Can't load Windows 7 (at least easily) on new Intel Skylake
-
@Dashrender said:
If Intel wasn't removing EHCI (which from this line EHCI host controller (USB 2.0 spec) and keeping only the xHCI host controller spec (also known as the universal USB 3.0 specification) reads to me that EHCI = USB 2.0) then there shouldn't be a problem with Windows 7 booting from a USB stick. Additionally the article said Windows 7 boot only supports EHCI (aka USB 2.0)
Windows 7 has a weird limitation on legacy tech that doesn't exist in Skylake. Not the same as removing and the limitation was always there.
The USB 2 specific controller is not included in Skylake, that much appears to be correct.
-
OK now you've lost me.
Did Broadwell include EHCI? from what I can tell, yes it does
Does Skylake include EHCI? no it does not.While technically you are correct, it wasn't removed as Skylake never had it, but your nitpicking. The general consideration is that the previous version had it and the new one does not.
@scottalanmiller said:
The USB 2 specific controller is not included in Skylake, that much appears to be correct.
Which means my statement
@Dashrender said:
no, it's USB 2.0 it's no longer there in the chip.
is correct.
-
@Dashrender said:
While technically you are correct, it wasn't removed as Skylake never had it, but your nitpicking. The general consideration is that the previous version had it and the new one does not.
Not nitpicking when the discussion is about vendors crippling things and making it sound like they have removed things to make it necessary to buy new products from other vendors. This is a normal, common, expect evolution of chip design. New chips are not iterations of old ones (tocks are, ticks are not) and saying that Intel is removing something leads people to a completely different place than saying the new chips don't include it. You can't remove what wasn't there. Skylake is a new chip.
-
@Dashrender said:
@scottalanmiller said:
The USB 2 specific controller is not included in Skylake, that much appears to be correct.
Which means my statement
@Dashrender said:
no, it's USB 2.0 it's no longer there in the chip.
is correct.
Except for the "no longer" statement, it appears to be true. Except doesn't USB 3 support USB 2? So USB 2 is still there, just not the USB 2 specific controller. Right?
-
@scottalanmiller said:
@Dashrender said:
@scottalanmiller said:
The USB 2 specific controller is not included in Skylake, that much appears to be correct.
Which means my statement
@Dashrender said:
no, it's USB 2.0 it's no longer there in the chip.
is correct.
Except for the "no longer" statement, it appears to be true. Except doesn't USB 3 support USB 2? So USB 2 is still there, just not the USB 2 specific controller. Right?
Yep, the no longer is wrong.
Though I wouldn't say USB 2.0 is in the chip since it's clearly not 100% backwards compatible, if it was it would respond to the EHCI requests from Windows 7.
-
Ah, was thinking it was a tick. It's still a new processor, though. The idea of removing implies for example that it could be returned. SOmeone mentioned that maybe this would happen, but it can't as there is no design for that. There is a concept of features being removed from chips, and could be returned on the manufacturing floor. But this is a core design change.
-
@Dashrender said:
Though I wouldn't say USB 2.0 is in the chip since it's clearly not 100% backwards compatible, if it was it would respond to the EHCI requests from Windows 7.
Being 100% compatible with past implementations and not being compatible with USB 2 are different things, though. I'm not savvy enough with the USB 2 spec to know if the one was doing something "extra", one is lacking something, etc.
-
I'd say it's clearly lacking something as you can't boot from a Windows 7 USB stick. If it wasn't missing, wouldn't you expect it to work?
-
@Dashrender said:
I'd say it's clearly lacking something as you can't boot from a Windows 7 USB stick. If it wasn't missing, wouldn't you expect it to work?
I have no reason to specifically have that expectation. Maybe you know more about USB 2 specs than me, but is that specific boot functionality part of the spec that is lacking later, or is Windows doing something odd outside of the USB 2 spec? The latter seems far more likely to me.
-
@scottalanmiller said:
Ah, was thinking it was a tick. It's still a new processor, though. The idea of removing implies for example that it could be returned. SOmeone mentioned that maybe this would happen, but it can't as there is no design for that. There is a concept of features being removed from chips, and could be returned on the manufacturing floor. But this is a core design change.
I see where you are going here - let's be as precise as we can be. I definitely appreciate the correction. But the general user/consumer doesn't see or understand or more importantly, care about these distinctions. All they know now is that something that was there in version versions is now missing, and as such something they have been doing for 5+ years they can no longer do. Is there a term or best way to describe this?
As for the article, while not emphasizing it, they did mention they didn't believe that Intel and Microsoft where in cahoots to force people to a new MS OS, but instead that this is a normal evolutionary trait. Old things get dropped as new things become mainstream and even newer things are added. There are many other articles around the web crying wolf about Intel and Microsoft being in bed together to force people to Windows 10, but not this one.
-
@scottalanmiller said:
@Dashrender said:
I'd say it's clearly lacking something as you can't boot from a Windows 7 USB stick. If it wasn't missing, wouldn't you expect it to work?
I have no reason to specifically have that expectation. Maybe you know more about USB 2 specs than me, but is that specific boot functionality part of the spec that is lacking later, or is Windows doing something odd outside of the USB 2 spec? The latter seems far more likely to me.
That is a great question, I have no idea.
-
@Dashrender said:
But the general user/consumer doesn't see or understand or more importantly, care about these distinctions.
And this matters why? Why are IT professionals and REALLY care about these things. When someone is implying that Intel is helping cripple Windows to aid adoption we need to know, immediately, that there is no reason to believe this. What consumers don't understand, it is our job to understand. If IT did what consumers did, we'd not be IT but end users.
-
@Dashrender said:
All they know now is that something that was there in version versions is now missing, and as such something they have been doing for 5+ years they can no longer do. Is there a term or best way to describe this?
Confusion. Because they've lost no functionality. Windows 7 never did, and was never going to do that, on this generation or later of chips. Nothing has been lost, just nothing new gained. Windows 7 still works as it always did on chips that still work as they always did.
If you want a more general term, it is called normal obsolescence. All software eventually doesn't work on newer hardware.
-
@Dashrender said:
As for the article, while not emphasizing it, they did mention they didn't believe that Intel and Microsoft where in cahoots to force people to a new MS OS, but instead that this is a normal evolutionary trait. Old things get dropped as new things become mainstream and even newer things are added. There are many other articles around the web crying wolf about Intel and Microsoft being in bed together to force people to Windows 10, but not this one.
Correct, this article makes it clear that they feel this was normal advancements and no reason to feel that anything was amiss. We've known this was coming for a long time, it would be ridiculous, IMHO, for new chips to have dedicated USB 2 controllers. That seems pretty wasteful. You don't want extra, unused hardware on your processors. This thread was started with a very strong theming of functionality being removed or crippled, of which there is no reason to suspect there ever was.
-
@scottalanmiller said:
@Dashrender said:
As for the article, while not emphasizing it, they did mention they didn't believe that Intel and Microsoft where in cahoots to force people to a new MS OS, but instead that this is a normal evolutionary trait. Old things get dropped as new things become mainstream and even newer things are added. There are many other articles around the web crying wolf about Intel and Microsoft being in bed together to force people to Windows 10, but not this one.
Correct, this article makes it clear that they feel this was normal advancements and no reason to feel that anything was amiss. We've known this was coming for a long time, it would be ridiculous, IMHO, for new chips to have dedicated USB 2 controllers. That seems pretty wasteful. You don't want extra, unused hardware on your processors. This thread was started with a very strong theming of functionality being removed or crippled, of which there is no reason to suspect there ever was.
When was the last time that Intel released hardware for the PC market that wouldn't run the currently supported OSes? Basically, during XP's support cycle was there ever Intel hardware you couldn't install XP on? I suppose you might say - but you can install Windows 7 on this hardware, but you can only do it using a DVD drive and have to have a PS2 port to do so.
This seems like a first, but your memory is better than mine, perhaps they've phased other technology out while it was still in mainstream use. But really I guess this boils down more to your earlier question.
Is the USB 3.0 spec missing something that the USB 2.0 spec included, or did MS find their own way to do something that just happened to work with the way Windows 7 used EHCI?
-
@Dashrender said:
When was the last time that Intel released hardware for the PC market that wouldn't run the currently supported OSes?
It does run Windows 7. You are being very dramatic. All that has happened here is that Windows 7 depended on a very strange, legacy hardware component for a very uncommon need that affects almost no one. To the same "normal consumers" you like to talk about, nothing has changed.
-
Part of the issue is that Microsoft "supports" older and older systems. When Windows XP came out, people were not broadly trying to use NT 3 series OSes. The desire to avoid current OSes (the only ones that can ship with Skylake) creates new issues. Why have consumers latched on to desired obsolescence? I have no idea. It's a problem with how consumers, and sadly a lot of IT, thinks. Windows 7 is three versions behind. New hardware and new software and created to leverage each other. Using old software on new hardware is not likely to give optimum results, it can't be tuned for it. For the hardware makers to make hardware based around obsolete software needs of a niche low end market makes no sense, we'd rarely see hardware advancements and the cost of hardware would skyrocket.
-
EHCI is not USB 2. EHCI is a specific implementation of USB 2. As Intel makes the specs for each, my guess is that they get them right (this isn't a guarantee, but it leans that way.)
Here is what Intel says about EHCI:
Enhanced Host Controller Interface for USB 2.0: Specification
The document describes the enhanced host controller interface for universal serial bus revision 2.0, including the interface between system software and host controller hardware interface.
Specification Features
• Full support for all USB 2.0 features.
• Low-risk support for full- and low-speed peripherals.
• Sys -
From what I can tell, both EHCI and xHCI implement USB 2 completely. The issue was that Windows 7 was written to depend on EHCI specifically rather than being written to the USB 2 spec alone.
-
Well, perhaps now that we see MS releasing OSs much more frequently (although that whole idea might be dead now considering the rumors that Windows 10 is the last new OS from MS and from now on it will just be continuously updated) we'll see the mainstream keep up a bit better.
That combined with the fact that MS is upgrading everyone on 7 or newer to Win 10 free, this might be a non issue. That is assuming of course that enterprises migrate corporate deployments away from Windows 7 (which I'll admit I personally think seems likely).
Since normal consumers themselves barely ever reinstall windows, you're right, this is not a real issue, as they will rely on their recovery solution or support personal to resolve their issues.
