BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0

[Leo Milano]

Binary package hint: acpi

I am using a Foxconn A7GM-S 2, with the newest BIOS: ID 872F1P02, Version 08.00.14 - At boot up, I get the following message:

####
[ 5.824140] [Firmware Bug]: powernow-k8: Your BIOS does not provide ACPI _PSS objects in a way that Linux understands. Please report this to the Linux ACPI maintainers and complain to your BIOS vendor.
####

Cool n' quiet is enabled in the BIOS. Here is the relevant info from lshal:

####
  system.board.product = 'A7GM-S 2.0' (string)
  system.board.serial = 'UY30907021058' (string)
  system.board.vendor = 'FOXCONN' (string)
  system.board.version = 'FAB-A' (string)
  system.chassis.manufacturer = 'FOXCONN' (string)
  system.chassis.type = 'Desktop' (string)
  system.firmware.release_date = '12/16/2008' (string)
  system.firmware.vendor = 'American Megatrends Inc.' (string)
  system.firmware.version = '080014' (string)
  system.formfactor = 'desktop' (string)
  system.hardware.primary_video.product = 38416 (0x9610) (int)
  system.hardware.primary_video.vendor = 4098 (0x1002) (int)
  system.hardware.product = 'A7GM-S 2.0' (string)
  system.hardware.serial = '' (string)
  system.hardware.uuid = '797F273E-1DE4-72C1-F33A-00226850C4D9' (string)
  system.hardware.vendor = 'FOXCONN' (string)
  system.hardware.version = 'FAB-A' (string)
  system.kernel.machine = 'i686' (string)
  system.kernel.name = 'Linux' (string)
  system.kernel.version = '2.6.28-11-server' (string)
  system.kernel.version.major = 2 (0x2) (int)
####

This happens in both 32 and 64 bits. This is not a kernel bug, but a BIOS bug. This has been well established. See here for a more generic bug report. See also below, it seems as if Foxconn is currently (2009 mid August) working on this.

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/251338

[Leo Milano (lmilano)]

Just an update: I reported this upstream and the bug report got quickly rejected as "invalid".

http://bugzilla.kernel.org/show_bug.cgi?id=13146

[brian mullan (bmullan)]

I've encountered the same bug but I'm using the Gigabyte MA78GM-s2h motherboard, AM2+.
For me I've only seen this if I've created a USB bootable 9.04 image using either the Ubuntu USB Startup Disk Creator -or- the PenDriveLinux tools.

In either case, using same flash media and procedures, 8.10 and 8.04 USB images boot fine. However, using Jaunty 9.04 the USB image starts to boot, gets to the spash screen then crashes with this error and ends in the BusyBox.

[Leo Milano (lmilano)]

Brian, this BIOS bug should not prevent you from booting, it just makes the powernow-k8 exit with an error, so you don't get cpu scaling. If you'd like to contribute to the report, please look at the info I provided above. What matters is the BIOS info. Just to make sure, you may try this (acpi=off is good for a start)

http://www.lesswatts.org/projects/acpi/debug.php

What's preventing you from booting from USB might be the same that's happening to me: an issue with the USB modules. Try this: boot in Recovery mode. If/when you see that after a few seconds the login is halted, remove and reinsert the usb drive. This happens to me at around 5 seconds. Reinserting the pendrive allows me to keep booting (the usb media is now properly recognized). But let's keep this bug report on topic. This USB issue must be reported separately. If you do, please notify me here and I'll post in the other bug report.

[brian mullan (bmullan)]

Leo

Thanks for the email...

First, I'll try turning off ACPI and see if it makes a difference and will email results later (hate to have to make that change tho to get this to work - if it does)

Second... so far its ONLY with Jaunty 9.04 that this happens and it does crash out into the busy-box.

If I use Jaunty 9.04 to create a USB thunb drive but use say Intrepid 8.10 or Hardy 8.04 --- the USB boots fine.

Brian Mullan
cell 919-345-5600

________________________________
From: Leo Milano <email address hidden>
To: <email address hidden>
Sent: Tuesday, April 28, 2009 8:00:48 AM
Subject: [Bug 363868] Re: BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0

Brian, this BIOS bug should not prevent you from booting, it just makes
the powernow-k8 exit with an error, so you don't get cpu scaling. If
you'd like to contribute to the report, please look at the info I
provided above. What matters is the BIOS info. Just to make sure, you
may try this (acpi=off is good for a start)

http://www.lesswatts.org/projects/acpi/debug.php

What's preventing you from booting from USB might be the same that's
happening to me: an issue with the USB modules. Try this: boot in
Recovery mode. If/when you see that after a few seconds the login is
halted, remove and reinsert the usb drive. This happens to me at around
5 seconds. Reinserting the pendrive allows me to keep booting (the usb
media is now properly recognized). But let's keep this bug report on
topic. This USB issue must be reported separately. If you do, please
notify me here and I'll post in the other bug report.

--
BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0
https://bugs.launchpad.net/bugs/363868
You received this bug notification because you are a direct subscriber
of the bug.

[brian mullan (bmullan)]

Ok tried turning off ACPI but no difference.

However, using Jaunty 9.04 as the installed bootable USB flash drive image... I can now get it to boot IF during the splash screen I briefly pull the USB flash out and then reinsert it.

If I don't pull the flash and reinsert... the boot fails into the busy box with the ACPI error message.

Again, Intrepid and Hardy images on the USB boot flash do not do this only Jaunty 9.04 seems to.

Brian

________________________________
From: brian mullan <email address hidden>
To: Bug 363868 <email address hidden>
Cc: Brian Mullan <email address hidden>
Sent: Tuesday, April 28, 2009 5:33:53 PM
Subject: Re: [Bug 363868] Re: BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0

Leo

Thanks for the email...

First, I'll try turning off ACPI and see if it makes a difference and will email results later (hate to have to make that change tho to get this to work - if it does)

Second... so far its ONLY with Jaunty 9.04 that this happens and it does crash out into the busy-box.

If I use Jaunty 9.04 to create a USB thunb drive but use say Intrepid 8.10 or Hardy 8.04 --- the USB boots fine.

Brian Mullan
cell 919-345-5600

________________________________
From: Leo Milano <email address hidden>
To: <email address hidden>
Sent: Tuesday, April 28, 2009 8:00:48 AM
Subject: [Bug 363868] Re: BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0

Brian, this BIOS bug should not prevent you from booting, it just makes
the powernow-k8 exit with an error, so you don't get cpu scaling. If
you'd like to contribute to the report, please look at the info I
provided above. What matters is the BIOS info. Just to make sure, you
may try this (acpi=off is good for a start)

http://www.lesswatts.org/projects/acpi/debug.php

What's preventing you from booting from USB might be the same that's
happening to me: an issue with the USB modules. Try this: boot in
Recovery mode. If/when you see that after a few seconds the login is
halted, remove and reinsert the usb drive. This happens to me at around
5 seconds. Reinserting the pendrive allows me to keep booting (the usb
media is now properly recognized). But let's keep this bug report on
topic. This USB issue must be reported separately. If you do, please
notify me here and I'll post in the other bug report.

--
BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0
https://bugs.launchpad.net/bugs/363868
You received this bug notification because you are a direct subscriber
of the bug.

[Robin.He (hechu)]

Hi, guys:

I have the same problem when I try to boot from a boot-able USB key and install a new ubuntu 9.04 system. It drop me to a busy box, and installation can not continue. I tried i386 and amd64 ISO, it's the same problem.

My motherboard is JETWAY HA03, used AMD 780G chipset, AMD AlthonX2 4800+ CPU.

An interesting thing is: before I decide to re-install a brand new system, I've upgrade my 8.10 to 9.04, and the system which I upgrade from 8.10 is working very fine.

Anybody know how can I skip this problem and install the system from a USB key?

Thanks.

Robin.He
2009/05/23

[Robin.He (hechu)]

Hi, guys:

Sorry I did not read the older post carefully.

It seems, Brian's temporary solution works very well... I don't know why, but it DID work.

Again, sorry to trouble you guys who's in the mail list.

Best wishes.

Robin.He
2009/05/23

[brian mullan (bmullan)]

One work-around I have found to this was alluded to in an earlier email in t his thread...

It worked for me when this happens.

If on booting your Ubuntu 9.04 release that you've installed on Bootable USB...

If during the boot process it gets to the Ubuntu "splash" screen showing the left/right orange/red flash
and seems to stop... pull the USB out for just a second and reinsert it.

For those USB's where I was getting the error mentioned... this fixed it. My guess is that the bug
is either a timing related bug or an interrupt issue.

 * Timing bug could be differences in single core, dual core, quad core my machine this happens on is a quad core... on dual core laptop I haven't see the problem.
 * If its an IRQ type of problem then pulling the USB flash and reinserting it would make some sort of state change which maybe that starts up the process again..In either case, for me, there was a workaround.

After pulling then reinserting the USB the Ubuntu "splash" screen would start to show the final displays prior to Ubuntu being fully functional.

Brian

________________________________
From: hechu <email address hidden>
To: <email address hidden>
Sent: Friday, May 22, 2009 2:06:57 PM
Subject: [Bug 363868] Re: BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0

Hi, guys:

I have the same problem when I try to boot from a boot-able USB key and
install a new ubuntu 9.04 system. It drop me to a busy box, and
installation can not continue. I tried i386 and amd64 ISO, it's the same
problem.

My motherboard is JETWAY HA03, used AMD 780G chipset, AMD AlthonX2 4800+
CPU.

An interesting thing is: before I decide to re-install a brand new
system, I've upgrade my 8.10 to 9.04, and the system which I upgrade
from 8.10 is working very fine.

Anybody know how can I skip this problem and install the system from a
USB key?

Thanks.

Robin.He
2009/05/23

--
BIOS does not provide ACPI _PSS objects in a way that Linux understands, Foxconn A7GM-S 2.0
https://bugs.launchpad.net/bugs/363868
You received this bug notification because you are a direct subscriber
of the bug.

[Leo Milano (lmilano)]

Hi Everyone

Please see my post #5:
https://bugs.launchpad.net/ubuntu/+source/acpi/+bug/363868/comments/5

This USB boot issue has nothing to do with the powernow-k8 / ACPI bug we are tracking here. I think it is a bug (and also a regression), and proposed a workaround in that post, just in passing, to help Brian.

Brian, Denis, or anyone having the USB boot-up issue, we need to report this separately, if anyone has the time please do, and probably just post a link here so anyone landing here by mistake goes to the right place :-) I can update the bug summary and clarify that, too.

Cheers
Leo

[Ivar Snaaijer (ivar-snaaijer)]

Same problem with Foxconn A74MX-K Bios dd 07/03/009 (r06) It did work before.
I have updated the Bios, then it stopped working. Restoring the BIOS to the original (backup) does not help.
I can not say for sure if the BIOS update did or did not coincide with a Kernel update.
I also send a question to Foxconn

[Leo Milano (lmilano)]

Hi Ivar

Regarding your observation: what kernel are you running now? It would be extremely beneficial if you could go back to an old kernel that works. What you are saying seems to mean that even if there is a BIOS bug (as claimed by the kernel acpi devs), scaling used to work for you in Linux. You could report it here:

http://bugzilla.kernel.org/show_bug.cgi?id=13146

On the other hand, if you don't give them away to reproduce it, it is pretty much useless from their perspective. But just for them to know might ring a bell. You would need to be a lot more specific though.

By the way: the kernel devs think now changed the error message in the kernel, and they send people to talk to their mobo manufacturer (they are _convinced_ this is a mobo bug; however, your case makes me wonder).

Best,
Leo

PD: My conversations with Foxconn unfortunately ended up in nothing useful. I spent a month exchanging messages in an online form (which seems to be the only means of communication with them). After lots of trivial exchanges (I never got them to understand what is going on despite my very technical and lenghty explanations), they sent me to download and apply the "amd software". Surprise: the "software" was the powernow-k8 module, which was posted by AMD a long time ago but has of course long been in the mainline tree. I asked them if they ever read what I was saying, since the very summary of the report I initiated with them has a log from powernow-k8. At that point, they stopped communication from their side. That was a month ago. If yhey don;t answer, you can't post again until they do. Needless to say, that was my last Foxconn product. But if you also bug them, maybe some day they'll do something.

[Ivar Snaaijer (ivar-snaaijer)]

I have tries one of the older kernels i still have running but they do not work anymore. so it must be the bios upgrade. darn update did not even solve the (unrelated) problem I had.
I have now also got access to the forum, and i'll try. but they seem quite MS centric, so i'm afraid I'll end up frustrated too, thanks for the warning.

[Ivar Snaaijer (ivar-snaaijer)]

I got his reply on 2009/07/09 :
> Please flash the BIOS to the previous version (R06)at present,
> we'll forward this information to our relative dept.

Patience seems to be a virtue ..

[Leo Milano (lmilano)]

Thanks, Ivar!

My request is still on stand-by (from the last time I reported). I may open a new "incident" with them, pointing out they left the other one unanswered. This is very frustrating.

In the meantime, can you flash the mobo in Linux? There is a new BIOS for my mobo. ASUS mobos typically can be flashed inserting a fat16 pendrive with the new firmware, and holding a special key combination at reboot. Maybe we should open an Ubuntuforums thread to discuss this sort of thing.

[AdrianBartel (bartelar)]

It may be worth checking whether there's an option to enable Cool 'n Quiet in the BIOS. Such an option may have been disabled after a BIOS upgrade.

This is discussed here for a different motherboard with similar symptoms:
http://ubuntuforums.org/showthread.php?p=7636364

[Ivar Snaaijer (ivar-snaaijer)]

re #16 : I have created a FAT partition on the harddrive and use a Freedos Bootdisk.
re #17 : Thanks, this was the reply from the FOXCONN people too, They currently nderstand it is a BIOS problem and seem to be willing to fix it. I just checked and there is a new version of the bios (07) (date 8/18/09). It fixes a Vista problem when Cool n' Quiet is disabled. I hope they have rumaged enough to fix the problem for Linux. I'll upgrade my BIOS and give Feedback.

[Leo Milano (lmilano)]

@Adrian: thanks, it is enabled. I pointed this out to the Foxconn representatives.
@Ivar. thanks for the update. Let's see how it goes with your upgrade.

Now I found this very relevant (and related) bug report:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/251338

Also this long thread:
http://ubuntu-virginia.ubuntuforums.org/showthread.php?t=869249

To put it straight, Foxconn has intentionally damaged the BIOS ACPI functionality for Linux (by putting OS dependent if() switches mostly to support windows, when the ACPI is a standard), and it took reverse engineering to unveil that. ONLY after that was made PUBLIC they started doing anything about this. They made me waste two months in a back and forth of online communications where it was clear they were not even reading what I was typing. When I pointed that out (politely, as whey were asking me to "install" something that was mentioned all over the place in my previous postings), they cut the communication with me, also two months ago now.

This is serious, they are not following an industry standard, just making sure their hardware works in Windows systems. Of course I'll stay away from Foxconn in the future, and I recommend this to anyone running Linux on their hardware. But I want this fixed, no so much for me, as for the community at large. Let's see how it goes.

Cheers,
Leo

[Ivar Snaaijer (ivar-snaaijer)]

I put this on the Foxconn feedback page for my incident :
--------------
I have installed the new bios that came available. The bios is still reporting :

Aug 21 22:48:28 home kernel: [ 0.520201] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
Aug 21 22:48:28 home kernel: [ 0.520487] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCE2._PRT]
Aug 21 22:48:28 home kernel: [ 0.520610] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0PC._PRT]
Aug 21 22:48:28 home kernel: [ 0.523898] ACPI: PCI Interrupt Link [LNKA] (IRQs 4 7 *10 11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524025] ACPI: PCI Interrupt Link [LNKB] (IRQs 4 7 10 *11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524148] ACPI: PCI Interrupt Link [LNKC] (IRQs 4 7 *10 11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524269] ACPI: PCI Interrupt Link [LNKD] (IRQs 4 7 *10 11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524390] ACPI: PCI Interrupt Link [LNKE] (IRQs 4 7 *10 11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524512] ACPI: PCI Interrupt Link [LNKF] (IRQs 4 7 *10 11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524633] ACPI: PCI Interrupt Link [LNKG] (IRQs 4 7 10 *11 12 14 15)
Aug 21 22:48:28 home kernel: [ 0.524754] ACPI: PCI Interrupt Link [LNKH] (IRQs 4 7 10 11 12 14 15) *0, disabled.
Aug 21 22:48:28 home kernel: [ 0.524871] ACPI Warning (tbutils-0217): Incorrect checksum in table [OEMB] - F0, should be E1 [20080926]
Aug 21 22:48:28 home kernel: [ 0.524902] ACPI: WMI: Mapper loaded

with this message i searched and came to this interesting post :
http://ubuntu-virginia.ubuntuforums.org/showthread.php?t=869249

so i tried to disassemble the bios and found this code :
If (MCTH (_OS, "Linux"))
{
Store (One, Local0)
}

much in the same line as described in the article. as the solution on the page does not work for me (the generated code does not compile) I have to check out how to solve that, any pointers are appreciated.
-----------

As the original forum is of last year, and they do fix stuff for Vista relating to exactly the same stuff (so fixing for Linux should be rather trivial) It looks like they still do not take Linux any more serious than a year ago.

on iasl not working when i run `iasl -dc dsdt.dat` i get error 4068. when i try to compile a decompiled piece of code i get a lot of errors. I'll run by a iasl forum or something.

[Darko Dimitrovski (darkodimitrovski)]

Hi all.

I have problem with ACPI_PSS too. I have Gigabyte MA78G-DS3H with 780G.

I am first time Ubuntu user and waited for the 9.10 release. Downloaded the images, burned the 64bit on CD and booted my PC. Selected the option to Try Unbuntu first, the black screen with white logo works fine and just when you expect to see the desktop/login screen, the PC reboots without any message. Tried 3-4 times, same result.

Second attempt, created a bootable USB stick with the 32bit image, booted the PC, it reboots at exactly the same point as above.

The same 32bit USB stick installed on my netbook without any problems, working perfect. So I've tried on the desktop two different types of media with two different images.

Third attempt, I downloaded the 9.04 64bit image. Created a USB stick on my netbook ubuntu using this new image. After booting the PC it gets to the same point as above but doesn't reboot, instaed giving the mentioned ACPI_PSS error. That was the first time that any error was mentioned. Goggling the error brought me here.

Fourth attempt. Started wubi.exe on my Vista, from the 9.10 64bit image. Waited the installer to download from the internet although the image was unpacked right there (???), the installation requested reboot and went fine. Looked like 9.10 is installed at last, at least under wubi. The next reboot to go into Ubuntu, selected Ubuntu from the windows loader, selected Ubuntu from the (virtual) grub, just when you expect the login screen, the PC reboots itself.

Should I give up? Is there no solution? New Ubuntu user and definitely out of ideas how to get it on my desktop. :(((

[Leo Milano (lmilano)]

Hi Darko

You are posting this in a bug report thread, which is intended to discuss that particular bug. If you need support, please see here:
http://www.ubuntulinux.org/support

Best,
Leo

[Darko Dimitrovski (darkodimitrovski)]

Sorry for the long post, I tried to explain everything that I tried. The bottom line is I can't install 9.10 or 9.04 because of this bug right? So I don't think the posting is wrong. Maybe just a bit long. :)

[Darko Dimitrovski (darkodimitrovski)]

UPDATE: I updated to the latest BIOS for my Gigabyte MA78G-DS3H ver1.0 but it didn't help. 9.10 live environments still reboots before the desktop shows up.

Before the BIOS update I downloaded 8.04 LTS to see it will go. The live environment runs fine, you can see the desktop and work in it. Haven't tried install.

Don't know what is different between 8.04 LTS and 9.04/9.10 but they don't want to run in live or install with the old and new BIOS.

[Darko Dimitrovski (darkodimitrovski)]

UPDATE 2: Sorry, I can't edit the last post and I just remembered something else.

In my last try to run the 9.10 live environment I pressed TAB to edit the command and at the end added "acpi=off noacpi". Didn't help.

[David Gaarenstroom (david-gaarenstroom)]

"Your BIOS does not provide ACPI _PSS objects in a way that Linux understands" almost always means that Cool 'n Quiet has been disabled in the BIOS. This kernel message also isn't really an error, it just tells that the kernel will not use CPUFREQ powernow-k8 support because Cool 'n Quiet has been disabled, so it's doing exactly what is requested.

Please verify if C'n Q has indeed been disabled in the BIOS, and then determine what you want. If you don't want C'n Q you can ignore this message (powernow-k8 has been built-in in the kernel so you cannot prevent this message from being displayed).

[David Gaarenstroom (david-gaarenstroom)]

@Darko:
What I meant was that not being able to install or run Ubuntu can not be caused by this kernel warning, but if you want to avoid the warning you may have to take a look at your BIOS. Gigabyte sometimes has hidden options that you can uncover by pressing <alt>-<F1> from the BIOS main screen.

What can be a problem is when you have enabled customized CPU frequency/voltage settings in your BIOS, because the powernow-k8 will undo these settings. This is often more harmful than leaving them be and you cannot avoid powernow-k8 doing so because since 9.04 it is built-in in the kernel...