Re: tlug: Sony VAIO & apm

[Frank Bennett <bennett@example.com>]

Short answers:

  (1) The PHDISK.EXE utility from the VAIO recovery CD can be
      used to re-create a deleted hibernation-mode partition.
      Without this partition, the power saving features of the
      BIOS will behave strangely, so it shouldn't be deleted.

  (2) The kernel for a VAIO 731 should be compiled with "Ignore User
      Suspend" turned "OFF".

  (3) The commands "apm -s" (or "apm --suspend") and "apm -S" (or
      "apm --standby") should NEVER be used on a VAIO.  These
      commands may be the cause of subsequent suspend or standby
      mode failures; they should be treated (on a VAIO 731) as
      EVIL.  Use the handy-dandy "processor snooze" and "disk
      snooze" keys instead; they at least MAY be safe.

Long-winded Posting in Full:

Okay, here's what I've derived from the day's experience (sigh).

I used the PHDISK.EXE utility from the recovery CD for the VAIO
to create the requisite partition at the tail end of the disk.
Once this was put in place, the stubborn spin-down and screen
blanking behavior that I had noted earlier went away, and the
machine began to behave in accordance with the BIOS settings.

However, the machine did lock up on my first attempt to allow a
timed backup, and the "apm" utility would put the machine into
"suspend" (i.e. memory kept alive on battery, with apm -s), but
not "standby" (i.e. memory saved to disk and system powered down,
with apm -S).[1] I also had a keyboard lockup after a restore and
a couple of keystrokes.

Andy Howell suggested that I have a look at the kernel options
relating to APM, which I did.  Two options seemed potentially
relevant to my troubles: Ignore User Suspend; and Enable APM at
Startup.  The former was "OFF" (i.e. don't ignore), and the
latter was "ON" (i.e. yes do initialize).

I recompiled the kernel with the options reversed (i.e. YES
ignore user suspend, and NO don't initialize APM).  The result of
that trial was to disable APM altogether, even via the "suspend"
and "standby" keys on the keyboard.  Oddly, "apm -s" and "apm -S"
still seemed to behave as they had with the previous kernel.

I guessed (correctly, it seems) that the Ignore User Suspend
option was the cause of the disabling of the keyboard options; I
turned that back OFF and recompiled.  I then torture-tested the
hibernate feature (hibernating during a kernel compile) about ten
times with no sign of failure.

I have concluded, tentatively, that "Enable APM at Startup" is
irrelevant to our problems, that "apm -s" and "apm -S" are evil
on the VAIO, and that the hardwired keys are safe.

The machine has also been through one timed hibernation in its
new configuration (which is just the old configuration, minus
Frank attempting to use those two commands).  The shortest fuse
that the BIOS allows for this is 5 minutes, so it will take some
time to work out whether it's truly stable from that angle as
well.

But (as Steve McQueen said in the Magnificent Seven, quoting some
poor fella in Denver who found himself falling out of a
seventh-story window) "So far, so good".


[1] Is this correct, or is the BIOS in the VAIO getting these
backwards?  The words seem to have a meaning opposite to their
natural sense here.

Cheers,
-- 
-x80
Frank G Bennett, Jr         @@
Faculty of Law, Nagoya Univ () email: bennett@example.com
Tel: +81[(0)52]789-2239     () WWW:   http://rumple.soas.ac.uk/~bennett/


----------------------------------------------------------------
Next Nomikai: 20 November, 19:30   Tengu TokyoEkiMae 03-3275-3691
Next Technical Meeting: 12 December, 12:30 HSBC Securities Office
----------------------------------------------------------------
more info: http://tlug.linux.or.jp Sponsors: PHT, HSBC Securities