A possible bug in Foxconn boards BIOS affects Linux ACPI

[themacmeister]

I am sure that Dell have done the same thing with some of their Optiplex computers. My computer has NEVER worked with ACPI suspend/hibernate from any distro I have installed, and I have installed quite a few recently.

"hold them to their box-specs"

[TheAlmightyCthulhu]

Unfortunately there are no such if statements in my DSDT table source. If anyone would like to have a look at it and see if you can make any sense of it, please feel free. It is attached below. Please note my BIOS ARE NOT up-to-date. I am using A07, not the new A08. The A08 causes my trackpad to malfunction. When I apply a fix, it becomes unconfigurable. I am sticking with A07 right not because it works.

I PM'd you, suddenly I don't feel so bad anymore.

I've fixed your BIOS, it had 8 warnings, and one error (causing it to not compile at all) caused by a typo, I left the warnings alone for now, I really don't like messing with variables any more than I have to, I'll email you the compiled result.

Here's what I did:

Remove all reference to Windows 98, NT, Me

Define _OS response "Linux" as "Windows 2006" (Vista!), point it to Store 20

Define _OSI response "Windows 2006" as synonymous with Linux, point it to Store 20

Create new _OSI response "Linux", and give it Store 20

Replaced: Name (_HID, "*pnp0c14") with Name (_HID, "pnp0c14")

In otherwords, like it or not, it's booting Linux.

Before removing typo that MS compiler ignores:

"Compilation complete. 1 Errors, 8 Warnings, 0 Remarks, 576 Optimizations"

(It says complete, but it won't compile if there are errors)

After:

"Compilation complete. 0 Errors, 8 Warnings, 0 Remarks, 576 Optimizations"

God I love Microsoft's compiler, leaving all those errors, warnings, and not optimizing ANYTHING. B-)

If anyone is interested, this is some kind of Dell BIOS I believe?

If you need me to fix your DSDT, dump it and PM me with your email address, we will do things THERE.

I _DO NOT_ want you to attach any source code or intellectual property of anyone's to this thread, but I will help you as a courtesy, just not here!

[Vadi]

The fact is that Foxconn isn't the only company that does this, as pointed out by some others and mj59 in this post: http://mjg59.livejournal.com/85923.html

Except he failed to get as much attention.

His message? "Message to vendors: If we're not behaving in the same way as Windows does, let us know. We'll fix it. This has the benefit of fixing not only your latest and greatest platform, but also all the older ones that expected the same behaviour. Please don't try to work around our bugs in your BIOS - it just means that we need to special case you forever, and in the long run your hardware won't work as well as it would if we just fixed things properly."

[vineh]

i find it quite surprising that there are administrators on these forums, a Linux forum, which is inherently reasonably technical, who presumably lack the technical understanding to comprehend what is being said here by our freind cthulu.

just to paraphrase, the bios is not just *defective* for linux.

you see in this instance linux actually impersonates windows for compatibility purposes.

so the bios wouldnt realy even NEED to have ANY reference to linux in it.

the fact that they have specifically put code that LOOKS FOR linux, and then sends an alternate response to a linux OS, to say it is not intentional, or that it could be a bug, well lets just say the term *heads in the sand* comes to mind. perhaps sand from the *lalala i cant hear you* dept... =P

[TheAlmightyCthulhu]

The fact is that Foxconn isn't the only company that does this, as pointed out by some others and mj59 in this post: http://mjg59.livejournal.com/85923.html

Except he failed to get as much attention.

His message? "Message to vendors: If we're not behaving in the same way as Windows does, let us know. We'll fix it. This has the benefit of fixing not only your latest and greatest platform, but also all the older ones that expected the same behaviour. Please don't try to work around our bugs in your BIOS - it just means that we need to special case you forever, and in the long run your hardware won't work as well as it would if we just fixed things properly."

In his position, he has to be professional, in my position, as a customer, I have a right to go in, bang my fist on the desk, and ask "Why the hell is this not working?"

I've been in contact with Foxconn and this has lit a fire under their ***, and I believe they are now committed to helping.

They have sent me a DSDT table that I intend to try out in a few minutes, cross your fingers.

Edit: STill doesn't work right, it fixes some of the more obvious errors though, I will apply my modifications to theirs and see if it helps.

[TheAlmightyCthulhu]

i find it quite surprising that there are administrators on these forums, a Linux forum, which is inherently reasonably technical, who presumably lack the technical understanding to comprehend what is being said here by our freind cthulu.

just to paraphrase, the bios is not just *defective* for linux.

you see in this instance linux actually impersonates windows for compatibility purposes.

so the bios wouldnt realy even NEED to have ANY reference to linux in it.

the fact that they have specifically put code that LOOKS FOR linux, and then sends an alternate response to a linux OS, to say it is not intentional, or that it could be a bug, well lets just say the term *heads in the sand* comes to mind. perhaps sand from the *lalala i cant hear you* dept... =P

Occom's Razor Time:

Theory: They were negligent wankers that didn't care and/or just a bug

Theory assumes many unlikely circumstances.

1. Foxconn accidentally put an entire DSDT table and a reference to it just for Linux, but thats a bug, they just didn't notice it before the BIOS shipped, oops.

2. Foxconn broke this table so that not even Windows could use it, much less anything else, but this is just a bug, they overlooked it before it shipped.

3. Foxconn claims not to support Linux, but has an entire table just for it, that is obviously broken and doesn't even pass a checksum test, but this is a bug you see, they didn't notice it.

4. Foxconn makes no mention of Linux on their site, but lets you create support tickets for it, to which they reply "Get Vista" or "Ship it back", this is a bug you see, the invisible man put FEDORA and GENTOO, and others in their support database.

5. (This is a joke, I have no proof this part happened) A large sum of money gets wired from Redmond to Shenzen, China, this was undoubtedly the work of Chinese h4x0r5 making Microsoft look bad, bad h4x0r5!

6. Being called out, publicly, Foxconn craps themselves and starts fixing the BIOS that they "accidentally" shipped with a "bug", being an entirely broken BIOS that won't even boot anything unless extreme hacks are used, especially for Linux.

Just a bug, my rear end, this was indeed deliberate, I have enough proof to satisfy any reasonable person.

As long as Foxconn fixes it, and does not introduce any more intentionally faulty products, I will be happy and go away.

[TheAlmightyCthulhu]

Carl,

I used the website, and I did not appreciate any of that at all, actually I was quite horrified that a motherboard manufacturer as large as Foxconn is has just totally torpedoed Linux and then blows off customers that are angry about it.

I will be more than happy, and in fact I am looking forward to writing a thorough and detailed review of your repaired BIOS, if it works reasonably well with operating systems that are not Windows, I will give it a glowing review, everywhere I have been thus far.

Furthermore, I think it shows a complete lack of professionalism, to treat millions of users, and potential customers (who may or may not be now) in this way, I know that if this issue is not resolved, I will not be the only one going to great lengths to avoid purchasing your company's products, Phoronix, Slashdot, and Digg, are probably the three places you guys really don't want to be, because people that buy products such as yours (including your line of Nvidia video cards, very popular with Linux users) read things like this and associate that with your brand.

So just telling 8% of the market, who does not use Windows to go kindly screw themselves is probably not the best thing you guys can do.

Now, I know you guys probably realize everything I've just said, I was throwing it ou there to make an example and for you to actually think about that when your boss is asking you why in the hell this justifies an expenditure, I centainly don't mean to be disrespectful, rude, or hostile, had I received any help without taking it to this level, I would have gone about my business.

As for ACPI, lots of luck with that, even just looking at the assembler, bluntly, makes me want to vomit, you should really be standardizing on the Intel reference specification and using Intel compiler tools, it would save you from having to fish for several versions of Windows, singling out Linux as an OS that supports ACPI features circa 10 years old (and broken), and totally ignoring everything else.

If you insist on basing it on this weird fishing behavior, I would suggest removing all references to Linux and removing the table it uses, removing all references to Windows NT and Me (This board wouldn't even boot them, it's kludge), and keeping the _OSI and tables for Windows Vista and XP, Linux will respond to an _OSI telling the BIOS it is Windows, so doing what I have said should cut the amount of crap code in the BIOS by half, or more.

Then, you can concentrate your efforts on fixing the crashing on reboot after having suspended, which seems either in the DSDT table, elsewhere in the BIOS, god only knows where, chances are this is a bug in any ACPI compliant OS, but XP and Vista simply ignore it, it's most likely a few lines of code somewhere that just needs tweaking, but for the love of God please get rid of all the legacy crap, it hurts my brain just looking at it.

When you post the fixed BIOS, please assemble the whole thing with Intel's tools, they work surprisingly well on pretty much any modern OS, even the piece of crap, Vista.

I know, it seems like a lot of work, it will pay off, you won't have to hack on bad, outdated code to support version X of operating system Y in the future, your code now is spaghetti code, scratch that.......Chef Boyardee Spaghetti code.

Anyhow, I appreciate Foxconn's new found willingness to help, I really really do. (I mean that!)

-Ryan

On Sat, Jul 26, 2008 at 6:57 AM, Carl Brunning <carl.brunning AT foxconn.com> wrote:

Hi Ryan

Thank you for all the information so far...

When you reported to foxconn did you email or use the website, sorry working out who give the bad answer to you and stop stuff like this from happening again.

On Monday, I will be sending a email to the bios team about this and having it fixed as well.

Time for me to read the acpi spec again last time I looked was over 4 years ago

Ouch going to be fun again lol


Anyway thanks again,


Kind Regards
Carl Brunning

Technical Manager for UK and Ireland
Foxconn CSDEU B.V (UK)
Tel +44 1908 276610
Tel +44 1908 276621(Direct)
Mobile 07870 622 398
VOIP : 528 2611 when it works

WEB http://www.foxconnchannel.com

[silkstone]

The BIOS coding is beyond me, but I understand the gist of this thread. I hope it now gets fixed and that what you've done here may deter people from making such 'mistakes' in future.

As a general comment, headlines such as 'XYZ tries to kill Linux' may be true, but may also leave 99% of people unimpressed as it doesn't apply to them. If I'm right in thinking that true acpi compliance will allow Linux to run, whether or not the board is officially 'approved for Linux', then perhaps the lack of compliance with acpi standards should be the main headline. Just a thought.

[phill0]

For those of you who are skeptic about Microsoft's involvement, I would like to post an email from Bill Gates to his staff:

From: Bill Gates
Sent: Sunday, January 24, 1999 8:41 AM
To: Jeff Westerinen; Ben Fathi
CC: Carl Stork (Exchange); Nathan Myhrvold; Eric Rudder
Subject: ACPI extensions

One thing I find myself wondering about is whether we shouldn’t try and make the "ACPI" extensions somehow Windows specific.

It seems unfortunate if we do this work and get our partners to do the work and the result is that Linux works great without having to do the work.

Maybe there is no way to avoid this problem but it does bother me.

Maybe we could define the APIs so that they work well with NT and not the others even if they are open.

Or maybe we could patent something related to this.

Source: http://antitrust.slated.org/www.iowa...00/PX03020.pdf
(The pdf was OCRed so I'm not certain that the names in "To" and "CC" headers are right.)

So judge for yourself. At the very least you have to agree that considering strong cooperation between Foxconn and Microsoft this looks fishy.

[autocrosser]

My note to FoxxConn website support:


I have considered FoxConn motherboards for myself & my customers in the past. I will consider them IF you will straighten up your Linux support. I would not be your largest customer (building only 8/10 computers per year), but I promise to not recommend FoxConn motherboards to my customers until I get a positive result from the Linux community.

Your management should look at this posting: http://ubuntuforums.org/showthread.php?t=869249

One unhappy linux customer will snowball into the community becoming very vocal & I might add that Ubuntu Linux has currently over 10 million users--with a fair amount for us being very tech-savy.

I don't expect a response from this, but if I do---I would like feedback from higher management.

Dean Loros
Performance by Design Ltd.