Given the shock they’ve had to endure as a result of the ongoing Secure Boot saga over the past year or so, Linux geeks may perhaps be forgiven if they’re a tad sensitive to apparent attempts to exclude Linux from other new technological developments as well.
It was perhaps less-than-entirely surprising, then, that FOSS fans’ reaction to the initial news about Intel’s Clover Trail Atom chip came so swiftly.
“IDF: Intel says Clover Trail will not work with Linux” was the headline that broke the news over at the Inquirer, and all heck broke loose after that.
In no time at all more than 400 comments appeared on Slashdot alone as geeks hastened to hash out what, exactly, the news really meant.
“I can’t see what possible benefit it is to Intel to deliberately limit the market for their processors,” wrote Slashdot blogger Simon Brooke, for example. “Unless they are doing this for Microsoft’s benefit, in which case, surely, there are anti-trust implications?”
Similarly, “please, notify the European Commission,” wrote G3ckoG33k. “I am positively sure they will not like this.”
Then again: “I’m sure there are hundreds out there that have already stood up and stated, quite boldly, ‘CHALLENGE ACCEPTED!’ and are already hard at work,” suggested blogger JonHerr in the comments over at PCWorld.
‘Intel Reverses Course’
The virtual ink had barely dried on those initial reports, however, when there appeared a new twist in the tale.
“Intel reverses course, bringing Android to Clover Trail chips for tablets,” was the headline at ITworld.
“Intel planning Clover Trail variant for Linux” was the word at The H.
Later in the week, the Inquirer offered more thoughts on the matter, suggesting that the Linux-friendly chip is actually intended for smartphones rather than tablets.
In any case, Linux fans down at the blogosphere’s Broken Windows Lounge had plenty to chew over.
‘Some Feature to Impede Linux’
“It is really quite obvious that Intel should make multiple versions of the same processor; when has a processor core only been used once since before x86 processors became internally RISCy?” suggested Hyperlogos blogger Martin Espinoza.
“On the other hand, it’s still difficult to see what Intel is playing at,” Espinoza added.
“Given that they’ve specifically taken the time to announce that there will be a Linux-compatible version, one might reasonably imagine that the Intel Clover Trail-based systems designed to run Windows may well have been designed with some feature designed to prohibit or impede Linux,” he pointed out.
‘There Must Be Some Other Reason’
Similarly, “I just can’t buy Intel’s explanation for this, since it would be easy for Intel’s own Linux kernel developers to add the needed support to Linux,” consultant and Slashdot blogger Gerhard Mack said.
“There must be some other reason, and I bet that reason is that Intel is desperately hoping that Windows 8 will extend the Intel/Microsoft juggernaut from the PC market into the tablet market,” Mack added.
Indeed, “Clover Trail is Intel’s hope to keep Atom relevant,” blogger Robert Pogson agreed. “M$ and tablets have killed off the netbooks, so Intel wants M$ to grow in the tablet-space, buying up Intel’s production.”
‘A Desperate Measure’
That strategy won’t work, however, “for a number of reasons, including being late to the game and costing more — a lot more than Android/Linux on ARM,” Pogson added.
“Foolishly, Wintel is trying to lock people into using WARM while driving them away with silly restrictions like secure boot and limitations on applications.
“Wintel on tablets is equally foolish,” he opined.
“Power consumption is huge for other Intel chips,” Pogson noted. “An Atom with M$-only features to manage power by shutting down unused parts of the chip is a desperate measure to make a bloated chip seem smaller.”
‘Cripple-ware on a Scale Never Seen Before’
A better approach would be to use ARM, Pogson suggested.
“When consumers see the prices for very limited performance increases compared to ARM, they will not be impressed,” he explained. “By making the chip M$-only even temporarily, Intel assures itself of a huge stock of unsold chips because ‘8’ is going to flop everywhere, not just on tablets, because it’s cripple-ware on a scale never seen before.
“Users will want a windowed experience, and they will have to use GNU/Linux to get it on new hardware,” Pogson concluded. “If Intel ships M$-only hardware that won’t run GNU/Linux, they won’t be able to sell much.”
‘An Uncertain Future’
The only reason Intel would need a variant to support Linux “is if the processor itself has some sort of cryptographic key relationship with the OS,” noted Chris Travers, a Slashdot blogger who works on the LedgerSMB project.
“Unless the chip itself is verifying the OS that’s running on it, there would be no reason to come up with a separate model,” he explained. “It would be the responsibility of the OS to support the chip, not the other way around.”
Of course, “given what is known about Microsoft’s requirements for Windows 8 tablets, this certainly seems likely,” Travers added. “I doubt they are placing all their bets on Windows — they are just making a product that makes it easier to meet Microsoft’s licensing demands.
“Since this is a new product with an uncertain future, however, it isn’t clear what difference this makes,” he concluded.
‘Two Losers Getting Together’
“This is a puzzling story because it gets everything backwards,” Google+ blogger Kevin O’Brien opined.
“Intel has never designed any chip ‘for Linux,'” he pointed out. “In the days of Wintel, it worked somewhat closely with Microsoft, but Linux was simply designed to work on x86 chips. If there is any need to do so, Linux will build in compatibility with these new Atom chips.”
The real question in O’Brien’s mind, however, “is why anyone wants to use Intel chips for low-power mobile applications,” he told Linux Girl. “Has Intel ever shown a competitive chip in this space?
“I read this as two losers in the mobility space getting together because no one else wanted to dance with them,” he concluded.
‘With or Without Intel’
“Intel’s support of Linux on their processors couldn’t matter to me less,” asserted Google+ blogger Linux Rants, who recently wrote a post of his own on the topic.
“There was no support for Linus when he was writing Linux, and no support from Intel is required to make Linux work,” Linux Rants explained. “I’m fully confident that Linux developers can make Linux work on my toaster if they so chose.”
It’s not clear if Intel “bowed to pressure from the community when it announced that there would be a Linux ‘version’ or if that was their plan all along,” he added, but “I don’t know that I care. Linux will continue on with or without Intel.
“All Intel can do is decide if they want to come along or be left behind,” he concluded.
‘You Should Buy AMD’
Last but not least, Slashdot blogger hairyfeet had a concrete suggestion.
“Here you have two companies: one has opened up their specs, just as the community asked (AMD), and the other… is buying PowerVR GPUs, which makes Lexmark look Linux-friendly,” he explained.
“The choice seems pretty simple to me,” hairyfeet concluded. “One is open source-friendlier than the other, so if you care about open source, you should buy AMD.”
I had moved on and was not paying much attention to the processor wars. I was back to a simple benefits vs cost evaluation when making a choice.
Intel has now turned the tables in favor of AMD. It is not the specifics of this discussion that has influenced me, but the character of a company that would make this choice.
I AM disappointed that Intel would continue collusion with M$.
btw…M$ is not the only beneficiary to this. Apple is more threatened by Linux than M$. The feature list between Linux and Mac is similar, and Apple also has a relationship with Intel.
that if the community truly cares about Open Source its time to put its money where its mouth is. Intel has cooked up UEFI which lets them lock down the system? AMD has gone Coreboot on their new chipsets. Intel using PowerVR on their low end mobile chips? AMD has opened the specs as it was asked to with the exception of the UVD engine which they legally can’t because it contains the HDCP code and that’s not their to give. You can still get full hardware acceleration, just use Catalyst drivers.
So the choice seems pretty simple to me, one company is being a LOT more FOSS friendly than the other so the choice should be a no brainer. This will be a good test of the community as well, as if AMD sales don’t go up after doing everything the community asked, why should other hardware companies support you? You asked for the docs? they gave them. You asked for the specs to make your own drivers? they not only gave all of them they legally could but actually hired devs to help out with the FOSS drivers to make them better.
So I’d say the answer is clear, if you want to support FOSS, be it server, desktop, or laptop, you should buy AMD chips and show hardware companies that opening specs and docs is good for business.