The officially official Devuan Forum!

You are not logged in.

#1 2022-10-12 20:50:00

pcalvert
Member
Registered: 2017-05-15
Posts: 129  

ACPI 2.0 question

I have a desktop computer with an Intel-based ASUS motherboard from 2009 that has the old-style BIOS. I was recently looking through the settings and noticed that ACPI 2.0 is disabled by default. Should I enable it or leave it the way it is? The computer shuts down, hibernates, and suspends normally, so I am inclined to leave it the way it is, but if there is a good reason to enable it, I will.


“Your only limitations are those you set up in your mind, or permit others to set up for you.”
— Og Mandino

Offline

#2 2022-10-13 01:45:14

Altoid
Member
Registered: 2017-05-07
Posts: 1,107  

Re: ACPI 2.0 question

Hello:

pcalvert wrote:

... noticed that ACPI 2.0 is disabled by default.
Should I enable it ...
... computer shuts down, hibernates, and suspends normally ...

If everything works properly ...  8^°

But just for curiosity's sake, you can have a good look at your dmesg printout and see if there are any ACPI errors.
Run this and see what you get:

sudo dmesg | grep -i "error\|warning\|fail\|segfault\|fatal"

Bear in mind that some ACPI errors are a permenent fixture in older boards but the kernel takes care of it:

user@devuan:~$ sudo dmesg | grep -i "error\|warning\|fail\|segfault\|fatal"
--- snip ---
[  0.021101] ACPI BIOS Warning (bug): 32/64X length mismatch in FADT/Gpe0Block: 128/64 (20200925/tbfadt-569)
[  1.318080] ACPI BIOS Error (bug): Could not resolve symbol [\_SB._OSC.SUPP], AE_NOT_FOUND (20200925/psargs-330)
[  1.318088] ACPI Error: Aborting method \_SB._OSC due to previous error (AE_NOT_FOUND) (20200925/psparse-531)
--- snip ---
user@devuan:~$ 

I don't use suspend or hibernate and as my Sun Ultra 24's BIOS is an absolute POS, I have to bear an ocassional, totally random and non-reproducible shutdown issue which I think is related to some internal temp sensor and how it interacts with the box's ventilation.

Best,

A.

Offline

#3 2022-10-20 21:23:54

pcalvert
Member
Registered: 2017-05-15
Posts: 129  

Re: ACPI 2.0 question

Thanks for the suggestion. It tried it and most of the messages are related to missing firmware. A couple are related to ACPI, though:

# dmesg | grep -i "error\|warning\|fail\|segfault\|fatal" | grep -i acpi
[    0.531264] acpi PNP0A08:00: _OSC failed (AE_NOT_FOUND); disabling ASPM
[    3.900266] ACPI Warning: SystemIO range 0x0000000000000828-0x000000000000082F conflicts with OpRegion 0x0000000000000800-0x000000000000084F (\PMRG) (20200925/utaddress-213)

“Your only limitations are those you set up in your mind, or permit others to set up for you.”
— Og Mandino

Offline

#4 2022-10-20 22:35:18

Altoid
Member
Registered: 2017-05-07
Posts: 1,107  

Re: ACPI 2.0 question

Hello:

pcalvert wrote:

Thanks ...

You're welcome.

pcalvert wrote:

... most ...  ... are related to missing firmware.

All the hardware works properly?

pcalvert wrote:

... couple are related to ACPI ...

Yes, I'd say those are more or less standard.
Like this one I have in my box:

[ 1.324984] acpi PNP0A08:00: ignoring host bridge window [mem 0x000d0000-0x000dffff window] (conflicts with Adapter ROM [mem 0x000ce000-0x000d3bff])
[ 1.327999] pci 0000:00:1f.0: quirk: [io  0x0800-0x087f] claimed by ICH6 ACPI/GPIO/TCO

This type of memory conflict is due (so I have had explained to me) to a badly written BIOS.

As the kernel evolves, these messages/warnings also evolve/change.
In time, some messages end up going away.

This is all crud generated by the fact that OEMs and BIOS writers are ruled by what MS wants and nothing else.

But this should not happen, much less in a box like mine, a Sun Microsystems Ultra24 WS which was released 15 years ago (10/2007) with the most basic configuration going for ~ US$1.000, a very hefty price tag at the time. (I purchased it used in 2011, of course)

OEMs (Sun included) didn't/don't give a monkey's toss if the BIOS they slapped together from bits and pieces of other older ones does or does not comply with what ACPI says in their specifications, as long as it certifies for hardware that ships with Windows OSs.

And if they have to do it the MS way because the ACPI way does not play ball with the currently shipping Windows OS, they just do it.
Otherwise their hardware will not be "Windows Certified", which means that it won't sell.

That's about it, you have to endure it.
Till we get open source BIOSes and that will be the end of it.

But I really don't see it coming any time soon.

Best,

A.

Offline

Board footer