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: 126 KB, 480x270, tinfoilhat.jpg [View same] [iqdb] [saucenao] [google] [report]
64134438 No.64134438 [Reply] [Original] [archived.moe] [rbt]

So basically every processor is just fucked right?

So what we do now?

>tfw the NSA backdoors wasn't a meme

>> No.64134451
File: 16 KB, 225x225, 5.jpg [View same] [iqdb] [saucenao] [google] [report]
64134451

Every INTEL processor, my dear intelshill.

>> No.64134510

>>64134451
Wrong.

>> No.64134521

>>64134510
Incorrect.

>> No.64134561

>>64134510
Only Intel processors are vulnerable to meltdown which is the easier to exploit and more harmful and that causes perf drops with the patch. AMD and ARM are only vulnerable to spectre which is harder to exploit and, from what is know thus far at least, less harmful(but also unpatchable)

>> No.64134604

>>64134561
>spectre
>harder to exploit
https://github.com/Eugnis/spectre-attack/

>> No.64134614

>>64134521
>>64134561
You are mistaken.

>> No.64134652

>>64134561
>harder to exploit
>less harmful
Source?

>> No.64134714

I can't wait for the day that the IME is completely broken

>> No.64134750

>>64134438
Buy Intel®'s latest Core™ processor with an accompanying motherboard plus DDR4 RAM modules. Buy a new VGA too, just to be sure.

>> No.64134751

>>64134652
>less harmful
The project zero blog post.
https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html
Can't be patched, though.

>> No.64134759

>>64134438
Buy an ARM license with all /g/, fix the bug in CPU design, produce this botnet-free CPU, I will be rich.

>> No.64134786

>>64134614
No. It is you who are mistaken, about a great many things.

>> No.64134795
File: 59 KB, 1196x300, 1515030456852.png [View same] [iqdb] [saucenao] [google] [report]
64134795

>>64134438
>So basically every processor is just fucked right?
No

>> No.64134844

>>64134795
Variant 2 has been demonstrated on AMD hardware.

>> No.64134864

>>64134844
Source

>> No.64134865

>>64134604
so much for "harder to execute"

>> No.64134883

>>64134864
Don't need one.

>> No.64134916

>>64134451
Wrong. There are older CPUs without cache prediction. Afaik this kind of shit was introduced in Pentium 3 or so

>> No.64134933

>>64134916
Yeah just everything from intel after 1999

>> No.64134935

>>64134916
That prevents Meltdown, but what about spectre?

>> No.64134948
File: 32 KB, 428x500, Born_to_feel.jpg [View same] [iqdb] [saucenao] [google] [report]
64134948

>>64134438
there's always a way in anon...

>> No.64134986

>>64134451
nope, spectre describes a general flaw in the way processors have been designed for decades. any cpu with branch prediction may have an exploit

>> No.64135010

>>64134986
eat shit nvidiot

>> No.64135037

>>64134883
Okay I'm convinced

>> No.64135040

>>64134604

nice

>> No.64135060

>>64134604
>>64134604
>>64134604
Someone with an amd processor run this and see if it works

>> No.64135064

windows users BTFO this is why you use linux

enjoy your slow-down patches and exploits!

>> No.64135070

>>64134714
fuck yes please. it would ruin everything and make every intel cpu for the past decade impossibly insecure when it is finally broken but hopefully it would force intel's hand, or at least their profits down below. I dream for the day that awful company is bankrupt

>> No.64135131

>even arm wont save
FUCK

>> No.64135134
File: 42 KB, 500x322, 1514831626563.png [View same] [iqdb] [saucenao] [google] [report]
64135134

>>64135060
>tfw noone uses AMD

>> No.64135143

>>64135010
no, unfortunately I have an intel i7. hopefully linux's new memory table fixes some things. I don't know the exact details of the meltdown hack, but I think that better virtual memory separation of processes and of kernel/userland is a good thing no matter what.

How would that even make me a nvidiot anyway saying everything is fucked? I would more likely be an inshill in that case. That's what Intel is arguing right now, or more accurately just trying to throw AMD under the bus for meltdown even though they have nothing to do with it.

>> No.64135163
File: 8 KB, 127x225, buros.jpg [View same] [iqdb] [saucenao] [google] [report]
64135163

SERIOUSLY, how can I also shill for Intel here? I need money and aparently it is pretty easy, since literally 60% of this thread are shills.

>> No.64135177

>>64134604
That's Spectre 1 which is easily patchable with no performance loss, Spectre 2 is the hard one which is also hard to fix and doesn't affect AMD unless you change a setting for the sole purpose of making it vulnerable.
Get your facts right retarded shills.

>> No.64135191

>>64135064
Are you fucking retarded?

>> No.64135209

>>64135064
Linux got the planned obsolescence patch before Windows lel
Also it's a hardware exploit and the OS doesn't matter.

>> No.64135334

>>64135060
what the fuck
glibc is installed

>> No.64135377

>>64135191
please don't disparage the differently abled
and no, i have Ubuntu so i'm at least smarter than every others windows cuck :-)

>> No.64135380
File: 13 KB, 647x404, 2018-01-03-191728_647x404_scrot.png [View same] [iqdb] [saucenao] [google] [report]
64135380

>>64135334
>>64134604
forgot pic

>> No.64135382
File: 39 KB, 1221x322, 423423453.png [View same] [iqdb] [saucenao] [google] [report]
64135382

>>64134561
>but also unpatchable
Nice try Brian:
>https://www.amd.com/en/corporate/speculative-execution
Variants 1 and 2 are part of the Spectre exploit, variant 3 is Meltdown.

>> No.64135416

>>64135380
font?

>> No.64135440

>>64135416
terminus

>> No.64135457
File: 54 KB, 735x517, ryzen 1800x.png [View same] [iqdb] [saucenao] [google] [report]
64135457

>>64135060
not sure what I'm supposed to be looking for here...I don't see the magic word in the output though

>> No.64135479

>>64135457
you compiled this right? what compiler and options did you use?

>> No.64135493

>>64135457
windows 10 patched out already, maybe yours already updated

>> No.64135502

>>64135457
>willingly installing exploits

>> No.64135503 [DELETED] 
File: 120 KB, 1306x519, back_to_housefires.jpg [View same] [iqdb] [saucenao] [google] [report]
64135503

>>64134438
TFW have a fully functional Pentium D sitting turned off because too much power consumption

Hyper Bipeline was the right way after all. I wonder if it was the special project Pentium M Haifa design jews that did this

>> No.64135525

>>64135502
Its literally one source file and you can see what it does pretty clearly

>> No.64135536
File: 122 KB, 900x344, unknown.png [View same] [iqdb] [saucenao] [google] [report]
64135536

>>64135382
Says AMD

>> No.64135573
File: 62 KB, 757x471, ryzen 1800x - no debug symbols.png [View same] [iqdb] [saucenao] [google] [report]
64135573

>>64135479
whatever the default for visual studio 2017 is for c/c++. No optimization, x64. I get a different result with a release config (no debugging symbols)

Open to suggestions

>> No.64135594
File: 70 KB, 735x517, ryzen 1800x.png [View same] [iqdb] [saucenao] [google] [report]
64135594

>>64135457
oh, duh

>> No.64135604

>>64135594
lmao

>> No.64135623
File: 47 KB, 649x280, committed.jpg [View same] [iqdb] [saucenao] [google] [report]
64135623

>> No.64135630

>>64135594
what... what am i looking at?

>> No.64135649

>>64135630

celery man

>> No.64135663
File: 190 KB, 853x480, 1454140123431.jpg [View same] [iqdb] [saucenao] [google] [report]
64135663

>>64135493
spectre is unpatched, kpti (or whatever is the windows equivalent) protects only against meltdown

>>64135594
neat

>> No.64135681

>>64135457
>>64135594
AMD rekt

>> No.64135728

>>64135663
Spectre 1 is patched and AMD isn't affected by Spectre 2.
This nomenclature is just one big PR ploy to rope AMD into this exploit fuckery even though it's not affected at all and won't have any performance loss either.

>> No.64135737
File: 207 KB, 1153x1187, Screenshot_1.png [View same] [iqdb] [saucenao] [google] [report]
64135737

funtimes

>> No.64135747

>>64135630
from the github:
>In this example we're putting text "The Magic Words are Squeamish Ossifrage." in memory and then we're trying to read it using exploit. If system is vulnerable, you'll see same text in output, readed from memory.

so I am vulnerable to this attack on a Ryzen 1800x, Win 10 LTSB (ver 1607). Should also be noted I'm about 3 months behind on security updates, not sure if it's been patched since about to test.

>> No.64135759

>>64135747
update and retry

>> No.64135781

>>64135759
Should also write a quick program that makes lots of sys calls and see how long it takes before and after updating

>> No.64136072

>>64134451
>Every INTEL processor, my dear intelshill.
hey nigger, the O.G. Pentium 1 is immune, along with Itanium and most older Atom processors.

>>64134916
>Afaik this kind of shit was introduced in Pentium 3 or so
It was the Pentium Pro/2 that introduced Out of Order and speculative execution.

>> No.64136190

>>64135380
>#define sscanf_s sscanf

>> No.64136312

I am glad I have some older computers lying around. I know for sure, I have at least one intel cpu that is pre-jewery.......
Definitely glad I moved my server over to an amd cpu powered tower a week before the news dropped..

>> No.64136331

>>64134438
AMD doesn't have this problem.

>> No.64136443

>>64135209
Linux added the nopti kernel parameter to disable it on boot idiot.

>> No.64136471

>>64134438
You can't escape the botnet. Open source firmware, OSes, and software won't solve the problem. You can try using network isolation, but have fun with airgap hacks... The only solution is the wrap yourself in tinfoil and hide in a faraday cage.

>> No.64136582

>all the cpus made to run with w7 are all fucked
>don't want to buy modern cpu nor downgrade to botnet 10
Perhaps it's time for me to unplug the Internet cable

>> No.64136611
File: 65 KB, 679x711, 1515035041416.jpg [View same] [iqdb] [saucenao] [google] [report]
64136611

>>64134614
Try again champ.

>> No.64136615

>>64136331
Yes it does
back to >>>/pol/ you

>> No.64136625
File: 350 KB, 672x794, 6541684984949948.png [View same] [iqdb] [saucenao] [google] [report]
64136625

>>64136615

>> No.64136629

>>64136615
>everything I dont like is /pol/

>> No.64136685

>>64136611
garbage table, author should kill themselves immediately

>> No.64136695

>>64136685
>I don't like it so its wrong REEEEEEE!!!
mmmk :^)

>> No.64136721

>>64135594
>>64135759

Just updated windows and I'm still vulnerable, I get the same results.

>> No.64136855

>>64135728
But I'm on AMD and am affected, see:

>>64135594
>>64136721

And I'm on windows so >>64136611 is wrong too.

Or the github code that was linked by >>64134604 is wrong.

>> No.64136960

>>64135663
>>64135493
So wait, are all machines already going to feel the slowdown?
How do we tell if our machine is patched or not?
Is leaving the machine unpatched a viable option?

>> No.64137299

>>64134933
1995, my dear friend

>> No.64137330

>>64136611
Is spectre 1 for amd on linux fixed with the nopti boot param. ?

>> No.64137465

>>64136629
Lately, yes actually

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