| Age | Commit message (Collapse) | Author | Lines |
|---|
|
miri: treat non-memory local variables properly for data race detection
Fixes https://github.com/rust-lang/miri/issues/3242
Miri has an optimization where some local variables are not represented in memory until something forces them to be stored in memory (most notably, creating a pointer/reference to the local will do that). However, for a subsystem triggering on memory accesses -- such as the data race detector -- this means that the memory access seems to happen only when the local is moved to memory, instead of at the time that it actually happens. This can lead to UB reports in programs that do not actually have UB.
This PR fixes that by adding machine hooks for reads and writes to such efficiently represented local variables. The data race system tracks those very similar to how it would track reads and writes to addressable memory, and when a local is moved to memory, the clocks get overwritten with the information stored for the local.
|
|
|
|
make basic allocation functions track_caller in Miri for nicer backtraces
This matches what we did with basic pointer and atomic operations.
|
|
|
|
|
|
|
|
|
|
Ban non-array SIMD
Nearing the end of https://github.com/rust-lang/compiler-team/issues/621 !
Currently blocked on ~~https://github.com/rust-lang/compiler-builtins/pull/673~~ ~~https://github.com/rust-lang/compiler-builtins/pull/674~~ ~~https://github.com/rust-lang/rust/pull/129400~~ ~~https://github.com/rust-lang/rust/pull/129481~~ for windows.
|
|
|
|
|
|
|
|
also add an explicit test for the fact that a Option<WidePtr> has padding when it is None
|
|
|
|
|
|
core: use `compare_bytes` for more slice element types
`bool`, `NonZero<u8>`, `Option<NonZero<u8>>` and `ascii::Char` can be compared the same way as `u8`.
|
|
|
|
|
|
|
|
|
|
std: move allocators to `sys`
Part of #117276.
|
|
In preparation for #3837, the tree traversal needs to be made bottom-up,
because the current top-down tree traversal, coupled with that PR's
changes to the garbage collector, can introduce non-deterministic error
messages if the GC removes a parent tag of the accessed tag that would
have triggered the error first.
This is a breaking change for the diagnostics emitted by TB. The
implemented semantics stay the same.
|
|
|
|
|
|
miri weak memory emulation: put previous value into initial store buffer
Fixes https://github.com/rust-lang/miri/issues/2164 by doing a read before each atomic write so that we can initialize the store buffer. The read suppresses memory access hooks and UB exceptions, to avoid otherwise influencing the program behavior. If the read fails, we store that as `None` in the store buffer, so that when an atomic read races with the first atomic write to some memory and previously the memory was uninitialized, we can report UB due to reading uninit memory.
``@cbeuw`` this changes a bit the way we initialize the store buffers. Not sure if you still remember all this code, but if you could have a look to make sure this still makes sense, that would be great. :)
r? ``@saethlin``
|
|
panicking: improve hint for Miri's RUST_BACKTRACE behavior
Should help with https://github.com/rust-lang/miri/issues/3838
|
|
|
|
|
|
|
|
Make unused states of Reserved unrepresentable
In the [previous TB update](https://github.com/rust-lang/miri/pull/3742) we discovered that the existence of `Reserved + !ty_is_freeze + protected` is undesirable.
This has the side effect of making `Reserved { conflicted: true, ty_is_freeze: false }` unreachable.
As such it is desirable that this state would also be unrepresentable.
This PR eliminates the unused configuration by changing
```rs
enum PermissionPriv {
Reserved { ty_is_freeze: bool, conflicted: bool },
...
}
```
into
```rs
enum PermissionPriv {
ReservedFrz { conflicted: bool },
ReservedIM,
...
}
```
but this is not the only solution and `Reserved(Activable | Conflicted | InteriorMut)` could be discussed.
In addition to making the unreachable state not representable anymore, this change has the nice side effect of enabling `foreign_read` to no longer depend explicitly on the `protected` flag.
Currently waiting for
- `@JoJoDeveloping` to confirm that this is the same representation of `Reserved` as what is being implemented in simuliris,
- `@RalfJung` to approve that this does not introduce too much overhead in the trusted codebase.
|
|
|
|
pre-fill with previous value
|
|
non-power-of-two alignment check
|
|
remove some SSE/SSE2 intrinsics that are no longer used by stdarch
Fixes https://github.com/rust-lang/miri/issues/3691
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Add `miri_start` support
This PR uses a function with the exported symbol `miri_start` as a drop-in alternative to `#[start]`. So the signature stays the same as suggested in [this comment](https://github.com/rust-lang/miri/issues/3498#issuecomment-2088560526). <del>I’ve also removed Miri’s restriction to only work on bin crates as I don’t think this is necessary anymore.</del>
Closes #3758
|
|
|
|
raw_eq: using it on bytes with provenance is not UB (outside const-eval)
The current behavior of raw_eq violates provenance monotonicity. See https://github.com/rust-lang/rust/pull/124921 for an explanation of provenance monotonicity. It is violated in raw_eq because comparing bytes without provenance is well-defined, but adding provenance makes the operation UB.
So remove the no-provenance requirement from raw_eq. However, the requirement stays in-place for compile-time invocations of raw_eq, that indeed cannot deal with provenance.
Cc `@rust-lang/opsem`
|
|
|
|
|
|
|
|
|
|
Miri subtree update
r? `@ghost`
|
|
|
|
|
