Google On-line Safety Weblog: Naked-metal Rust in Android
Final 12 months we wrote about how moving native code in Android from C++ to Rust has resulted in fewer security vulnerabilities. Many of the parts we talked about then had been system providers in userspace (operating underneath Linux), however these aren’t the one parts sometimes written in memory-unsafe languages. Many security-critical parts of an Android system run in a “bare-metal” atmosphere, outdoors of the Linux kernel, and these are traditionally written in C. As a part of our efforts to harden firmware on Android devices, we’re more and more utilizing Rust in these bare-metal environments too.
To that finish, we now have rewritten the Android Virtualization Framework’s protected VM (pVM) firmware in Rust to offer a reminiscence protected basis for the pVM root of belief. This firmware performs an analogous perform to a bootloader, and was initially constructed on high of U-Boot, a broadly used open supply bootloader. Nevertheless, U-Boot was not designed with safety in a hostile atmosphere in thoughts, and there have been numerous security vulnerabilities present in it as a consequence of out of bounds reminiscence entry, integer underflow and reminiscence corruption. Its VirtIO drivers specifically had a lot of missing or problematic bounds checks. We mounted the precise points we present in U-Boot, however by leveraging Rust we are able to keep away from these kinds of memory-safety vulnerabilities in future. The brand new Rust pVM firmware was launched in Android 14.
As a part of this effort, we contributed again to the Rust group through the use of and contributing to present crates the place attainable, and publishing a lot of new crates as nicely. For instance, for VirtIO in pVM firmware we’ve hung out fixing bugs and soundness points within the present virtio-drivers crate, in addition to including new performance, and at the moment are serving to preserve this crate. We’ve printed crates for making PSCI and other Arm SMCCC calls, and for managing page tables. These are only a begin; we plan to launch extra Rust crates to help bare-metal programming on a variety of platforms. These crates are additionally getting used outdoors of Android, equivalent to in Project Oak and the bare-metal section of our Comprehensive Rust course.
Coaching engineers
Many engineers have been positively shocked by how productive and nice Rust is to work with, offering good high-level options even in low-level environments. The engineers engaged on these tasks come from a variety of backgrounds. Our complete Rust course has helped skilled and novice programmers rapidly come up to the mark. Anecdotally the Rust sort system (together with the borrow checker and lifetimes) helps keep away from making errors which might be simply made in C or C++, equivalent to leaking tips that could stack-allocated values out of scope.
One in all our bare-metal Rust course attendees had this to say:
"varieties could be constructed that herald all of Rust's niceties and safeties and but nonetheless compile right down to extraordinarily environment friendly code like writes of constants to memory-mapped IO."
97% of attendees that accomplished a survey agreed the course was value their time.
Benefits and challenges
Gadget drivers are sometimes written in an object-oriented style for flexibility, even in C. Rust traits, which could be seen as a type of compile-time polymorphism, present a helpful high-level abstraction for this. In lots of circumstances this may be resolved completely at compile time, with no runtime overhead of dynamic dispatch by way of vtables or structs of perform pointers.
There have been some challenges. Secure Rust’s sort system is designed with an implicit assumption that the one reminiscence this system must care about is allotted by this system (be it on the stack, the heap, or statically), and solely utilized by this system. Naked-metal packages usually should cope with MMIO and shared reminiscence, which break this assumption. This tends to require numerous unsafe code and uncooked pointers, with restricted instruments for encapsulation. There’s some disagreement within the Rust group concerning the soundness of references to MMIO house, and the services for working with uncooked pointers in secure Rust are at the moment considerably restricted. The stabilisation of offset_of, slice_ptr_get, slice_ptr_len, offset_of and different nightly options will enhance this, however it’s nonetheless difficult to encapsulate cleanly. Higher syntax for accessing struct fields and array indices by way of uncooked pointers with out creating references would even be useful.
The concurrency launched by interrupt and exception handlers may also be awkward, as they usually have to entry shared mutable state however can’t depend on having the ability to take locks. Higher abstractions for vital sections will assist considerably, however there are some exceptions that may’t virtually be disabled, equivalent to web page faults used to implement copy-on-write or different on-demand web page mapping methods.
One other subject we’ve had is that some unsafe operations, equivalent to manipulating the web page desk, can’t be encapsulated cleanly as they’ve security implications for the entire program. Often in Rust we’re capable of encapsulate unsafe operations (operations which can trigger undefined behaviour in some circumstances, as a result of they’ve contracts which the compiler can’t verify) in protected wrappers the place we guarantee the required preconditions in order that it’s not attainable for any caller to trigger undefined behaviour. Nevertheless, mapping or unmapping pages in a single a part of this system could make different components of this system invalid, so we haven’t discovered a means to offer a completely normal protected interface to this. It ought to be famous that the identical considerations apply to a program written in C, the place the programmer at all times has to purpose concerning the security of the entire program.
Some individuals adopting Rust for bare-metal use circumstances have raised considerations about binary dimension. We have now seen this in some circumstances; for instance our Rust pVM firmware binary is round 460 kB in comparison with 220 kB for the sooner C model. Nevertheless, this isn’t a good comparability as we additionally added extra performance which allowed us to take away different parts from the boot chain, so the general dimension of all VM boot chain parts was comparable. We additionally weren’t notably optimizing for binary dimension on this case; velocity and correctness had been extra vital. In circumstances the place binary dimension is vital, compiling with size optimization, being cautious about dependencies, and avoiding Rust’s string formatting equipment in launch builds often permits comparable outcomes to C.
Architectural help is one other concern. Rust is mostly well supported on the Arm and RISC-V cores that we see most frequently, however help for extra esoteric architectures (for instance, the Qualcomm Hexagon DSP included in lots of Qualcomm SoCs utilized in Android telephones) could be missing in comparison with C.
The way forward for bare-metal Rust
General, regardless of these challenges and limitations, we’ve nonetheless discovered Rust to be a major enchancment over C (or C++), each when it comes to security and productiveness, in all of the bare-metal use circumstances the place we’ve tried it thus far. We plan to make use of it wherever sensible.
In addition to the work within the Android Virtualization Framework, the group engaged on Trusty (the open-source Trusted Execution Setting used on Pixel telephones, amongst others) have been onerous at work including help for Trusted Functions written in Rust. For instance, the reference KeyMint Trusted Application implementation is now in Rust. And there’s extra to come back in future Android gadgets, as we proceed to make use of Rust to enhance safety of the gadgets you belief.