I really want and need a shutdown command to turn off the Pegasos from MorphOs instead of keeping down the power button for 4 secs... Does the command exist any any other form than "wait for MorphOS 1.5" ??

No.


- CISC

You can shutdown from the firmware, though.
It is possible to put a shutdown entry in your "boot" menu.

Out of interest, whats the technical reason for it not be impletented? Short version will do

Quote:

MorphOS 1.4 doesn't support it.


- CISC

Oh how much you're right :(

I simply solved this annoyiance by powering off with
the main switch (well we're not under windows so we
can ;)

After all, perhaps there's some location in the PCI address
space where to rudely POKE some value to have the
system shutting down B-)

Quote:

You are completely right but this doesn't work for some strange reason, i've tried it. I've succeeded in accessing 8235 ACPI controller, i can even detect "power" button presses , but when i try to switch to soft-off state as described in the datasheet it simply doesn't work. The system simply ignores setting to '1' that bit. I've studied the question more carefully and came to a conclusion that before doing this i need to set up interrupts state and do some other machine-dependent magic which i can't know. So this probably really can't be done. For those who want to learn more, read ACPI specifications and Linux kernel ACPI driver source code.

[ Edited by Sonic on 2006/7/3 8:47 ]

Linux (eg.Suse 10) can shutdown the system without probs (speaking of a Peg2, not Peg1).

I'm not enought in the world of powermanagement, to understand that, but for a working example this might help ...
bye

Quote:

No, this can't help.
Linux on PPC doesn't use ACPI because in fact ACPI is Intel-only thing. The fact that Pegasos has 8235 ACPI controller doesn't say that Pegasos implements ACPI standard. There are no configuration tables.
Linux on CHRP uses RTAS for rebooting and shutting down. RTAS is a component of OpenFirmware and in order to use it you need to be able to call OpenFirmware. Under MorphOS this is impossible since MorphOS of course doesn't present OpenFirmware entry point address to the public.

Cool. Will MOS ever have such OF entry point?

What about OF entry point bounty ? :)

Quote:

Quote:

No, MorphOS kills OF at init.

MorphOS 1.5+ has an appropriate API for shutdown.


- CISC

@CISC

> No, MorphOS kills OF at init.

Why does MorphOS need to kill the OF at all?

Quote:

Because it's a waste of space and there's no point in having it around.


- CISC

Yeah, so we have a real KICK-start! :D

Thanks to Sonic for having tried my stupid hint...
I told is almost for kidding, even if not without
a slight hope of success

I seem to understand that power management is a bit
messy :( So once again let's wait desperately for
this 1.5 to come........

Oh, warm reboot would be nice aswell (OS4 does have it).
A pity that a fast system like MOS cannot have a
fast reset too.

&

by CISC on 2006/7/3 10:11:50

Quote:


Why does MorphOS need to kill the OF at all?



Because it's a waste of space and there's no point in having it around.


- CISC

So does Mos 1.5 keep the OF alive after boot up, or does it use a different technique? Also would it be possible with Mos 1.4 to use the RTAS route that Linux uses? or does it need the OF active to use this technique as well??

Quote:

Ofcourse it doesn't.

MorphOS gets all the necessary info from OF before killing it and runs its own RTAS-server (qbox), however 1.4 simply doesn't support shutdown, like I must have said a million times already in this thread. ;)

Quote:

No, that won't work, just as Sonic experienced (instead of believing us in the first place).


- CISC

Is it possible to bypass all the OF and OS-stuff and talk directly to
the chipset to get an reboot or shutdown ?
There must be some bit which pulls the plug of the maching...

"Oh, warm reboot would be nice aswell (OS4 does have it).
A pity that a fast system like MOS cannot have a
fast reset too.
"
exactly! i was so jelaous when i saw os4 warm reboot for the first time! my morphos experience is, that i have to reboot it so many times (because of crashed apps or annoying shellls on desktop), and it is not so comfortable to wait the whole boot process every time (but of course its still 10-15 secs, which is far better than winsuxx' bootime or x0secs hang ups)

is it posible to have warm rebbot on mos 1.4 (or 1.5 at least)?

a stupid question to core mos devs: what is the amount of money do you need/want to release mos 1.5? or you dont even care 'bout $ and theres not such amount of $ on earth?:)

@ CISC,

I guess I didn't explain my question well enough...but you answered it
well... I just wanted to know why... isn't enough to tell me that it
doesn't work... I'm the guy who will always ask why...

Saddam,

I really wonder what's so cool about a warm reboot hack that works only maybe 8 times out of 10 (depending on the moon phase).

Quote:

Quote:

No, and I believe Sonic already answered that question .. read the thread before you ask questions please. ;)

Quote:

Something like that .. I guess time will tell...


- CISC

Quote:
Read my message before. This is exactly what i was trying to do and this fails. Setting this bit simply does nothing because some machine-specific magic should be done before and there is no information about this. If you read the specs: you need to deactivate all SMI and SCI events and activate only some specific wakeup events.

If there are more non-beleiving people i could publish my experimental source code here. Try this yourself.

[ Edited by Sonic on 2006/7/4 8:38 ]

This is an unexpectedly interesting thread! Not only from the technical side, also from the fact that all is about a tiny function that everyone else has, and that, actually, should never be used if thinghs work fine.
MorphOS is different, isn't it.
I'm a bit worried about the fact (which I didn't know until now) that MorphOS doesn't use SmartFirmware at all. It means that, counter to Genesi's philosophy, it won't work on different computers that are able to run SmartFirmware. Which is very "amigan", by the way.
MorphOS is, in all, very interesting. It's a pity that their main programmers can't find much more resources in order to give users a new release. O r perhaps, releasing to the users is not really important for them! Sure, being inside the development team must be fantastic. But there are not many people with the required capabilities.
How about a MorphOS team member course (and exam!)?