[ltp] Reliable Suspend/Resume (Was:Filesystem choice?)

[Michael Selway]

Richard Griffith writes:
 > Recently I was reading some recommendations about adding a kernel option
 > (CONFIG_APM_ALLOW_INTS) which sounded similar, and worth looking into. A
 > quick check in the kernel source showed that yes there was such an
 > option, but better still, you can pass an option at boot time that will
 > do the same thing! Much quicker that recompiling the kernel (I love my
 > 760xd, but speed is not its primary virtue.) By adding the option
 > 'apm=allow_ints' to the boot, my problems seem to have disappeared.

I don't have a 2.4.9 kernel source to hand at the moment, but in
2.4.18, the CONFIG_APM_ALLOW_INTS setting is ignored when you boot
a thinkpad, as is the kernel boot time options no-allow-ints and
allow-ints.  If you see a message during boot saying

  IBM machine detected. Enabling interrupts during APM calls

that means it's forcing the allow-ints parameter ON.  At least
according to my reading of the code.  I've been experimenting with
overriding this (by editing the kernel source - see
arch/i386/kernel/dmi_scan.c) and running with no-allow-ints.  My
conclusion is that it makes no difference on my t21.  Either
setting works just as well.  And I still fail to recover from
resume a couple of times a month.


I agree with your comments on pcmcia, although again, my
conclusion is that it isn't the root problem here.  I now do not
use pcmcia cards and I still have the resume problem.


I'm interested in your comments about external power.  For me, at
one time, plugging/un-plugging external power lead *during
shutdown* increased the chances of a crash during resume.
However, that is no longer true.  I have now had crashes when the
suspend and the resume were carried out far from any external
power; and I've suspended many times while plugging/un-plugging
the power lead and it's resumed fine.  I think this might have
been related to the pcmcia card which was certainly making resumes
fail frequently under 2.4.18.

If D.Sen is right, I've got dodgy hardware and I need to wait
until something else fails and then get IBM to replace the
motherboard.  Not good.

Hmmmm.
Michael.