> CVE-2017-6074: DCCP double-free vulnerability (local root)
seclists.org
So it's 2017 and we are being cucked by a fucking double-free vulnerability.
C is a cancer that plagues the modern software industry. If you want guaranteed memory exploits and security vulnerabilities in your program with timesink development period then use Assembly, not C.
And fuck Linus too. NetBSD written in Rust when?
Other urls found in this thread:
>Look mom, I posted it again!
Are you upset?
> NetBSD written in Rust when?
Now: redox-os.org
That's actually nice. Does it support Vim and GCC? Please say yes
No, it doesn't even have rustc working on it yet.
Hmm, I have to wait then
the exploit was patched a day after it was discovered nigger
You missed the point. This isn't a le linux suckzz!!11 XDD thread
>You shat your pants in public!
>Yes, but this isn't a problem since I've changed them eventually.
>then use Assembly, not C.
C is platform-independent Assembly.
how difficult would it be to write an os in Ada?
i mean you could basically just rewrite gnu/linux in Ada right?
False equivalence and you know it.
If you prefer slow programs why not do java? It doesn't have vulnerabilities of this nature due to the VM.
Nobody writes rust for good reason.
Rust is pretty fast. By your logic C is slow as well.
Not really user. I wish it were though.
>a day after it was discovered
We don't know when it was first discovered actually, it may have been used in attacks in secret.
And people may not have updated their systems yet so there's still vulnerable people.
Java is too high-level, C is used mainly for low level things, MCU l, talking to metal etc.
>by your logic
What logic? I made claims. What your post would imply is that C is the same speed as Rust. Which is simply false if rust is to offer extra protection. For instance here with the double free you couldn't protect against that for free.
It's pretty close
I think that most of the protection is checked while compiling, not in run time
> For instance here with the double free you couldn't protect against that for free
Rust has no runtime checks for double free, it's all checked at compile-time. The only checks I know it does at run-time in release builds is bound-checking, it's highly optimized and you can disable it if you want.
Reminder that 1 Rust compilation = 1 formal verification and Rust programs are guaranteed to be bugfree. Another benefit of Rust is that it helps to smash the patriarchy.
>microkernel
why
Hell yeah I love smashing imaginary figures from candyland
zcat /proc/config.gz | grep DCCP
# CONFIG_IP_DCCP is not set
phew
>it was discovered
you mean officially or unofficially, you silly excuse of a CIA nigger?
kill yourself, now
I doubt they would have caught this at compile time.
>benchmarksgame
I'll keep saying it. It's not a realistic test of programming languages at all. Especially for many of the less low level languages because they're all attempting to emulate the lower level implementation. That's how you get speed, you program for the hardware you're on. So you find plenty of python examples that aim to call out to precompiled C code as quickly as possible. Is it realistic to do that? Similarly with Rust I'm sure you're using unsafe to get where you are in that chart and you didn't even get all the way there. That's a major problem. Even in this contrived example you don't measure up.
>that damage control
>I doubt they would have caught this at compile time.
I doubt you know what you are talking about
I should add that there's plenty of places where languages like rust (languages with focus on safety) are appropriate. But in a kernel is probably not the place. Unless it's exclusively for home users
>someone shouts "nigger" in a crowded theater
>OP gets *triggered*
>demands that we ban the word "nigger"
>later, goes home, checks his email, sees a new CVE bulletin in the mailing list
>it's a C double-free vulnerability
>OP gets *triggered*
>demands that we ban the C programming language
> I doubt they would have caught this at compile time.
What do you mean you "doubt"? Why don't you read on how Rust actually works before making such statements?
Borrow checker and move-by-default guarantee you never have two owners of the same piece of memory, neither can you access memory in any way after freeing it, thus making double frees impossible in safe Rust.
Who are you quoting?
but who are you quoting?
go back to your garden, /jp/
Who are you talking to?
Because microkernels are the future.
t. Andrew Tanenbaum ca. 1987
Well there's tools like valgrind especially designed for this sort of thing that wouldn't have to my knowledge (which isn't perfect). It's a qualified guess user. Just like op does his own guess that Rust would have solved it.
>in safe rust
But it's a kernel user. It'd be outrageous to apply the performance penalties of Rust to the entity of the Linux kernel. The fact that there is an unsafe keyword makes it clear that the Rust community understands the limitations.
>write everything in rust
>get rekt anyway
>stroke out before fixing anything
>Nobody writes rust for good reason.
because whatever you wrote won't fucking work in a week or two when they change the ABI again
And I doubt you can write an OS without unsafe code, so your hole starting point is moot. If you don't model the way you use your pointers correctly within Rust's type system (and no one forces you to do it), you get your double free soon enough.
Why is rust being shilled? It's an SJW language
you cannot use valgrind with a kernel
> It's a qualified guess user.
Well, it's not, since you don't know how Rust works and just assume it works like C.
> It'd be outrageous to apply the performance penalties of Rust to the entity of the Linux kernel.
Yes, but it's better to have 5% unsafe code than 100% unsafe code. See for example, most of the kernel code is written in safe Rust.
Why is C being shilled? It's an NSA backdoor language
Any language can be turned into a backdoor. Retard.
>C is backdoored
>still uses a computer
kek
> he thinks a language developed by > is free from NSA
C apologists are NSA shills plain and simple. Rust makes it harder to produce vulnerable programs as it actually implements safety precautions like Bounds checking, dangling pointer prevention, data race prevention, lifetimes etc. This is against the interest of the NSA and so they are shilling C so that people continue to produce exploit prone, unsafe and vulnerable programs.
Ignore the NSA/C shills. Programming in C should be banned and outlawed.
Say no to security exploits, say no to the NSA. Your data and your privacy is worth preserving.
Networking was a mistake.
t. SJW
Nice try, CIA. No one wants your weekly NSA backdoors anymore
>be clueless Rust idiot
>have Github filled with Ruby, Python, JS projects and ofc the type-safest FizzBuzz in Rust
>hear of vulnerability in age-old C project with a trillion contributers
>security expert mode: activated
t. CIA Nigger
Shy why are you anti-NSA? Do you have something to hide? Are you a terrorist?
Who said these?
No, I am a programmer who doesn't want vulnerabilities in my programs.
Who are you quoting?
>assume it works like C
I clearly didn't. There's no way in hell C catches this. I don't know what unsafe rust does though. If it's as worthless as C at catching these vulnerabilities then it's useless.
>it's better to have some amount of safe code
No. It isn't. Facebook considers 1% performance gain in their backend a massive win. If you're slowing them down this way then they won't use your software. Facebook is an extreme example but really anyone with performance concerns will tell you they prefer a mild increase in performance over the mild security gains of moving from a mature kernel like Linux to your rust copy. As I've said. If you're a home user you can go with the gimped version.
Why don't you rewrite the Linux kernel in rust and see how many people use it before you start making wild speculations about how good or bad it is.
>see
And? How many serious customers use redox? How many people moved when they got the news of this amazing change? I'm sure you can find some marketing wank on that somewhere if it's such a great idea. Likely you will find that there's a small group of people with all the resources in the world to throw around that like it on the concept level and have no proof that it was a good idea.
t. SJW who wants a compiler to shame you for being a white male.
t. CIA Nigger who wants his shitty compiler to allow buggy piece of shit programs to be published
>How many serious customers use redox?
They don't use Linux either. How did you miss the point so bad?
>they don't use Linux either
You have to be joking.
>YOU'RE A FUCKING WHITE MALE
Not an argument
I'm not. Linux was always a joke, and people laugh at the codebase as new vulnerabilities come out every single week
>A
>FUCKING
>DOUBLE FREE
Who are you quoting?
what do serious customers use then?
No one said this ITT
Windows
>Windows
KEK alright good bait you got me
That's what you're saying when you shill rust. Only someone who wishes jamal would fuck his wife use it.
Ever had a job?
>No. It isn't.
Yes, it is, you're just being contrarian for no reason. Besides, safe rust already has C-tier performance, so int's not like you'll be trading safety for 50% slowdown. Besides, the point of unsafe in Rust isn't performance, the only runtime check Rust has is bound-checking, and you can disable it. The point of unsafe Rust is low-level stuff like hardware and direct memory allocation.
> no one uses redox
Ofc no one does, it's a hobby OS being written by a single guy. Redox is a proof of concept, not a production-ready OS. Still, it proves the concept well enough.
No, actually C stands for Cucks. C is a cancerous. A Communist agenda if you ask me.
>BAN C
...but how?
That's why an anti white communist organization is rewriting Firefox in rust, a language they created.
Ban C programming books
Seize C documentation
Fire C developers
Outlaw C book publications
>zcat /proc/config.gz | grep DCCP
gzip: /proc/config.gz: No such file or directory
phew
C.ucks getting cucked by bulls. That's to be expected.
That seems a bit authoritarian.
By requiring that all new government-funded project be written in a safe language.
By banning C language classes in colleges.
You mean rust cucks get cucked by Jamal, which is part of the rust language.
You don't need to. Given how quickly C is dying it'll be forsaken in the next decade.
No one really uses C anymore for new projects.
>tfw Tanenbaum was right and Linus was wrong
No I meant C.ucks. C literally stands for Cucks and Communism.
This is why we all use Minix 3.
Have you?
>Google uses Linux
>Amazon uses Linux
>Twitter uses Linux
>>no srs company uses linux
Spitting out lies isn't recommended desu
>safe language
first you would have to make one.
Most of the world uses Windows anyway :^)
This, also it must have a COC so that women/PoC/trans* people feel safe and included.
>Redox
I'm sorry to have to tell you this user, but that's not NetBSD.
>By requiring that all new government-funded project be written in a safe language.
This doesn't fix projects that are not (primarily) government-funded, like Linux or OpenSSL.
>By banning C language classes in colleges.
Are there any? The only C classes I ever did were under the electrical engineering faculty for an 8-bit microcontroller. We were never told how to use C for anything with an MMU, yet alone a full PC. It was just C++ and Java.
C can't be when a literal commie cuck organization is writing rust.
>Check your privilege C shitlord
Not an argument
>Are there any?
Yea, very common in EE, probably only language they learn, but then again they don't write linux kernels.
Actually, most of the world uses Linux-based OS.
gj, you're safe
Just in time for Google to abandon the Linux kernel in Android, in favor of the Magenta kernel in Fuchsia.
Never coming out
Just like a good Linux distribution, then.
But android has been out for 10 years.