Author Topic: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)  (Read 6332 times)

0 Members and 1 Guest are viewing this topic.

Offline jr2

  • The Mail Man
  • 212
  • It's prounounced jayartoo 0x6A7232
    • Steam
Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

http://reddit.com/r/intel/comments/7npcfx/kernel_memory_leaking_intel_processor_design_flaw/

EDIT:

Looks like we've got initial fix benchmarks:

Hardware Unboxed initially benchmarked the fix: https://www.youtube.com/watch?v=_qZksorJAuY.

Also, there's a second vulnerability, Spectre, that affects Intel, AMD, and ARM (smartphone processors):

Guess what: it's not just Intel users.

https://www.nytimes.com/2018/01/03/business/computer-flaws.html

We're all ****ed.

AMD response to Spectre (they say risk for AMD chips for the two Spectre attacks is "zero, and almost-zero" because of the way their processors are designed - we'll see):

https://www.amd.com/en/corporate/speculative-execution


EDIT2:

Linux excludes AMD from Meltdown patch: (as they claim they aren't affected by it)

https://twitter.com/phoronix/status/948725135971897345

http://reddit.com/r/Amd/comments/7nzts4/linux_accepts_excluding_amd_from_pti_if_amd_is_so/
« Last Edit: January 04, 2018, 07:49:03 pm by jr2 »

 

Offline The E

  • He's Ebeneezer Goode
  • 213
  • Nothing personal, just tech support.
    • Steam
    • Twitter
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
This is bad news.

For those of us running data centers and cloud applications, that is; gaming performance, at least on Linux, does not seem to be affected.

If I'm just aching this can't go on
I came from chasing dreams to feel alone
There must be changes, miss to feel strong
I really need lifе to touch me
--Evergrey, Where August Mourns

 

Offline JSRNerdo

  • [`_`]/
  • 29
  • Gone!
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
The moment I saw

Quote
Think of the kernel as God sitting on a cloud, looking down on Earth. It's there, and no normal being can see it, yet they can pray to it.

I immediately made my mind up

that this is a non-issue
Former Inferno lead, BTA fredder-ish and DE fredder. Driven out by ordinary fascists the_e, aesaar and general battuta. Will return if they're ever removed.

 
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
lol, it's definitely not a non-issue though; a 15% performance hit on database queries, which is what some of the benchmarks are showing, is a big ****ing deal
The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell.

 
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
initial speculation was that the (still very hush-hush) exploit potentially involved direct circumvention of memory protection through preemptive execution, far more serious than just compromising KASLR
The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell.

 

Offline MP-Ryan

  • Makes General Discussion Make Sense.
  • Global Moderator
  • 210
  • Keyboard > Pen > Sword
    • Twitter
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Well, isn't that lovely.  Suddenly the performance benefits of Intel chips over the last several years aren't looking as great as they once did.  What a screwup.
"In the beginning, the Universe was created.  This made a lot of people very angry and has widely been regarded as a bad move."  [Douglas Adams]

 
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
It should be noted that the performance hit from this is more nuanced than '15% slower'; it's heavily dependent on the task. If your main interest in CPU performance is for gaming, it looks like this isn't going to be a very big deal. If you're a database admin, it might well be.
The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell.

 

Offline Mika

  • 28
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Now that could explain why I'm seeing Kernel level crashes occasionally. Especially if this "feature" allows accessing protected memory just like that.

But it could be that Windows 7 just doesn't like 4K too.
Relaxed movement is always more effective than forced movement.

 

Offline MP-Ryan

  • Makes General Discussion Make Sense.
  • Global Moderator
  • 210
  • Keyboard > Pen > Sword
    • Twitter
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Guess what: it's not just Intel users.

https://www.nytimes.com/2018/01/03/business/computer-flaws.html

We're all ****ed.
"In the beginning, the Universe was created.  This made a lot of people very angry and has widely been regarded as a bad move."  [Douglas Adams]

 

Offline jr2

  • The Mail Man
  • 212
  • It's prounounced jayartoo 0x6A7232
    • Steam
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Guess what: it's not just Intel users.

https://www.nytimes.com/2018/01/03/business/computer-flaws.html

We're all ****ed.

To be clear, there's two separate flaws; Meltdown only affects Intel CPUs, and requires a performance-draining patch (which will also be applied to the non-affected AMD chips, because "glory to Intel!" (no, really!)) and Spectre, which affects all (to include the ARM processors in your smartphones) systems.  There is no known fix for Spectre, besides new hardware.

So yeah, we're screwed once Spectre leaks.

 

Offline MP-Ryan

  • Makes General Discussion Make Sense.
  • Global Moderator
  • 210
  • Keyboard > Pen > Sword
    • Twitter
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Exactly.  Meltdown is the least of our problems.
"In the beginning, the Universe was created.  This made a lot of people very angry and has widely been regarded as a bad move."  [Douglas Adams]

 

Offline The E

  • He's Ebeneezer Goode
  • 213
  • Nothing personal, just tech support.
    • Steam
    • Twitter
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
To be clear, there's two separate flaws; Meltdown only affects Intel CPUs, and requires a performance-draining patch (which will also be applied to the non-affected AMD chips, because "glory to Intel!" (no, really!))

You're so very wrong sometimes. All we know is that AMD has said that their chips aren't affected by Meltdown and that the Linux Kernel devs have decided to err on the side of caution and not accept a patch that would disable the Meltdown fix on AMD processors at this time. This isn't about "glory to intel", this is about not trusting a company blindly when they say something.
If I'm just aching this can't go on
I came from chasing dreams to feel alone
There must be changes, miss to feel strong
I really need lifе to touch me
--Evergrey, Where August Mourns

 

Offline jr2

  • The Mail Man
  • 212
  • It's prounounced jayartoo 0x6A7232
    • Steam
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
To be clear, there's two separate flaws; Meltdown only affects Intel CPUs, and requires a performance-draining patch (which will also be applied to the non-affected AMD chips, because "glory to Intel!" (no, really!))

You're so very wrong sometimes. All we know is that AMD has said that their chips aren't affected by Meltdown and that the Linux Kernel devs have decided to err on the side of caution and not accept a patch that would disable the Meltdown fix on AMD processors at this time. This isn't about "glory to intel", this is about not trusting a company blindly when they say something.

 We'll see. (I heard that line, not sure if I buy it.  While initially true perhaps, my prediction is Intel will be pushing very hard to leave AMD in the same basket as them until they have refreshed the market with their own CPUs again).

Let me re-read the article and see if I can find the part where AMD not being affected is pointed out (made me pretty sure there's no way around it).

Edit: also, see the Linus comments at the end of this article.

Edit2: alright, in addition to Linus complaints, here's the AMD response in the article.  This should be easy: if we know or can test for the the veracity of the claim by AMD about how their processors work, then AMD systems should not be patched.  If we don't know, err on the side of caution.  The claim:

Quote
However, it may be that the vulnerability in Intel's chips is worse than the above mitigation bypass. In an email to the Linux kernel mailing list over Christmas, AMD said it is not affected. The wording of that message, though, rather gives the game away as to what the underlying cockup is:

Quote
AMD processors are not subject to the types of attacks that the kernel page table isolation feature protects against. The AMD microarchitecture does not allow memory references, including speculative references, that access higher privileged data when running in a lesser privileged mode when that access would result in a page fault.


So what you're saying is, we don't know if that statement is true, and we cant verify its accuracy in any reasonable time to not patch AMD systems as well?  Also, Linus Torvalds is a pretty smart guy.  Why is he upset?  I kind of gathered the patch isn't being made easy to disable on AMD systems or something:

Quote from: Linus Torvalds
From Linus Torvalds <>

Date Wed, 3 Jan 2018 15:51:35 -0800

Subject Re: Avoid speculative indirect calls in kernel

On Wed, Jan 3, 2018 at 3:09 PM, Andi Kleen wrote:

> This is a fix for Variant 2 in https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html

> Any speculative indirect calls in the kernel can be tricked to execute any kernel code, which may allow side channel attacks that can leak arbitrary kernel data.

Why is this all done without any configuration options?

A *competent* CPU engineer would fix this by making sure speculation doesn't happen across protection domains. Maybe even a L1 I$ that is keyed by CPL.

I think somebody inside of Intel needs to really take a long hard look at their CPU's, and actually admit that they have issues instead of writing PR blurbs that say that everything works as designed.

.. and that really means that all these mitigation patches should be written with "not all CPU's are crap" in mind.

Or is Intel basically saying "we are committed to selling you **** forever and ever, and never fixing anything"?

Because if that's the case, maybe we should start looking towards the ARM64 people more.

Please talk to management. Because I really see exactly two possibibilities:

Intel never intends to fix anything

OR

these workarounds should have a way to disable them.

Which of the two is it?

Linus
« Last Edit: January 04, 2018, 10:20:42 am by jr2 »

 
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
jr2 do you even use linux
The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell.

 

Offline jr2

  • The Mail Man
  • 212
  • It's prounounced jayartoo 0x6A7232
    • Steam
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
jr2 do you even use linux

1) Not relevant
2) Yes
3) Google: MOS 6694

Edit: I took that as some form of dismissal and responded to that.  Apologies if I'm mistaken.  My response is a bit twatish if you didn't mean yours as a putdown.
« Last Edit: January 04, 2018, 10:25:01 am by jr2 »

 

Offline Mito [PL]

  • 210
  • Proud Member of Slavicus Mechanicus
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Hardware Unboxed initially benchmarked the fix: https://www.youtube.com/watch?v=_qZksorJAuY.

A tl;dw version: Most user tasks and benchmarks aren't affected beyond the margin of error, including rendering and gaming.
« Last Edit: January 04, 2018, 11:24:19 am by MitoPL »
How do you kill a hydra?

You starve it to death.

 
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
The reason I asked, jr2, is that desktop users mostly use Windows and we have no idea if Windows has applied the slowdown to AMD chips. Also I read somewhere (lwn, probably) that AMD’s pull request has been officially merged into a future kernel.
The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell.

 

Offline jr2

  • The Mail Man
  • 212
  • It's prounounced jayartoo 0x6A7232
    • Steam
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
The reason I asked, jr2, is that desktop users mostly use Windows and we have no idea if Windows has applied the slowdown to AMD chips. Also I read somewhere (lwn, probably) that AMD’s pull request has been officially merged into a future kernel.

So why is Linus getting twitchy (well, he always is, but why is he pointing at Intel)?  And again, my personal use of Linux has no bearing.  Also, FYI, I run Intel (Ryzen wasn't out yet when I built this, my wife's MacBook is of course Intel, and my two PC laptops were i7-2670s - AMD wasn't competing well then either) so I'm affected regardless (although as a personal user it's probably not significant, or so they say... gotta check out those banchmarks Mito posted).

Sorry for the misunderstanding.  I read that as basically "stop talking about things you know nothing about".  I'm not an expert by any means, but I can understand concepts, and I can understand that actual experts are perturbed (Linus Torvalds, kinda hope he knows what he's talking about).

Of course this is all developing and we'll have a clearer picture when the dust settles.

 

Offline jr2

  • The Mail Man
  • 212
  • It's prounounced jayartoo 0x6A7232
    • Steam
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
AMD response to Spectre (they say risk for AMD chips for the two Spectre attacks is "zero, and almost-zero" because of the way their processors are designed - we'll see):

https://www.amd.com/en/corporate/speculative-execution

 

Offline Mito [PL]

  • 210
  • Proud Member of Slavicus Mechanicus
Re: Intel users, prepare for a 5-30 % slowdown (NOT microcode patchable)
Added a tl;dw piece to my previous post. I guess not everyone has the time to watch a 10min long video, so that could be useful.

I am not very surprised by how Linus acts. This is a goddamn mess from his point of view, I guess, and I somehow get the feeling that it isn't the first or second time when his hair is going gray over what Intel is doing with their dominant in the market devices.

BTW, what's about that MOS thing? What did you refer to?
How do you kill a hydra?

You starve it to death.