[ 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: 65 KB, 726x781, IntelRektByLinus.png [View same] [iqdb] [saucenao] [google] [report]
64133720 No.64133720 [Reply] [Original] [archived.moe] [rbt]

>> No.64133749

Linus is the best,

>> No.64133760


>> No.64133772


>> No.64133776
File: 225 KB, 657x527, 1514979273472.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64133780

intel In meltdown, trying to drag amd to hell with them

>> No.64133808

>Doesn't even suggest AMD as an alternative
>Jumps straight to ARM64 as the alternative
Even Linus knows AYYYMD isn't worth anything.

>> No.64133842

He would be call "Amd shill" if he do that

>> No.64133852

>Acorn anything

>> No.64133886

AMD doesn't need the Meltdown patches, now kys.

>> No.64133906

Yeah it was pretty disgusting of Krzanich to try and implicate AMD by association today. He was a fucking mess today

>> No.64133908
File: 58 KB, 700x700, 1514990703031.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64133923

RISC-V should be the future.

>> No.64133924

Is PPC secure?

>> No.64133929

Printing this out to hang in my cubicle

>> No.64133930


But seriously fucking lel

>> No.64133960

Fuck, my laptop and desktop both use Intel processors. Next build is gonna be AMD only, AMD cpu and AMD gpu. Intel has gone from being the obvious best to being a shitty company that keeps fucking up and refuses to clean up. This plus the IM Engine is enough for me to just switch next time I'm buying a computer.

>> No.64133963


I wish more people in the tech world were like Linus, just fucking wrecking cunts and doesn't afraid of anything

>> No.64133968


>> No.64133981

This is what happens when you hire H1Bs

>> No.64133990
File: 184 KB, 1125x1119, 1482515713843.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64133992

He may be coarse, but he's never wrong!

>> No.64133997

Holy fucking shit I love this man

>> No.64134005


Linus can just not accept the shitty intel patch into mainline, leaving affected people to apply the patch themselves until the code is good enough to make the grade. I don't care I run AMD.

>> No.64134031


>> No.64134064
File: 117 KB, 1920x1080, 1463446606016.jpg [View same] [iqdb] [saucenao] [google] [report]

Based linus. Seriously, intel has been screwing their customers hard for a long time but no one cares until it hits them on "muh performance", the sad part is that most people will still not care until it hits them hard on "muh FPS".

>> No.64134066
File: 32 KB, 410x266, needful.jpg [View same] [iqdb] [saucenao] [google] [report]

>poo on cpu

>> No.64134071

Fuck off.

>> No.64134101

why isn't linus the /g/od of /g/ but that fat fuck is instead?

>> No.64134113

AMD is vulnerable to Spectre attacks you faggots

>> No.64134130

Only to the first one, and it's mitigatable.
Intel has just patched the second one.

>> No.64134131
File: 24 KB, 680x430, l56.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64134133
File: 115 KB, 845x740, smug.jpg [View same] [iqdb] [saucenao] [google] [report]

Someone's got some hurt feelings

>> No.64134142

Richard cares about our f-freedom.

>> No.64134150
File: 54 KB, 679x758, 1489898569240.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64134193
File: 29 KB, 556x263, theo.jpg [View same] [iqdb] [saucenao] [google] [report]

>> No.64134216


>> No.64134228

So capitalism is to blame.

>> No.64134230


>> No.64134254
File: 41 KB, 400x579, 1497629135187.jpg [View same] [iqdb] [saucenao] [google] [report]

You have earned a free helicopter ride!

>> No.64134271

I understand that reference xd

>> No.64134273


What is meant by

>l1 i$ that is keyed by cpl

>> No.64134280

>source: my arse
AMD is virtually safe, retard.

>> No.64134345

>>>64133720 (OP)
>Printing this out to hang in my cubicle
I did that today man, the best idea I never had.

>> No.64134392

A company that cared would be open to full disclosure and protecting their customers. But intel is all about blowing smoke and covering their ass.

>> No.64134408

>A company that cared would be open to full disclosure and protecting their customers.
ARM just did that.

>> No.64134434

>intel commits


>> No.64134457

The iOS/A chip/M processor, Russia's Elbrus CPU, ARM, et cetera are the future. Goodbye Intel, and possibly AMD.

>> No.64134488

>applel chips
>can't even type text without lagging
Yeah no.

Neither Intel nor AMD are going anywhere for decades.

>> No.64134638
File: 109 KB, 460x1061, xoNZbS9.jpg [View same] [iqdb] [saucenao] [google] [report]

Linus is ok now.

>> No.64134663

I actually own 5th generation, 2017 iPads, iPhones, and the latest Macbook (no Intel). No type lagging here, whatsoever.

This incident will further development in other areas. This is just the beginning for the end of Intel. AMD will survive, but they will drastically change.

I hope everyone trashed their Intel CPUs and/or computers.

>> No.64134741

No but it looks like you do you fucking man baby.

>> No.64134743

>Only owns 5th generation iPhone
>Not owning an iPhone X

Poorfag detected.

>> No.64134784

RISC-V is the future. CISC is a fucking meme that just doesn't want to die.

>> No.64134820



>> No.64134835

This. Apple is a fucking joke, and so is ARM.

>> No.64134849
File: 11 KB, 240x240, 1510143073562.jpg [View same] [iqdb] [saucenao] [google] [report]

what did he mean by this

>> No.64134855

literally, the definition of a cuck

>> No.64134859

>le epic hate apple maymay xDDDD

>> No.64134861
File: 212 KB, 793x307, 1515020554678.png [View same] [iqdb] [saucenao] [google] [report]

From the google report they only got it working on some FX bulldozer APU with default settings turned off

>> No.64134875

>>le epic apple worship maymay xDDDD

>> No.64134900

Linux Torvalds doesn't know shit.

>> No.64134961
File: 42 KB, 649x280, 1.jpg [View same] [iqdb] [saucenao] [google] [report]




post yfw

>> No.64134987


>> No.64135054

They merged it?

>> No.64135089
File: 135 KB, 700x934, 1514843019598.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64135099

I don't get why /g/ is so wet for eat-from-his-own-foot instead of Linus.

>> No.64135103

AMD can into ARM, so nonono

>> No.64135106


>> No.64135107
File: 1.62 MB, 229x250, 1355696736911.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.64135114

because he's le epic autist meme man

>> No.64135130

How op proves that this email is legit?

>> No.64135165

How old are you? I've seen it before with CPUID and LOLNGPIPES. Everyone should thank AMD from saving us from late 90s and early 2000s Intel. It seems the cycle is upon us again.

>> No.64135180

What does this mean?

>> No.64135182

Are you a retard?
The kernel mail list is public. You can search it.

>> No.64135208

>Wanted to buy a new laptop
>Already got myself 1500$
>It all happend
>Still don't have normal in-one-piece laptop.

What to do /g/uys?

>> No.64135239
File: 641 KB, 500x364, 1504991659012.gif [View same] [iqdb] [saucenao] [google] [report]

Linus Torvalds just killed Intel with a single git merge.

>> No.64135244
File: 42 KB, 369x468, 1393869469465.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64135250

>ernel mail list
Here -> https://lkml.org/lkml/2018/1/3/797

And stop being a newfag/retard.

>> No.64135255

Go Apple. Mac OS X was patched against this since last December without any reports of performance hits.

>> No.64135257

It means that the performance won't be gimped on loonix if you have AMD.

>> No.64135269


Intel CPUs are affected by an exploit because of a hardware bug. This kernel patch (merged 2 hours ago) activates a software workaround for that hardware bug on all X86 systems except AMD.

I've read that some tests confirm that there is performance degradation when this fix is enabled.

google "x86_bug_cpu_insecure" for more info

>> No.64135283

Lol who the fuck uses liuux.

>> No.64135292

ahhhhhh yes that's the good shit right there mmmmmmmmmmmffff open me THE FUCK up daddy make me free

>> No.64135304


>> No.64135319

Certainly not those running Intel.

>> No.64135365

I am already, and this shit drives me mad. Damaged drywall with this shit.
Good that it is 2012 MacBook Pro and not that modern with Ivy Bridge.

>> No.64135414

I am. But I will not upgrade the kernel, until they fix it normally.

>> No.64135458

>windows experiences 5% degradation
>linux experiences 50% degradation

who really is btfo here?

>> No.64135470

Did Windows even fixed it?

>> No.64135472

>idiots buy bad products
>"we need communism"
yes surely we should make our entire political and economic systems cater to retards

>> No.64135501

Literally a newfag retard, kill yourself

>> No.64135519

!= means no

>> No.64135552
File: 368 KB, 819x380, fuckwit.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64135567


Intel are literally a laughing stock now, the only way I would ever take them seriously again is if they took ownership of their problems and remediation.

>> No.64135595

>2 hours ago



>> No.64135651


I'm be so happy to have an instruction set that I don't have to sign an NDA to code against. Arm is a step in the right direction, but why stop there?

>> No.64135678

A fucking lot of devices and servers you dumb shits

>> No.64135694

> - Exclude AMD from the PTI enforcement. Not necessarily a fix, but if
AMD is so confident that they are not affected, then we should not
burden users with the overhead"

>> No.64135725


Permavirgin, get back to fapping to traps, and stop shilling for money to pay for your wife's son. She still won't fuck you.

>> No.64135756
File: 13 KB, 457x101, IMG_20180104_025011_419.jpg [View same] [iqdb] [saucenao] [google] [report]

>> No.64135765

So that means VIA x86 platforms are also assumed insecure?


>> No.64135782
File: 68 KB, 960x760, 1509058564912.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64135833

Degradation will go way up when we crack open your system and suprise sex you up the pooper. At least Linux is down 50% and then it's done.

>> No.64135899

Yep. AMD is kinda did it bad, because nobody says, that VIA is affected by this crap..

>> No.64136044

Literally everything except "personal" computers.

>> No.64136059


Fuck off /v/.

>> No.64136102

what browser is that?

>> No.64136173

>what browser is that?

the current state of /g/.

>> No.64136211

VIA is confirmed to have this issue or its simply a security measure?

>> No.64136316


>> No.64136351

Everything is assumed rotten unless otherwise noted. AMD noted, VIA hasn't yet.

>> No.64136365


Yes, it's only a security measure. Kernel developers are very conservative when it comes to this. For them, everything is considered insecure until you prove that it's not.
That's why the X86_BUG_CPU_INSECURE flag was initially applied to all CPUs.

Because of the low marketshare of VIA processors, I can imagine that they haven't been tested yet.

>> No.64136967

>reddit spacing

>> No.64137026


>> No.64137035
File: 3.51 MB, 298x224, 1465517403569.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.64137085
File: 81 KB, 500x482, 1455067410601.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64137099
File: 49 KB, 600x906, lul.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64137120
File: 124 KB, 5000x2571, 1441508177732.png [View same] [iqdb] [saucenao] [google] [report]

Where were you the day Linus stood up to Intel's bullshit and absolutely destroyed their market edge with a four line git commit?

>> No.64137128


>> No.64137253

Feels good sitting on delay slots that haven't been relevant since the 80s.
Register windows were an interesting experiment, which we can all agree turned out a negative result.
>Power Arch
Nice speshul-purpose registers for speshul looping instructions. One could almost be fooled into thinking it's an especially ugly CISC architecture.
>No mention of Alpha
Otherwise known as the only good proprietary RISC.

>> No.64137266

reported for cyberbulling

>> No.64137270

Man I really wish I jumped to the AMD ship instead of spending $500 to upgrade my Intel CPU and motherboard a few months ago.

>> No.64137297

>CISC is a fucking meme that just doesn't want to die.
If you just disregard the legacy stuff in it, x86 has actually been a fairly regular and nice architecture for many years now. The only really bad thing about it is the instruction encoding, but not so much because it's variable-length, but because of prefixes, variable-length arguments and just cruft. s390 also does variable-length encoding, but does it good.

>> No.64137325

i hope your mom dies tonight in her sleep

>> No.64137342


>Cache Userspace Crosstalk Kernel

>> No.64137343
File: 226 KB, 631x362, rms-gigolo.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64137345

capitalism is built around exploiting retards instead of educating those retards and fixing that problem.
it's more profitable if population is full of uneducated idiots who buy anything that you advertise
it's not profitable if the population is being picky

>> No.64137365

so do i

>> No.64137372

>capitalism is built around exploiting retards instead of educating those retards and fixing that problem.
All the best universities in the world are in capitalist countries. Communist China sends hundreds of thousands of students a year to attend them.

>> No.64137397

sent from my iPhone

>> No.64137400

The best malware installs antivir. Keeps the competition out.

>> No.64137403

Well, the bright side he know how to use the "green text"

>> No.64137445


they send them over to steal industry secrets to grow faster themself. China has the cleanest capitalism. They just don't buy into your western ideologies, they do whatever benefits them most while using your own laws against you. I.E. fuck your IP laws we take what we want but still use your patent system for their stuff inside your country.

>> No.64137454

So....for real though. Is that Firefox?

>> No.64137484

>they do whatever benefits them most while using your own laws against you. I.E. fuck your IP laws we take what we want but still use your patent system for their stuff inside your country.
The US Trade Representative office is investigating them for that now. I predict large tariffs in the future if they don't cut that shit out.

>> No.64137555
File: 1021 KB, 500x281, 1448063336779.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.64137607

Unless the cost is higher than the benefits it's just taxation. Not to mention they produce almost all our electronic shit wich gives them power.

>> No.64137616

Underrated post

>> No.64137618

We're talking tariffs high enough to make electronics manufacturing in the US a viable alternative again. China should be very concerned.

>> No.64137634

You are not welcome here, get out living cancer

>> No.64137646

That's a good shop holy shit. If it weren't for the thumbnail I wouldn't have noticed.

>> No.64137660

This is never gonna happen. China holds too much US debt and there's too much trade happening. Any sanctions will fuck US companies just as hard, which means that the most that anybody in the government will do is wag a finger at them very sternly and then point that finger at another crisis elsewhere so the media stop asking him about China for a while.

>> No.64137669

It's literally already happening. You don't understand how Trump works.

>> No.64137681

>tfw intel bounced back up and amd is currently bouncing back down
Capitalism was a mistake

>> No.64137699

No, brainlet, you don't understand how Trump works. Lots of hot air and not much done, where is my promised industrial revival? Nowhere, instead Bannon just got rekt off of him in a press release or something, and we all know Bannon was the only one there actually driving through the point that China is a threat.

>> No.64137716

>where is my promised industrial revival? Nowhere
It's happening right now. Over a million jobs created, half a trillion in trade deals brought back from overseas, tax reform injecting a ton of money into the economy and repatriating trillions of dollars held overseas, unemployment falling through the floor. Stop following kike media that screams about TWO SCOOPS and look at the actual numbers sometime.

>> No.64137723

oh I forgot. HP WHERE ARE YOU

>> No.64137736

The whole infrastructure is missing and wages would be much higher in US requiring a singificant higher price. Who is going to buy that? Not to mention that Apple&co have to compete internationally.

You have already chinese companies producing phones at 0 profit just to get more market share. They tell it as it is even themself. They don't care about profit. They first go for market domination.

>> No.64137739

They're in the process of actively killing Itanic.

>> No.64137748



>> No.64137755

>The whole infrastructure is missing
Oh good, more construction jobs.
> wages would be much higher in US requiring a singificant higher price.
Tariffs will be raising prices that high anyways.

>> No.64137760

Where can I get these non-jewish numbers that you speak of, because none of the ones I've seen align with this vision

>> No.64137785

/pol/, pls

>> No.64137814

>employment numbers

>half a trillion trade deals
Trump has been over this several times.
>$400B arms deal with Saudi Arabia
>$253B in trade deals from China alone on the Asia trip

>tax reform
Look at all the companies that have announced moves made as a direct result of the bill.

That's leftypo

>> No.64137826

The point is if you raise the price you can't compete internationally. Just winning on your local market isn't enough. When they dominate the world market they will still win.

Secondary longterm all are fucked because of modern pure robot factories. If you produce locally they become even more desirable since the profit replacing high wage workers will be higher.

>> No.64137833

TFW Linus is finally actually concerned about security issues. I wonder how hard Theo De Raadt rolled his eyes after reading that.

>> No.64137835

China is not communist and they come in droves to steal your tech. They did the same to Soviet Union
USSR had one of the best if not the best STEM university system in the world, their matematicians, physicists and programmers are top tier even today
After that crash of the union that system crashed, thanks capitalism
I don't use that cancer

>> No.64137853

>muh robot workers
Robot factories still produce an absolute fuckton of jobs, just not on the factory floor. Ever tried automating a plumber or an HVAC tech?

>muh international competition
Losing access to the US market is a death blow to any export focused company. That's why we're able to extract good terms in trade deals.

>> No.64137880

>China is not communist

>> No.64137903

If you believe automation creates not significantly less job you're beyond saving. Especially as we can build them somewhere else cheaper unlike your local workers.

>> No.64137904

Anon why. Please, why.

>> No.64137907

Learn the difference between communism and state capitalism, the chinese state actually wants their private sector to grow and they're using every protectionist trick in the book to make it so. In comparison in communism there is no private sector.

>> No.64137908

>Losing access to the US market is a death blow to any export focused company.
Today. You are losing that influence as Chinese market becomes more and more profitable, and Indian is going to develop too. Both of those countries dwarf your market. You are going to have a tough wake up call once you realize that more and more companies start to ignore western markets in favor of Asian ones.
And once that happens they will no longer have any use for USD as a reserve currency. So, good luck.

>> No.64137924

Means of production are owned by workers in China? News to me.

>> No.64137930

So why isn't AMD gloating about this more when they aren't affected by Meltdown?

>> No.64137932

>Robot factories still produce an absolute fuckton of jobs
Yea for college educated people mostly. Compare number of those vs number of plebs anywhere in the world, including US. Making plebs unemployed is literally asking for a jamal led uprising.

>> No.64137949

no, it is also affected by spectre
> Additional exploits for other architectures are also known to exist. These include IBM System Z, POWER8 (Big Endian and Little Endian), and POWER9 (Little Endian).

>> No.64137960


>> No.64137983

>Learn the difference between communism and state capitalism
Sure, also North Korea is a democracy, because they say, so.

>> No.64138032

ye shall know them by their fruits
stop paying attention to how governments arbitrarily name themselves

>> No.64138038

I thought AMD is also affected by this, but less of their CPUs have the problem? How deep does this problem go? Is it a flaw in how x86 works or did Intel just fuck up and its a bug?

>> No.64138044
File: 495 KB, 1070x601, 1514989572614.png [View same] [iqdb] [saucenao] [google] [report]

AMD! The only one x86 to use to be secure, right now in the Linux source code.

> - /* 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);

>> No.64138062

So x86-64 is fine?

Non-issue then

>> No.64138083

>stop paying attention to how governments arbitrarily name themselves
No shit wise anon.

>> No.64138088

That's Intel PR doing their job.

From the whitepaper:
>Meltdown exploits a privilege escalation vulnerability specific to Intel processors, due to which speculatively executed instructions can bypass memory protection.


Meltdown is why you need the performance killing KPTI patch.

>> No.64138101


Intel fucked up.

>> No.64138161

>That's Intel PR doing their job.
I read that there were 3 different vulnerabilities identified, and that amd wasn't affected by all 3 at the same time (and neither was intel cpus if that matters). Supposedly they are (intel, and amd) are working together for whatever reason.
All I remember was yesterday anons were spamming how only intel was affected, and amd was completely immune though.

>> No.64138207
File: 240 KB, 465x435, 1502145015315.png [View same] [iqdb] [saucenao] [google] [report]

RIP Intel.

>> No.64138260

No who told you that?

>> No.64138287



First 64-bit best 64-bit

>> No.64138329



>> No.64138450

There's two spectre bugs and one meltdown bug.
Google said AMD is vulnerable but they tested using a FX and a FX Pro CPU. Ryzen doesn't have all the bugs that the FX CPUs have so I'm not sure if Ryzen has none of the bugs and FX has one of them or if Ryzen has one of the spectre bugs and FX has both bugs.

>> No.64138549


AMD64 is not vulnerable to any of the bugs certainly not able to read protected pages like Intel, which is why the Linux patch only applies to non AMD64 CPUs like intel.

>> No.64138696

>Intel has gone to a shitty company that keeps fucking
>ftfy, also
how gnu r u

>> No.64138746

Intel is a platinum Linux Foundation member at $500,000.00 per year. AMD is silver at $20,000 or less.

>> No.64138800


>> No.64138829

Haha, you really showed him. XD

>> No.64138869

What sort of grown man uses the word shit on the internet while discussing a tech problem?

>> No.64138914
File: 62 KB, 827x511, 1477948178800.png [View same] [iqdb] [saucenao] [google] [report]

its obvious, isnt it?

>> No.64138937

It affects x86-64 architecture as a whole jumping to AMD will do you no good.

>> No.64138960
File: 201 KB, 1345x960, gjPnHPXOLZEN8Hrm[1].jpg [View same] [iqdb] [saucenao] [google] [report]

>jumping to AMD will do you no good.
>v1 confirmed on shitty old AMD processors that nobody uses
>v2 might affect AMD, nobody's proven it yet
>v3 doesn't even work on AMD

>> No.64138977


>incuck lies

>> No.64138985
File: 81 KB, 985x813, INTEL IS FINISHED.png [View same] [iqdb] [saucenao] [google] [report]

>still in denial

>> No.64139007
File: 141 KB, 439x290, 1455687281173.png [View same] [iqdb] [saucenao] [google] [report]

>massive bug confirmed to affect Intel (Meltdown)
>AMD explicitly not affected
>/g/ doubles down o nthe retardation and denies 1) that the bug is serious and 2) that only Intel is affected

If there was any doubt this place is infested by shills Meltdown should remove it.

>> No.64139012
File: 36 KB, 604x453, 1422077649658.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64139023


>> No.64139065

Capitalism, a system that's only purpose is free trade of goods and services.

You're like one of those creationists who argue about evolution/natural selection and come up with all this nonsense about it.

>> No.64139073

Have you never had a job? Christ, I'd be fucking worried if my colleagues didn't swear. Unless you're sucking up hard to someone who has a stick up her ass or a religious yank twat, swearing is fine. Linus doesn't need to suck up to anyone on the face of the planet.

>> No.64139080

One year in and you're expecting an entire turn-around? What spoiled, unrealistic fantasy land do you live in?

>> No.64139143

The kind that has been writing and maintaining his own kernel for over two decades and still retains control of the code. Why would he mince words?

>> No.64139185


>It is possible to have a capitalist economy without complete free enterprise, and possible to have a free market without capitalism. Any economy is capitalist as long as the factors of production are controlled by private individuals.

>> No.64139206


So only religious or strange people don't curse?

How sad.

>> No.64139224

I'm confused as to why ARM64 is being recommended as the alternative. It's affected by the same attack vectors as Intel.


>> No.64139244

Intel CPUs are also AMD64

>> No.64139306
File: 71 KB, 1696x610, AMD_confident.png [View same] [iqdb] [saucenao] [google] [report]

We did it guys. AMD is saved.

>> No.64139325

But they are handling properly, unlike Intel who keeps on denying their responsability.

>> No.64139365

Any loss of x86 market share to arm64 is much worse for intel than losing market share to amd.

Linus is just being savage

>> No.64139387
File: 8 KB, 241x228, 1468443368760.jpg [View same] [iqdb] [saucenao] [google] [report]

>tfw I have a FX

>> No.64139400

Don't worry, it's nowhere near as bad as Intel's fuckup.

>> No.64139427

This is true, but his argument leads me to believe he was under the impression only intel was affected.

The crux of the flaws relies on speculation and branch prediction, these are not features limited intel,

There are many other CPU architecture options out there. Google, Amazon and Apple are all easily capable of producing custom CPUs for their own purposes.

Google already has it's own asic for AI, and Apple builds it's own arm processors.

His entire argument is based the idea that no other attack vectors will come about that could make use of these methods.

>> No.64139451

AMD has PR people?

>> No.64139459

Because it's a PR debacle if it turns out AMD is vulnerable too. It's bedt to just shut the fuck up.

>> No.64139466

Intel uses their own 64 bit implementation called intel64.

>> No.64139474

>fix it normally

In any case you can disable it by doing this https://askubuntu.com/questions/991874/how-to-disable-page-table-isolation-to-regain-performance-lost-due-to-intel-cpu

>> No.64139478


>> No.64139496
File: 68 KB, 1154x971, rip.png [View same] [iqdb] [saucenao] [google] [report]

The Spectre White Paper (https://spectreattack.com/spectre.pdf) contains an PoC at the end.

Here's the code, in non-pdf format: https://pastebin.com/Txidik7V

Screen shot of it working on Ryzen (in a VM)

>> No.64139512

>current state of /g/

>> No.64139524

Greatest day of my life, so far. Next greatest day would be Intel imploding.

>> No.64139561

You'd have be a diehard cuckbrain to continue to use Intel after this. Along with the passwordless ME access and remote JTAG attack on the same, who knows what other sloppy obscure fuckup lurks under the hood.

>> No.64139563

Intel "mostly" follows the AMD64 specifications, but there have been (and still are) some various discrepancies between the two companies implementations.

One example: https://media.blackhat.com/bh-us-12/Briefings/Wojtczuk/BH_US_12_Wojtczuk_A_Stitch_In_Time_Slides.pdf

Intel's sysret instruction can cause an exception in ring0 where AMDs cannot. This is still the case IIRC.


> Because Intel’s implementation of SYSRET matches their published spec, they consider their processors to be behaving correctly. However, the SYSRET instruction was defined by AMD as part of the x86-64 architecture, and Intel’s version is obviously intended to be compatible with AMD’s version. If the majority of operating systems (acting independently) managed to “not properly handle uncanonical return addresses on Intel EM64T CPUs”, it’s hard not to conclude that Intel made a mistake when designing their specification.

>> No.64139568

I know you're too buttmad to admit it but he DID get you pretty well there

>> No.64139577

>posts image for attention
>cries when somebody gives him attention

the absolute state of /g/

>> No.64139587


Not quite the same attack, as this only shows access to the process' own pages unlike on intel where you can access kernel and other process pages.

>> No.64139639

This is just "Spectre". All it shows is that it's possible to use timing attacks to read memory that's elsewhere in the system.

That's all "Spectre" is - it's a class of bugs, not a bug in of itself.

So there's really nothing to do about Spectre unless we have an option to disable speculative execution, which would result in fairly substantial performance penalties everywhere.

This PoC is done in a users process, but there's nothing stopping you from performing it across kernel / userspace boundaries as far as I can tell.

>> No.64139645


And pay more for it too xD.

>> No.64139651

>I know you're too buttmad to admit it but he DID get you pretty well there
I'm not him..

>> No.64139661

Yes they do although they are nearly identical to AMD64 there are some differences but yes Intel has their own implementation of 64 bit called intel64
You dumbfucks.

>> No.64139675


Yes there is, on AMD you cannot view any pages other than the process' own, unlike on Intel where you can look at whatever you like.

>> No.64139686

Clearly you've never worked with computer hardware at a low level. The closer to the silicon you get the more awful bugs like this you find and the more time you spend swearing at the designers.

>> No.64139700
File: 65 KB, 679x711, 15150528681880.jpg [View same] [iqdb] [saucenao] [google] [report]

Best timeline ever.

>> No.64139708

does this mean we (intel) will also have performance drop once distros roll it out?

wasn't it patched?

>> No.64139711

Correct. However, what I'm saying is that you can perform the attack in the KERNEL from USERSPACE, and leak the data back out into USERSPACE which you can then use to craft a real exploit.

The trick is finding a place in the kernel where you can reliably get the state into something you can work with. I'm not saying the user process can read kernel mode memory or other process memory directly. But you could use this attack to leak kernel memory into userspace.

>> No.64139713

Intel gets performance drops. AMD does not.

>> No.64139720

> Linus Torvalds

>> No.64139722

So I was planning on upgrading my PC this year... am I fucked? Are Ryzen prices going to shoot through the roof over this.

>> No.64139725
File: 67 KB, 895x770, pajeet.png [View same] [iqdb] [saucenao] [google] [report]

>This is what happens when you hire H1Bs
More H1Bs than you would imagine...

>> No.64139728

The kind of man who built a kernel dubbed after his own name and who's a fervent chauvinist. Torvalds has never cared about professionalism.

>> No.64139729

that guy that has the youtube channel. hes one of the bigger tech channels but a pretty clear shill, doesnt really matter dont see how he even fucking matters here, unless hes lying for intel again

>> No.64139735

RISC will change everything.

>> No.64139743

Order early.

>> No.64139744

Ryzen prices aren't going to rise. The profit margins are already comfortable, it takes them what, only 40 dollars for a single CPU.
It's the EBIN chips that are going to be short on stock high demand because they're meant for the big boys.

>> No.64139748


No, there has been no proof that works on AMD.

>> No.64139757

What's with his childish dialect?
Is this really a professional programmer?
He comes off incredibly salty.

>> No.64139760
File: 61 KB, 1000x800, 1477584575140.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64139763

24 million richer though.

>> No.64139766


It was applied by a intel engineer even after a amd one said it doesn't affect amd.

It was not being conservative you stupid shill

>> No.64139775
File: 31 KB, 351x395, 1514626217693.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64139779 [DELETED] 

I know. This is not what I was asking.

>> No.64139791
File: 375 KB, 572x464, 65465.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64139794
File: 1.99 MB, 320x240, 1512276149710.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.64139800

>if vendor not amd.

>> No.64139815

I know. This is not what I was asking.

>> No.64139833

I think you're too retarded to be involved in this conversation. You clearly lack the cognitive capacity to think about the larger picture and how this could be leveraged on other parts of the system.

Spectre itself is not an exploit. It is a technique that can be used to determine if data was pulled from the cache or from RAM. You can time these accesses to determine with a strong degree of certainty where the byte came from.

Now. Meltdown is an EXPLOIT which leverages ideas from Spectre - which is just a technique for determining whether a byte was read from cache or RAM. Spectre should not be thought of as an attack by itself. When leveraged properly it will be usable on AMD as well as Intel systems (and anything else that uses memory caching and branch prediction).

Now. Overwriting the Branch table is a different story, and may not work on newer AMD chips - I don't know. However, Spectre is in fact usable to an extent on modern AMD platforms.

>> No.64139835

and no proof that AMD is exempt. Researches say that the reason for the failed attack on AMD could be
and this include a possible shortcoming in their own PoC.

>> No.64139851
File: 56 KB, 150x150, 15147376001.gif [View same] [iqdb] [saucenao] [google] [report]

The amazing thing is the speculative side channel exploit has been known and has been used since at least July 2017.

>> No.64139860


They were being polite, it would have been reasonable to:

if (c->x86_vendor == X86_VENDOR_INTEL)

>> No.64139871


Proofs or STFU. AMD themselves have clearly stated this isn't a big deal.

>> No.64139905

maybe but i don't think so
arm is affected also
which leaves only amd
nothing polite about it
just fact

>> No.64139919


>trying to imply that the basic spectre attack implies amd is open to meltdown despite it being easy to show on intel and impossible on amd


>> No.64139923

>arm is affected also
By Spectre (parts of which do work on AMD). This patch is for mitigating Meltdown.

>> No.64139933


>> No.64139945


>> No.64139959
File: 390 KB, 1529x1352, rulethirtydebian.png [View same] [iqdb] [saucenao] [google] [report]

Needs moar Super-H

>> No.64139980

The flaws being exploited here are not just bugs. They are inherit flaws in the concepts of speculation and branch prediction in the context of security regardless of what CPU architecture is used.

The only current proofs for intel and arm are just proof of concepts, they may not be the only ways to utilize these flaws on CPU that support said features.

Further, going forward, the only hardware solution around this will require a performance regression.

>> No.64139993


Not on AMD, see >>64134961

>> No.64140010
File: 212 KB, 500x580, niggered2.png [View same] [iqdb] [saucenao] [google] [report]

niggers buy tarded shit

>> No.64140019

Don't bother trying to explain it.

People here are fucking retarded and don't understand that there are two different things be discussed. It's partly the fault of the groups publishing the Spectre and Meltdown details. They're advertising Spectre as a vulnerability when it's really just a technique, and the real vuln is being able to overwrite entries in the Branch Prediction Table.

>> No.64140051

>intel and ayyymd BTFO simultaneously

>> No.64140052

we've got a /v/irgin here

>> No.64140067

>AMD themselves have clearly stated
Press releases from both Intel and AMD are irrelevant.
We waited at least 2 years for a public PoC on Intel. The availability of a speculation attack has been long... speculated. The theoretical availability of this bug affect any OoO CPU respecting the original AMD64 spec. Researchers have not been able to argue that AMD is fundamentally exempt from this flaw, nor we have any pointers to argue this. We at best can say that *this* CVE affects Intel.
I don't need to imply anything. It's all in the papers.

>> No.64140073
File: 20 KB, 300x400, 1488754927177.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64140075
File: 41 KB, 301x301, rulethirtyarch2.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64140085

That image should be changed to YES (but only FX series)

>> No.64140111

Will I die while browsing net now or what is this?

>> No.64140115

Variant 1 works on Debian out of the box in a Ryzen 1700 VM (Windows 10 Host)


>> No.64140124

Man, I've read the papers on the flaws. If Linus was really so confident that AMD was unaffected he wouldn't off load credibilty onto AMDs PR in his GIT commit message.

>> - Exclude AMD from the PTI enforcement. Not necessarily a fix, but if
>> AMD is so confident that they are not affected, then we should not
>> burden users with the overhead"

>> No.64140131
File: 243 KB, 1024x768, trumpian-lion.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64140135


Is this correct? Can't you can leak from process to process as all user processes would share the same protection domain?

>> No.64140137
File: 48 KB, 819x593, pre.png [View same] [iqdb] [saucenao] [google] [report]

ivy bridge pre patch

>> No.64140150
File: 47 KB, 786x593, post.png [View same] [iqdb] [saucenao] [google] [report]

ivy bridge post patch

>> No.64140168

>Can't you can leak from process to process
Not on AMD. They don't work that way.

>> No.64140170

It's better to offload this kind of decision on the hardware manufacturer in any case so they cannot come crying to him if they're affected after all.

AMD engineer posted the patch so why not pull since they're so confident.

>> No.64140171
File: 28 KB, 600x320, f.jpg [View same] [iqdb] [saucenao] [google] [report]

I'm so sorry

>> No.64140173

Benchmarking tools are going to be a lot slower since they rely on random data which fucks branch prediction. It's unlikely that consumers will see much variance in day to day operations. Gaming might get hit harder, but it likely won't be as bad as benchmarks suggest.

What will be affected are cloud clusters which ingest huge quantities of untrusted data.

>> No.64140181

what the fuck? That's almost cut down in half.

>> No.64140183
File: 10 KB, 236x199, 1513734109791.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64140213


This makes no sense as you are not bouncing between user and system code in a cpu benchmark. You are not doing syscalls in those tests.

>> No.64140214
File: 32 KB, 688x578, 1506942072546.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64140215

CISC died. Even x86 is a RISC processor, we're just not allowed access to it. All instructions get passed through a CISC -> RISC interpreter

>> No.64140223

It's Windows, he don't gotta explain shit.

>> No.64140244

didn't they enable it for AMD in this patch

>> No.64140255

I agree but everyone is using AMDs word as proof there is no issue. We simply lack a proof of concept or an extant malware.

If they leave the fix in and there is no exploit, then AMD gets a performance degradation. If they leave it out and an exploit exists, then for however long until its discovered AMD CPUs could be exploited with no means of detection.

Really, it's just waiting for the other shoe to drop.

>> No.64140259

Where is patch for windows what is KB # ?

>> No.64140260


>incucks desperately implying that this is a serious security flaw on AMD64

>> No.64140261

I was just thinking, couldn't we use meltdown on intel to our advantage to exfiltrate decryption keys (like aacs, hdcp, denuvo gaymen whatever) to free content from the clutches of malicious and/or obfuscated drm?

>> No.64140268

Finally somebody with a brain.

>> No.64140270

Intel's damage control is insane and a thing of beauty.

>Conflate two unrelated bugs.
>One is the issue and only touch Intel cpus, the other is nebulous and not important, but could potentially touch all CPUs.
>Also you won't lose performance. Only in some workload would you lose performance, like reading files and connecting to internet or DirectX calls, not all workloads.

Shit. This is absolutely genius.

>> No.64140276


>> No.64140285


Windows will get a patch as well. And as far as all news sites report, the linux patch will be the best, performance-wise, by only taking 5% performance. On Windows, you can expect up to 30% performance decrease, depending on your CPU

>> No.64140298
File: 7 KB, 146x199, segald.jpg [View same] [iqdb] [saucenao] [google] [report]

Literally fapping to this now

>> No.64140310

This is only the beginning. Usercode can read kernel memory. You've got access to any protected memory you've ever wanted with a bit of time and code.
Chaos is on its way for the tech world.

>> No.64140314

Linus invented Linux processors so he knows something. Luckily, Apple don't have any issues with bugs for now. Apple is the best.

>> No.64140315

You know, considering that, since everyone seems to be pushing for ARM desktops:
Would it be possible to make a ARM CPU that has an x86 translation layer?
That way we get to keep all of our x86 programs and don't suffer any performance losses brought about by emulation.

Of course, this kind of thing is why so many people hate x86, having decades of old instructions mucking the whole thing up

>> No.64140319


It gets worse kek

>> No.64140335

>Where is patch for windows what is KB # ?


>> No.64140337


>doesn't affect artificial benchmarks which is all that matters to intel customers

>> No.64140344

>Would it be possible to make a ARM CPU that has an x86 translation layer?
Of course. There's the Transmeta Crusoe, a VLIW processor that implements x86, I believe Nvidia's project Denver does the same but with ARM and is capable of x86 as well. There's also the Loongson processors that have some die dedicated to speeding up x86 instructions in QEMU.

>> No.64140350
File: 232 KB, 640x510, 1426448251001.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64140351


True that Linus invented 32-bit protected multitasking OS before Microsoft did.

>> No.64140352

So for anyone into cracking you shouldn't upgrade/patch your CPU at all since everything has been un-cockblocked.

>> No.64140360

Who the fuck is this retard?

>> No.64140375

son of the head of the european union

>> No.64140376
File: 8 KB, 246x278, 1513577441036.jpg [View same] [iqdb] [saucenao] [google] [report]

>first 64-bit
that was the MIPS R4000
alpha shit was overpriced and useless trash with no software

>> No.64140379

No, you should, but you should keep a separate unpatched partition or just a separate machine for it.

Would the exploit still work within a VM with a patched hypervisor?

>> No.64140380


In fact referring to "Linux" the kernel is correct as this is the only GPL free software in the GNU/Linux operating system which interacts with the CPU at this level.

>> No.64140387
File: 191 KB, 512x468, 1238964782.png [View same] [iqdb] [saucenao] [google] [report]

>unironically being this retarded

>> No.64140400

Any decryption key in user memory would already be vulnerable due to being store on hard disk in some way. Denuvo is an obfuscation software and already fucks with branch prediction badly.

The real risk is to data that should never be viewed by anything but the process owner, such as passwords and other user data. This is why VMs are shitting themselves right now, since they are sharing CPUs and have no idea if other processes could be trying the exploit.

>> No.64140403

>there are some differences
One of those differences causes AMD processors to be unaffected.

>> No.64140404

>a separate machine for it.
I thought that would be obvious.

>> No.64140415

They're not using KPTI on other CPU architectures either so why should AMD use it just because there are vulnerable x86 by Intel?

Besides not that KPTI exist it's just a matter of flipping the support on if AMD turns out to be vulnerable after all.

>> No.64140434


In the context of the kernel in the VM, yes.

>> No.64140485
File: 3 KB, 112x76, Cerebrate_SC1_Game1.png [View same] [iqdb] [saucenao] [google] [report]

Here are the updates for Windows.

KB4056897 - Win 7 & Server 2008 R2
KB4056898 - Win 8.1 & Server 2012 R2
KB4056893 - Win 10 1507
KB4056888 - Win 10 1511
KB4056890 - Win 10 1607 & Server 2016
KB4056891 - Win 10 1703 & Server 1703
KB4056892 - Win 10 1709 & Server 1709

>> No.64140517


It'd be interesting to confirm the "fix" doesn't gimp AMD. Any winbros to A/B some benchmarks?

>> No.64141287

Delay the update for a week and wait for the inevitable benches.

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.