Re: grub 0.90 prevents Standby in Windows

[dman]

On Fri, Oct 26, 2001 at 11:50:57AM -0700, address@hidden wrote:
| 
| dman <address@hidden> wrote:
| 
| > On Thu, Oct 25, 2001 at 05:03:59PM -0700, address@hidden wrote:
| > ...
| >  
| > | I think it's the presence of APM checking code in GRUB itself 
| > ...
| > | The current version of GRUB appears to use an APM checking/initialization
| > | routing similar to that of the Linux kernel's setup code.
| > | 
| > | I know for a fact that the Linux kernel's APM setup code is busted in
| > | the sense that it leaves the APM in a "connected" state, and the
| > ...
| > ...
| > | it will cause random crashes for no apparent reason.
| > ...
| > 
| > These quoted lines here struck me as particularly interesting.  I use
| > a Dell Inspiron laptop with grub (not sure on version, probably 0.90)
| > and linux (2.4.8).  Sometimes when I close the cover (suspend-to-ram)
| > then open it later I get lots of "hda: lost interrupt" messages and
| > all I can do is hard-reset the machine.  Could this be caused by the
| > above mentioned interactions?  I have the APM stuff in the kernel and
| > I like it because the machine shuts off automatically when I tell it
| > too.  If this is indeed (or might be) the problem, how can I correct
| > it?  Do I simply need to remove the apm sutff from my kernel?
| 
| Hmm.  There are machines with broken APM, but in general the Linux
| kernel is supposed to deal with that decently.
| 
| You may want to try the patch I just posted and see if it solves
| your problem (i.e. maybe it was confusing Linux's APM code a bit...
| who knows?  though I tend to think not if it worked at all).
| 
| The problems I mentioned above were with:  APIC/IO-APIC use along
| with APM, and with SMP/APIC/IO-APIC and APM, on only some machines
| (for example, disabling the APIC for main interrupt delivery on those
| SMP machines fixed the problem).
| 
| So, if the patch I sent doesn't resolve your problem (and like I said,
| I tend to think it won't), then see if you're using the UP-APIC and
| UP-IO-APIC configuration in your kernel, and disable them, then try
| that.

This is a uniprocessor machine, and I no nothing about *APIC stuff.
I'll have to read up on it sometime.

| If you disable APM, I think bad things might happen if you use suspend/
| resume.  I've never tried it, but there's a reason they have APM support
| in the kernel.

Closing the cover on the laptop causes the BIOS to put the machine in
suspend mode.  The BIOS only allows "Suspend-to-RAM" or
"Suspend-to-Disk".  The suspend-to-disk requires a primary partition
and takes so long to restore (256MB RAM, 4MB video) that a reboot is
better.  (I'd rather it just blanked the display and left the machine
on, but oh well)  I don't do anything special with linux itself as far
as APM goes other than to turn off the machine when I run
'/sbin/halt'.

Thanks for your comments!
-D