Hacker News Reader

Comment by khuey | Hacker News Reader

> Just want to remind everyone that only 1% of vulnerabilities are memory related in the average Joe's code.

Unless your point is merely that average Joes write such terrible code that you don't even need memory safety issues to exploit their software, [citation needed]

Google says memory safety issues are 75% of exploited zero days. (https://security.googleblog.com/2024/10/safer-with-google-ad...)

View on HN

And at least in Chromium project, half of those memory safety issues are use-after-free: https://www.chromium.org/Home/chromium-security/memory-safet...

Then you're looking at very old data, recent top 500 CVE's 100% of the use-after-free exploits where escalation to Admin on Windows.

Most memory bugs in Chromium are in V8, either entirely in the JIT or at the boundary with C++. Rust wouldn't help here because the borrow checker can't see through these boundaries, and it's precisely this opacity where the developers also lose track of things.

Which isn't to say Rust wouldn't have caught many of the other memory safety issues, but 75% is horribly misleading.

In Chromium which is the condom of the internet, already millions of developer hours of the most highly paid developers have spent their days trying to make it water tight. This is not your average software project.

My understanding is that they claim that the average Joe writes code in a garbage-collected memory-safe language.

Which is... true? but irrelevant. Such applications are not suggested to be ported to Rust. Of course, some people still do that, because they like Rust; but that's their personal choice.

PHP is garbage collected for reference. So rewriting wordpress plugins in Rust would not make them more secure but rewriting HAProxy or Nginx would? It's delusional.

If the class of bugs WordPress has is different from the class of bugs HAProxy/Nginx, eg SQl injection vs buffer overflow, then why is a different mehtoyof fixing "delusional"?

The point is that memory issues are a smallish number of issue compared to the larger ecosystem of vulnerabilities, and choosing to port everything to Rust is like over-optimizing. Well, that’s my 2 cents.

For a language as ugly as Rust, my thought is that people should actually be using Ada, and have a mathematically provable correctness angle; not just a replacement for C/C++ with memory safety.

> The point is that memory issues are a smallish number of issue compared to the larger ecosystem of vulnerabilities

If memory safety issues are 75% of exploited zero days it sounds to me like they're the biggest issue in the ecosystem by far.

Perhaps. But what percentage of exploits are actually zero days versus, say, 10 days or 100 days?

Most exploited code probably exists in the application layer in a high-level, memory safe language. I would wager that but I don’t have time to cite ten papers on HN.

In highly specific programs like browser sandboxes. Not in things like wordpress plugins or nodejs servers.

Rust and Ada are about equally safe, both have advantages and disadvantages. Perhaps you're thinking about SPARK ADA, but that's a different kettle of fish.

It's a bit like saying you should program in C, because formal verification tool X generates C code hence C is safe.

Yeah SPARK ADA is what I meant :)

I think formal verification is the way to go with AI moving forward.

Rust and non-SPARK Ada are not equally safe. Ada is unsafe in the presence of data races, and also has runtime checks that slow it down, or you disable them and then it's even less safe.

Rust is a beautiful language. Gorgeous.

Google Chromium but not it's not even in the top 10 of vulnerabilities in Github

Reply on news.ycombinator.com