Faulty Microsoft Driver Saps Intel Core Duo power

Critical_ writes "Tom's Hardware recently discovered a bug in Microsoft's ACPI driver implementation under Windows XP SP2 that causes a loss of more than one hour of battery time when connecting any USB 2.0 device to an Intel Core Duo based system. Apparently Microsoft, Intel and ODMs have known of this problem under a confidentiality agreement since July 12, 2005 via (a still private) Knowledge Base article KB899179. The bug lies in the asynchronous scheduler component inadvertently being left running causing Windows' internal task scheduler (ITS) to treat it as a running process involving the attached device. This in turn prevents the ITS from powering down the processor into one of the ACPI sleep states causing the system to use more battery power. At this time there seems to be no fix. Strangely, single-core systems and AMD systems are not affected. This leads one to wonder if it is truely a software problem or if there a much larger hardware problem that may affect Core Duo equipped Apple systems."

Comments

And thanks to the confidiality agreement

[Anonymous Coward]

You'll never know if you're being asked to buy broken hardware or broken software.

Seems best to stay away from both companies.
Why can't they just be honest and say "this is the problem and this is what we're doing about it"

Re:And thanks to the confidiality agreement

[MadTinfoilHatter]

Why can't they just be honest and say "this is the problem and this is what we're doing about it"

Because they don't want people to know there is a problem, and that they're not doing anything about it, maybe?

Why?

[Stefman]

Why does the last phrase target specifically Apple computers since the beginning mentions Win XP. Obviously, this affects XP laptops with a core duo.

Re:Why?

[smalljs]

Well, if the problem exists in Apple's hardware as well, it just would lead one to believe that it's a problem with the processor rather than Windows. I didn't take it as a shot at Apple...

Re:Why?

[goldn_64]

Because the article also hints that it might not be only be a software problem, but that maybe there's a hardware problem too.

Re:Why?

[nurb432]

Beacuse if its actually a hardware issue, as the story suggests it may be, it may equally effect other OS's that follow the proper ACPI rules. Espcially if Intel and Microsoft are hiding these facts from other manufacturers.

That, and who around here cares about problems that effect only XP ? :)

Re:Why?

[peragrin]

The problem is where does the failure lie?

My bet the problem is in BIOS, and not EFI. Since this affects only XP computers and those require bios to function. BIOS with ACPI has always been a poor hack. Windows Computers have always had a hard time returning from sleep with 100% accuracy. Maybe it wasn't windows fault but the bios underneath.

Wait did I just say it wasn't windows fault? damn I have got to get some sleep.

Re:Why?

[Overly Critical Guy]

Because Apple's got the buzz, and randomly tying into it for no reason at all generates more page hits for Slashdot. Accuracy isn't a concern anymore; bombast is.

Submitter didn't RFTA

[rosalindavenue]

Quote: "Since Microsoft's drivers are now believed to be directly involved, then all of Apple's upcoming MacBook Pro systems - which use the Core Duo processor and 945 chipset - should be unaffected by this issue. We have yet to attain access to a MacBook Pro to verify this." Why bring Apple into a conversation about a defective XP driver?

Re:Submitter didn't RFTA

[Richard_at_work]

There are already none Apple Core Duo laptops around (but they are expensive), why not see if the problem exists under Linux on these devices? If it doesnt, then it isnt the hardware. Infact, you should be able to put the Intel iMacs into the same state and see if a larger than expected current is drawn from the mains.

Comon..

[evilNomad]

"Strangely, single-core systems and AMD systems are not affected."

So once again we have a chance to bash Intel, perfect!

Did you ever stop to consider that maybe that specific state, which cannot be reached, is only utilized by the Core Duo? Maybe if AMD had a laptop dual core chip we'd see the same behavior.. But hey, if we can make Intel look bad because of a Microsoft bug, then we are two for two!

This is great news.

[Anonymous Coward]

Great for Apple that is.
Who wants to bet their next marketing campaign is going to take advantage of this deficiency in Windows, like they did when W95 came out?

Re:AMD Looking Good

[goldn_64]

Seeing that this problem is caused by Windows, I don't see the point of dumping Intel over it ;) It's like saying, hey that guy can't handle his car so his car must be useless.

Re:confidentiality agreement

[dysk]

Isnt this a basis for a class action fraud suit? If not, it should be investigated by the SEC at least.

The SEC investigates fraud which victimizes shareholders. This is fraud against consumers, a much less important group.

What about linux?

[maynard]

Our quad opterons (two dualcore) appear perfectly stable under Linux. Have you tried that as a test to rule out hardware?

Apple using sockets?

[Sad Loser]

I understood that macbooks will have socketed processeors (?? for upgrade).

So if I have to send my new macbook (delivery 15th Feb haha believe it when I see it) back for a newer faster processor sometime down the track I won't lose any sleep over it.

Re:confidentiality agreement

[shaka999]

Would you sue ATI or NVIDIA for updating their drivers and getting more out of the hardware? Obviously there was a problem where the hardware wasn't being used to its full potential?

How does the shorter battery life make this defective? If the company had sold this as having a much longer battery life then failed to live up to it then that would be a problem. Just because the software (or hardware bug) isn't shutting down a processor doesn't make this a legal issue.

YEAAAAHHHHHH...

[CODiNE]

This article really pinpoints to me one of the best things of all about Apple switching to Intel, REAL OS COMPETITION. Now at last if Windows seems dog slow, you can't claim it's the Intel chip... or when the Finder seems the suckage, you KNOW it's the Finder and not the PPC chip running at a lower MHz. Once we start seeing these systems getting into people's hands and they notice a real difference between the two OS' on the same hardware you can bet they'll be whining about it and performance will definitely be a focus for both OS venders. In the past they could have been lax about it... thinking "But what can they compare it to?" but now if they can show that Quartz drawing is 3x slower than DirectX or vice versa, you can bet there will be performance updates in the near future. This is better for all of us.

P.S. Linux doesn't really count in this manner because it gets ignored as a "geek OS" and not really something anybody can run.

Re:Yawn, non free sucks.

[Critical_]

Critical_ sees a typical Wintel bug and thinks Apple has a problem. It's an interesting thought, but not one to publish without checking.

I never concluded Apple had a problem. Rather I suggest it could be a problem because Microsoft's ACPI driver communicates with the ICH7-M Southbridge. If I am not mistaken, Apple uses the same southbridge on it's hardware. As the article repeatedly states, this issue can be anywhere on the chain from the southbridge, the Microsoft driver or even the attach peripheral. If it's purely a driver problem then why has it taken Microsoft and Intel 6 months of a non-working fix? Why are single core systems not affected by the same driver? Could this issue affect Linux or Mac OSX users on those platforms? Sure it could be a state-based issue but no one can really know until further testing takes place and Intel/Microsoft release more details.

heh

[nexcomlink]

No offense while I am not a intel supporter or a mac zealot I find it quite funny that every time there is a problem with Intel everyone begins to bash them simply because of a list of problems. No problem simply do a recall. It's quite obvious to me that the Yonah chips where rushed because of Apple. Plus it states XP driver. Nothing related to Apple because they don't use "XP drivers". Or maybe these flaws are intentional so that Windows runs poorly on them. But I seriously doubt they would sacrifice there marketshare just for that. Also AMD has it's own flaws as well, instead of keeping your eye on one thing why not both? From what I read in another /. comment which gave a link to the AMD Duron with a list of 124 flaws. But correct me if I am wrong. Nothing is ever perfect and for you AMD lovers your chip is not perfect either. So live with it. I am pretty sure when you bought a piece of software you just never had to upgrade it right?

full disclosure of bugs

[l2718]

I'm not sure you can label the product as "defective". Software is too complicated to be labelled "defective" just becuase it has bugs. Moreover, I'm not sure you could legally require Microsoft to reveal every bug they know about, especially since the software you bought carried a prominent notice in the EULA saying, roughly "This software is not guaranteed to work; if it fails to function in some way it's not our problem -- you shouldn't have relied on it in the first place". They never promised the ACPI driver will actually work. Note that the GPL carries a similar clause.

That said, I'd rather rely on free software to function as advertized. When the big pieces fail (kernel, web broswer, ...) fixes are usually quick since many experts are working transparently. When small pieces fail (my favorite editor) I can fix them myself and submit a patch.

The other solution, of course, is to pay for warranty. The problem is that no-one is willing to guarantee Windows will work, and that includes the hardware OEM -- I'm sure the people who make the laptop will say that they can't warranty someone else's OS.

Re:full disclosure of bugs

[SilverspurG]

With the very important distinction that I paid $200 for MS EULA to tell me that they are not responsible for broken software. The GPL didn't charge me anything. Even if Debian had cost $20 I still wouldn't be really concerned about bugs. $200, though, is a significant portion of most people's monthly incomes.

Re:full disclosure of bugs

[klingens]

If something isn't working as intended,it's buggy and defective. There is no way around that even if it's just a minor defect.

And yes, the GPL has a similar clause to the MS EULA, but with the GPL and my OS (Debian) I can be sure that the Debian devs don't withhold information about bugs affecting me (bugs.debian.org and Debain Social Contract clause 3).

The problem here is not that bugs in a driver/CPU combo exist. There are many of those: recently there was a bug with AMD dualcores and Cool'n'Quiet under Windows and Linux. There was no fix but it was disclosed. The problem here is that neither MS nor Intel told us about it, and for this they should get their well deserved bad press and a boot to the head

Re:full disclosure of bugs

[kfg]

Software is too complicated to be labelled "defective" just becuase it has bugs.

Just about everything is "too complicated" to be labeled "defective" if you're going to take that approach.

But "defective" is simply as defective does.

If your car engine tears itself apart after 10k miles because a piston was made out of tolerance that is a defect because the part is a piston, not due any actual property of the object itself. It's role is to play a part in a system, and it is the system that defines the defect.

Perhaps you are laboring under the misapprehension that that warranty guaruntees function. This is obvious nonsense as just about everything can only be guarunteed to not function under certain circumstances. Perfection does not exist if only because perfection is defined by the environment.

A warranty simply states that the producing takes responsibility for certain failures.

KFG

Re:Not Intel bashing at all.

[CyricZ]

I'm not sure where you're getting the idea I "hate" Intel. Frankly, I don't have any feelings or emotions towards them. All I care about is providing stable systems for my clients. If a company puts out products that don't work, be it because of their hardware design or the software that runs on their hardware, I will have to avoid such products until I am sure that the problems have been resolved.

I think the problem might be that you don't have much experience when it comes to real-world systems. We're not talking about some college student writing an English report. Often times there are situations where executives are making presentations to potential clients for contracts worth tens of millions of dollars. It could be financially dangerous for such a company if a laptop that's supposed to give them 5 hours of life unexpectedly only gives 4 hours, especially while sending important emails or while doing a presentation.

The minor savings in terms of reduced power usage would never make up for a lost $45 million contract.

WTF?

[Anonymous Coward]

This:

I dual boot between Windows XP Pro SP2 for gaming and Windows XP Pro x64 for work, and both work absolutely perfectly.

and this:

I have been running this system since November with only one or two reboots.

does not compute.

Who would have thought?

[Zencyde]

So, it seems that mixing bad hardware with bad software is bad for the environment? That is too funny...