Slashdot Mirror
Intel Says Clover Trail Atom CPU Won't Work With Linux
girlmad tips this news from the Inquirer:
"Intel's Clover Trail Atom processor can be seen in various non-descript laptops around IDF and the firm provided a lot of architectural details on the chip, confirming details such as dual-core and a number of power states. However Intel said Clover Trail 'is a Windows 8 chip' and that 'the chip cannot run Linux.' While Intel's claim that Clover Trail won't run Linux is not quite true — after all, it is an x86 instruction set, so there is no major reason why the Linux kernel and userland will not run — given that the firm will not support it, device makers are unlikely to produce Linux Clover Trail devices for their own support reasons."
Chips aren't exactly designed to "run Linux" or any other OS. It's Linux that supports CPUs.. NOT the other way around.
All this means, is that Intel doesn't want to help. It does not mean it won't run Linux. Linux always finds a way to work.
I can't see what possible benefit it is to Intel to deliberately limit the market for their processors. Unless they are doing this for Microsoft's benefit, in which case, surely, there are anti-trust implications?
I'm old enough to remember when discussions on Slashdot were well informed.
Doubt it. They can choose to support or not support whatever they want. They just can't actively use their current monopoly position to harm competition in another market (operating systems). If they put in some special instructions that actively sabotage the Linux kernel from running, that would be one thing. From what it sounds like though, they are merely not providing drivers/source code for Linux for some of the CPU features for this platform. Of course since a lot of geeks will try to get Linux running on a toaster for the lulz, I expect this to only be a short-term hindrance.
If there is one thing life has taught me it is that anything can run Linux. All intel has done now is simply issued a challenge which my guess is won't take long for some skilled hardware hacker.
So, as an aside, isn't the entire point of a tech aggregator to provide a technical summary? Not just copy and paste the article's summary... anyway...
FTFA:
Intel went to great lengths to highlight the new P-states and C-states in which it can completely shut down the clock of a core. The firm said the operating system needs to provide "hints" to the processor in order to make use of power states and it seems likely that such hints are presently not provided by the Linux kernel in order to properly make use of Clover Trail.
In other words, Intel has added new capabilities to Clover Trail that allow enhanced power management, and Linux doesn't currently support it. Anyone who thinks that this will continue to be the case for much longer is a moron, especially if Intel continues to release its architecture datasheets, which we have no reason to think that they won't.
The article really says: It can't run Linux because there's no support for it in Linux, and there's no support for it because it's literally brand-new.
IANAL, but I recall that Microsoft got in a bit of trouble because early versions of Windows were designed NOT to run on top of Digital Research DOS. Not going out of your way to support something is one thing, being exclusionary and abusing a monopoly position is another
90% of everything is crap. Also, crap is relative.
Sigh. Why is this one of the first reactions when a manufacturer doesn't do something you want them to do? Seriously, Intel not only does not have a monopoly of tablet processors, I would say they don't even have a majority. ARM processors power the vast majority of tablets. Intel is only hurting themselves by not supporting Linux.
Well, there's spam egg sausage and spam, that's not got much spam in it.
"Not supported" is very different from "Can't run Linux". I would call this monopoly abuse.
Nae king! Nae laird! Nae yurrupiean pressedent! We willna be fooled again!
It is the first time ever that Intel announced direct hostility toward some piece of software -- I hope, it's just someone's fuckup and not a policy change.
Contrary to the popular belief, there indeed is no God.
...into the Appelsk walled garden that Windows 8 appears to be heralding in (Windows Store only apps, "for your own security, comfort and ease of use", coming to you in Windows 8.5/9). Last thing our walled gardeners want is an alternative OS weed like Linux, working perfectly on the same hardware...
Please, notify the European Commission.
I am positively sure they will not like this.
(I don't have any appropriate channels, otherwise I would have done it)
how, in a media swamped with Apple mania, do you get attention for a processor launch?
They never actually did it, and they never got into trouble for considering it either.
There was a beta of one version of Windows 3.x that put up a message along the lines of "This software is unreliable and unstable and will EAT YOUR BABY if you run it over DR-DOS" (well, words to that effect.) IIRC it was only in the beta, the version that was sold would run on everything.
You are not alone. This is not normal. None of this is normal.
Is there any source for this statement besides The Inquirer? They're basically a tech tabloid and have gotten a lot of things wrong (or overly sensationalized) in the past. I checked Anandtech and Tom's Hardware, both of which covered Intel's presentations this week. No mention of this. I did a Google search for "clover trail" "Windows 8 chip" and found ONLY the Inquirer article and other articles and blog posts directly quoting and linking to it. No reliable third-party tech sites saying the same thing.
This doesn't make sense in terms of Intel's overall philosophy. They have always been good about Linux support for nearly everything else – they don't want to get themselves tied in too closely with Microsoft, for fear that this would reduce their leverage.
I think this story is bullshit. A generous interpretation would be that the reporter heard that the chip ran Windows 8 and that Linux *currently* did not have the necessary support for the "new P-states and C-states" in Clover Trail, and misinterpreted that as saying that only Windows 8 will ever be officially supported. A less generous interpretation is that the Inquirer knowingly made up this crap to get more page hits. In any case, I expect Intel to make their actual position clear soon enough, now that this story seems to have gone viral.
Let's say the reason is either one.
Let's say that Intel wants to limit the audience for the chip, and cut their own sales. Let's say that AMD, VIA, and the ARMs makers will be delighted to fill in any vacuum.
Do.We.Care?
---- Teach Peace. It's Cheaper Than War.
Actually it means Intel won't support running Linux on it.
Apparently Wintel is alive and well.
Again, Intel is only hurting themselves. The vast majority of tablets run ARM and not Windows. It's not anti-trust for Intel to make bad strategic decisions in a market that they have very little influence.
Well, there's spam egg sausage and spam, that's not got much spam in it.
if it's a win8 only chip, it quite possibly includes the damn "Trusted Boot" feature that MS wants on all hardware with UEFI. This means that unless the fucking bootloader and OS is blessed by MS, it wont run.
Now if this is only for those Win8 Tablets and such that MS wants to get to market, then the chip may well run Linux or any other OS that's x86 based though w/o the blessings of Intel. Another issue is that Intel May restrict sales of this to OEM's with a minimum of 10k per order. Another possibility is that Intel could limit this chip's initial sales due to fab problems and/or to prevent canabilizing higher profit chip sales if this chip is as big an improvement as the C2D was over the Netburst/P4 architecture.
Mod me up/Mod me down: I wont frown as I've no crown
Interesting perspective, but Microsoft obviously did something bad enough to compel them to pay Caldera at least 155 million (http://www.windowsitpro.com/article/news2/microsoft-settles-dr-dos-antitrust-lawsuit) . Something tells me if this was just a blip in a beta, they'd take it to court.
90% of everything is crap. Also, crap is relative.
And, yes, I know OS X and Linux share a common heritage.
What color is the sky in your world? They're both Unixlike but that's as far as that goes. Linux is descended from Minix and runs a GNU userland. OSX is descended from NeXTStep which is descended from BSD. They are almost as different and differently-descended as they could be and still be able to run much of the same code without major changes.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
actually, parent is correct. There are antitrust issues to be investigated here. This has nothing to do with choosing to support something or not, and is a very bad move by intel. If it were choosing they would say "this processor is not designed for linux" not "this processor will not *work* with linux". Sounds small, but it's of critical importance. The reality is that the antitrust issue is not with Intel - it is with Microsoft.
Why should anyone use the x86 instruction set if they're explicitly saying that things are not compatible? All they're trying to say to people is "please use ARM", which is not the smartest idea. That is entirely different from what intel is implying, which is that the BIOS issues regarding windows 8 preventing other operating systems from running...that issue from before.
So all this is, is basically antitrust fodder against MS.
This is completely irrelevant, and AMD does not make low-power x86 chips anymore.
They most certainly do. On the very low side, they make Geode. In the middle end (for low power) they have Semprons at 8 watts. They also have a varieety of SOC.
There aren't any antitrust issues here. Intel can do whatever it wants with it's processors so long as it doesn't use it's processors (I'm not even sure you could call Intel a monopoly in processors, but that would be for a court to decide) to give another of their products an unfair advantage.
Pretty much the same for Microsoft. Unless you think somehow Microsoft strong armed Intel into it, and can prove it in court. Even then it would be difficult. You would have to prove that Microsoft abused it's monopoly position in OSes to do so in a way that harms consumers. Good luck with that.
I can tell you that I ran windows 3.0 and 3.1 on DRDos 6 with no problems whatsoever. I never owned or used Microsoft DOS. So if there was some compatibility or stability problem I never saw it.
"Intel is just not going the extra step to allow Linux compatibility"
According to TFA: However Intel said Clover Trail "is a Windows 8 chip" and that "the chip cannot run Linux".
That's not saying "We won't support it" that's LYING IN MARKET about the capabilities of its chip and causing direct harm to a competing kernel and subset of operating systems based upon that kernel.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
"There aren't any antitrust issues here."
Bullshit, Intel is falsely advertising that a chip with all the standard (for today) x86 instructions will not run Linux, which is an x86 compatible kernel, and says that the chip is for Windows 8. Intel is colluding with Microsoft in this instance to create an anticompetitive market.
FALSE ADVERTISING IS STILL ILLEGAL AND AN ANTITRUST ISSUE WHEN A CONVICTED MONOPOLIST IS INVOLVED.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
I think this mostly due to the PowerVR SGX graphics engine (remember the gma500 poulsbo). for the gma500 intel made a binary linux driver that did not impress anyone. I guess for clovertrail they are just not bothering with releasing a binary driver.
So it might work fine as a CPU, but have no graphics acceleration. however for a tablet chip that cannot play video or composite a desktop in software, it might be effectively useless.
Well, I never thought I'd be standing up for Microsoft, at least a little, But IMHO they had at least a LITTLE justification for putting up the warning message. Old Windows HAD to make MANY patches into the DOS resident code, and it depended on MANY undocumented data areas inside the DOS resident code. Any DOS clone, if it was to have a chance of running Windows, had to be very carefully engineered to match all those undocumented locations in DOS. The odds of Digital Research being able to guess all the exact locations that Windows depends on, and will depend on, is somewhat slight.
You know that this is a processor, right? A processor is something that you use at both phones,tablets, netbooks, notebooks, desktops, workstations and servers. Also, all of those categories are fuzzy, and processors do leak to the neigboring ones.
For the looks of it, this one is a tablet's processor. On tablets, iOS has most of the market, Linux is a minority and Windows does not even mark outside of the error margin, that last OS is the one Intel is going to support. Of course, it will leak to netbooks and notebooks, where Windows rules (but is losing space fast for OS-X).
I have no idea why Intel would even make such a decision, and I doubt AMD, VIA, or ARM management agree with it. From the public info it just doesn't make any sense, there must be something Intel is hidding.
Rethinking email
They tried writing drivers themselves and again they sucked.
Dead wrong. Intel drivers are excellent and I and many others have had great success with them. They also usually work quite closely with the kernel community as a whole to make sure things work as expected; that's why what this article is saying seems to out of character for Intel. For instance, try searching for "intel.com" in the git commit log. Lots of kernel developers are on Intel's payroll, including core people like Alan Cox.
No, that's not what TFA says:
"The firm said the operating system needs to provide "hints" to the processor in order to make use of power states and it seems likely that such hints are presently not provided by the Linux kernel in order to properly make use of Clover Trail."
I doubt this will be very difficult for Linux to put into the kernel.
There are two types of people in the world: Those who crave closure
There aren't any antitrust issues here. Intel can do whatever it wants with it's processors so long as it doesn't use it's processors ... to give another of their products an unfair advantage.
Not true. The range of Anti-Trust includes oligarchic rings. If Intel purposely altered the chip by examining instructions that Linux calls but Windows 8 Metro doesn't, and "conveniently" removes/alters those instructions, it's anti-trust. It's a variant on collusion. If you have multiple parties of overlapping business sectors working to block other products, it triggers legal implications, just different ones.
As stated elsewhere, the geeks will get it working again, but the move would result in lost development resource just fixing the problem Intel made.
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine
http://www.phoronix.com/scan.php?page=news_item&px=MTE1ODA
Maybe it's intentional, maybe it's not, but it does seem to be a trend.
From what I can tell Intel is creating API calls to these new processors to shutdown cores. These APIs are being provided to MS only. Best case is that Linux on these processors will be less efficient. Worse case is that the processors don't run properly (until it is reverse engineered) in Linux.
Well, there's spam egg sausage and spam, that's not got much spam in it.
I think I could count on one hand the number of Windows 8 users worldwide and I think that 7 of them are bored with it.
My name is Inigo Montoya. You killed my father. Prepare to die.
Yeah they FUD it to death ! They did it verbally using that line from the beta and by paying industry magazines to repeat that message even thought it was a lie.
Jehovah be praised, Oracle was not selected
No, I remember it well, and it was a specific check to print the scary message and do nothing else. It was quite clearly marketing driven, not a lick of technological reasoning behind it.
Infuriate left and right
If it has an incompatible memory model, it wouldn't be x86 anymore, would it?
You could do that because that code was disabled in the production version - though still present.
For the technical details see this fine article.
Real life is overrated.
Your comprehension must not be up to par, sir. Anyone should know I'm talking about Intel's claim that "This will not run Linux" and "This is a Windows 8 Chip" means linux kernel vs Microsoft kernel.
Somebody's skills aren't up to par; I'd vote for the person who spoke of "a competing kernel" without indicating why Intel would care about two competing kernels when they make neither and have supported both and without bringing up collusion - the only thing that could be at issue here would be collusion between Microsoft and Intel, so I don't think anybody could go after Intel alone on an anti-trust issue, they'd have to go after both Intel and Microsoft.
(I'm also a bit skeptical that any antitrust authority would give a rat's ass about Intel just stating that "this chip will not run Linux" - assuming they really did say that; the Ars Technica piece merely says that the Inquirer is claiming that they said that, and are only themselves reporting what David Perlmutter said. I rather doubt Linux developers would be dissuaded from trying to make Clover Trail work merely because Intel says it won't - heck, it might encourage them. It's only a big issue if Intel are telling the truth, e.g. if you have to support the new power management features to run on Clover Trail at all and if Intel are only documenting those features in non-public documents.)
Intel is using a dominant market position
So this chip expected to be used for desktop/laptop computing to a significant degree? Intel hardly has a dominant market position for tablet computing (and Microsoft doesn't have one, either).
to say "This won't run Linux, use Windows 8 instead!"
...which means that, if there are any antitrust issues, they'd have to involve collusion, given that Windows 8 isn't an Intel operating system.
Some where, some Linux developer just yelled "Challenge Accepted" at his computer screen. I've got money that it'll be working before this thing hits the market.
Any programmer worth their salt can count to at least 31 on one hand, and 1023 on two hands.
For a comparable example, look at the Raspberry Pi. On one hand, it's a cheap SBC intended to run Linux. But the SoC itself has huge parts without public documentation, and anybody trying to write open drivers for it is going to have a huge challenge ahead. For now, it's "Broadcom's buggy .ko binaries, or nothing at all."
It's probably safe to say that someone will almost certainly find a way to run Linux on this chip. What they probably WON'T be able to do is take advantage of the chip's full capabilities, and Linux running on it will always be crippled or compromised in some way (at least, until long after nearly everyone has ceased to care about this specific chip).