Resume does not work

[Lars Erdmann]

1) running debug ACPI.PSD without any switches 2) single core board with AWARD BIOS, see attached DST file 3) 14.104a SMP debug kernel 4) having PS/2 keyboard available

The system will suspend. The system will never properly resume. The keyboard LEDs will get to state "After restoring 1MB of RAM" (NumLock = oFF, Scroll Lock = ON, Caps Lock = ON) but will not get any further. Nonetheless I can hear the 4 beeps of increasing frequency (where the first beep is shorter then the other 3 beeps, is this observation correct ?).

[Lars Erdmann]

By the way: it's a fairly old MOBO. At best it supports ACPI 1.0 (I seem to remember). It therefore does not support any of the ACPI methods that where invented after ACPI 1.0 was released.

[David Azarewicz]

The keyboard LED thing is not very reliable, but since you saw something means that the PSD got control on wakeup which is correct. The 4 tones should all be the same length, but this also can get changed if something else tries to beep. When the first tone occurs, the PSD is almost done restarting the system, and when the last tone occurs the PSD is completely done. The only thing left to do at this point is call all the drivers to resume. If the system is not functional then one (or more) of the drivers didn't do its job. Do you see anything on the screen?

[Lars Erdmann]

1) on resume I am left with a completely black screen with the (textmode) cursor blinking in the upper left corner
2) the system is completely unresponsive, no PS/2 keyboard, no USB devices whatever

I had been waking up the system via the USB OHCI wakeup capability (my USB mouse supports wakeup cap.). I also tried with the power switch with the same result.

It's a long time I had been working on this but at some point back in time (Pashas driver versions) the video would restore but the MAC driver would be dead (I doubt that driver does any suspend/resume whatsoever).
I also seem to remember that I had modified acpidaemon.exe so that resume would work. Back at the time I had shifted around statements in "acpidaemon.c" in the "DoSuspend" thread. I don't remember exactly.

Curiosity: ACPI spec 4.0a lists in chapter 15.1.6 a whole bunch of stuff that needs to be done to suspend the system. Is this all done through the calls invoked from "DoSuspend" ? And how is it handled that older ACPI versions (like my system with ACPI 1.0b) do not implement all of the methods known in ACPI 4.0a ?

[Lars Erdmann]

"acpistat.exe CheckDrivers" output

[Lars Erdmann]

Forgot to say: I remember (from log outputs from Pashas ACPI versions) that UDF.IFS also registers with APM !!! Even though this does not show up in acpistat output (as it only scans drivers).
And UDF.IFS obviously is not a classic device driver. I don't know if this has any impact.
I will try and temporarily remove UDF.IFS from config.sys. While I am already at it I can also remove the MAC driver GENM320$ (and GENMDB0$) ,which by the way turns out to support APM juding from the acpistat output, and USBD.SYS.

[David Azarewicz]

Actually your problem is probably video. That is something that is still on my list to work on.

[Lars Erdmann]

"acpistat.exe CheckDrivers" output with reduced driver set

[Lars Erdmann]

For the record: I removed UFS.IFS,GENM32.OS2 (the MAC driver), USBD.SYS from config.sys.
Essentially I was running with the "standard" set of drivers (apart from IBM1FLPY.ADD maybe ...)

The result was exactly the same. A black screen on resume with the textmode cursor blinking in the upper left corner.

[David Azarewicz]

There is nothing unique about your system. This is the same problem everyone else has. All "resume does not work" tickets are being consolidated into one ticket #585.