Foxconn Releases Test BIOS Fixing Linux Crashes

Ryan1984 writes "Only a week after the bad press coverage regarding the Linux-related bugs in a number of motherboards released by Foxconn (which turned out to be the AMI BIOS that several board makers use), Foxconn is the first vendor out with a publicly released test patch that fixes the bulk of the problems, allowing kernel 2.6.26 to run well on the afflicted boards. The remaining issues appear to either be kernel bugs in builds earlier than 2.6.26, issues with the Intel chipset itself, or minor annoyances that Foxconn is still working to resolve. Foxconn representative Heart Zhang has posted on the Ubuntu forums (where the situation began), apologizing for the issues, thanking Foxconn customers and the community at-large for their feedback, and promising that Foxconn will take Linux support and testing seriously, going forward."

When a mobo manufacturer supports linux publicly..

[SpzToid]

I'll consider their stuff. What I can't accept is non-acknowledgment, ostrich-style. That just loses me.

Re:Good sign

[ratboy666]

So, you think it would have been fixed if there had not been angry, almost rabid, users? You know, the ones you refer to as "crazy zealots"?

I don't believe so. I believe the issue would have been ignored, and Linux would have been patched in some obscene manner to "work around" the issue. Giving a bad reputation to Linux; "it doesn't work -- what kind of fucking shit is THIS?". Hurting the reputations of many developers.

Sometimes, the only sane response is to be angry and rabid.

Was it a bug? Was it deliberate? Who knows. That debate is still open. What IS important is that there is at least ONE open source OS with the clout to keep vendors honest.

Re:Theyre fixing it

[Gazzonyx]

No doubt... I don't care who they are, low level hardware fixes against an OS in a week is impressive. I think Linus' window on RC's for fixing this stuff is two weeks, and the kernel team has been moving at breakneck speeds lately (averaging 4 LOC/hour, every hour, every day).

Re:When a mobo manufacturer supports linux publicl

[capnkr]

Willy - Take a look at what the original poster at Ubuntu forums said:

Ubuntu forum thread [ubuntuforums.org]. Starts at post #114.

If he is correct in what he writes, then it doesn't seem much like speculation.

Perhaps if someone else has linkage to a sound refutation of his claims, it would be a good thing to post here. I've seen comments that TheAlmightyCthulu's claims were 'debunked', but the comments didn't say where, or have links.

companies often contradictory on Linux support

[bcrowell]

Perhaps instead of malice or stupidity, it was simply "taking care of the biggest customer pool first."

Yeah. This also seems to be an example of a more general phenomenon with Linux support, which is that the same company will make completely contradictory statements about their own Linux support. In the earlier slashdot story [slashdot.org], someone from Foxconn is directly quoted as saying 'it doesn't support Linux;' now they say they always intended to support Linux. The truth is probably that they never even thought about Linux support, and then when the issue was brought to their attention random representatives started saying random things off the cuff.

I've had a similar experience with Amazon's MP3 store. If you want to buy entire albums (as opposed to individual tracks), you have to use special downloading software that they supply. The software was initially only available in Windows and Mac versions, but pretty quickly they brought out Linux versions as well. Nowadays when you use your Linux box to shop for albumbs on their site, if you don't have the software installed your browser will detect that, and detect your OS as linux, and they'll generate a page for you offering links to download a linux version of the downloader. In fact, they even have it available in multiple versions for different linux distros. However, the linux downloader has been pretty buggy for me (and was also hard to get working properly on x64). I've had it working, then it broke, etc. I've done two calls to Amazon's tech support about this, and in both cases, the initial reaction was to tell me to do a bunch of stuff (with the usual confusion because the Indian tech support person gives Windows+IE instructions, and has never heard of Linux), and then when that didn't help they checked with someone else, who told them Linux wasn't supported. Never mind that they've had Linux versions of the software up on the site for months now.

I think part of the problem is that so many people in the hardware and software industries live in a 100%-Windows environment. It honestly never even occurs to them that anyone is running any other OS. (In the case of Foxconn, they're not making mac-compatible boards, so it's probably true that 99% of their boards are being used with Windows.) Then when the issue comes up, they just deal with it off the cuff. It's like asking them what their policy is on recycling cardboard -- they probably don't have one, and they don't see why it's important.

Another problem may be that in a Windows monoculture environment, many people don't understand what a standard really is. They think Windows and Word and IE are standards. Instead of developing for the relevant standard, some PHB makes the decision that they're going to target something proprietary, calling that a "standard," and they think of it as extra work to add support for anything else -- when in fact, it would have made more sense just to support the standard properly in the first place.

Re:Theyre fixing it

[hedwards]

I believe we have that, it's called ACPI. And if motherboard manufacturers are having a hard time properly coding things due to the spec, the spec needs to be fixed. If they're being lazy about it, then people should avoid buying their products.

But what really needs to happen is for MS to stop accepting broken implementations. I don't know for sure, but I'm sure that the broken ACPI implementations are a headache for those writing the parts of Windows that have to interact or take results from the ACPI, requiring a proper adherence would make it less of a headache for everybody.

No issue, only bad press from one guy

[Anonymous Coward]

http://mjg59.livejournal.com/

Quote :

Take home messages? There's no evidence whatsoever that the BIOS is deliberately targeting Linux. There's also no obvious spec violations, but some further investigation would be required to determine for sure whether the runtime errors are due to a Linux bug or a firmware bug. Ryan's modifications should result in precisely no reasonable functional change to the firmware (if it's ever hitting the mutex timeout, something has already gone horribly wrong), and if they do then it's because Linux isn't working as it's intended to. I can't find any way in which the code Foxconn are shipping is worse than any other typical vendor. This entire controversy is entirely unjustified.

Re:When a mobo manufacturer supports linux publicl

[sumdumass]

Well, it is definitely written into the APCI 2.0 specs. When implemented, the bios can check the OS running and give specific tables to the OS that ease compliance and nuances that are different among other operating systems.

If this is the case in which it happened that way, then it can be as simple as other mainboard manufacturers not using specific DSDT tables or referring all non recognized or handled returns as NT and providing NT versions of the DSDT tables. When Foxxcom's programmer decided not to acknowledge linux and/or forgot to point it back to the windows tables, we see an issue that is specific to one instance and manufacturer.

Make no mistake, they didn't write anything specific into the bios to check for linux. That is already there and part of the spec. Failing to handle the return properly is a mistake but it doesn't imply the malice accusations that are going around. And no, this first appearance doesn't mean it can't happen with the other manufacturers, it just means that it hasn't because they did something different or were more thorough. Linux mimics windows in a lot of ways on these levels and for the most part, can handle the windows DSDT returns. If a mainboad simply passes the windows tables on a linux return, you would likely never know the difference without the source or decompiling the bios.

Re:Theyre fixing it

[Fred_A]

An even better thing to do would be to standardize one API that the Linux kernel uses and give that to manufacturers so they can support all Linuxes, rather than masquerading as Windows.

We already do that, Linux implements ACPI. However Windows doesn't.
So motherboard makers theoretically would have to accommodate the Windows oddities *and* support the standard (which wouldn't work in Windows). Fun huh ? How many are prepared to support the extra cost ?

Or of course the kernel developers can make do with the broken implementation of ACPI that are seen in the wild and that do work with Windows. In practice it's the only way to make sure the system will work on a random x86 type of machine.

AARD

[paradigm82]

I'm surprised noone is comparing this saga to the AARD scandal that ultimately resulted in Microsoft having to pay a settlement to Caldera. you can read about it here http://en.wikipedia.org/wiki/AARD_code [wikipedia.org] but the case was about encrypted & obfuscated code inserted in Windows 3.1 to detect DR-DOS and preventing Windows from running on it. Internal Microsoft memos revealed the intention of the code: At one point, Microsoft CEO Bill Gates sent a memo to a number of employees, reading "You never sent me a response on the question of what things an app would do that would make it run with MSDOS and not run with DR-DOS. Is there [sic] feature they have that might get in our way?"[1] Microsoft Senior Vice President Brad Silverberg later sent another memo, reading "What the [user] is supposed to do is feel uncomfortable, and when he has bugs, suspect that the problem is DR-DOS and then go out to buy MS-DOS"[1] Later, after DR-DOS had been purchased by Novell and renamed "Novell DOS", Co-President Jim Allchin stated in a memo, "If you're going to kill someone there isn't much reason to get all worked up about it and angry. Any discussions beforehand are a waste of time. We need to smile at Novell while we pull the trigger."[1] The lawsuit was later settled.[1][2] Compare this to: "One thing I find myself about is whether we shouldn't try and make the "ACPI" extensions somehow Windows specific. If 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 Io avoid this problem but it does bother me. Maybe we couid 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 relaled to this." In both cases it was Bill Himself that suggested to employees that they threw a wrench into something to prevent competing o/s'es from interoperating properly. Many of you probably know about the AARD scandal for I wanted to post this for those who don't :)

Re:This is Foxconn's BIOS problems, not generic to

[Anonymous Coward]

That's right -- and AMI makes their money by intentionally NOT taking fixes from their customers and integrating them into their core (so that when problems crop up, they can offer their "services" to help fix them) -- it's like pulling teeth getting them to take a fix upstream, believe me. AMI sucks just as much as any other BIOS vendor.

Re:They're fixing themselves all else is incidenta

[Daengbo]

In the original Ubuntu Forums thread, some people said the story needed to be Dugg (which it was). Later, a voice of reason said that the story should be submitted to Slashdot so that they could find out what was really going on in the BIOS.

I'm proud that Slashdot has the rep of having really smart posters who know their shit.

BTW, I was always in the "bad copy-paste" camp.