Re: Hibarnation / lphdisk / Destroyed MBR (solved)
Hi Nils Anders, Ceasr and whoever's interested !
Here's the stuff:
Dell I5000 PIII 600 Mhz
Phoenix BIOS 4.0 Release 6.0, Dell-Update revA08
BIOS APM suspend set to 'disk'
12 GB FUJITSU MHK21DAT
Though dma enabled, it's far not as fast as my deskbox seagate.
( but does BIOS hib use dma here ? )
Spread over several weeks, that were the steps to get it work:
1) Created a spare partition of Ram (MB) + VideoRam (MB) + 2 MB,
schemed to future expansions.
2) Configured the kernel for APM
Here are the Kernel 2.4.18 options:
XXXCONFIG_APM=y
# CONFIG_APM_IGNORE_USER_SUSPEND is not set
CONFIG_APM_DO_ENABLE=y
# CONFIG_APM_CPU_IDLE is not set
CONFIG_APM_DISPLAY_BLANK=y
CONFIG_APM_RTC_IS_GMT=y
CONFIG_APM_ALLOW_INTS=y
CONFIG_APM_REAL_MODE_POWER_OFF=y
3) Installed powemanagement-base 1.6 with apmd 3.0.2-1 from stable (woody).
As Thomas H. suggested, apm from unstable might be the better choice.
4) Installed lphdisk and rtfm ;-)
5) With cfdisk, made the partition of type IBM Hibernation 'A0' .
6) After a reboot to reread the paartition table, ran 'lphdisk' which
autodetects the A0 one.
7) Enabled 'suspend to disk' in the BIOS.
It turned out that linux apm didn't regard those settings;
however, i suppose without linux apm BIOS would do the job.
That was it.
Hibernation is released with Fn-q or Fn-a.
Fn-suspend does suspend-to-Ram.
On Win2000 the Fn-q does nothing, instead Fn-suspend key executes the choosen
suspend mode (disk or Ram) correctly.
----------------------------------------------------------
Hibernation speed
Given in [seconds saving | loading ]:
Saving from pressing Fn-q
Loading from pressing the power button
----------------------------------------------------------
Sorry i go disappoint you now :(
I thought it's rather fast cause I compared it to a Win98 which was
preinstalled (now thrown out). Did need far more then a minute.
However, my first try with hibernation was on an service-extensive 'testing'
runlevel without X, just on textconsole after login. Speed was 30 | 20
there. I was happy with that.
Hib shows up a BIOS screen with 3 levelmeters for different kind of data,
two of them very quick filled up and related to BIOS itself, i think.
Same screen is shown up when awaking.
Later i did a test from a gnome2-session with openoffice writer (empty
document): 80 | 30.
That's quite similar to Win98.
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.
To me this all looks as if Win2000 ditched the old BIOS based method at all,
and does it all within it's own partition, making use of all opportunities of
it's full environment to speed up the thing:
- A specific driver and maybe dma at all
( Provided that BIOS only uses implemented slow 'generic' features)
- Comprimizing data, and not simply the whole used Ram but selective data
As you know from my previous posting, i ran into trouble since BIOS
hibernation on an 'ouside' partition immediateley after the MasterBootRecord
( that was hda1, here) prevented the bootloader to do it's work,
giving an error message like 'Insert the original HD and reboot'....
.
It's a grub here with 'grub-install /dev/hda' settled directly into the mbr.
After i made a 7 MB 'spacer' between mbr and hda1, it works fine.
I am wondering if the error message could have been been from grub, not from
the BIOS. I won't investigate it furthermore, since with experimental stuff
i had a lot of filsys crashes here at all, and it's not my own lap at all.
Have to go on, now.
However, it would be interesting to know if it works with a usual linux hda1
partition and the hib partition elsewhere. Additional, if it would have
worked with hib on hda1 and a Windows9x specific bootloader.
I hope i can use Hibernation to prevent battery powerfail crashing. Just this
moment I'm checking out if this will work here ( it's at 20 % now ;-).
I'll post some related stuff to another topic.
See 'APM and Hibernation on battery power'
-- micha.
Reply to: