[ddevault]

>But if sudo were written in Rust, it could have the same level of complexity and not be vulnerable.

This is not true. Complexity breeds bugs, including security bugs, and memory safety doesn't change that. Your example is a good one - here's another: doas once failed to limit the environment variables which are passed to the child process, which could be used to nefariously influence the program running (e.g. with LD_PRELOAD). How would Rust prevent that oversight? It wouldn't.

A simpler program will generally be more secure than a complicated one, no matter what language either is written in. Furthermore, rewriting an established program from one language to another will always introduce more bugs than it fixes, and more severely the more complex the program is. The single best way to improve security is to reduce the attack surface, and the single best way to do that is to reduce the complexity of your system.

[zamadatix]

If you go a little further with the quote:

"Yes, it would still be vulnerable to logic errors... But it wouldn't be vulnerable to this. "

I think you'll find in disagreeing with the comment on logic errors you just said the same thing the comment did about logic errors.

Also I think the generalization that rewriting an established bit of code in a new language in a secure language is a bit too general. clearly Firefox not only set out to make Rust for this purpose but it's not had an explosion in vulnerabilities with the modules it has replaced. Quite the opposite actually. Nor has every tool or app rewritten become a security failure compared to the original. I do think it's something that can easily be screwed up though, especially if someone rushes through by focusing on functionality duplication instead of building a more secure version of something.

Regardless, both "using a memory safe language results in a more safe program" and "having a minimum attack sufrace results in a more safe program" can be true. There is no need to make it a choice of A or B.

[ddevault]

>I think you'll find in disagreeing with the comment on logic errors you just said the same thing the comment did about logic errors.

I think you'll find that my comment explicitly acknowledges this and expands on it with another example. Are we done telling each other to read the things we're writing?

>Firefox not only set out to make Rust for this purpose but it's not had an explosion in vulnerabilities with the modules it has replaced.

You're setting the bar pretty high with an "explosion" of vulnerabilities here. Rust programs have vulnerabilities, including rewrites. They also have other kinds of bugs, often ones which were not present in the code that they're replacing. You need only browse your nearest convenient RiiR bug tracker to find evidence of this.

Let me restate my thesis in mathematical terms. If we presume that 1 in 100 lines of production code has a bug in it, regardless of language (generous, I know), and that 1 in 10 bugs in C programs are memory corruption related, then saving 10% of those bugs by rewriting it in Rust would take a 10,000 line codebase from 100 bugs to 90 bugs. A 1,000 line codebase, still written in C and without the advantage of memory safety, would have only 10.

In today's example, sudo is a caricature of runaway complexity. Rust is often touted as a panacea, but C has very little to do with why sudo is insecure. Sudo is comically overengineered and that level of overengineering has no place in a security context. This is the larger issue that needs to be addressed, not Rust.

[vitno]

I agree Rust is not a panacea and that rewrites create their own set of problems, the only issue with this analysis is assuming 1/10 bugs are memory corruption related.

Both Chrome & Microsoft found about 70% of bugs to be memory safety related. I've heard similar numbers out of FB as well. The math looks a little different with that data.

https://www.chromium.org/Home/chromium-security/memory-safet...

https://www.zdnet.com/article/microsoft-70-percent-of-all-se...

[ddevault]

Even if we run the same math with 7 out of 10 bugs being memory safety related, and assuming that Rust prevents all of them, those same example programs end up with 30 bugs in Rust and 10 bugs in C.

There's another argument I could make, too. Look at the bug tracker for the program you want to rewrite in Rust, examining the historical bugs. You'll find that there are often hundreds or thousands of mistakes that they made and already fixed in the original codebase. If you're rewriting it from scratch, can you be sure you won't make just as many? A stable, maintained codebase with a low throughput of changes tends to have fewer bugs over time, as the lack of churn avoids introducing new bugs and the application of time susses out all of the existing bugs. Rewriting the whole thing from scratch has a very high rate of churn, introducing a whole new slew of bugs on its own.

Now, a small codebase, focused on delivering its key value-adds without distractions, kept stable and at a low-churn rate over a long period of time: no matter what language you use, this is the best recipe for reliability and security.

[zamadatix]

So again why does it have to be "rewrite at 1/10th the complexity in <Language A>" (10%) vs "rewrite in <Language B> at full complexity" (30%)? What's preventing using Language B for the complexity rewrite and getting 0.1 * (1 - 0.7) = 3%?

Rewrites do bring the chance to Royally Screw it Up™ so it's certainly not simply a product of "it is now written in <Language X> therefore safe" but as it said not only have projects shown the security didn't fall apart but they have shown the opposite.

I agree you don't get there by a bunch of yolo rewrites to whatever is hip though, it has to be a planned effort that isn't rushed. Much in the same way quickly writing a small replacement utility does not inherently make it more secure or reliable than an existing significantly more complex utility. Even just trying to shave some functionality off the existing code is rife with "but how does removing this piece affect the app remaining logic" and takes time and effort to do right.

Both methods do have to be done right and both do greatly help security but there is nothing about picking a memory safe language or making a significantly narrower focused utility that preclude each other.

[ddevault]

You can do both! But because simplicity has a substantially greater impact than the language choice, I think it's better to focus on that. Right now, the ecosystem is focusing more on the language choice, and hardly talking about simplicity at all. And particularly in the case of Rust, I think it fails simplicity a lot in its own ways - in the stdlib, the compiler and toolchain, the language design - and the trade-offs don't really make sense for a lot of use cases that people are pining for it over anyway.

[myrrlyn]

helps that a 10kloc c program getting riir'd probably won't be a 10kloc rust program, because c doesn't have libraries and rust does.

it is literally impossible to write "a small codebase focused on its key value-adds without distractions" in a language that doesn't have strings and requires you to build a dictionary from scratch

[ddevault]

>helps that a 10kloc c program getting riir'd probably won't be a 10kloc rust program, because c doesn't have libraries and rust does.

What? Rust has so few libraries of significance that it still depends on C for security-critical areas like SSL.

>it is literally impossible to write "a small codebase focused on its key value-adds without distractions" in a language that doesn't have strings and requires you to build a dictionary from scratch

Strings are misunderstood, I'm not going to get into it here. My dictionaries in C usually clock in at about two dozen lines of code. The complexity doesn't go away because your language does it for you.

[vitno]

Except that Rust is also a much much more expressive language. Even ignoring things like solid module support and libraries you'll find your Rust programs to be much fewer LoC (assuming bug/LoC is the right metric) for equivalent functionality.

I agree that rewrites have the serious potential to introduce new bugs and the cost is rarely worth it if the codebase is actually that stable and low througput, but the reality is that most aren't. A one time high cost in exchange for introducing 70% less bugs over a period of N years starts to look like a good trade off.

Yes, complexity is the root of all evil. I can get onboard with the whole statement except the "no matter what language you use". If you have the ability to use any language that enforces memory safety, we should use it.

[ddevault]

Lines of code is a poor approximation for complexity. Rust programs are shorter, but they are not less complex. The AST is similar and the graph of relationships between different parts of the code is much more complex than in C. Overall I'd say it balances at best, if not that Rust is more complex.

[jodrellblank]

> "Even if we run the same math with 7 out of 10 bugs being memory safety related, and assuming that Rust prevents all of them, those same example programs end up with 30 bugs in Rust and 10 bugs in C."

Maybe but not necessarily; it's reasonable to assume that Microsoft and FaceBook put non-zero effort into designing around, programming around, testing, looking for and fixing memory safety related issues in their C code. It could be the case that not having to care so much about those frees up some non-trivial amount of attention and time which could be spent on the other classes of problems.

[joveian]

Similarly, it is possible that they would use the time to add new features with new bugs. I'd personally suspect that to be the more common outcome.

[jodrellblank]

> "Furthermore, rewriting an established program from one language to another will always introduce more bugs than it fixes"

Here[1] is a link to a slideshow of a talk on the F# language, with a case study from EON PowerGen company rewriting the core of an application evaluating revenue due from their balancing services contracts nationwide in the UK. It was originally 350,000 lines of C# developed by 8 people in 5 years and incomplete. It was redeveloped by 3 people (2 had never used F# before) in 30,000 lines, complete in 1 year.

They claim zero bugs in the F# redeveloped system (page 29). This example also gets a mention in a Don Syme (F# language designer) talk in 2018[2] with the PowerGen employee in the audience.

The PDF cites a testimonial from Kaggle saying they're moving more and more of their application into F# which is "shorter, easier to read, easier to refactor, and because of strong typing, contains far fewer bugs".

[1] https://www.microsoft.com/en-us/research/wp-content/uploads/...

[2] https://www.youtube.com/watch?v=kU13g_noAQM

[comex]

Incidentally, after inspecting doas for a few minutes, I found two near-vulnerability bugs in it.

The first bug lets any user cause doas to read out of bounds of an array, though not in a way that's exploitable.

Well, it's arguably a bug in libc. If you run doas with a completely empty argv (argc = 0, so not even an executable name; the two systems I tried, Linux and macOS, both let you do this), getopt will exit with optind = 1. Then when doas does;

    argv += optind;
    argc -= optind;

`argc` will become negative, and `argv` will advance past the null terminator. On most OSes, the `argv` array is immediately followed in memory by `environ`, so argv will now point to the list of environment variables.

doas will then dereference argv, and generally act as if you tried to execute a command consisting of the environment variables. However, the environment variables are not secret, and doas doesn't behave any differently than if you just passed the environment variables as normal command-line arguments, so this is not exploitable.

On an OS where argv is not followed by environ or a similar array of character pointers, doas might crash instead, although since it only reads from those pointers rather than writing to them, this still probably wouldn't be exploitable.

The second bug would compromise memory safety if things were slightly different. The bug is in configuration file parsing. Even if it did compromise memory safety, it would not actually be exploitable, because doas normally only parses the trusted systemwide configuration file. It can be asked to parse a configuration file passed on the command line, but it drops privileges before doing so. This is a good example of layered defense, so kudos to doas for that! Still, I thought the bug was worth mentioning.

The bug is a traditional sort of integer overflow. parse.y grows the array of rules with

    maxrules *= 2;

but maxrules is an int, so this will eventually overflow if the configuration file is large enough.

However, because maxrules happens to be signed, before doubling produces a smaller-than-expected positive value, it will first produce a negative value. This will then get sign-extended when converting to size_t (assuming 32-bit int and 64-bit size_t), and reallocarray's overflow check will trigger, causing reallocarray to return NULL. doas interprets that as out-of-memory and handles it cleanly.

(On a system where sizeof(int) == sizeof(size_t), things are a bit different, but it will just run out of memory before maxrules gets that high.)

Moral of the story? Well, as I see it:

Simplicity and layered defense, both featured in doas, are both effective ways to avoid vulnerabilities. But guaranteed memory safety, which would require a different implementation language, is also an effective way to avoid vulnerabilities. You aren't forced to pick and choose. Why not demand all three?

[ben_bai]

The argv += optind; is a standard pattern. I have never considered argc=0 case to be possible. I need to read some more on this.

As for your second find. It already got fixed: https://marc.info/?l=openbsd-cvs&m=161176698927944&w=2

[ddevault]

Nice finds. I would agree that that's more arguably a bug in libc than in doas, but also note that the startup code for any language has to consider this case. As far as theoretical operating systems are concerned, this is a consequence of the System-V ABI, so any OS compatible with it would have the same issue.

As for the integer overflow case, it's also highly unlikely to be exploitable, even if it were unsigned - the system would have to, as I'm sure you can infer, have tens of millions of rules before this was an issue. It's quite within the realm of reason, in my opinion, to declare this an acceptable trade-off. The rest of your explanation shows that even if this weren't the case, the bug wouldn't be exploitable.

Anyway, I like your comment, but I'd recommend a different moral to this story: in the space of 47 minutes you were able to conduct a reasonably thorough audit on the doas codebase. Wanna give that a shot for sudo now?

[comex]

> I would agree that that's more arguably a bug in libc than in doas, but also note that the startup code for any language has to consider this case.

This is true, but for a language where dynamically sized arrays are a standard data type, the most natural thing to do is to start by collecting the arguments into an array (maybe copying the strings at this point, maybe not). All further argument parsing is done with the array and is thus bounds-checked. I checked Rust's standard library and sure enough, it follows this pattern. Though, I could imagine some hypothetical startup code messing up the argc=0 case if it tried to separate argv[0] from the rest of the arguments while constructing the array.

> Anyway, I like your comment, but I'd recommend a different moral to this story: in the space of 47 minutes you were able to conduct a reasonably thorough audit on the doas codebase. Wanna give that a shot for sudo now?

Fair point. (And I didn't downvote you.) But in my opinion, that just confirms my view: ideally you want both simplicity and memory safety.

[ddevault]

Aye, I agree. But if we consider that case, a similar mistake could be made: hard-coding argv[0]. The result is different, in that the program just aborts, but it's still the Wrong Thing To Do, and in both cases it never leads to anything exploitable. Bugs are bugs, no matter what language. We could come up with examples all day. Just head to your nearest Rust program's bug tracker :)

[gpm]

Aborting when argv[0] doesn't exist... is a perfectly reasonable thing to do? Someone called the program with arguments severely out of spec, crashing is fine.

[ddevault]

It's actually within spec, in this case. Still reasonable?

[staticassertion]

A lot of your statements are pretty strong, and imo totally incorrect.

> Complexity breeds bugs, including security bugs, and memory safety doesn't change that.

Yes, memory safety changes that radically.

> A simpler program will generally be more secure than a complicated one, no matter what language either is written in.

Disagree, but the statement is really weak anyways, especially since 'complexity' is an ill-defined term. More features? Cyclomatic?

> urthermore, rewriting an established program from one language to another will always introduce more bugs than it fixes, and more severely the more complex the program is.

Should be obvious to anyone that this isn't true.

> The single best way to improve security is to reduce the attack surface,

Not true, but it's a great way to start.

[ddevault]

>Disagree, but the statement is really weak anyways, especially since 'complexity' is an ill-defined term. More features? Cyclomatic?

I'm not sure of any definition of complexity you could appeal to which makes my argument weak.

>>rewriting an established program from one language to another will always introduce more bugs than it fixes, and more severely the more complex the program is.

>Should be obvious to anyone that this isn't true.

The opposite is painfully obvious: (1) Writing code causes bugs. (2) Rewriting an established project involves writing more code than leaving it would. (3) Writing all of that new code will introduce new bugs which were not present in the original.

[staticassertion]

Yeah I think that's an absurd reduction. Rewriting code means that you can solve fundamental architectural issues, that you can start fresh with better tooling, that you have the lessons learned without the technical debt, etc.

[ddevault]

Yeah, but why the assumption that those things are an issue? We're talking about mature codebases. Rewriting it again in C would also give you a chance to start fresh with better tooling, lessons learned, paying back tech debt, etc. Even still, you're going to introduce new bugs in the process. You might fix a few hard-to-address architectural issues, but all of the other bugs would be easier to fix in the original codebase than by rewriting the whole thing.

I'm not saying that a rewrite is never justified, but rather that the argument that we should rewrite in Rust simply to avoid bugs has little weight.