written to, therefore no special partitions are needed. I think Windows 2000 is the same.
Windows ME (if anyone out there is daft enough to run it) had some hibernation support but needs extra help from all the drivers, Usually unless you used only Microsoft certified drivers the hibernation feature will be silently disabled and a file (nohiber.txt) will be written into c:\windows telling you which driver is the cause.
Windows 98 as far as I can remember had some built in hibernation support but it didn't work very well so often manufacturers fell back to the BIOS and special partition and then had a special hibernation driver/utility.
Windows 95 and before had no direct hibernation support and needed the special partition and BIOS support you sometimes see on older laptops.
Linux (I presume) just writes to the swap partition ? I wonder how this works if swap is heavily utilised at the point of hibernation and there is not enough free to write the memory contents ?