Is XP hibernate flag stored in MBR?

[Mark M]

I had some unusual system behavior which made me think about where
WinXP Pro stores the flag(s) which tell it whether to boot into a
hiberanated system or to boo tinto a fresh system.

Is the hibernate flag stored in the MBR or somewhere outside the XP
partition?

 

[Jim]

AFAIK, hibernation is nothing special, heck, I can hibernate W2K Pro on my 6
yr old AT machine (w/ only ATM support, not even ACPI) without a problem.
All hibernation does is dump your memory to a file, and upon rebooting, W2K
(or WinXP) sees the hibernation file on the root C:\ and uses it to restore.
I'm no expert on the topic, but given the fact it works on my old AT system
w/ W2K under ATM, I suspect there really isn't all that much too it. The
fact it's listed under "Power Options" I think is misleading since it
implies some dependence on ATM or ACPI functionality, but given how severely
limited my own AT machine is in this area, the fact hiberation works even
for me leads me to believe the process has only the simplest of
requirements, a) enough storage to save the hibernation file and b) perhaps
a means to idle the CPU so a snapshot remains vaild throughout file creation
(i.e., a function of the CPU). But all and all, seems to me the process is
incredibly simple, much more so than people suspect.

So certainly nothing to do w/ the MBR, nothing do to w/ the BIOS, maybe
nothing all that much to do w/ the registry other than for bookeeping
purposes. The OS may merely look for the hibernation file on reboot and
decide THEN AND THERE to restore from hibernation!

HTH

Jim

 

[J.Clarke]

I still don't think it has anything to do w/ APM/ACPI, it may indeed
read the registry to trigger the process, or there may be some flag in
the hiberate file itself (nothing says it has to be a pure copy of
memory, it may in fact be structured), but just thinking about it
logically and observing the behavior, it's hard to imagine WHAT would
require APM/ACPI dependence. Again, I'm not claiming expertise here,
merely observing. Consider what happens, the CPU is (probably) idled,
the memory contents copied to a file, (maybe a flag set), and
shutdown. On reboot, the process is reversed. It's not like other
power options that need, for example, to rely on the BIOS/APM/ACPI to
"wake" a PC from sleep, e.g., WOL (Wake On LAN) network adapter, or
switch to Standy. All of these involve true "power management"
directly, the quiescing of various components, and support by the
hardware (mobo, power supply, and OS) to make it possible. I see
NOTHING in the act of hibernation that requires any of these features.
It's
just an "image", a snapshot of memory.

Not quite--it also has to store the CPU state. And in order to work
reliably it needs to pretty much shut down everything except disk I/O so
that the state doesn't get changed during the write.

That said, I suspect that the only way APM is involved is that hibernate
is one of the options available for a timed shutdown.

 

[Jim]

I still don't think it has anything to do w/ APM/ACPI, it may indeed read
the registry to trigger the process, or there may be some flag in the
hiberate file itself (nothing says it has to be a pure copy of memory, it
may in fact be structured), but just thinking about it logically and
observing the behavior, it's hard to imagine WHAT would require APM/ACPI
dependence. Again, I'm not claiming expertise here, merely observing.
Consider what happens, the CPU is (probably) idled, the memory contents
copied to a file, (maybe a flag set), and shutdown. On reboot, the process
is reversed. It's not like other power options that need, for example, to
rely on the BIOS/APM/ACPI to "wake" a PC from sleep, e.g., WOL (Wake On LAN)
network adapter, or switch to Standy. All of these involve true "power
management" directly, the quiescing of various components, and support by
the hardware (mobo, power supply, and OS) to make it possible. I see
NOTHING in the act of hibernation that requires any of these features. It's
just an "image", a snapshot of memory. As far as the system is concerned, a
clean reboot or reboot to a hibernated file appear the same.

Again, I could be wrong, but I just don't see how APM/ACPI are involved.
The clincher for me was when I was able to hibernate W2K on my old VA-503+
motherboard w/ K6-2+ 450 CPU and AT power supply using APM. You can get
much more lowly in the power management arena than that. That convinced me
this was all a function of the OS, nothing more.

Jim