Corrupt LINUX ACPI tables in BIOS

[D_A]

VERY interesting post on the Ubuntu forums.

It seems that Foxconn have deliberately set up their BIOS to cripple linux users.

They have several different tables, a group for Windws XP and Vista, a group for 2000, a group for NT, Me, 95, 98, etc. that just errors out, and one for LINUX.

The one for Linux points to a badly written table that does not correspond to the board's ACPI implementation, causing weird kernel errors, strange system freezing, no suspend or hibernate, and other problems, using my modifications below, I've gotten it down to just crashing on the next reboot after having suspended, the horrible thing about disassembling any program is that you have no commenting, so it's hard to tell which does what, but I'll be damned if I'm going to buy a copy of Vista just to get the crashing caused by Foxconn's BIOS to stop, I am not going to be terrorized.

The poster also offers a possible fix that people can try, if they're game.

-------------------------------------------------------------------------

EDIT:
It seems the same code is used by Asus and probably other mainboard mfanucturers as well, and its not crippled on purpose by the mainboard manufacturers to cause Linux users and developers problems

-Saaya

[NaeKuh]

rawr

stupid foxconn.

unfortunately there one of the largest motherboard makers in the business.

[D_A]

Not for long

[pwolfe]

Not for long

unfortunately, and I say this as someone who dropped ms 9 years ago and used linux exclusively since then, I do not think we represent enough of the market share to cause them to stop doing this. But who knows, maybe if there is enough uproar they'll do something about it.

side note: this hit the front pages of digg and reddit today

[RAMMIE]

Why should foxconn or any other MB assembler care about Linux when the user base is less than .1 percent of all desktop users?Microsoft has a stranglehold on the public and it isn't going to change anytime soon.If I was a hardware manufacturer I would first care about what the majority of users use.When Linux becomes mainstream,they will be catered to.Gotta make the money to stay in business.

[fart_plume]

Now that this out, it could prompt government investigations, and if it's found MS has anything to do with it, it will be anti-trust city for them..............

[D_A]

RAMMIE, you misunderstand. The BIOS is written SPECIFICALLY to disadvantage linux. It's not a matter or apathy, it's DELIBERATE. The BIOS even goes to the trouble of using multiple redundant checks to make sure the OS is or is not Windows Vista/XP rather than just using the standard checks. It is very, very deliberate.

[D_A]

One thing I find myself wondering about is whether we shouldn’t try to make the “ACPI” extentions 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 API’s 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.

Full text PDF here

[Mr.Guvernment]

the mobo was marketed as Windows certified, NOT linux certified? can foxconn not make ONE motherboard that works specifically for windows?

[STEvil]

current news has it as a bug.

[saaya]

It seems that Foxconn have deliberately set up their BIOS to cripple linux users.

RAMMIE, you misunderstand. The BIOS is written SPECIFICALLY to disadvantage linux. It's not a matter or apathy, it's DELIBERATE. The BIOS even goes to the trouble of using multiple redundant checks to make sure the OS is or is not Windows Vista/XP rather than just using the standard checks. It is very, very deliberate.

dont you think this sounds "a little" paranoid ?
i work for foxconn and never noticed that any of our engineers, hardware or software, have anything against linux, on the contrary...

and even if they had any problem with linux, why would we do this?
why would we make such an effort as the article suggest, of doing multiple checks and whatnot and then deliberately create false tables, to cause a very small percentage of the users of that product to have a hard time?

just think about it, it doesnt make any sense whatsoever...

people just love the idea of a big evil organisation trying to enslave the human race, and consipracy theories and whatnot... but isnt it pretty obvious that this is just a bug or simply badly written acpi tables?

i dont know that much about BIOS programming, so i might be wrong...
im on a business trip atm, but when im back in taipei ill check with the engineers to see what this is about, ok?

[Solarfall]

dont you think this sounds "a little" paranoid ?
i work for foxconn and never noticed that any of our engineers, hardware or software, have anything against linux, on the contrary...

and even if they had any problem with linux, why would we do this?
why would we make such an effort as the article suggest, of doing multiple checks and whatnot and then deliberately create false tables, to cause a very small percentage of the users of that product to have a hard time?

just think about it, it doesnt make any sense whatsoever...

people just love the idea of a big evil organisation trying to enslave the human race, and consipracy theories and whatnot... but isnt it pretty obvious that this is just a bug or simply badly written acpi tables?

i dont know that much about BIOS programming, so i might be wrong...
im on a business trip atm, but when im back in taipei ill check with the engineers to see what this is about, ok?

im with you on this 1 saaya.. it dont seem to make any sense what so ever..

lol the reason i dont use linux is that it sucks on 3Dmark

[64dragon]

lol the reason i dont use linux is that it sucks on 3Dmark

imo, thats a lame excuse to stick with an OS, bc of a benchmark. there is a version of Pi for linux and my times are 7 seconds quicker. if 3dmark's got ported to linux i can't imagine how high new world record scores would be.

[Solarfall]

imo, thats a lame excuse to stick with an OS, bc of a benchmark. there is a version of Pi for linux and my times are 7 seconds quicker. if 3dmark's got ported to linux i can't imagine how high new world record scores would be.

lol sorry if im little bit addicted to benching.. but honestly that was a just a silly joke dude dont take stuff so seriously
btw what super pi version are you talking about cos if thats true i could get my e8500 chip to 1.235s with it if it shaves 7 secs off hahahahah

[64dragon]

lol sorry if im little bit addicted to benching.. but honestly that was a just a silly joke dude dont take stuff so seriously
btw what super pi version are you talking about cos if thats true i could get my e8500 chip to 1.235s with it if it shaves 7 secs off hahahahah

no worries, i knew you were joking.

im running an amd x2 and only run 64bit OSs' and comparing XP 64bit and kubuntu 8.04 64bit. i expect the time difference is less for intel's. heres where i got the info about linux pi from http://blog.linuxoss.com/2008/04/sup...your-computer/

screenshot in XP, 1M time = 34 seconds
http://i37.photobucket.com/albums/e9.../pc/1Mpi-1.jpg

screenshot in linux. computed pi while crunching and watching a movie ~1M time = 27 seconds
http://i37.photobucket.com/albums/e9...Mlinuxpi-1.jpg

[D_A]

Let's see ... Unkie Bill has already publicly stated that he wants to find a way to stop linux users from benifiting from ACPI (an open standard, btw)

The BIOS in question makes MULTIPLE checks to verify what OS is loaded on the system, regardless of what is returned to it via the normal channels.

When it identifies the OS as Linux based, it direct to a table that LOOKS ok, but which causes a Linux OS to have all manner of weird ACPI related issues.

Microsoft has "certified" that the ACPI implementation on this board meets the standard, which it clearly does not.

Mate, it's not paranoia when they REALLY ARE out to get you!

[64dragon]

Originally Posted by Bill Gates
One thing I find myself wondering about is whether we shouldn’t try to make the “ACPI” extentions 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 API’s 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.

sadly, that doesn't surprise me

the mobo was marketed as Windows certified, NOT linux certified? can foxconn not make ONE motherboard that works specifically for windows?

last i checked but i could be wrong, NO board is marked linux certified and they shouldn't need to be. you dont see them marked as mac certified either

im on a business trip atm, but when im back in taipei ill check with the engineers to see what this is about, ok?

please do

[saaya]

When it identifies the OS as Linux based, it direct to a table that LOOKS ok, but which causes a Linux OS to have all manner of weird ACPI related issues.

so there are problems when the linux acpi table is being used... that doesnt necessarily mean the table is bad, as you said, it looks ok, so couldnt it be that there is a problem in how the linux version you use reads or interprets the table? I dont know that much about linux or acpi, but for two things not working well together you are very fast in judging which of the two is two blame.

Microsoft has "certified" that the ACPI implementation on this board meets the standard, which it clearly does not.

it works for them, so thats probably all they care about?
maybe microsoft wants a seperate acpi table for linux and windows, and our bios engineers accidently borked the linux table... at most i could imagine that microsoft MIGHT ask us and other mobo makers to use a certain way of handling acpi tables that causes trouble for linux... thats already quite paranoid but i could actually believe something like that is possible... but thinking that us or any other mainboard maker creates a new way of handling acpi tables, risking to bork acpi compatibility with windows AND linux and putting a lot of time and energy into this, ONLY to cause Linux developers and users a headache... just doesnt make sense...

Sure, large corporations tend to be evil cause after all they only exist to make money, and sooner or later all the ideology fades and the management does quite immoral things to make more money or to secure their market position.

But then think about it, how would foxconn secure its market position or make any money by crippling linux acpi tables? that doesnt really make sense.. agreed? the one who profits from this is microsoft, to some extent, and microsoft sure as h3ll wont pay us or other mobo makers to cripple linux support, that would be way too dodgy for one of the largest corporations in the world and theyd get caught quickly i think.

its more realistically that microsoft wrote this code and is asking us and other mainboard manufacturers to use it. And even then, im not convinced that they wrote the code to cripple linux acpi support on purpose. id rather think its likely they want to seperate acpi tables for linux and windows, and then came up with some code to do this, and we and maybe other mainboard makers as well didnt implement it 100% correctly, which results in the linux acpi tables causing problems. again, we might have had the same issues with the windows tables too, but either us or microsoft noticed it and we then fixed it, while we definately dont test that much with linux as we do with windows, so we overlooked the problems for the linux acpi tables.

im just speculating...

Mate, it's not paranoia when they REALLY ARE out to get you!

dont you think microsoft can think of better ways to cause the open source community trouble than by crippling acpi tables?
its like calling "the 3rd world war is near, china is gonna get us" just cause china is exporting children toys that contain higher than allowed chemical levels that can hurt our kids health, and then say this is an attack by china and they are out to get us and they are doing this on purpose... :P

Read the article, fanboy. The BIOS directs Linux to a DIFFERENT TABLE than it does Windows. It does not use the same ACPI tables for both and goes to some lengths to make sure of it.

dude, chill out... please dont call people names just cause you dont agree with them... and as i said, i dont know that much about acpi and linux, but is it maybe possible for linux to use the windows acpi tables anyways?
my guess is microsoft is trying to seperate windows and linux acpi tables, but how can the bios know for sure if the os is windows or linux? it cant right... so then wouldnt it be the best for linux to just use the windows acpi tables? cause those tables will work for sure, as we test them and have the certified all the time...

its similar to IE vs FF and Opera, where microsoft pushed websites to check if they are beeing viewed by IE or FF, and when the browser identified itself as FF it might get different code that causes problems. The fix was easy back there as well, FF and Opera just pretend to be IE and there is no problem...

Because MS told you to..... Oh no that wouldn't happen! never has!

i dont think i ever talked to any microsoft employee in my entire life... at least not knowingly. so to me personally, no, microsoft didnt tell me to do anything, and to foxconn as a whole, i couldnt know, but im 100% sure microsoft didnt tell any of our engineers to purposely sabotage linux... thats just silly...

What is this? 20 mindless questions?

would you mind taking a chill pill?

Think for yourself for a change. The 'why?' is simple. $$$$$$$ Microsoft has a known history of handing massive kickbacks to hardware vendors who give them exclusive deals, just like they gave out massive bribes and 'salted' voting memberships in the recent OOXML standards scandal. Don't the words "anti trust" ring a bell with you?

so you think a microsoft guy hands over a suitcase of money to a foxconn guy and says "make sure you cripple the acpi tables for linux real good" ?

you honestly dont think thats paranoid?
im sure microsoft is trying to cause the open source community trouble, they keep doing it, and if you read posts of mine youll see im not a fan of this either, but the way you imagine how they do it is just really really unrealistic... you picture it quite james bond and hollywood like...

You work for who exactly?

*puts on tinfoil hat*
and you work for who? maybe asus or gigabyte and you only came up with this to try and make foxconn look bad?

dont you see how ridiculous all this paranoia is?

Exactly. Foxconn claim that the board meets the ACPI standard when if fact it does not. Even the implementation for Vista has dodgy workarounds built in to cater for Microsoft's sloppy work. When the guy who found the issue queried Foxconn about it they said that it was fine because MS said it was fine.

thats the real problem here, you ended up with a tech support employee who didnt really understand the problem.
how can you certify the acpi standard? with microsoft right? is there anybody else who certifies acpi tables? correct me if im wrong, but no, there isnt.
did microsoft certify the acpi tables? yes they did!
are they acpi tables faulty? the linux ones possibly are...
so how is it foxconns fault if the tables are not working fine?
If your buddy brings you car to a garage to have it checked, and they say the car is fine and the next day it breaks down, who do you blame, the garage or your buddy? you blaming your buddy atm... which doesnt make sense!

pwolfe, it's interesting you deliberately cost your customers more than necessary. Do you get kick-backs too? Or is a Linux based system just too hard for you to learn?

can you PLEASE mellow out and finally stop insulting people and stay on the topic?

Linux IS perfect for some situation, while Windows never is.

and this has what to do with this topic?
and you call other people fanboys?

I'm sorry, Ady, but I fail to see how being "vehemently opposed to using linux" on his own machines and being a "linux user for 9 years" constitutes congruent comments. One or the other is an outright lie. This is just another Microsoft fanboy that can't keep his story straight.

why do you feel the need to insult people left and right? you seriously need to take some time off and relax and work on your anger managment :P

Well, it looks like at least one Foxconn employee places the blame on it just being a bug, but I guess we'll have to wait until Monday for the official response. It's possible that the ACPI implementation was written by a group not so knowledgeable, at least according to the Foxconn employee in that thread, I wouldn't be surprised.

The part that gets me is Foxconn is being accused of specifically looking for Linux in a different manner, when the person in this post showed code from his ASUS board that's the exact same. While other motives "could" be in play here, I don't see why a motherboard manufacturer would try to limit their market.

havent had the chance to talk to our bios engineers yet, i might have time in the afternoon to catch one of them and ask him about the acpi tables. and im not surprised asus is using the same multiple table thing, im pretty sure its a microsoft recommended way of handling acpi tables. im curious if the code was indeed supplied by MS, and if it was already faulty when we and others got it, or if our and the asus bios engineers broke it somehow causing the linux acpi problems... its unlikely that asus and our engineers broke it in the same way, so most likely this is the code we all get from MS and its already broken...

I logged in to check for updates on the drama and noticed that "TheAlmightyCthulhu" had lost his avatar and his user title was somewhat uncomplimentary. I also discovered that my own user title is now "spilled the beans".

It seems that a particular admin has had it in for Ryan (TheAlmightyCthulhu) and has been looking for a way to get at him. It looks like he found something he could ponce up and use, too.

and MAYBE its just cause you and this other guy really enjoy all the attention you created with this, but the ubuntu community or at least the admin does NOT like that you cause a mayor manufacturer BAD PRESS that it actually doesnt deserve?
did you or this other guy ever think that far of what your actually trying to achieve?
I dont think you did, cause you jumped the gun quite quickly to blam foxconn and claim that this is done on purpose, when you had 0 proof, and now it actually looks like it has nothing to do with foxconn at all but is a common problem, either a bug or badly written code from microsoft.

thank you so very very much to you and TheAlmightyCthulhu for throwing tomatoes at foxconn and making the company look bad to thousands of people, for a problem that doesnt have anything to do with foxconn

[Mr.Guvernment]

If it isn't ACPI "standard" why does windows not have any issues with ACPI then?, as foxconn said it meets ACPI standards for MS OS, and this board was marketed as a windows certified board, not a linux certified board...

[littleowl]

If it isn't ACPI "standard" why does windows not have any issues with ACPI then?, as foxconn said it meets ACPI standards for MS OS, and this board was marketed as a windows certified board, not a linux certified board...

IMO
That is the problem!!! The way MS works, mobo companies will start using linux for marketing in a way that you will have to buy special stuff no matter what. hum makes me think of apple

why would we make such an effort as the article suggest, of doing multiple checks and whatnot and then deliberately create false tables, to cause a very small percentage of the users of that product to have a hard time?

Because MS told you to..... Oh no that wouldn't happen! never has!

[D_A]

Read the article, fanboy. The BIOS directs Linux to a DIFFERENT TABLE than it does Windows. It does not use the same ACPI tables for both and goes to some lengths to make sure of it.

[Mr.Guvernment]

Read the article, fanboy. The BIOS directs Linux to a DIFFERENT TABLE than it does Windows. It does not use the same ACPI tables for both and goes to some lengths to make sure of it.

Little immature to be calling people names aren't we...

considering the amount of motherboards foxconn makes, has a part in (foxconn makes dell motherboards and they sell them with linux...)

This looks like a large over step for foxconn, considering the size of the company and what they do, i dont see them purposely blocking out linux market either way, not that it would damage their sales anyways since they are so huge and have their foot in everything.

Linux might not be perfect for every situation out of the box, but neither is Windows. On the other hand, Linux IS perfect for some situation, while Windows never is.

that is personal opinion, giving someone linux who has used windows their entire life is not the perfect situation or answer. Giving someone linux who has to work with Access Databases - is not, that is a perfect situation for windows.... the list can go on and on back and forth for BOTH side, neither OS is perfect , as you said, only in some situations and yes, both OS's are perfect in their own situations. but obviously not all.

Foxconn said they are going to be fixing this, so as said, they fix it, i guess they told MS to go shove it, since apparently this was all some big Foxconn in MS pocket fiasco...

[BenchZowner]

I take it nobody has a normal and exciting life nowadays, and because of that everybody's looking and bringing up ( or seeing ) conspiracy theories everywhere.
Have you ( personally ) experienced any of the issues reported by that site ?
If it's a conspiracy as you suggest, then wouldn't it apply to all the boards released by Foxconn recently ?
What does Foxconn have to win out of this ?
Why would they do something like that, especially now that except from the OEM manufacturing that they always had ( you don't know how big Foxconn is, most of you I'd say ) and "their own branded" products on the market ?

So...by your saying, if the issue is true and affects everybody and every board, and Foxconn introduce a new BIOS to fix that, would that mean that their deal with the devil ( Microsoft per se ) didn't work out ?

[D_A]

...
What does Foxconn have to win out of this ?
Why would they do something like that ...

What is this? 20 mindless questions? Think for yourself for a change. The 'why?' is simple. $$$$$$$ Microsoft has a known history of handing massive kickbacks to hardware vendors who give them exclusive deals, just like they gave out massive bribes and 'salted' voting memberships in the recent OOXML standards scandal. Don't the words "anti trust" ring a bell with you?

If they don't comply their customers (Dell HP etc) start getting polite "suggestions" that other manufacturers support Vista "better" and it could be in their interests to change suppliers. Then they mention how hard it is to get "Vista Ready" certification without "adequate" hardware. If a manufacturer can't get certification for the latest monopolistic OS, they sell NOTHING.

Microsoft isn't THE Evil Empire, it's just another big corporation and behaves just like most other big corporations.

[alucasa]

Microsoft isn't THE Evil Empire, it's just another big corporation and behaves just like most other big corporations.

Well..., Microsoft is the evil empire, at least to you, it seems. No offense, but that is the impression I've gotten from your posts and the link you've posted.

Note : I am a heavy windows user, but I also deal with Ubuntu (for my Yonah/Merom crunchers) and CentOS due to my dedicated hosting server.