Only compute recursive callees once.
Inlining MIR in a cyclic call graph may create query cycles, which are ICEs. The current implementation `mir_callgraph_reachable(inlining_candidate, being_optimized)` checks if calling `inlining_candidate` may cycle back to `being_optimized` that we are currently inlining into.
This PR replaces this device with query `mir_callgraph_cyclic(being_optimized)` which searches the call graph for all cycles going back to `being_optimized`, and returns the set of functions involved in those cycles.
This is a tradeoff:
- in the current implementation, we perform more walks, but shallower;
- in this new implementation, we perform fewer walks, but exhaust the graph.
I'd have liked to compute this using some kind of SCC, but generic parameters make resolution path-dependent, so usual graph algorithms do not apply.
Insert checks for enum discriminants when debug assertions are enabled
Similar to the existing null-pointer and alignment checks, this checks for valid enum discriminants on creation of enums through unsafe transmutes. Essentially this sanitizes patterns like the following:
```rust
let val: MyEnum = unsafe { std::mem::transmute<u32, MyEnum>(42) };
```
An extension of this check will be done in a follow-up that explicitly sanitizes for extern enum values that come into Rust from e.g. C/C++.
This check is similar to Miri's capabilities of checking for valid construction of enum values.
This PR is inspired by saethlin@'s PR
https://github.com/rust-lang/rust/pull/104862. Thank you so much for keeping this code up and the detailed comments!
I also pair-programmed large parts of this together with vabr-g@.
r? `@saethlin`
Similar to the existing nullpointer and alignment checks, this checks
for valid enum discriminants on creation of enums through unsafe
transmutes. Essentially this sanitizes patterns like the following:
```rust
let val: MyEnum = unsafe { std::mem::transmute<u32, MyEnum>(42) };
```
An extension of this check will be done in a follow-up that explicitly
sanitizes for extern enum values that come into Rust from e.g. C/C++.
This check is similar to Miri's capabilities of checking for valid
construction of enum values.
This PR is inspired by saethlin@'s PR
https://github.com/rust-lang/rust/pull/104862. Thank you so much for
keeping this code up and the detailed comments!
I also pair-programmed large parts of this together with vabr-g@.
Add runtime check to avoid overwrite arg in `Diag`
## Origin PR description
At first, I set up a `debug_assert` check for the arg method to make sure that `args` in `Diag` aren't easily overwritten, and I added the `remove_arg()` method, so that if you do need to overwrite an arg, then you can explicitly call `remove_arg()` to remove it first, then call `arg()` to overwrite it.
For the code before the rust-lang/rust#142015 change, it won't compile because it will report an error
```
arg `instance`already exists.
```
This PR also modifies all diagnostics that fail the check to pass the check. There are two cases of check failure:
1. ~~Between *the parent diagnostic and the subdiagnostic*, or *between the subdiagnostics* have the same field between them. In this case, I renamed the conflicting fields.~~
2. ~~For subdiagnostics stored in `Vec`, the rendering may iteratively write the same arg over and over again. In this case, I changed the auto-generation with `derive(SubDiagnostic)` to manually implementing `SubDiagnostic` and manually rendered it with `eagerly_translate()`, similar to https://github.com/rust-lang/rust/issues/142031#issuecomment-2984812090, and after rendering it I manually deleted useless arg with the newly added `remove_arg` method.~~
## Final Decision
After trying and discussing, we made a final decision.
For `#[derive(Subdiagnostic)]`, This PR made two changes:
1. After the subdiagnostic is rendered, remove all args of this subdiagnostic, which allows for usage like `Vec<Subdiag>`.
2. Store `diag.args` before setting arguments, so that you can restore the contents of the main diagnostic after deleting the arguments after subdiagnostic is rendered, to avoid deleting the main diagnostic's arg when they have the same name args.
Only traverse reachable blocks in JumpThreading.
Fixes https://github.com/rust-lang/rust/issues/131451
We only compute loop headers for reachable blocks. We shouldn't try to perform an opt on unreachable blocks anyway.
Reason about borrowed classes in CopyProp.
Fixes https://github.com/rust-lang/rust/issues/141122
The current implementation of `CopyProp` avoids unifying two borrowed locals, as this would change the result of address comparison.
However, the implementation was inconsistent with the general algorithm, which identifies equivalence classes of locals and then replaces all locals by a single representative of their equivalence class.
This PR fixes it by forbidding the unification of two *classes* if any of those contain a borrowed local.
Add `-Z hint-mostly-unused` to tell rustc that most of a crate will go unused
This hint allows the compiler to optimize its operation based on this assumption, in order to compile faster. This is a hint, and does not guarantee any particular behavior.
This option can substantially speed up compilation if applied to a large dependency where the majority of the dependency does not get used. This flag may slow down compilation in other cases.
Currently, this option makes the compiler defer as much code generation as possible from functions in the crate, until later crates invoke those functions. Functions that never get invoked will never have code generated for them. For instance, if a crate provides thousands of functions, but only a few of them will get called, this flag will result in the compiler only doing code generation for the called functions. (This uses the same mechanisms as cross-crate inlining of functions.) This does not affect `extern` functions, or functions marked as `#[inline(never)]`.
This option has already existed in nightly as `-Zcross-crate-inline-threshold=always` for some time, and has gotten testing in that form. However, this option is still unstable, to give an opportunity for wider testing in this form.
Some performance numbers, based on a crate with many dependencies having just *one* large dependency set to `-Z hint-mostly-unused` (using Cargo's `profile-rustflags` option):
A release build went from 4m07s to 2m04s.
A non-release build went from 2m26s to 1m28s.
Manually invalidate caches in SimplifyCfg.
The current `SimplifyCfg` pass unconditionally invalidates CFG caches. This is unfortunate if there are no modifications that require this invalidation.
Unimplement unsized_locals
Implements https://github.com/rust-lang/compiler-team/issues/630
Tracking issue here: https://github.com/rust-lang/rust/issues/111942
Note that this just removes the feature, not the implementation, and does not touch `unsized_fn_params`. This is because it is required to support `Box<dyn FnOnce()>: FnOnce()`.
There may be more that should be removed (possibly in follow up prs)
- the `forget_unsized` function and `forget` intrinsic.
- the `unsized_locals` test directory; I've just fixed up the tests for now
- various codegen support for unsized values and allocas
cc ``@JakobDegen`` ``@oli-obk`` ``@Noratrieb`` ``@programmerjake`` ``@bjorn3``
``@rustbot`` label F-unsized_locals
Fixesrust-lang/rust#79409
mir-opt: Do not create storage marks in EarlyOtherwiseBranch
Fixes#141212.
The first commit add `StorageDead` by creating new indirect BB that makes CFG more complicated, but I think it's better to just not create storage marks.
r? mir-opt
This hint allows the compiler to optimize its operation based on this
assumption, in order to compile faster. This is a hint, and does not
guarantee any particular behavior.
This option can substantially speed up compilation if applied to a large
dependency where the majority of the dependency does not get used. This flag
may slow down compilation in other cases.
Currently, this option makes the compiler defer as much code generation as
possible from functions in the crate, until later crates invoke those
functions. Functions that never get invoked will never have code generated for
them. For instance, if a crate provides thousands of functions, but only a few
of them will get called, this flag will result in the compiler only doing code
generation for the called functions. (This uses the same mechanisms as
cross-crate inlining of functions.) This does not affect `extern` functions, or
functions marked as `#[inline(never)]`.
Some performance numbers, based on a crate with many dependencies having
just *one* large dependency set to `-Z hint-mostly-unused` (using
Cargo's `profile-rustflags` option):
A release build went from 4m07s to 2m04s.
A non-release build went from 2m26s to 1m28s.
Fix for async drop inside async gen fn
Return value (for yield) is corrected for async drop inside async gen function.
In CFG, when internal async drop future is polled and returned `Poll<()>::Pending`, then async gen resume function returns `Poll<(OptRet)>::Pending`.
Fixesrust-lang/rust#140530
Async drop - type instead of async drop fn, fixes#140484Fixes: rust-lang/rust#140484Fixes: rust-lang/rust#140500
Fixes ICE, when type is provided in AsyncDrop trait instead of `async fn drop()`.
Fixes ICE, when async drop fn has wrong signature.
`slice.get(i)` should use a slice projection in MIR, like `slice[i]` does
`slice[i]` is built-in magic, so ends up being quite different from `slice.get(i)` in MIR, even though they're both doing nearly identical operations -- checking the length of the slice then getting a ref/ptr to the element if it's in-bounds.
This PR adds a `slice_get_unchecked` intrinsic for `impl SliceIndex for usize` to use to fix that, so it no longer needs to do a bunch of lines of pointer math and instead just gets the obvious single statement. (This is *not* used for the range versions, since `slice[i..]` and `slice[..k]` can't use the mir Slice projection as they're using fenceposts, not indices.)
I originally tried to do this with some kind of GVN pattern, but realized that I'm pretty sure it's not legal to optimize `BinOp::Offset` to `PlaceElem::Index` without an extremely complicated condition. Basically, the problem is that the `Index` projection on a dereferenced slice pointer *cares about the metadata*, since it's UB to `PlaceElem::Index` outside the range described by the metadata. But then you cast the fat pointer to a thin pointer then offset it, that *ignores* the slice length metadata, so it's possible to write things that are legal with `Offset` but would be UB if translated in the obvious way to `Index`. Checking (or even determining) the necessary conditions for that would be complicated and error-prone, whereas this intrinsic-based approach is quite straight-forward.
Zero backend changes, because it just lowers to MIR, so it's already supported naturally by CTFE/Miri/cg_llvm/cg_clif.
Make two transmute-related MIR lints into HIR lint
Make `PTR_TO_INTEGER_TRANSMUTE_IN_CONSTS` (rust-lang/rust#130540) and `UNNECESSARY_TRANSMUTES` (rust-lang/rust#136083) into "normal" HIR-based lints.
Funny enough this came up in the review of the latter (https://github.com/rust-lang/rust/pull/136083#issuecomment-2614301413), but I guess it just was overlooked.
But anywyas, there's no reason for these to be MIR lints; in fact, it makes the suggestions for them a bit more complicated than necessary.
Note that there's probably a few more simplifications and improvements to be done here. Follow-ups can be done in a separate PR, especially if they're about the messaging and suggestions themselves, which I didn't write.
gvn: bail out unavoidable non-ssa locals in repeat
Fixes#141251.
We cannot transform `*elem` to `array[idx1]` in the following code, as `idx1` has already been modified.
```rust
mir! {
let array;
let elem;
{
array = [*val; 5];
elem = &array[idx1];
idx1 = idx2;
RET = *elem;
Return()
}
}
```
Perhaps I could transform it to `array[0]`, but I prefer the conservative approach.
r? mir-opt
