Quantcast
[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology


View post   

[ Toggle deleted replies ]
File: 64 KB, 291x357, bug inside.gif [View same] [iqdb] [saucenao] [google] [report]
64102534 No.64102534 [Reply] [Original] [archived.moe] [rbt]

Remaking >>64097853 as it flew past the bump limit.

Please keep the shitposting and memery to a minimum.

My questions:
>How much do we know about this potential bug? Does it allow unprivilidged code to escape its VM? Can it access the host's kernel?
>When does this “embargo” on the official details end?
>What part of the CPU is affected? Is this a VT-d bug?
>How does this affect home users? Are they at risk? What non-VM software might be affected?
>Do we all need to use the Linux patch that slashes performance ~30-50%?
>What CPUs do we think are affected? Just the latest ones or further back?

>> No.64102574

first for AMD Threadripper 1950x > *

Intel shills btfo

>> No.64102601
File: 258 KB, 653x533, devape dis.png [View same] [iqdb] [saucenao] [google] [report]
64102601

trying to screenshot the old thread but a ~350 post page is surprisingly hard to actually capture and then upload to 4chin

>> No.64102763
File: 109 KB, 100x5329, intel bug thread tiny.png [View same] [iqdb] [saucenao] [google] [report]
64102763

Holy shit this link is hard to post without it being spam

ibb dot co / hG30Zw
screenshot of old thread

>> No.64102790

>>64102763
Archive.is and warosu aren't considered spam, why do you even use ibb?

>> No.64102815

>even SELinux folks think this is not worth patching
lol just some nobody dev trying to create a problem that doesn't exist so he can get famous

>> No.64102826

>>64102763
>>64102790
Here, because the last portion of your image is missing, here is a different archive
https://warosu.org/g/thread/S64097853

>> No.64102968

I work @ Intel
Can confirm, but performance hit is overblown. Worst case 17% hit, but most only take a hit of 3-5%
Affects CPUs since 2012
Ama

>> No.64102994

>>64102968
Is the vulnerability itself overblow?

>> No.64103012

>>64102968
hey silicon merchant why can't I buy the 8700 anywhere? it never came into stock but 8700K and every other 8th gen did

>> No.64103027

>>64102968
Will this allow government agencies to breach things like tails and qubes?

Would buying a laptop with Intel processor made earlier than 2012 stop this? Lots of people use pre '08 processor's to avoid management engine exploits.

>> No.64103150

>>64102968
Post your kippah or lies.

>> No.64103183

>>64102994
Wouldn't say that, it's still serious
>>64103012
Sorry, don't really know about that. I have a low level job. Heard details about this bug over lunch a month ago
>>64103027
iirc there was a way to breach tails in 2015, but it's fixed now. Unrelated to this
>>64103150
That's funny because I am jewish

>> No.64103196

>>64103027
Yes only post 2011 affected

>> No.64103256

>>64102968
Would disabling hardware virtualisation remove the vuln?
What about questions from OP?
Thank you for your time, merchant.

>> No.64103280

>On Nov. 29, Brian Krzanich, the CEO of chip giant Intel (NASDAQ:INTC), reported several transactions in Intel stock in a Form 4 filing with the SEC.

Most of the transactions involved Krzanich exercising employee stock options (these options allowed Krzanich to purchase Intel shares at prices significantly below where they are currently trading) and then immediately selling those shares that he bought at a discount on the open market.
HAHAHAHA. FUCKING BRIAN. HE KNEW

>> No.64103289
File: 40 KB, 468x585, 21211-7212-12994.jpg [View same] [iqdb] [saucenao] [google] [report]
64103289

>>64103183
OY VEY

>> No.64103333

>>64103280
Nah, this won't tank the stock, investors and 'analysts' will bring out every excuse in the book on why this is just a minor hickup

>> No.64103342

>>64103333
Their morbid failure of a 10nm node will tank the stock.

>> No.64103343

>X86_BUG_CPU_INSECURE

>> No.64103394

>>64103342
Until Intel themself says (fat chance) that 10nm is a failure it won't affect the stock as everything else is SPECULATION AND CONJECTURE, regardless of how shitty and pathethic than 2 core cannonlake is(which they can claim and lie is a specialized product which doesn't reflect 10nm current state)

tl;dr a bunch of smoke and mirrors and everyone will fall for it because they're dumber than shit

>> No.64103445

>>64103394
Intel is practically immune to fuckups, they can get away with murder.

That is until their revenue starts going down, then you'll get a chain reaction landslide

>> No.64103543
File: 302 KB, 450x488, 1385824940405.png [View same] [iqdb] [saucenao] [google] [report]
64103543

>>64103183

>> No.64103644

x86 was a mistake

>> No.64103673
File: 944 KB, 1228x1502, 1495812218308.png [View same] [iqdb] [saucenao] [google] [report]
64103673

>>64103626
>>64103626
>>64103626
COFFEE LAKE X CANCELED, INTELKEKS ON SUICIDE WATCH

>> No.64103683
File: 26 KB, 200x200, Holy+i+didn+t+have+that+one+_27f4ea5616d2dbe9b98e1e04f0cfb675.png [View same] [iqdb] [saucenao] [google] [report]
64103683

>>64102968
Thank you for correcting the record shlomo

>> No.64103767

>>64102815
are you sure you mean SElinux and not grsecurity?

>> No.64103896

>>64103767
>63% performance hit on Skylake i7-6700
HAHAHAHAHAHAHA

>> No.64103944
File: 246 KB, 882x758, 1498229459814.png [View same] [iqdb] [saucenao] [google] [report]
64103944

>>64103896
DELID DIS YOU FUCKING AMDRONE

>> No.64103952

>>64103896
>It's TLB bug but ABSOLUTELY ANALDEVASTATING
HOLY SHIT INTEL

>> No.64103977

>https://grsecurity.net/~spender/exploits/wait_for_kaslr_to_be_effective.c
really makes you think

>> No.64104024

When does the bug embargo end?

>> No.64104063

>>64102534
(((they))) cant keep getting away with this

>> No.64104141

Thread theme
https://hooktube.com/watch?v=oIscL-Bjsq4

>> No.64104221

I really hope this bug plus all the negative press about intel management engine encourages everyone switching to non x86 and independent architectures

>> No.64104243

>>64104221
>switching to non x86 and independent architectures
Hahahaha.

>> No.64104251
File: 24 KB, 500x283, armisgay.png [View same] [iqdb] [saucenao] [google] [report]
64104251

>>64104221
> inb4

>> No.64104259

>>64104251
Didn't there used to be a dozen competing architectures during the 70s and 80s hence why C was created

>> No.64104277

>>64104259
Every single manufacturer did their own architecture. I don't want to go back to those times. My wallet isn't ready for 50 different computers.

>> No.64104284

>>64104277
As long as the compiler works for the architecture any C software will work for it

>> No.64104303

>>64104277
you wont need 50 computers, just software makers will have to compile for 50 different ones

>> No.64104304

>>64104284
I'm worried about games / commercial software / "architecture exclusives" by jews

>> No.64104351

>>64104221
that's what I've been saying for ages. you need diversity in an ecosystem. if everyone goes wintel (or soon gnulintel), that's a recipe for disaster once a single, impossible-to-fix or hard-to-fix bug wipes out the ecosystem. all you'd need is a wormable killer poke.

>> No.64104373

>>64102968
A) Does it affect Sandy-Bridge-E?
B) Do you know any details on how the bug works, or what it involves?

>> No.64104375

Embargo ends 4 jan. It's happening.

>> No.64104414
File: 84 KB, 600x587, e8e.png [View same] [iqdb] [saucenao] [google] [report]
64104414

>>64102968
proof?

>> No.64104447
File: 3.10 MB, 3968x2976, X86_BUG_CPU_INSECURE.jpg [View same] [iqdb] [saucenao] [google] [report]
64104447

>>64104373
Not Intel dude but
>https://lkml.org/lkml/2017/12/27/426
>Linux kernel treats all non-AMD x86 processors as X86_BUG_CPU_INSECURE
>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.

>> No.64104455

>>64104375
Anyone with the knowhow already fucking knows about it, why aren't sites reporting it instead of waiting for Intel damage control preperation?

>> No.64104470
File: 32 KB, 160x150, 1477688826056.jpg [View same] [iqdb] [saucenao] [google] [report]
64104470

>>64104447
that means my VIA CPU will take a performance hit even though it's not affected by this bug?INACCEPTABLE

>> No.64104484

>>64104447
So the Intel Architecture can perform speculative accesses on memory that the CPU shouldn't have access to. Huh. Wish we could all figure this out.

>> No.64104492
File: 4 KB, 452x523, Mysterious Merchant.gif [View same] [iqdb] [saucenao] [google] [report]
64104492

>>64104484
Shhhhhh.

>> No.64104511

>>64104484
the process that the speculative prefetching is pulling data for, not the CPU.

>> No.64104512
File: 124 KB, 382x491, proofs.png [View same] [iqdb] [saucenao] [google] [report]
64104512

>>64102968

>> No.64104514

This is fucking insane, why isn't any site reporting it? I mean the fucking kernel flag is called X86_BUG_CPU_INSECURE, they should at least advance it until the embargo is lifted.

>> No.64104538

>>64104514
Because most sites that do report news usually shill for Intel or are owned by Purch (that shills for Intel).

>> No.64104542

>>64104511
I mean it is part of the CPU.
Sorry for my lack of knowledge. Which makes me wonder, wouldn't this cause a TLB miss if it wasn't in the TLB? And then that causes an exception right?

>> No.64104569

>>64104542
the CPU has and needs access to all the adress space. The problem is that the speculative prefetching is allowing other processes to pull data from adress space allocated to a different process, something it shouldn't have acess to. At least thats how i understand it. How that happens, no idea. Do they get that adress space information from the TLB?

>> No.64104589

>>64104569
Either that or the page table, since all addresses are going to virtual AFAIK(in the program).

>> No.64104613

When does embargo end

>> No.64104681

>>64104447
>/* Assume for now that ALL x86 CPUs are insecure */
>Assume
>for now

>> No.64104701

>>64104681
The linux devcucks are trying really hard to not pin this on one kind of CPU. Unfortunately for them, word is already out that this doesn't affect AMD.

>> No.64104707

>>64104681
other implementations status are unknown so yeah. i bet that it will end up being a rule targeting intel only, instead of all non-ayymd

>> No.64104722

>>64104681
The patch was submitted by a developer with @amd.com address.
Would AMD lie to Linus?

>> No.64104734

>>64104722
They wouldn't, given they spent half a fucking year to make Linus stop ranting about AMDGPU-DRM.

>> No.64104756

>>64104613
IIRC, January 4th.

Until then, I'd hold off on patching in mysterymeat.

>> No.64104829

>>64104756
>patching in mysterymeat
?

>> No.64104836

>>64104829
Codeword for big fat surprise dick

>> No.64104878

>>64104756
Where are you getting this from? Is this when the linux update drops?

>> No.64104995

>mandatory reboots on xen and azure
IT'S HAPPENING.

>> No.64105049

>>64104995
Isn't it the 10th they want to do the reboot

>> No.64105120

>>64104470
You can send in the patch yourself Kevin

>> No.64105232

maybe now i'll be able to make back the money i lost last time i bought amd stock
i doubt it though

>> No.64105233

In a strange twist of fate Bulldozer might end up aging better than Intel's stuff

>> No.64105292

So would AMD CPUs be faster if they did this speculative prediction thingy that Inte does?

>> No.64105317

>>64105292
AMD does speculative execution too, but apparently it manages privileges correctly so it's not affected by this.

>> No.64105394

>>64103673
Holy shit this is a disaster, and 2018 has just begun

>> No.64105408

There is a Xen security advisory embargo'd until Thursday, maybe it's related to this?

https://xenbits.xen.org/xsa

Also apparently all Azure instances and older AWS instances are going to be forcibly rebooted on early January according to the emails people are getting.

>> No.64105413

INTEL IS CRASHING

SELL SELL SELL

>> No.64105423

>>64105317
So AMD makes me check my privilege? Good thing I'm Intel and not a dumb SJW like AMD

>> No.64105428
File: 614 KB, 992x1043, 1506383390040.png [View same] [iqdb] [saucenao] [google] [report]
64105428

>>64103896
THIS CAN'T BE HAPPENING

>> No.64105441

2018 already off to a bang, this will be a very exciting year of happenings

>> No.64105463
File: 127 KB, 600x600, 1502378315806.gif [View same] [iqdb] [saucenao] [google] [report]
64105463

>>64104447
>Linux kernel treats all non-AMD x86 processors as X86_BUG_CPU_INSECURE

>> No.64105480

>>64104447
>>Linux kernel treats all non-AMD x86 processors as X86_BUG_CPU_INSECURE
PLEASE DELET.

>> No.64105493
File: 23 KB, 490x480, x86_btfo.jpg [View same] [iqdb] [saucenao] [google] [report]
64105493

2018 will be the year of the RISC desktop

>> No.64105511

>>64105493
More like the year of AMD unexpectedly painting the datacenter market red.
Even before fucking Rome.
That's quite an unexpected turn of events.

>> No.64105526

>>64105511
But Zen is orange.

>> No.64105533

>>64105526
And AMD is red/white/black.

>> No.64105545

>>64105533
I didn't realize that AMD is an American rock duo formed in 1997 in Detroit, Michigan.

>> No.64105551

>>64105533
>datacenter
And Epyc is blue.

>> No.64105574

Will I have to get new machines after this is all official? I just bought a new used thinkpad that has an intel CPU a week ago.

>> No.64105589

>>64105574
>intel CPU
RIP.

>> No.64105623

I have i7-2600K, am I affected by this?

>> No.64105634

>>64105623
Probably.

>> No.64105660

>>64104829

Incomplete docs + bug embargo == mysterymeat.

>> No.64105661

>>64102534
>tfw haswell
fucking intel

class action when?

>> No.64105678

>>64105463

You can always patch it out and live dangerously.

>> No.64105711

>>64105678
>server side javascript pwning your entire machine

It will be glorious

>> No.64105715

wait. so this will affect everyone? even gaymen kiddos running windows 10? wtf thank god i saw this, ryzen here i come

>> No.64105729

>>64105574
>TPM
You fucked up buddo.

>> No.64105742

>>64105715
Windows 10 uses virtualization in Device Guard and Credential Guard.

>> No.64105748

>>64105715
Every Intel CPU since 2011

>> No.64105773

>>64105748
My Q6600 is still faster than any AMD Poocessor made so i don't see a problem.

>> No.64105804

>>64105773
Unironically this, 2017 AMD has same IPC as 2008 Intel

>> No.64105807

>>64105715
The Linux patch applies to ALL Intel platforms. We still don't know if old Intel CPUs are safe, but assume the worst.

Likewise, people are assuming the bug affects virtualization because that could be the reason why they're being so secretive and rushing so much (to avoid planetary-scale meltdown if cloud providers like Amazon, Google and Microsoft are attacked via privilege escalation from customer VMs), but it could also affect regular users, which I think it's likely given the magnitude the whole thing seems to have. Imagine the vulnerability somehow allows for unprivileged code to read kernel memory, just browsing a site and executing JavaScript in the browser would make you vulnerable as sites could read all sensitive data that is currently on memory (including passwords, data from other websites, private documents and photos...). In the worst case the bug might allow code execution with kernel privileges (full system compromise).

>> No.64105817
File: 55 KB, 640x640, 1480250776729 (2).jpg [View same] [iqdb] [saucenao] [google] [report]
64105817

>>64102534
heh, my ivy bridge based servers are fine :^)
Also my laptop and desktop. All IB powered.

>> No.64105824

>>64105804
Yes.... Really sad....

>> No.64105843

>>64105715

> Windows 10

They already received a patch a few weeks ago.

>> No.64105848

>2015: HTTPS is literally useless
>2016: Monitors allow remote code execution on phones even when the phones have all network services disabled
>2017: WPA2 is one hundred percent compromised, all wifi networks are basically public and nearly unsecurable
>2018: All intel processors allow undefined access to kernelspace memory and potentially Ring-1 code execution even from web browsers

What's next, are we going to suddenly learn that USB ports come alive at night and slaughter people? Why was this the decade that all technology suddenly became completely insecure?

>> No.64105854

>>64105773
>>64105804
>the current city of incels

>> No.64105858

>>64105715
i was planning to buy a used i7 2600k system, does this mean i should buy ryzen instead before prices rise? also, wtf? how did they let this happen i mean this will affect the whole world once the public finds out.

so how will you guys prepare for the widespread panic?

>> No.64105862

>>64105807
It certainly affects regular users - by affecting everything it makes virtualization insecure (you can read other VM's memory or possibly even escape), which is why they're so panicked. I can't tell you how I know that though :^)

>> No.64105866
File: 66 KB, 1250x533, ayy.png [View same] [iqdb] [saucenao] [google] [report]
64105866

>>64105773
>>64105804
not even that retard userbench agree with you

>> No.64105868

>>64105661
>class action when?
Never, even suicidal Atoms were swept under the rug.

>> No.64105877

>>64105862
>tfw jailbreaking vm software becomes a reality

>> No.64105882

>>64105848
Most of those bugs are actually very old - WPA2 was broken by design from day one.
Technology didn't become insecure, it has been from the beginning (and that's how we've been fucked by NSA and the like). We're just looking more closely now.

>> No.64105883
File: 123 KB, 638x599, 1497689967244.jpg [View same] [iqdb] [saucenao] [google] [report]
64105883

>>64105807
>full system compromise

>> No.64105891

>>64105848
>What's next, are we going to suddenly learn that USB ports come alive at night and slaughter people?
Yes, and hopefully they are voiced by Ellen McLain.

>> No.64105922

>>64105858
Yes by ryzen before price falls. Also drop $11 into some AMD stock

>> No.64105923

>>64105868
Suicidal Atoms fucked consumers, this fucks with datacenters owned by big corps, intel will get fucked one way or another.

>> No.64105936

>>64105923
>Suicidal Atoms fucked consumers
Suicidal Avoton/Rangeley Atoms (C2000) fucked up explicitly datacenter folks. Fuckoff expensive switched say hello!

>> No.64105959
File: 698 KB, 600x337, 1495560538692.gif [View same] [iqdb] [saucenao] [google] [report]
64105959

Are we finally going to see the long overdue end of x86?

>> No.64105969

>>64105959
No, AMD is still around, violently killing any even remote attempt for ARM to get into the datacenter.

>> No.64106036

>>64105969
AMD already makes ARM chips, they can kill x86 if Intel loses the server market

>> No.64106047

>>64106036
>AMD already makes ARM chips
Their only ARM core is eternally shelved.
>they can kill x86 if Intel loses the server market
Why the fuck would they do that? They will enjoy the de-facto monopoly much like Intel did.

>> No.64106065 [DELETED] 
File: 114 KB, 680x358, 845526B0-84F2-4376-BDDA-D52CA8BA956B.jpg [View same] [iqdb] [saucenao] [google] [report]
64106065

>yfw based Apple uses this opportunity to make an ARM Mac that dominates the competition with its powerful A11 processor

BASED APPLE
A
S
E
D

A
P
P
L
E

>> No.64106067

>>64106047
They have arm servers out, but it's the a57 ones. K12 is shelved for now. Fact remains they have shown they can do it

>> No.64106077

>>64106067
>They have arm servers out, but it's the a57 ones. K12 is shelved for now.
A1100 was a morbid failure that was murdered by Avoton looooooooong before the launch.
>Fact remains they have shown they can do it
*Why* would they do that?

>> No.64106082

>>64106065
> using rigged benchmark to show his point

>> No.64106088

>>64106065
They won't do this. Switching from PowerPC to Intel was such a fiasco the only reason they salvaged it is because of how many people switched to macs after the release of the iPhone and had no use for older software.

>> No.64106091
File: 1.87 MB, 331x197, 1.gif [View same] [iqdb] [saucenao] [google] [report]
64106091

>>64106065
you're fucking retarded if you compare scores between architectures. This is not how it works, dumbfuck.

>> No.64106115

>almost fell for the quad-core i3 meme
Thank god this news came out. This is worse than 3.5 nvidia-tier disaster. I would've hanged myself on the spot for being a gullible goy

>> No.64106139 [DELETED] 

>>64106091
>>64106088
>>64106082

Intel shills BTFO. The A11 is as powerful as an i5 while fitting inside a phone. Once intel gets its shit throttled 30% Apple will dominate.

APPLE A BASED

>> No.64106153

>>64106139
Wilco, fuck off to your shithole already.

>> No.64106160

LWN has made their subscription-only article about Linux kernel PTI available to everyone: https://lwn.net/SubscriberLink/741878/1a52f79ffe567125/

(If you like, consider subscribing since it's often original, in-depth content)

>> No.64106167 [DELETED] 

>>64106153
Who is Wilco?

>> No.64106168
File: 9 KB, 300x300, 8x8.jpg [View same] [iqdb] [saucenao] [google] [report]
64106168

>>64103196
>tfw still using a first gen i7 without VT-d

>> No.64106184

>>64106077
>why would they do that?
Maybe because x86 is a power hungry zombie kept alive by Intel's extensions? x86 can't even compete with mediatek in the mobile market.

>> No.64106185

>>64106167
The ARM shill that was spamming the fucking RWT forums until Linus entered the chad mode and told him to fuck off.

>> No.64106202

>>64106184
>Maybe because x86 is a power hungry zombie kept alive by Intel's extensions?
By "power hungry" you mean "it can't scale into the same low power envelopes as ARM trash"?

>> No.64106204

>>64106139
>yes goyim this <1w chip in this tiny phone is more powerful than a 90w desktop chip
>w-why don't you believe me

>> No.64106210 [DELETED] 
File: 112 KB, 587x801, 97A54EB8-E5AE-459A-971D-1B2E680BB2E4.jpg [View same] [iqdb] [saucenao] [google] [report]
64106210

BASED APPLE
A
S
E
D

A
P
P
L
E

>> No.64106232

>>64106210
Do I have to repeat myself?

>> No.64106241

4770k here, maybe ryzen 2 can save me

>> No.64106253

German article that summarizes everything known so far (Google Translate makes it readable): https://translate.google.com/translate?sl=de&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=https%3A%2F%2Fwww.computerbase.de%2F2018-01%2Fintel-cpu-pti-sicherheitsluecke%2F&edit-text=&act=url

>> No.64106288

>>64105773
>Not Q9550
Top Pleb

>> No.64106306

>>64104447
Time to buy a G34 dual socket board

>> No.64106474

>>64102534
is there a list of processors being affected? i'm a layman

>> No.64106481

>>64106474
Looks like everything Westmere and onwards.

>> No.64106502

>>64106481
holy shit, so basically everything i3/5/7/9???

>> No.64106507
File: 253 KB, 1000x1000, 1503848584264.jpg [View same] [iqdb] [saucenao] [google] [report]
64106507

>>64106481

>> No.64106520

>>64106502
YES!

>> No.64106597

>>64106520
then fuck me sideways i'm buying b350 board

>> No.64106613

>>64106597
https://www.asrock.com/mb/AMD/AB350M%20Pro4/index.asp

>> No.64106634

>>64106597
I'd suggest waiting a little for Ryzen Refresh.
New CPUs, new boards.
Better!

>> No.64106655

>>64106634
wouldn't they be AM4 compatible anyways?

>> No.64106664

>>64106655
Yes, but new boards may bring new goodies.

>> No.64106666

>>64106634
I highly doubt it would be worth it,Ryzen refresh will not have many new features, just general improvements, and new motherboard features would be for high-end anyways. Though waiting for Ryzen refresh itself is definitely worth it.

>> No.64106679

>>64106666
The board vendors are definetly more interested in anything Zen-based than they were a year ago.
So, why wouldn't one wait?

>> No.64106694

>>64106664
i may wait for the new boards to come out, so b350 price can tank even more

>> No.64106700

>>64106679
Go right ahead.

>> No.64106712

>>64106700
Imagine the other vendors (and not just ANUS) putting some actual effort into low-end AM4 boards?
Sounds nice.

>> No.64106745

>>64105862

This is 4chan. Embargoes just mean you have to write it a bit better.

>> No.64106754
File: 1.38 MB, 4032x3024, 1513739420961.jpg [View same] [iqdb] [saucenao] [google] [report]
64106754

>>64102534

How does it feel being fucked by Intel anal beads?

>> No.64106772
File: 369 KB, 1280x853, 5522123133_e9d3d3c221_o.jpg [View same] [iqdb] [saucenao] [google] [report]
64106772

welp guess it's time to shelve my intel boxes

>> No.64106776

>>64106754
>How does it feel being fucked by Intel anal beads?
I use Ryzen and want a 7401P box, so I don't feel anything.

>> No.64106778

>>64106065
>>64106139
>>64106210
Please fuck off, we're having a nice thread and you already attempted to derail the previous one with the exact same shit. Just go and make your own thread if you want replies.

>> No.64106782

im a brainlet that just plays games all day and has intel, should i worry?

>> No.64106785

>>64106754
Wait, THIS is 4chan??

>> No.64106792
File: 185 KB, 1762x858, ibm.png [View same] [iqdb] [saucenao] [google] [report]
64106792

>>64104995
>mandatory reboots on xen and azure

General emergency security update planned on ibm cloud already... for general reboot of their hosts

They are running xen here is the maintenance note :

>> No.64106801

>>64106792
OHHHHHHHHHHHHHHHHHHHHHHHHBOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOYYYYYYYYYYYYYY
But scholomo said Xen wasn't affected.

>> No.64106807

>>64104470
>VIA CPU
is this 2003?

>> No.64106822
File: 246 KB, 800x612, PeUcX0U.png [View same] [iqdb] [saucenao] [google] [report]
64106822

>>64105463
Intel fuck boi's on suicide watch

>> No.64106828

>>64106807
>he doesn't have at least one via epia board still running
newfags out

>> No.64106843

>>64106782
Yes.

>> No.64106854

>>64106772

I bet you have some awesome velour seats.

>> No.64106871

>>64106807

The chinks still make the. The newest architecture from Via came out last year iirc.

>> No.64106877

>>64106782
Yes, for 2 reasons:
- It's very likely the Windows fix will be even slower than the Linux fix
- DirectX has lots of context switching, which is what becomes slower with PTI

>> No.64106899

>>64105858
>does this mean i should buy ryzen instead before prices rise?
You should have always bought Ryzen even without this bug, Ryzen 7 1700 is king for performance vs buck and only 3% slower

>> No.64106913
File: 27 KB, 331x473, 1494162984990.jpg [View same] [iqdb] [saucenao] [google] [report]
64106913

>>64106899
>tfw, just upgraded to a Ryzen 7 1700

>> No.64106914

>>64106877
So what you're saying is is that once Microsoft rams the patch "fixing" this down everyone's throat, everyone running an Intel CPU will notice a drop in overall performance, right?

>> No.64106919

>>64106914
Yes.

>> No.64106933

>>64106877

>- DirectX has lots of context switching, which is what becomes slower with PTI

Even a 10% uniform performance hit would drastically change the face of vidya performance for many, many recent big titles.

>> No.64106935

So, if I'm not doing any virtualization, I shouldn't update. Alright, I'm used to that.

>> No.64106965

>>64106935
Completely false.

>> No.64106977

>>64106935
>So, if I'm not doing any virtualization, I shouldn't update. Alright, I'm used to that.
A lot of software uses processor virtualization extensions, so everyone will be affected, not just people running VM's

>> No.64106986

>>64106965
I don't give a shit about """security""", why would I willingly destroy my machine's performance.

>> No.64106996
File: 1.70 MB, 2592x1936, IMG_0790.jpg [View same] [iqdb] [saucenao] [google] [report]
64106996

/non-x86 general/?

>> No.64106999

>>64106935
Virtualization is simply assumed to be affected because it's consistent with what little technical details are known and the embargo/rush would make sense as it's critical to big corps.

However, there's no way to know for sure until the embargo is lifted, and it might as well affect regular-day computing (for example by allowing system compromise via mere JavaScript).

>> No.64107004

>>64106996
More like /buy $AMD general/.

>> No.64107019

>>64106996
>no pizzabox sparcstation
fucking out

>> No.64107024

>>64106986
so that cheng can't steal the CS:GO skins you consider as your most valued possession after you accidentally opened a page with javascript while searching for new trap porn

>> No.64107036

>>64107024
I pirate everything and don't play online.

>> No.64107106
File: 495 KB, 1070x601, 1495358050273.png [View same] [iqdb] [saucenao] [google] [report]
64107106

>>64106822

>> No.64107113

>>64106996
AMD x86 doesn't have this problem, only in intel

>> No.64107114

>>64107019
Not OC btw, but sold my last one(a SS20) for like ten bucks to another collector a few months ago as I don't have the space. My sparc daily driver is a Ultra 10 nowadays.

>> No.64107147

>>64105848
>>2015: HTTPS is literally useless
Wait, when did this happen? I don't remember that

>> No.64107169

>>64105533
These are also the colors of the Nazis.
Are AMD /ourguys/ trying to stop jewtel?

>> No.64107204
File: 140 KB, 548x424, 1508752957941.png [View same] [iqdb] [saucenao] [google] [report]
64107204

>>64107169
yes

>> No.64107236
File: 162 KB, 633x900, 1494356435910.png [View same] [iqdb] [saucenao] [google] [report]
64107236

>>64107106

>> No.64107309

>>64106772
i have dat fuggin shelf to
ikea man

>> No.64107356

>>64107169
Black, white and red are the three colors that people tend to see and describe before any other.

https://www.youtube.com/watch?v=gMqZR3pqMjg

>> No.64107405

>>64107204

Seems Hans has picked up on this.

https://www.computerbase.de/2018-01/intel-cpu-pti-sicherheitsluecke/

>> No.64107534

>>64107405
> - /* Assume for now that ALL x86 CPUs are insecure */
> - setup_force_cpu_bug(X86_BUG_CPU_INSECURE);
> + if (c->x86_vendor != X86_VENDOR_AMD)
> + setup_force_cpu_bug(X86_BUG_CPU_INSECURE);
>
> fpu__init_system(c);


HAHAHAHAHAHAHA

>> No.64107584
File: 318 KB, 882x758, 1490730560541.jpg [View same] [iqdb] [saucenao] [google] [report]
64107584

>>64107534
DELID DIS

>> No.64107701

So... When will the Epyc processor scarse will begin?

>> No.64107705

>>64107701
now

>> No.64107729
File: 136 KB, 1473x707, firefox_2018-01-02_13-46-41.png [View same] [iqdb] [saucenao] [google] [report]
64107729

>>64107701
Never, they are in good enough supply now that they are listed on newegg in boxed editions

>> No.64107744

>>64107729
Well, I guess will begin to save.

>> No.64107756

>>64107729
>no 7401P
What a bunch of niggers.

>> No.64107776

>>64107756
they just got listed recently be patient they will be there

>> No.64107848

The Register article: https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

>> No.64107862

>>64107036
>I pirate everything
In that case you should DEFINITELY update, because exploits for this hole are going to be the new fizzbuzz.

>> No.64107865
File: 234 KB, 882x758, 1490730560540.png [View same] [iqdb] [saucenao] [google] [report]
64107865

>>64107848
I FUCKING HATE AMD SO MUCH

>> No.64107927
File: 21 KB, 650x650, 1499791161072.png [View same] [iqdb] [saucenao] [google] [report]
64107927

So does this affect my 4670k on windows 7?

>> No.64107952

>>64107927
Read >>64107848
It affects at least post-2011 Intel CPUs, and possibly even older ones. This is basically an Intel holocaust.

>> No.64107967

>>64107848
>https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
holy fuck, this might be the worst bug in Intels history

>> No.64107977

>>64107967
Nothing not-recall-worthy beats FDIV bug.

>> No.64108017

>>64102534
AAHHHAHAHAHHAHAHA
This is the dumbest thing that's ever happened in IT, hands on.
Fucking Intel, they screwed up hard this time.
2018 man, current year is already looking awesome

>> No.64108018

>>64107977
30% performance hit for the big names below. If you think they will sit still without huge compensation you are mistaken

"The bug will impact big-name cloud computing environments including Amazon EC2, Microsoft Azure, and Google Compute Engine, said a software developer blogging as Python Sweetness in this heavily shared and tweeted article on Monday:"

>> No.64108027

>At one point, Forcefully Unmap Complete Kernel With Interrupt Trampolines, aka FUCKWIT, was mulled by the Linux kernel team, giving you an idea of how annoying this has been for the developers.
Lost.

>> No.64108033

>>64106933
Good. Fuck 'em for not even trying to make their games runnable on a laptop.
This is just the chemo the industry needs.

>> No.64108040

>>64108017
This scene was pretty boring, but now it actually got some interesting action.

>> No.64108075

>>64108018
>30% performance hit for the big names below
It's probably a ~30% hit for some edge cases.
But yeah, EPYC sales are going to skyrocket.

>> No.64108095
File: 70 KB, 627x627, AbAwkYZ.jpg [View same] [iqdb] [saucenao] [google] [report]
64108095

>shitwrecker knew this was coming and designed around

how does he do it?

>> No.64108115

>>64108095
>how does he do it?
I dunno, probs by being a man that co-designed MOTHERFUCKING amd64.
Probably.
Maybe.

>> No.64108124
File: 50 KB, 1280x720, it's the future zura.jpg [View same] [iqdb] [saucenao] [google] [report]
64108124

>>64108095
THANK YOU KELLER
BASED SHITWRECKER

>> No.64108140

>>64108095
Post more Nazi husbandos.

>> No.64108171
File: 116 KB, 548x424, 1508746876535.png [View same] [iqdb] [saucenao] [google] [report]
64108171

>>64108140

>> No.64108174

>>64107701

Best case scenario for Intel is maybe 5-15% performance hit but let's go middle road and say 20%. Cloud providers, perhaps anyone who has Intel servers for any reason will need to suffer a performance hit or buy more servers to make it up (and the scaling would be shit). Do you buy new Intel servers at a high cost, after you just got fucked by them? Or do you take a chance with AMD to find out now that they have guaranteed better performance and cost, with more features and so far rock solid security features? "No body was fired for buying Intel" won't hold true for long when they see the clusterfuck this will cause in their depts.

>> No.64108187
File: 240 KB, 320x320, 047.gif [View same] [iqdb] [saucenao] [google] [report]
64108187

I have an i5 currently running on windows 10. What precautions should I take against Intel's stupidity here? And no, something other than selling the card and getting something else. AMD is being full retard by refusing compatibility with old games, NVIDIA is still garbage, and I'm on a budget right now so I can't get anything more top of the line.

>> No.64108215

>>64108187
Just take the performance loss then.

>> No.64108217

>>64108187
Nvidia and AMD graphics cards have literally nothing to do with this situation, only Intel CPUs are affected.
That being said, your ONLY options are to either accept a 20% reduction in performance or buy an AMD processor.

>> No.64108234

>>64108187
>AMD is being full retard by refusing compatibility with old games
thats gpus and the only games that don't work are C&C 3 and 4, ive played both and belive me you aren't missing out

>> No.64108237

Is the 2500k affected?

>> No.64108240

>>64108174
People aren't even fucking account for the OTHER side of the coin: The niggas that run the apps on the cloud to begin with. It's not just providers that are gonna suffer. Major companies like Adobe, Oracle, etc who run applications on top of google cloud and amazon web services are gonna see their CPU usage spike. You fucking KNOW those cucklords wrote their software as unoptimally as possible (which means they're already using a ton of CPU). Imagine what a 5-35% CPU increase means to them. It means they're going to be forced to spend more jewgold on servers.

>> No.64108251

>>64106168
My core2duo system already had vt-d

>> No.64108256

>>64108187
>I have an i5 currently running on windows 10. What precautions should I take against Intel's stupidity here? And no, something other than selling the card and getting something else. AMD is being full retard by refusing compatibility with old games, NVIDIA is still garbage, and I'm on a budget right now so I can't get anything more top of the line.

You either take the 20 to 38% performance hit giving you below fx8150 performance or you buy a Ryzen.

AMD will fix directx 9 issues see https://twitter.com/CatalystMaker/status/948227463188017152

>> No.64108257

>>64108237
Everything past Westmere seems affected.

>> No.64108269

>>64108217
>>64108215
So I assume they'll be rolling out a patch for this in the next few days? I have windows update disabled, so I guess I'll have to update them manually? That sucks.

>performance loss

Just curious, how will this performance loss affect games I currently play? Like Team Fortress 2 or Counterstrike? Will I get less FPS?

>> No.64108271

>>64108237
Yes. According to El Reg, all Intel processors from this decade are affected. According to the Linux patches, possibly even older ones might also be affected.

>> No.64108272

>>64108237
Yes. All Intel Core i processors are affected.

>> No.64108277

>>64102534
>>64106935
>>64106977
Virtualization is only brought up because big virtual server providers are rushing in a fundamental, performance impacting change on how Linux manages virtual memory. The patch was even pulled in outside the kernel merge window.

Virtual memory is a fundamental feature of modern systems and was introduced to x86 architecture back in the 386 days. It's about making memory management easier and isolating user processes from messing with memory that doesn't belong to them.

Since memory addresses are now virtual there needs to be a lookup table for translating virtual addresses into physical memory addresses. This data structure is in the kernel memory and isn't readable by user processes.

Linux (I guess Windows too) traditionally kept this memory page table in a single data structure with both kernel and user stuff mixed together. Users can't read it so no problem, right? Well, now they are moving the kernel memory page table into it's own data structure even though it causes a performance hit.

There seems to be an exploit which allows user process to get information about the memory page table contents on Intel CPUs. Additionally there seems to be a way to exploit this information.

>> No.64108283

>>64108237
Yes according to the table it has a 35,45% performance hit

>> No.64108317

>>64102601
fuck ure retarded kys fgt

>> No.64108319

>>64108240
You mean switch to Epyc

>> No.64108322

Time to disable updates :^)

>> No.64108335

>>64108322
Yeah I too love getting pwned by pajeetscript.

>> No.64108339

How would this get solved? Do I need to buy a new Intel CPU?

>> No.64108353
File: 94 KB, 500x275, 1466800367189.gif [View same] [iqdb] [saucenao] [google] [report]
64108353

I play videogames on an i7 6700k. Will this affect my dwarf fortress, my We Are Billions or my Eve Online?

>> No.64108358

>>64108319
Or that.
That works.
>>64108339
>Do I need to buy a new Intel CPU?
Your only choice is to buy AMD CPU.
>>64108353
YES!

>> No.64108364

>>64108269
You will notice overall slower performance on your whole system. On 3D games, it'll be more noticeable (because DirectX has many context switches), and CPU-intensive 3D games will be the most affected.

>> No.64108370

>>64108339
That won't help. You need to buy an (any) AMD cpu. Newer Intel CPUs are just as affected by this as older models.

>> No.64108380

>>64102534
This is why I have a Ryzen CPU.

>> No.64108386

>>64108364
What do you mean "slower performance"? Can I bypass it by simply lowering the graphics settings?

>> No.64108406

>>64107952
If it's related to speculative execution, does that mean everything after pentium II when "dynamic execution" was implemented is affected?

>> No.64108410

2020 will the the year of ARM.
Screencap this.

>> No.64108413

>>64108358
>Your only choice is to buy AMD CPU.
2018 is starting of great kek

>> No.64108415

>>64108406
Probably.
Though it's probably Nehalem-onwards.

>> No.64108417

>>64108406
We still don't know, the bug details are under embargo. Even the Linux patch code comments are redacted.

>> No.64108425
File: 280 KB, 752x548, 1412582918709.png [View same] [iqdb] [saucenao] [google] [report]
64108425

we might see a shortage of FX chips

>> No.64108432

>>64108319
Yup. Couldn't come at a better time for AMD. This is like that time a computer security firm demo'd their network intrusion system for the US Govt. During that demo they discovered the breaching and theft of millions of security clearance holder records from OPM.

You literally can't generate more confidence in your product in any other way. For AMD, this is a major opportunity. If they don't capitalize on this, they deserve to burn.

>> No.64108434

>>64108370
New CPUs should have this fixed, not CPUs being currently sold since they have already been made but if it's indeed a hardware bug it's something they should be able to correct at the fabs.

>> No.64108437

>>64102534
Intel themesong: https://www.youtube.com/watch?v=0c7UFLD7Ka8

>> No.64108439

>>64108386
Technically yes, but a 35% drop in performance means your CPU will be so slow that a settings drop may not even be enough.

>> No.64108448
File: 19 KB, 720x720, 1507551007298.jpg [View same] [iqdb] [saucenao] [google] [report]
64108448

does this affect K-series cpus too?

>> No.64108450

>>64108187
don't let windows update so you don't get the 30% less performance hit
>Windows 10
oh wait, better accept the loss in performance, at least you'll be better protected against stuff :^)

>> No.64108457

>>64108432
Oh believe they'll capitalize on this. Epyc just had its initial launch, when they ship out the lower priced models you bet your ass they'll be plugging those benchmarks like no tomorrow.

>> No.64108463

>>64102763
>screenshot of old thread
>posts huge unreadable picture
why are you like this

>> No.64108471

>>64108434
>New CPUs should have this fixed
this can't be fixed with microcode, this will require major redesigns, expect at least 1.5 to 2 years for an intel chip without this flaw

>> No.64108475

>>64108448
It affects anything Intel, Westmere-onwards.

>> No.64108481
File: 1.25 MB, 1002x1480, 1492962036185.png [View same] [iqdb] [saucenao] [google] [report]
64108481

I want Intel to crash and burn so fucking hard.

>> No.64108488

>>64108448

>>64106481

>> No.64108505
File: 13 KB, 226x223, 1512690593823.jpg [View same] [iqdb] [saucenao] [google] [report]
64108505

>>64108488
holy shit this is a complete and udder holocaust

>> No.64108507

>>64108450
As long as I have javascript disabled on most sites, would it be worth avoiding the update?

>> No.64108528

>>64108507
No, this has nothing to do with Javascript.

>> No.64108537

>>64108507
No. JS is just a popular vector for such attacks.

>> No.64108541

>>64108471
>this can't be fixed with microcode, this will require major redesigns
I said nothing about microcode, I meant fixing it for new silicon (perhaps a new stepping) and if this "bug" requires major redesigns it's not a hardware bug but a outright design flaw. Calling it a bug just downplays it.

>> No.64108542

Reminder that Brian Krzanich sold more than half his stake in Intel leaving him with 250,000 shares...which is the contractually obligated minimum # of shares he has to hold in order to remain CEO. Selling that many shares is abnormal. Shares were sold mid december 2017.

Totally not fucking suspicious.

Means the scope of this problem is fucking huge enough to make him shit.

>> No.64108544
File: 90 KB, 400x400, 80149182.jpg [View same] [iqdb] [saucenao] [google] [report]
64108544

>>64108507
>As long as I have javascript disabled on most sites, would it be worth avoiding the update?

You can't be serious..........

>> No.64108567

>>64106204
Yeah it's like saying that diesel motor are better than electric because it makes more noise, uses a ton of power just to keep itself running, requires huge radiators to dissipate all the energy it wasted as a porta-heater, and is not usable without an auxiliary transmission system bigger than itself.

>> No.64108569

>>64108542
Dude Intel started FUDing on EPYC way-way BEFORE it hit the market.
They are panicking.
They are fucked beyond repair, and their 10nm node is a fucking mess.

>> No.64108575

>>64108542
>Means the scope of this problem is fucking huge enough to make him shit.
could be as seriously seen as insider trading

>> No.64108586
File: 3.16 MB, 1172x1416, 1501032598329.png [View same] [iqdb] [saucenao] [google] [report]
64108586

>>64108542
Resignation soon?

>> No.64108589

Intel validation team is fucked. HAHAHAHAHA

>> No.64108596

>>64108507
why the fuck would you avoid the update. the vulnerability doesn't require pajeetscript. you got 2 options, buy amd or let papa krzanich make you his bitch

>> No.64108604

>>64108541
>Calling it a bug just downplays it.
No shit.
Panic would only make it worse.
>>64108589
>dead Atoms
>now something that is worse than the TLB bug
The fuck.

>> No.64108626

>>64108283
Link to the table, kind anon?

>> No.64108633

>>64108604
>https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
SELL SELL SELL

>> No.64108639

:√)

>> No.64108641
File: 257 KB, 562x567, 1501450636151.png [View same] [iqdb] [saucenao] [google] [report]
64108641

>tfw my Core Duo and Core2 Quad systems are unaffected

>> No.64108645

>>64108633
You can't sell what you don't own.

>> No.64108648

>>64108569
Intel did that shit PRIOR to EPYC's launch because its just good business. Then they realized that it fucking nuked their offerings (once it actually came out). This is when they shit themselves.

If they weren't panicking before, they're sure are now.

>>64108575
Possibly. But look at equifax. Nobody cared. Trump's SEC ain't gonna fine anyone for this shit.

>>64108586
If the 10nm process is as fucked as people claim it is, expect his resignation this year. Sooner rather than later. Intel losing datacenter marketshare to AMD (when they're holding 99%) means they fucked up big.

>> No.64108649

>>64108641
>related to speculative execution
Even fucking PII is affected.

>> No.64108663
File: 256 KB, 500x500, u3o1_500.png [View same] [iqdb] [saucenao] [google] [report]
64108663

>>64108649
>tfw my Pentium 1 is unaffected

>> No.64108681

>>64108663
don't be too sure, this functionality was already in 486

>> No.64108685

>>64108663
anime

>> No.64108687

>>64105817
>Ivy Bridge
So glad I didn't get a newer Shintel CPeePoo.

>> No.64108693

holy fuck.
does this affect celeron too?

>> No.64108694
File: 365 KB, 2000x1153, 1497658837285.gif [View same] [iqdb] [saucenao] [google] [report]
64108694

I own 10k intc.

>> No.64108695

it's a conspiracy to raise prices of used c2q/c2d

>> No.64108699

>>64108648
>If the 10nm process is as fucked as people claim it is
DUDE THE ONLY PRQ'D SKU IS 2+0 2.2GHz TURBO LMAO
It's as dead as it can get, Charlie was fucking right.

>> No.64108700
File: 2.27 MB, 1920x1080, [HorribleSubs] Animegataris - 11 [1080p].mkv_snapshot_14.30_[2017.12.19_03.19.40].png [View same] [iqdb] [saucenao] [google] [report]
64108700

>>64108685

>> No.64108703
File: 258 KB, 640x360, tsuchinoko3.png [View same] [iqdb] [saucenao] [google] [report]
64108703

>>64108681
>tfw my 386 is unaffected

>> No.64108709

> 30% hit
> can heavily slow down typical workloads

This fucking hurt, lads.
What do we do? How do we fix this?
I really fucking hope we will figure out which update/patch is it and we can remove it/block it.

I will return to Windows 8.1 just to get back the lost performance for fucks sake.

>> No.64108716

>>64108699

>Charlie was fucking right.

To be fair, that would be a first.

>> No.64108726
File: 325 KB, 1296x968, photo-4.jpg [View same] [iqdb] [saucenao] [google] [report]
64108726

>>64108681
>tfw my 386 is unaffected

>> No.64108728

>>64108700
WEEB POWER!

>> No.64108734

>>64108277
>Virtual memory is a fundamental feature of modern systems and was introduced to x86 architecture back in the 386 days. It's about making memory management easier and isolating user processes from messing with memory that doesn't belong to them.
surely this does not mean that all Intel CPU's since the 386 are affected...?

>> No.64108738

Just disable Intel VT if you don't need it, problem solved.

>> No.64108753
File: 341 KB, 1920x1080, debiru.jpg [View same] [iqdb] [saucenao] [google] [report]
64108753

>>64108726
So, how does it feel to be 386masterrace?

>> No.64108755

What does this mean for the average rural and suburban retard?

Does it affect my 4790k?

>> No.64108758

>>64108734
>tfw my 286 is unaffected
Looks like it's time for me to upgrade to the secure windows 3.1

>> No.64108767

>>64108681
Source?
Does the DX stand for Dynamic eXecution?

>> No.64108769

>yfw eBay will be flooded with haswell and broadwell cheap xeons'
>they are now worthless

>> No.64108770

>>64108755
It appears to affect every Intel CPU within the last decade and possibly beyond but no one knows for sure yet because its embargoed.

>> No.64108772

>>64108755
If you have an Intel CPU from this decade, your computer will now be slower.

>> No.64108773

>>64108481
well, it's your lucky day.
problem is, all providers will also suffer. Amazon, DigitalOcean, Vultr, Microsoft. literally every fucking one. shit's not even funny, it's ridiculously bad.

> inb4 intel owner
I do own one, but it's not about me, or just a silly AMD vs. intel purchase. Datacenters will slow down a lot because of this stupid bullshit.

Worst part? 99.99999999% of people would not even be affected by such an exploit/breach. So why punish everyone?

And yes, I am going to buy a Ryzen.

>> No.64108774

>>64108755
>What does this mean for the average rural and suburban retard?
Either shitty performance or shitty security

>Does it affect my 4790k?
Yes

>> No.64108775
File: 23 KB, 662x967, xuo3mdl2gdsz.png [View same] [iqdb] [saucenao] [google] [report]
64108775

Times are getting hard boys
Times don't get no better boys
Gonna leave this place
Say goodbye to everyone
Goodbye to everyone

>> No.64108776

>>64108755
>I'm retarded pls tell me
Yes you are affected, everyone not using AMD or an ancient system is

>> No.64108778

>>64108755
>What does this mean for the average rural and suburban retard?
Buy AMD OR lose performance OR get pwned by scriptkiddies.
>Does it affect my 4790k?
Read the thread; Westmere - onwards.

>> No.64108781

>>64108758
>tfw my Amiga is unaffected
Those crazy furfags knew all along.

>> No.64108795
File: 109 KB, 1920x1080, 1493961226443.jpg [View same] [iqdb] [saucenao] [google] [report]
64108795

>>64108681
Oh. I thought it was a fundamental flaw in Out Of Order Execution.

>> No.64108796

>>64108772
>installing the slowdown update

>> No.64108802

>>64108322
It says the NT kernel has been getting patched since November or so. No idea what that means, I do hope researchers will test around.

Also I am not sure you can disable such an update. Microsoft can just silent patch. On Linux, you can just use an older kernel - for a while I guess.

>>64108335
that's bullshit though. regular folks would never get infected by this. datacenters, vps providers and such? hell yes, it's a very important fix.

>> No.64108807

>>64108386
graphics settings don't affect GPU performance

>> No.64108813

>>64108796
What if my CPUs are already horribly underutilized? Would installing the update make any tangible difference?

>> No.64108819

>>64108772
the linux patch actually makes all x86 cpu's slower except AMD, so even a c2d or p4 will be slower because of the patch, even though it might not be affected

>> No.64108825

>>64108802
>It says the NT kernel has been getting patched since November or so. No idea what that means, I do hope researchers will test around.
It's not active right now, though Fast Ring builds do have it enabled.

>> No.64108836
File: 172 KB, 260x325, 1471144297233.png [View same] [iqdb] [saucenao] [google] [report]
64108836

Can Windows force this update, or is it a motherboard driver thing?

Will
apt-get update
apt-get upgrade

cause my Ubuntu server to be affected, or do I need to manually install this "update" somehow?

>> No.64108842

So I get that it's mainly related to enterprise virtualization but will it actually fuck Intel in everything or is this fix only going to affect specific use cases?

>> No.64108844

>>64108693
Yep.

>> No.64108846

>>64108802
>that's bullshit though. regular folks would never get infected by this.
ITS RELATED
TO
VIRTUAL FUCKING
MEMORY
YOUR PEESEE USES IT QUITE EXTENSIVELY IN DAILY FUCKING TASKS

>> No.64108849

>>64108017
well it's the biggest blunder of the past 20 years IMO. there were a few failed and badly designed products but this fucking tops them all.

>>64108825
So, we have the patch but it's waiting for activation?
Also it said "NT Kernels". Honestly if someone finds the fucking update and we can disable it, I am on my way to 8.1. My CPU is already slow as it is (4700HQ).

Did look at the new Ryzen ROG, but it's hot and loud as shit.

>> No.64108854

>>64107848
Damn. Unprivileged code that can snoop on protected kernel memory. Sounds like a blast.

>> No.64108855

>>64108836
It's a kernel-level update, so updating Windows or Ubuntu after patches have rolled out to production will fix the bug.

>> No.64108861

>>64108709
Then expect your OS to get nuked by script kiddies running wild with this flaw.

>> No.64108863

You think Intel will be forced to offer a partial refund on all of the affected chips? I have to imagine they'll be forced to for the last 2-3 years worth of chips at least.

>> No.64108865

>>64108836
>apt-get
>not using the superior apt
git gud

>> No.64108874

>>64108846
nigguh I did not mean the performance hit but the security hole / impact. A regular guy will get infected by "hot ladies in your area", no one needs such a sophisticated hack for them.

>>64108836
You can just use an older kernel. On Windows, it's much harder. We have to wait and see which updates are they, and if we can ignore/skip them.

>>64108861
It's just an exploit and there is a dozen out there. There is a reason why we run antivirus products.

>> No.64108878

>>64108836
The fixes are OS updates, so updating Windows and updating your Linux kernel will be enough.

>> No.64108882
File: 913 KB, 580x799, 1461197692692.png [View same] [iqdb] [saucenao] [google] [report]
64108882

>>64108795
this

>> No.64108883

Holy mother of TLB bugs, this is hilariously bad.

>> No.64108884

>>64108836
You can block all kernel updates, but that won't save you in the long run, or you change the code like AMD dev did and compile it yourself

>> No.64108886

>>64108807
>>64108386
CPU limit i meant

>> No.64108888

>>64108802

So being able to just read protected kernel memory from userspace is something that regular folks would never get infected by?

You sure famalam?

>> No.64108901
File: 914 KB, 500x350, aaa.gif [View same] [iqdb] [saucenao] [google] [report]
64108901

>tfw my RM7000 is unaffected

>> No.64108907

>>64108874
>There is a reason why we run antivirus products.
HOW THE FUCK WOULD THAT STOP SOMETHING FROM "ACCIDENTALY" READING/WRITING INTO YOUR PROTECTED FUCKING KERNEL MEMORY YOU FUCKING TODDLER?

>> No.64108912

>all thinkpads are now unusable

>> No.64108913

>>64108874
>We have to wait and see which updates are they, and if we can ignore/skip them.
Windows 7/8 and earlier version can selectively refuse updates and even disable them altogether, though I don't recommend doing so for this bug unless you know what you're putting yourself at risk of.
>There is a reason why we run antivirus products.
Useless in the wake of the bug. It's a kernel-level vulnerability that no software can mitigate.

>> No.64108917
File: 241 KB, 1920x1080, prad3-11.jpg [View same] [iqdb] [saucenao] [google] [report]
64108917

>>64108781
>tfw Amiga didn't even have protected memory
She lets you touch whatever you want in her address space

>> No.64108925

>>64108912
A275/A475 will be refreshed with Ryzen Mobile (and probably presented at CES), so no worries.

>> No.64108930

>>64108863
> entire last decade
> partial refund
Keep dreaming

>>64108888
Nice quads, Satan.
No I am saying regular folks can get infected much easier than this.
>>64108907
Okay, you have a malicious file. It's blocked by AV.
You have a malicious website. Ok, blocked by Chrome, Firefox.

So you have... a malicious javascript. Can't just Chrome sandbox that? How would it even escape?

>> No.64108933

>>64108874
being this retarded, you deserve what is coming to you

>> No.64108936

>>64108912
Just install Linux sweetie.

.

Linux doesn't have virus :^)

>> No.64108942

>>64108925
>A series

>> No.64108944
File: 40 KB, 651x496, INTC.png [View same] [iqdb] [saucenao] [google] [report]
64108944

>>64108694
Well this certainly doesn't seem like the time to hodl

>> No.64108955

>>64108942
Which is X/T with AMD APUs.
A475 is T470 with an APU.

>> No.64108960

AMD's H1B visa ''engineers'' and an obese welfare case working on grsecurity are not reliable sources. This is an obvious attempt by AMD to discredit Intel because of bad CPU sales pretty much across the board.

>> No.64108965

>>64108738
The design flaw has nothing to do with virtualization.

>> No.64108967

>>64108913
>can selectively refuse updates
yeah but Microsoft pushed down the Windows 10 update forcefully as well.
disabling it all might help, but then you have a bunch other problems to worry about.

>>64108933
Well, for one, I sure as hell won't update my server host, as I only use it to virtualize and I am the sole user.
The guests will get patched but that's all.
Desktop wise... I don't know. It's insecure maybe, but fuck me, my PC is already goddamn slow.

>> No.64108968

>>64108930
Agreed, it's just another pile of shit added on top.

But after this without PTI exploits are going to be much easier to write, simpler to use and less prone to crashing the target system.

>> No.64108974
File: 554 KB, 637x1430, 1514816714531.png [View same] [iqdb] [saucenao] [google] [report]
64108974

>>64108917
Memory protection is strange. My Win3.1 install never crashed. Win95 crashed at every other mouse movement or keystroke.

>> No.64108978

>>64108960
>H1B visa ''engineers''
That's Intel's entire validation team.

>> No.64108984

>all these claims of 30% worse performance

As far as I can tell that's how much slower a syscall will be. Code that isn't constantly switching contexts will only see a small performance impact.
For example it might have a noticeable impact on games but it won't make file compression slower.

Correct me if I'm wrong.

>> No.64108987
File: 234 KB, 800x612, 1493932977399.png [View same] [iqdb] [saucenao] [google] [report]
64108987

>>64108960

>> No.64108996
File: 232 KB, 800x449, intel ceo yells at clouds.png [View same] [iqdb] [saucenao] [google] [report]
64108996

>>64108960

>> No.64108997

>>64108967
>Microsoft pushed down the Windows 10 update forcefully
Only if you were a retard that blindly installed every update or even worse, had Windows Update set to automatic.

>> No.64109001

>>64108984
Wait two days and you'll know for sure.

>> No.64109021

>>64108984
Yeah, the average performance impact should be less, somewhere in 5-15% range, which is still fucking noticeable.

>> No.64109033

>>64108984
>won't make file compression slower
Reading and writing files incurs massive amounts of context switches. Compression (where data is read from disk to be processed by userspace code) will be severely affected.

>> No.64109035

>>64108813
Not really.

>> No.64109043

>>64108968
I mean, for one thing, theoretically - and with our KNAWLEDGE - we know that a payload can be prepared using such an exploit.
On the other fucking hand. About 50% of my country uses NO (or cracked) antivirus, and has updates disabled, and use 7 or XP still.
There is still a 20% that also has no updates but have cracked OR Defender AV.

And they get NO serious infection besides these search site replacement crap - which is just adware basically.
So you get the idea. I always believe (honestly) that these security issues are kinda overblown. Especially in this case.
IMO it should be seeded to Amazon, Microsoft and such companies ASAP. But regular customers? I don't know.

(I kinda feel sorry for companies. Imagine you paid X$ to plan out things, you have a system load, etc. And now you are going to lose up to 30% load capacity. I mean HOW FUCKING BAD IS THAT?)

>> No.64109045
File: 164 KB, 577x354, amada tiles.png [View same] [iqdb] [saucenao] [google] [report]
64109045

>>64108984
>>64109021
>>64109033
Amada wins again.

>> No.64109049
File: 89 KB, 1280x720, maxresdefault[1].jpg [View same] [iqdb] [saucenao] [google] [report]
64109049

>TFW i5 2500k
>Was about to get intel but ordered Ryzen just yesterday

>> No.64109053

>>64109035
Well I'll see what happens next Tuesday.

>> No.64109060

>>64108960
Yeah, I bet Linus Torvalds is disregarding the entire dev process he's been so anal about for the past 30 years just to discredit Intel. It's obviously that and not that this is the worst hardware bug of this century.

>> No.64109064

>>64108984
Yes, it depends on the workload.
But the funny thing is how this will affect graphics drivers. Remember how nVidia implements a lot of shit in driver (instead of hardware like AMD)? I guess Intel just fucked the only remaining market they are the leader of.

>> No.64109074

>>64108844
welp time to install gentoo

>> No.64109080

>>64108930
>you have a malicious file. It's blocked by AV
define malicious
>How would it even escape?
it doesn't have to

>> No.64109082

>>64109064
>Jensen's face when
Oh shi~

>> No.64109097
File: 15 KB, 267x297, 7c2.jpg [View same] [iqdb] [saucenao] [google] [report]
64109097

I own only one Intel CPU, it is in a original Xbox p3 733Mhz, is it affected

>> No.64109103

>>64109033
>Reading and writing files incurs massive amounts of context switches
A context switch is several orders of magnitude faster than disk access. It shouldn't make much of a difference.
Most of your time is spent churning the bytes, not reading them.

>> No.64109105

>>64109097
>related to speculative execution
Probably.

>> No.64109106

>>64109103
Fair enough.

>> No.64109113
File: 56 KB, 625x473, yoda.jpg [View same] [iqdb] [saucenao] [google] [report]
64109113

>>64109049

>> No.64109124

>>64109097
no and consoles are shit

>> No.64109127

>>64108716
>first

What is Optane?
What is SKL-SP?
What is Fermi?
What is Kepler?

>> No.64109160

>>64109064
>nVidia implements a lot of shit in driver (instead of hardware like AMD)
So that's why nvidia drivers are more stable

>> No.64109172

>>64109160
No, nVidia simply has more monkeys for QC.

>> No.64109181

>>64109043

The best hacks are the ones you never know are there. Could be part of a widely distributed botnet and you'd never know.

>> No.64109186

>>64108965
But you guys said it's a VT bug

>> No.64109194

>>64109186
Virtual memory, pal.

>> No.64109204

>>64109186
No, it's a virtual memory bug, which just happens to affects virtualisation and Big 7 the most.

>> No.64109215

we need a new thread goy

>> No.64109225

>>64109215
No, we need a sticky.

>> No.64109238

>>64109225
you want my cummies?

>> No.64109243

>>64109215
NEW THREAD >>64108951
NEW THREAD >>64108951
NEW THREAD >>64108951
NEW THREAD >>64108951
NEW THREAD >>64108951
NEW THREAD >>64108951

>> No.64109260

>>64109160
No, that's a different issue.
This is a fundamental difference in design, where AMD implements scheduling (and other things) in hardware and nVidia in software. The hardware approach is more reliable if done right, but increases the complexity of the hardware. The software approach is more flexible and cheaper, but offloads the functions to the driver (which depends on CPU and OS) so that's another variable that affects the final performance.

>> No.64109268

>I have never seen so many important Linux developers pull all-nighters to get such fixes done, and with complete radio silence especially to mid-level developers and enthusiasts (ie. me), it must be super critical.


Absolutely fucking miserable.

>> No.64109270

>>64109243
Nice try, schlomo.

>> No.64109319

>>64109243
>>64109225
>>64109215
NEW BREAD >>64109308
NEW BREAD >>64109308
NEW BREAD >>64109308
NEW BREAD >>64109308
NEW BREAD >>64109308

>> No.64109420

https://lwn.net/SubscriberLink/741878/1a52f79ffe567125/

KPTI, in other words, has all the markings of a security patch being readied under pressure from a deadline. Just in case there are any smug ARM-based readers out there, it's worth noting that there is an equivalent patch set for arm64 in the works.


ARM might also be affected

>> No.64109443

>>64109420
Awww fuck, is AMD the only one doing security seriously?

How will Apple explain two performance degrading 'features' in the span of a month?

>> No.64109452

>>64109420
I think the patch might be to have the same mitigations/feature set on all relevant archs, but this bug seems Intel-specific.

>> No.64109456

>>64109443
JIM FUCKING KELLER

>> No.64109471
File: 116 KB, 800x600, 1505487410115.jpg [View same] [iqdb] [saucenao] [google] [report]
64109471

intel i7 920 nehalem affected?

>> No.64109472

>>64109456
Jim isn't involved in this part, he's purely a high level design architect, not a debug/QC/validation codemonkey that caught this before tapeout

>> No.64109478

>>64109471
YES.

>> No.64109590

>>64109478
but https://github.com/mjg59/mei-amt-check/ said no. Is that unrelated?

>> No.64109615

>>64108930
>So you have... a malicious javascript. Can't just Chrome sandbox that? How would it even escape?
Literally the entire reason Intel is scrambling to silently patch this shit is because it allows code to break out of VMs and sandboxes.

>> No.64109693

>>64109186
Virtual memory is not related to virtualization. Your processor uses virtual memory for the vast majority of applications you run.

>> No.64109727

Just sold my intel rig on facebook earlier, it's getting picked up by some sucker tomorrow.

>> No.64109731

>>64107169
Nazism was a jewish scheme to base judaism+zionism even more.
t.rothschild

>> No.64109745

THIS IS A FUCKING GLOBAL DISASTER HOLY SHIT AHAHAHAAHAHAHH

>> No.64109781

>>64108425
FX-8350 owner here, shedding tears of joy at my toasty warm workstation.

>> No.64109811

>>64109590
AMT isnt' related to this flaw

>> No.64109871

>tfw my pocket calculator is unaffected
Feels good man

>> No.64110028

>>64105959
That would only happen once local computing is completely dead. I don't think you want that.

>> No.64110218

Welp, time to go back to harpertown

>> No.64110233

>>64110218
It's also affected.

>> No.64110260

>>64102968
good thing that i didnt fall for the consumerism meme so im still safe. this thing has no uefi no management engine or other shit that i dont need.

>> No.64110311

>>64110233
It's before westmeme, so it should be fine.

>> No.64110352

>>64110311
No-no, the kernel will flag your CPU as insecure and apply the workaround.
And it's still probably affected, nobody simply tested it with pre-Westmere CPUs.

>> No.64110375

Oh so this is why libreboot breaks vtx

>> No.64110396

>>64110352
On Linux this seems to be configurable on boot by passing pti=off to the kernel.

>> No.64110409

>>64110396
Your CPU may still be affected.
Careful.

>> No.64110416

>>64110375
virtual memory != virtualization

>> No.64110438

>>64110416
Well it came out in 08 so i doubt core 2 duo is effected

>> No.64110450

>>64110352
But Penryn doesn't have vt-d
That's where the issue is, no?
I don't think Nehalem supported vt-d either for that matter. So maybe Gainestown is good too?

>> No.64110451

>>64110438
Virtual memory has been used since 386.

>> No.64110470

>>64110450
ITS UNRELATED TO VIRTUALIZATION

>> No.64110472

>>64110451
Yes but from what we know its 2011-present cpus

>> No.64110483

>>64110438
Virtual memory was implemented in the 80386. It predates the Core 2 Duo by a decade.

>>64110450
No, this flaw has nothing to do with vt-d or virtualization in general.

>> No.64110487

>>64110472
Because no one tested anything older than a fucking Westmere.

>> No.64110497

>>64110451
Time to dust off the ol 286 then

>> No.64110508

>>64110472
No, it's Nehalem-to-present CPUs.
And that's "at least." The Linux patch team has implied that it's possible the flaw goes all the way back to the Pentium II, maybe further, but they don't know because nobody has tested it yet.

>> No.64110510

>>64110487
Ah fuck can it be fixed theroeticly in an open bios

>> No.64110529

>>64110510
It's an actual CPU desing flaw.
The only fix is (FUCKING UNIRONICALLY I SWEAR) AMD.

>> No.64110532

>>64110508
>pentium 2
Well just fucing kill me then

>> No.64110550

>>64110529
So whens windows 7 getting the patch and how much of a hit can i expect

>> No.64110552

>>64110438
>>64110450
>>64110451
Fucking retards the lot of you.
The bug seems to be Intel CPUs speculatively executing kernel-mode code while in user mode which leaks into TLB and at the very least exposes addresses of kernel datastructures.
If your CPU does this sort of reckless speculative execution then it's fucked and needs the page table isolation. I'm very skeptic i386 came anywhere near the sophistication needed for this bug to exist.

>> No.64110614

>>64110552
When will we know the damages

>> No.64110638

>>64110614
bout tree fiddy

>> No.64110706

>>64106977
just disable them in the bios?

>> No.64110741

>>64110706
read the fucking thread dumbass

>> No.64110774

>>64108187
>AMD is being full retard by refusing compatibility with old games
That was just sensationalist noise blowing stuff out of proportion and not waiting for confirmation, AMD already promised fix ASAP.

>> No.64110803

>>64110529
Probably within the next month. According to Linus Torvalds you can expect an average of 5% performance drop for common applications, increasing to double-digit performance drops in syscall-heavy applications.

>> No.64110860

>>64108925
A285/A485 model numbers leaked in some Lenovo PDF I think.

>> No.64110899

>>64110860
Oh, so they are going to actually bother differenciating it.
Both are probably going to be at CES in AMD's booth.

>> No.64111125

>>64109225
those just get filled with spam and the 1000 post cutoff sucks.

>> No.64111237

>>64110774
>MUH C&C 3 AND 4

>>
Name (leave empty)
Comment (leave empty)
Name
E-mail
Subject
Comment
Password [?]Password used for file deletion.
Captcha
Action