Linux Creator Linus Torvalds recently expressed frustration with change proposals to the Linux kernel that are only “required” due to faulty hardware. Phoronix notes that Torvalds ranted on the Linux kernel mailing list over the last week, saying that buggy hardware should not be Linux’s problem, suggesting kernel developers just say “no” and see what the “hardware people” say.
Yes, but did he slam the hardware manufacturers?
Give them credit. They actually used his words instead of the hyperbole. That should be encouraged.
He’s got a good point, and it’s true for anything. If you try to adapt or cover up an issue, that doesn’t fix or address the problems, and it encourages more problems in the future to deal with (that will probably multiply into more) because you always take care of it. Throw it back at them explaining they have to meet the same standards that everyone does, and it’s not your problem anymore.
Can Linus get Microsoft to follow? If Microsoft follows they can make Windows update give a message “Windows has detected you have defective hardware, contact [manufacturer] and demand a fixed replacement”. This would force it.
I but Microsoft won’t (in part because Windows and Office has many more bugs that they are constantly trying to fix and so a showdown will be bad for Microsoft)
The problem is that it takes months (in the absolute best case) to years to fix hardware bugs. And every piece of hardware has bugs, just like all software has bugs. So does everyone always have known compromised computers?
The timelines for software vs hardware are just so different. After they both figure out the problem, find a viable solution and test that solution, there’s still so much time left for the hardware fix to come out. With current pocesses, it’s like 4-5 months to get silicon back. And that’s if you’ve already reserved the time for them to make the masks and the new chips. And that’s just to get the first chips back. You still have to get them to consumers.
There are chipset design issues and there are firmware issues. The former is much more difficult to address quickly than the latter, sure.
Torvalds’s point, though, is that hardware developers (Intel specifically) keep making changes that “fix” imaginary problems while screwing over compatibility, and trying to shift the onus of making it work to the volunteers who contribute to open source instead of just paying their engineers to produce working firmware.
If the problem were only with defective silicon, I’d agree with you (to an extent), but this is not really an issue with the circuits.
