3 [![Actions build status][actions-badge]][actions-url]
5 [actions-badge]: https://github.com/rust-lang/miri/workflows/CI/badge.svg?branch=master
6 [actions-url]: https://github.com/rust-lang/miri/actions
8 An experimental interpreter for [Rust][rust]'s
9 [mid-level intermediate representation][mir] (MIR). It can run binaries and
10 test suites of cargo projects and detect certain classes of
11 [undefined behavior](https://doc.rust-lang.org/reference/behavior-considered-undefined.html),
14 * Out-of-bounds memory accesses and use-after-free
15 * Invalid use of uninitialized data
16 * Violation of intrinsic preconditions (an [`unreachable_unchecked`] being
17 reached, calling [`copy_nonoverlapping`] with overlapping ranges, ...)
18 * Not sufficiently aligned memory accesses and references
19 * Violation of *some* basic type invariants (a `bool` that is not 0 or 1, for example,
20 or an invalid enum discriminant)
21 * **Experimental**: Violations of the [Stacked Borrows] rules governing aliasing
24 On top of that, Miri will also tell you about memory leaks: when there is memory
25 still allocated at the end of the execution, and that memory is not reachable
26 from a global `static`, Miri will raise an error.
28 Miri has already discovered some [real-world bugs](#bugs-found-by-miri). If you
29 found a bug with Miri, we'd appreciate if you tell us and we'll add it to the
32 However, be aware that Miri will **not catch all cases of undefined behavior**
33 in your program, and cannot run all programs:
35 * There are still plenty of open questions around the basic invariants for some
36 types and when these invariants even have to hold. Miri tries to avoid false
37 positives here, so if you program runs fine in Miri right now that is by no
38 means a guarantee that it is UB-free when these questions get answered.
40 In particular, Miri does currently not check that integers/floats are
41 initialized or that references point to valid data.
42 * If the program relies on unspecified details of how data is laid out, it will
43 still run fine in Miri -- but might break (including causing UB) on different
44 compiler versions or different platforms.
45 * Program execution is non-deterministic when it depends, for example, on where
46 exactly in memory allocations end up. Miri tests one of many possible
47 executions of your program. If your code is sensitive to allocation base
48 addresses or other non-deterministic data, try running Miri with different
49 values for `-Zmiri-seed` to test different executions.
50 * Miri runs the program as a platform-independent interpreter, so the program
51 has no access to most platform-specific APIs or FFI. A few APIs have been
52 implemented (such as printing to stdout) but most have not: for example, Miri
53 currently does not support SIMD or networking.
54 * Miri currently does not check for data-races and most other concurrency-related
57 [rust]: https://www.rust-lang.org/
58 [mir]: https://github.com/rust-lang/rfcs/blob/master/text/1211-mir.md
59 [`unreachable_unchecked`]: https://doc.rust-lang.org/stable/std/hint/fn.unreachable_unchecked.html
60 [`copy_nonoverlapping`]: https://doc.rust-lang.org/stable/std/ptr/fn.copy_nonoverlapping.html
61 [Stacked Borrows]: https://github.com/rust-lang/unsafe-code-guidelines/blob/master/wip/stacked-borrows.md
66 Install Miri on Rust nightly via `rustup`:
69 rustup +nightly component add miri
72 If `rustup` says the `miri` component is unavailable, that's because not all
73 nightly releases come with all tools. Check out
74 [this website](https://rust-lang.github.io/rustup-components-history) to
75 determine a nightly version that comes with Miri and install that using
76 `rustup toolchain install nightly-YYYY-MM-DD`.
78 Now you can run your project in Miri:
80 1. Run `cargo clean` to eliminate any cached dependencies. Miri needs your
81 dependencies to be compiled the right way, that would not happen if they have
82 previously already been compiled.
83 2. To run all tests in your project through Miri, use `cargo miri test`.
84 3. If you have a binary project, you can run it through Miri using `cargo miri run`.
86 The first time you run Miri, it will perform some extra setup and install some
87 dependencies. It will ask you for confirmation before installing anything.
89 `cargo miri run/test` supports the exact same flags as `cargo run/test`. You
90 can pass arguments to Miri via `MIRIFLAGS`. For example,
91 `MIRIFLAGS="-Zmiri-disable-stacked-borrows" cargo miri run` runs the program
92 without checking the aliasing of references.
94 When compiling code via `cargo miri`, the `cfg(miri)` config flag is set. You
95 can use this to ignore test cases that fail under Miri because they do things
96 Miri does not support:
100 #[cfg_attr(miri, ignore)]
101 fn does_not_work_on_miri() {
102 std::thread::spawn(|| println!("Hello Thread!"))
108 There is no way to list all the infinite things Miri cannot do, but the
109 interpreter will explicitly tell you when it finds something unsupported:
112 error: unsupported operation: can't call foreign function: bind
114 = help: this is likely not a bug in the program; it indicates that the program \
115 performed an operation that the interpreter does not support
118 ### Cross-interpretation: running for different targets
120 Miri can not only run a binary or test suite for your host target, it can also
121 perform cross-interpretation for arbitrary foreign targets: `cargo miri run
122 --target x86_64-unknown-linux-gnu` will run your program as if it was a Linux
123 program, no matter your host OS. This is particularly useful if you are using
124 Windows, as the Linux target is much better supported than Windows targets.
126 You can also use this to test platforms with different properties than your host
127 platform. For example `cargo miri test --target mips64-unknown-linux-gnuabi64`
128 will run your test suite on a big-endian target, which is useful for testing
129 endian-sensitive code.
131 ### Running Miri on CI
133 To run Miri on CI, make sure that you handle the case where the latest nightly
134 does not ship the Miri component because it currently does not build. For
135 example, you can use the following snippet to always test with the latest
136 nightly that *does* come with Miri:
139 MIRI_NIGHTLY=nightly-$(curl -s https://rust-lang.github.io/rustup-components-history/x86_64-unknown-linux-gnu/miri)
140 echo "Installing latest nightly with Miri: $MIRI_NIGHTLY"
141 rustup set profile minimal
142 rustup default "$MIRI_NIGHTLY"
143 rustup component add miri
150 When using the above instructions, you may encounter a number of confusing compiler
153 ### "note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace"
155 You may see this when trying to get Miri to display a backtrace. By default, Miri
156 doesn't expose any environment to the program, so running
157 `RUST_BACKTRACE=1 cargo miri test` will not do what you expect.
159 To get a backtrace, you need to disable isolation
160 [using `-Zmiri-disable-isolation`](#miri-flags):
163 RUST_BACKTRACE=1 MIRIFLAGS="-Zmiri-disable-isolation" cargo miri test
166 #### "found possibly newer version of crate `std` which `<dependency>` depends on"
168 Your build directory may contain artifacts from an earlier build that have/have
169 not been built for Miri. Run `cargo clean` before switching from non-Miri to
170 Miri builds and vice-versa.
172 #### "found crate `std` compiled by an incompatible version of rustc"
174 You may be running `cargo miri` with a different compiler version than the one
175 used to build the custom libstd that Miri uses, and Miri failed to detect that.
176 Try deleting `~/.cache/miri`.
178 #### "no mir for `std::rt::lang_start_internal`"
180 This means the sysroot you are using was not compiled with Miri in mind. This
181 should never happen when you use `cargo miri` because that takes care of setting
182 up the sysroot. If you are using `miri` (the Miri driver) directly, see the
183 [contributors' guide](CONTRIBUTING.md) for how to use `./miri` to best do that.
186 ## Miri `-Z` flags and environment variables
187 [miri-flags]: #miri--z-flags-and-environment-variables
189 Miri adds its own set of `-Z` flags, which are usually set via the `MIRIFLAGS`
190 environment variable:
192 * `-Zmiri-disable-alignment-check` disables checking pointer alignment, so you
193 can focus on other failures, but it means Miri can miss bugs in your program.
194 Using this flag is **unsound**.
195 * `-Zmiri-disable-stacked-borrows` disables checking the experimental
196 [Stacked Borrows] aliasing rules. This can make Miri run faster, but it also
197 means no aliasing violations will be detected. Using this flag is **unsound**
198 (but the affected soundness rules are experimental).
199 * `-Zmiri-disable-validation` disables enforcing validity invariants, which are
200 enforced by default. This is mostly useful to focus on other failures (such
201 as out-of-bounds accesses) first. Setting this flag means Miri can miss bugs
202 in your program. However, this can also help to make Miri run faster. Using
203 this flag is **unsound**.
204 * `-Zmiri-disable-isolation` disables host isolation. As a consequence,
205 the program has access to host resources such as environment variables, file
206 systems, and randomness.
207 * `-Zmiri-env-exclude=<var>` keeps the `var` environment variable isolated from
208 the host so that it cannot be accessed by the program. Can be used multiple
209 times to exclude several variables. On Windows, the `TERM` environment
210 variable is excluded by default.
211 * `-Zmiri-ignore-leaks` disables the memory leak checker.
212 * `-Zmiri-seed=<hex>` configures the seed of the RNG that Miri uses to resolve
213 non-determinism. This RNG is used to pick base addresses for allocations.
214 When isolation is enabled (the default), this is also used to emulate system
215 entropy. The default seed is 0. **NOTE**: This entropy is not good enough
216 for cryptographic use! Do not generate secret keys in Miri or perform other
217 kinds of cryptographic operations that rely on proper random numbers.
218 * `-Zmiri-symbolic-alignment-check` makes the alignment check more strict. By
219 default, alignment is checked by casting the pointer to an integer, and making
220 sure that is a multiple of the alignment. This can lead to cases where a
221 program passes the alignment check by pure chance, because things "happened to
222 be" sufficiently aligned -- there is no UB in this execution but there would
223 be UB in others. To avoid such cases, the symbolic alignment check only takes
224 into account the requested alignment of the relevant allocation, and the
225 offset into that allocation. This avoids missing such bugs, but it also
226 incurs some false positives when the code does manual integer arithmetic to
227 ensure alignment. (The standard library `align_to` method works fine in both
228 modes; under symbolic alignment it only fills the middle slice when the
229 allocation guarantees sufficient alignment.)
230 * `-Zmiri-track-alloc-id=<id>` shows a backtrace when the given allocation is
231 being allocated or freed. This helps in debugging memory leaks and
233 * `-Zmiri-track-pointer-tag=<tag>` shows a backtrace when the given pointer tag
234 is popped from a borrow stack (which is where the tag becomes invalid and any
235 future use of it will error). This helps you in finding out why UB is
236 happening and where in your code would be a good place to look for it.
237 * `-Zmiri-track-call-id=<id>` shows a backtrace when the given call id is
238 assigned to a stack frame. This helps in debugging UB related to Stacked
239 Borrows "protectors".
241 Some native rustc `-Z` flags are also very relevant for Miri:
243 * `-Zmir-opt-level` controls how many MIR optimizations are performed. Miri
244 overrides the default to be `0`; be advised that using any higher level can
245 make Miri miss bugs in your program because they got optimized away.
246 * `-Zalways-encode-mir` makes rustc dump MIR even for completely monomorphic
247 functions. This is needed so that Miri can execute such functions, so Miri
248 sets this flag per default.
249 * `-Zmir-emit-retag` controls whether `Retag` statements are emitted. Miri
250 enables this per default because it is needed for [Stacked Borrows].
252 Moreover, Miri recognizes some environment variables:
254 * `MIRI_LOG`, `MIRI_BACKTRACE` control logging and backtrace printing during
255 Miri executions, also [see above][testing-miri].
256 * `MIRIFLAGS` (recognized by `cargo miri` and the test suite) defines extra
257 flags to be passed to Miri.
258 * `MIRI_SYSROOT` (recognized by `cargo miri` and the test suite)
259 indicates the sysroot to use. To do the same thing with `miri`
260 directly, use the `--sysroot` flag.
261 * `MIRI_TEST_TARGET` (recognized by the test suite) indicates which target
262 architecture to test against. `miri` and `cargo miri` accept the `--target`
263 flag for the same purpose.
265 The following environment variables are internal, but used to communicate between
266 different Miri binaries, and as such worth documenting:
268 * `MIRI_BE_RUSTC` when set to any value tells the Miri driver to actually not
269 interpret the code but compile it like rustc would. This is useful to be sure
270 that the compiled `rlib`s are compatible with Miri.
271 When set while running `cargo-miri`, it indicates that we are part of a sysroot
272 build (for which some crates need special treatment).
273 * `MIRI_CWD` when set to any value tells the Miri driver to change to the given
274 directory after loading all the source files, but before commencing
275 interpretation. This is useful if the interpreted program wants a different
276 working directory at run-time than at build-time.
278 ## Miri `extern` functions
280 Miri provides some `extern` functions that programs can import to access
281 Miri-specific functionality:
286 /// Miri-provided extern function to mark the block `ptr` points to as a "root"
287 /// for some static memory. This memory and everything reachable by it is not
288 /// considered leaking even if it still exists when the program terminates.
290 /// `ptr` has to point to the beginning of an allocated block.
291 fn miri_static_root(ptr: *const u8);
293 /// Miri-provided extern function to obtain a backtrace of the current call stack.
294 /// This returns a boxed slice of pointers - each pointer is an opaque value
295 /// that is only useful when passed to `miri_resolve_frame`
296 /// The `flags` argument must be `0`.
297 fn miri_get_backtrace(flags: u64) -> Box<[*mut ()]>;
299 /// Miri-provided extern function to resolve a frame pointer obtained
300 /// from `miri_get_backtrace`. The `flags` argument must be `0`,
301 /// and `MiriFrame` should be declared as follows:
305 /// struct MiriFrame {
306 /// // The name of the function being executed, encoded in UTF-8
308 /// // The filename of the function being executed, encoded in UTF-8
309 /// filename: Box<[u8]>,
310 /// // The line number currently being executed in `filename`, starting from '1'.
312 /// // The column number currently being executed in `filename`, starting from '1'.
314 /// // The function pointer to the function currently being executed.
315 /// // This can be compared against function pointers obtained by
316 /// // casting a function (e.g. `my_fn as *mut ()`)
321 /// The fields must be declared in exactly the same order as they appear in `MiriFrame` above.
322 /// This function can be called on any thread (not just the one which obtained `frame`).
323 fn miri_resolve_frame(frame: *mut (), flags: u64) -> MiriFrame;
325 /// Miri-provided extern function to begin unwinding with the given payload.
327 /// This is internal and unstable and should not be used; we give it here
328 /// just to be complete.
329 fn miri_start_panic(payload: *mut u8) -> !;
333 ## Contributing and getting help
335 If you want to contribute to Miri, great! Please check out our
336 [contribution guide](CONTRIBUTING.md).
338 For help with running Miri, you can open an issue here on
339 GitHub or contact us (`oli-obk` and `RalfJ`) on the [Rust Zulip].
341 [Rust Zulip]: https://rust-lang.zulipchat.com
345 This project began as part of an undergraduate research course in 2015 by
346 @solson at the [University of Saskatchewan][usask]. There are [slides] and a
347 [report] available from that project. In 2016, @oli-obk joined to prepare miri
348 for eventually being used as const evaluator in the Rust compiler itself
349 (basically, for `const` and `static` stuff), replacing the old evaluator that
350 worked directly on the AST. In 2017, @RalfJung did an internship with Mozilla
351 and began developing miri towards a tool for detecting undefined behavior, and
352 also using miri as a way to explore the consequences of various possible
353 definitions for undefined behavior in Rust. @oli-obk's move of the miri engine
354 into the compiler finally came to completion in early 2018. Meanwhile, later
355 that year, @RalfJung did a second internship, developing miri further with
356 support for checking basic type invariants and verifying that references are
357 used according to their aliasing restrictions.
359 [usask]: https://www.usask.ca/
360 [slides]: https://solson.me/miri-slides.pdf
361 [report]: https://solson.me/miri-report.pdf
363 ## Bugs found by Miri
365 Miri has already found a number of bugs in the Rust standard library and beyond, which we collect here.
369 * [`Debug for vec_deque::Iter` accessing uninitialized memory](https://github.com/rust-lang/rust/issues/53566)
370 * [`Vec::into_iter` doing an unaligned ZST read](https://github.com/rust-lang/rust/pull/53804)
371 * [`From<&[T]> for Rc` creating a not sufficiently aligned reference](https://github.com/rust-lang/rust/issues/54908)
372 * [`BTreeMap` creating a shared reference pointing to a too small allocation](https://github.com/rust-lang/rust/issues/54957)
373 * [`Vec::append` creating a dangling reference](https://github.com/rust-lang/rust/pull/61082)
374 * [Futures turning a shared reference into a mutable one](https://github.com/rust-lang/rust/pull/56319)
375 * [`str` turning a shared reference into a mutable one](https://github.com/rust-lang/rust/pull/58200)
376 * [`rand` performing unaligned reads](https://github.com/rust-random/rand/issues/779)
377 * [The Unix allocator calling `posix_memalign` in an invalid way](https://github.com/rust-lang/rust/issues/62251)
378 * [`getrandom` calling the `getrandom` syscall in an invalid way](https://github.com/rust-random/getrandom/pull/73)
379 * [`Vec`](https://github.com/rust-lang/rust/issues/69770) and [`BTreeMap`](https://github.com/rust-lang/rust/issues/69769) leaking memory under some (panicky) conditions
380 * [`beef` leaking memory](https://github.com/maciejhirsz/beef/issues/12)
381 * [`EbrCell` using uninitialized memory incorrectly](https://github.com/Firstyear/concread/commit/b15be53b6ec076acb295a5c0483cdb4bf9be838f#diff-6282b2fc8e98bd089a1f0c86f648157cR229)
382 * [TiKV performing an unaligned pointer access](https://github.com/tikv/tikv/issues/7613)
383 * [`servo_arc` creating a dangling shared reference](https://github.com/servo/servo/issues/26357)
384 * [TiKV constructing out-of-bounds pointers (and overlapping mutable references)](https://github.com/tikv/tikv/pull/7751)
385 * [`encoding_rs` doing out-of-bounds pointer arithmetic](https://github.com/hsivonen/encoding_rs/pull/53)
386 * [TiKV using `Vec::from_raw_parts` incorrectly](https://github.com/tikv/agatedb/pull/24)
388 Violations of [Stacked Borrows] found that are likely bugs (but Stacked Borrows is currently just an experiment):
390 * [`VecDeque::drain` creating overlapping mutable references](https://github.com/rust-lang/rust/pull/56161)
391 * Various `BTreeMap` problems
392 * [`BTreeMap` iterators creating mutable references that overlap with shared references](https://github.com/rust-lang/rust/pull/58431)
393 * [`BTreeMap::iter_mut` creating overlapping mutable references](https://github.com/rust-lang/rust/issues/73915)
394 * [`BTreeMap` node insertion using raw pointers outside their valid memory area](https://github.com/rust-lang/rust/issues/78477)
395 * [`LinkedList` cursor insertion creating overlapping mutable references](https://github.com/rust-lang/rust/pull/60072)
396 * [`Vec::push` invalidating existing references into the vector](https://github.com/rust-lang/rust/issues/60847)
397 * [`align_to_mut` violating uniqueness of mutable references](https://github.com/rust-lang/rust/issues/68549)
398 * [`sized-chunks` creating aliasing mutable references](https://github.com/bodil/sized-chunks/issues/8)
399 * [`String::push_str` invalidating existing references into the string](https://github.com/rust-lang/rust/issues/70301)
400 * [`ryu` using raw pointers outside their valid memory area](https://github.com/dtolnay/ryu/issues/24)
401 * [ink! creating overlapping mutable references](https://github.com/rust-lang/miri/issues/1364)
402 * [TiKV creating overlapping mutable reference and raw pointer](https://github.com/tikv/tikv/pull/7709)
403 * [Windows `Env` iterator using a raw pointer outside its valid memory area](https://github.com/rust-lang/rust/pull/70479)
404 * [`VecDeque::iter_mut` creating overlapping mutable references](https://github.com/rust-lang/rust/issues/74029)
405 * [Standard library `SipHasher` using a raw pointer outside its valid memory area](https://github.com/rust-lang/rust/pull/78484)
409 Licensed under either of
411 * Apache License, Version 2.0 ([LICENSE-APACHE](LICENSE-APACHE) or
412 http://www.apache.org/licenses/LICENSE-2.0)
413 * MIT license ([LICENSE-MIT](LICENSE-MIT) or
414 http://opensource.org/licenses/MIT)
420 Unless you explicitly state otherwise, any contribution intentionally submitted
421 for inclusion in the work by you shall be dual licensed as above, without any
422 additional terms or conditions.