| Age | Commit message (Collapse) | Author | Lines |
|---|
|
|
|
|
|
r=workingjubilee
Consistently use the highest bit of vector masks when converting to i1 vectors
This improves the codegen for vector `select`, `gather`, `scatter` and boolean reduction intrinsics and fixes rust-lang/portable-simd#316.
The current behavior of most mask operations during llvm codegen is to truncate the mask vector to <N x i1>, telling llvm to use the least significat bit. The exception is the `simd_bitmask` intrinsics, which already used the most signifiant bit.
Since sse/avx instructions are defined to use the most significant bit, truncating means that llvm has to insert a left shift to move the bit into the most significant position, before the mask can actually be used.
Similarly on aarch64, mask operations like blend work bit by bit, repeating the least significant bit across the whole lane involves shifting it into the sign position and then comparing against zero.
By shifting before truncating to <N x i1>, we tell llvm that we only consider the most significant bit, removing the need for additional shift instructions in the assembly.
|
|
Add cache to `FoldEscapingRegions`
Fixes #135780
ty `@lqd` for the tests
|
|
|
|
This improves the codegen for vector `select`, `gather`, `scatter` and
boolean reduction intrinsics and fixes rust-lang/portable-simd#316.
The current behavior of most mask operations during llvm codegen is to
truncate the mask vector to <N x i1>, telling llvm to use the least
significat bit. The exception is the `simd_bitmask` intrinsics, which
already used the most signifiant bit.
Since sse/avx instructions are defined to use the most significant bit,
truncating means that llvm has to insert a left shift to move the bit
into the most significant position, before the mask can actually be
used.
Similarly on aarch64, mask operations like blend work bit by bit,
repeating the least significant bit across the whole lane involves
shifting it into the sign position and then comparing against zero.
By shifting before truncating to <N x i1>, we tell llvm that we only
consider the most significant bit, removing the need for additional
shift instructions in the assembly.
|
|
Fix tests on LLVM 20
For sparcv8plus.rs, duplicate the test for LLVM 19 and LLVM 20. LLVM 20 resolves one of the FIXME in the test.
For x86_64-bigint-add.rs split the check lines for LLVM 19 and LLVM 20. The difference in codegen here is due to a difference in unroll factor, which I believe is not what the test is interested in.
Fixes https://github.com/rust-lang/rust/issues/132957.
Fixes https://github.com/rust-lang/rust/issues/133754.
|
|
Rollup of 7 pull requests
Successful merges:
- #133631 (Support QNX 7.1 with `io-sock`+libstd and QNX 8.0 (`no_std` only))
- #134358 (compiler: Set `target_abi = "ilp32e"` on all riscv32e targets)
- #135812 (Fix GDB `OsString` provider on Windows )
- #135842 (TRPL: more backward-compatible Edition changes)
- #135946 (Remove extra whitespace from rustdoc breadcrumbs for copypasting)
- #135953 (ci.py: check the return code in `run-local`)
- #136019 (Add an `unchecked_div` alias to the `Div<NonZero<_>>` impls)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Get rid of `mir::Const::from_ty_const`
This function is strange, because it turns valtrees into `mir::Const::Value`, but the rest of the const variants stay as type system consts.
All of the callsites except for one in `instsimplify` (array length simplification of `ptr_metadata` call) just go through the valtree arm of the function, so it's easier to just create a `mir::Const` directly for those.
For the instsimplify case, if we have a type system const we should *keep* having a type system const, rather than turning it into a `mir::Const::Value`; it doesn't really matter in practice, though, bc `usize` has no padding, but it feels more principled.
|
|
Add an `unchecked_div` alias to the `Div<NonZero<_>>` impls
Inspired by https://github.com/rust-lang/libs-team/issues/526, if people are looking for `unchecked_div`, point them to `u32: Div<NonZero<u32>>` and friends which do no runtime checks -- and are safe! -- rather than today's behaviour of [the intrinsic being the top result](https://doc.rust-lang.org/std/?search=unchecked_div).
|
|
ci.py: check the return code in `run-local`
If the run fails, it should report that and return a non-zero exit
status. The simplest way to do that is with `run(..., check=True)`,
which raises a `CalledProcessError`.
|
|
Remove extra whitespace from rustdoc breadcrumbs for copypasting
The docs header used to display [item names with their full path](https://doc.rust-lang.org/1.82.0/std/os/unix/ffi/trait.OsStrExt.html), but a [recent design change](https://doc.rust-lang.org/1.83.0/std/os/unix/ffi/trait.OsStrExt.html) has split the path and added extra styling to it.
The problem is the new styling affects how this text is copied to clipboard. I used to copy and paste the paths into `use` statements in the code, but the new styling has extra formatting and whitespace that makes copied text unusable in Rust source code.
Instead of:
> std::os::unix::ffi::OsStrExt
I now get:
> std
> ::
> os
> ::
> unix
> ::
> ffi
> Trait OsStrExt
This change removes extra whitespace from the markup, and removes `display: flex`. Paths (now in small text) are unlikely to be that long to wrap, and even then regular text wrapping should be sufficient.
|
|
TRPL: more backward-compatible Edition changes
- Improve the discussion of `unsafe` blocks within `unsafe` functions.
- Fix formatting in Appendix A
- Incorporate line edits to Chapter 17 from NoStarch.
|
|
Fix GDB `OsString` provider on Windows
It would throw an exception due to trying to look up `Wtf8Buf.__0`. The field it actually wants is called [`bytes`](https://github.com/rust-lang/rust/blob/b605c65b6eb5fa71783f8e26df69975f9f1680ee/library/std/src/sys_common/wtf8.rs#L134).
|
|
compiler: Set `target_abi = "ilp32e"` on all riscv32e targets
This allows compile-time configuration based on this. In the near future we should do this across all RISCV targets, probably, but this cfg is essential for building software usable on these targets, and they are tier 3 so it seems less of a concern to tweak their definition thusly.
|
|
Support QNX 7.1 with `io-sock`+libstd and QNX 8.0 (`no_std` only)
Changes of this pull request:
1. Refactor code for qnx nto targets to share more code in file `nto_qnx.rs`
1. Add support for an additional network stack on nto qnx 7.1.
QNX 7.1 supports two network stacks:
1. `io-pkt`, which is default
2. `io-sock`, which is optional on 7.1 but default in QNX 8.0
As one can see in the [io-sock migration notes](https://www.qnx.com/developers/docs/7.1/index.html#com.qnx.doc.neutrino.io_sock/topic/migrate_app.html), this changes the libc API in a way similar to e.g. linux-gnu vs. linux-musl.
This change adds a new target which has a different value for `target_env`, so that e.g. libc can distinguish between both APIs.
2. Add initial support for QNX 8.0, thanks to AkhilTThomas. As it turned out, the problem with forking many processes still exists in QNX 8.0. Because if this, we are now using it for any QNX version (i.e. not check for `target_env` anymore).
|
|
Rollup of 7 pull requests
Successful merges:
- #133951 (Make the wasm_c_abi future compat warning a hard error)
- #134283 (fix(libtest): Deprecate '--logfile')
- #135785 (use `PassMode::Direct` for vector types on `s390x`)
- #135948 (Update emscripten std tests)
- #135951 (Use `fmt::from_fn` in more places in the compiler)
- #136031 (Expand polonius MIR dump)
- #136032 (Account for mutable borrow in argument suggestion)
Failed merges:
- #135635 (Move `std::io::pipe` code into its own file)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Account for mutable borrow in argument suggestion
```
error: value assigned to `object` is never read
--> $DIR/mut-arg-of-borrowed-type-meant-to-be-arg-of-mut-borrow.rs:21:5
|
LL | object = &mut object2;
| ^^^^^^
|
help: you might have meant to mutate the pointed at value being passed in, instead of changing the reference in the local binding
|
LL ~ fn change_object3(object: &mut Object) {
LL |
LL | let object2 = Object;
LL ~ *object = object2;
|
```
instead of
```
error: value assigned to `object` is never read
--> $DIR/mut-arg-of-borrowed-type-meant-to-be-arg-of-mut-borrow.rs:21:5
|
LL | object = &mut object2;
| ^^^^^^
|
help: you might have meant to mutate the pointed at value being passed in, instead of changing the reference in the local binding
|
LL ~ fn change_object3(object: &mut mut Object) {
LL |
LL | let object2 = Object;
LL ~ *object = object2;
|
```
Fix #136028.
|
|
Expand polonius MIR dump
This PR starts expanding the polonius MIR:
- switches to an HTML file, to show graphs in the same document as the MIR dump, share them more easily since it's a single file that can be hosted as a gist, and also to allow for interactivity in the near future.
- adds the regular NLL MIR + polonius constraints
- embeds a mermaid version of the CFG, similar to the graphviz one, but that needs a smaller js than `dot`'s emscripten js from graphvizonline
[Here's an example](https://gistpreview.github.io/?0c18f2a59b5e24ac0f96447aa34ffe00) of how it looks.
---
In future PRs: mermaid graphs of the NLL region graph, of the NLL SCCs, of the polonius localized outlives constraints, and the interactive polonius MIR dump.
r? ```@matthewjasper```
|
|
Use `fmt::from_fn` in more places in the compiler
Use the unstable functions from #117729 in more places in the compiler, follow up to #135494
|
|
Update emscripten std tests
This disables a bunch of emscripten tests that test things emscripten doesn't support and re-enables a whole bunch of tests which now work just fine on emscripten.
Tested with `EMCC_CFLAGS="-s MAXIMUM_MEMORY=2GB" ./x.py test library/ --target wasm32-unknown-emscripten`.
|
|
use `PassMode::Direct` for vector types on `s390x`
closes https://github.com/rust-lang/rust/issues/135744
tracking issue: https://github.com/rust-lang/rust/issues/130869
Previously, all vector types were type erased to `Ni8`, now we pass non-wrapped vector types directly. That skips emitting a bunch of casting logic in rustc, that LLVM then has to clean up. The initial LLVM IR is also a bit more readable.
This calling convention is tested extensively in `tests/assembly/s390x-vector-abi.rs`, showing that this change has no impact on the ABI in practice.
r? ````@taiki-e````
|
|
fix(libtest): Deprecate '--logfile'
rust-lang/testing-devex-team#9 proposed changing the behavior of `--logfile`. The given reasons were:
(1) Bazel can't programmatically process stdout. This seems like a limitation in Bazel and we recommend focusing on that. If we look at the wider Rust ecosystem, Rustc and Cargo don't support any such mechanism and the Cargo team rejected having one. Expecting this in libtest when its not supported elsewhere seems too specialized.
(2) Tests that leak out non-programmatic output that intermixes with programmatic output. We acknowledge this is a problem to be evaluated but we need to make sure we are stepping back and gathering requirements, rather than assuming `--logfile` will fit the needs.
Independent of the motive, regarding using or changing `--logfile`
(1) Most ways to do it would be a breaking change, like if we respect any stable `--format`. As suggested above, we could specialize this to new `--format` values but that would be confusing for some values to apply but not others.
(2) Other ways of solving this add new features to lib`test` when we are instead wanting to limit the feature set it has to minimize the compatibility surface that has to be maintained and the burden it would put on third party harnesses which are a focus area. Examples include `--format compact` or a `--log-format` flag
(3) The existence of `--logfile` dates back quite a ways (https://github.com/rust-lang/rust/commit/5cc050b265509c19717e11e12dd785d8c73f5b11, rust-lang/rust#2127) and the history gives the
impression this more of slipped through rather than being an intended feature (see also
https://github.com/rust-lang/rust/pull/82350#discussion_r579732071). Deprecation would better match to how it has been treated. By deprecating this, we do not expect custom test harnesses (rust-lang/testing-devex-team#2) to implement this.
T-testing-devex held an FCP for deprecating in rust-lang/testing-devex-team#9 though according to
[RFC #3455](https://rust-lang.github.io/rfcs/3455-t-test.html), this is still subject to final approval from T-libs-api.
Closes rust-lang/testing-devex-team#9
|
|
Make the wasm_c_abi future compat warning a hard error
This is the next step in getting rid of the broken C abi for wasm32-unknown-unknown.
The lint was made deny-by-default in https://github.com/rust-lang/rust/pull/129534 3 months ago. This still keeps the `-Zwasm-c-abi` flag set to `legacy` by default. It will be flipped in a future PR.
cc https://github.com/rust-lang/rust/issues/122532
|
|
Rollup of 7 pull requests
Successful merges:
- #134300 (remove long-deprecated no-op attributes no_start and crate_id)
- #134373 (Improve and expand documentation of pipes)
- #135934 (Include missing item in the 1.81 release notes)
- #136005 (ports last few library files to new intrinsic style)
- #136016 (Improve check-cfg expected names diagnostic)
- #136039 (docs: fix typo in std::pin overview)
- #136056 (Fix typo in const stability error message)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Fix typo in const stability error message
|
|
docs: fix typo in std::pin overview
Unimportant fix in `std::pin` documentation.
|
|
Improve check-cfg expected names diagnostic
This PR improves the check-cfg `allow-same-level` test by ~~normalizing it's output and by~~ adding more context to the test.
It also filters the well known cfgs from the `expected names are` note, as to reduce the size of the diagnostic. Users can still find the full list on the [rustc book](https://doc.rust-lang.org/nightly/rustc/check-cfg.html#well-known-names-and-values), which is reinforced for Cargo users by adding a note in the Cargo check-cfg specific section.
Fixes https://github.com/rust-lang/rust/issues/135995
r? `@jieyouxu`
|
|
ports last few library files to new intrinsic style
This pr ports the last 2 library files to new intrinsic style this pr is part of issue #132735
|
|
Include missing item in the 1.81 release notes
It was pointed out to me that when I prepared the CVE-2024-43402 fix in the stable branch, I added the release notes in the stable PR (https://github.com/rust-lang/rust/pull/129960), but I forgot to do so in the beta or nightly PR. Because of that, the relnotes line only appeared in 1.81, and disappeared afterwards.
|
|
Improve and expand documentation of pipes
- Reference UNIX, not just Linux
- Simplify some of the language
- Don't imply that pipes *only* work across multiple processes; instead,
*suggest* that they're typically used across two or more separate
processes.
- Specify that portable applications cannot use multiple readers or
multiple writers for messages larger than a byte, due to potential
interleaving.
Tracking issue for anonymous pipes:
https://github.com/rust-lang/rust/issues/127154
|
|
remove long-deprecated no-op attributes no_start and crate_id
These have emitted a deprecation warning since forever (https://github.com/rust-lang/rust/pull/64471) and they already don't do anything. In fact they [apparently](https://github.com/rust-lang/rust/pull/64471#issuecomment-531517332) have done nothing since pre-1.0, so... do we even need a crater run? Doesn't seem worth it.
|
|
Shorten linker output even more when `--verbose` is not present
- Don't show environment variables. Seeing PATH is almost never useful, and it can be extremely long.
- For .rlibs in the sysroot, replace crate hashes with a `"-*"` string. This will expand to the full crate name when pasted into the shell.
- Move `.rlib` to outside the glob.
- Abbreviate the sysroot path to `<sysroot>` wherever it appears in the arguments.
This also adds an example of the linker output as a run-make test. Currently it only runs on x86_64-unknown-linux-gnu, because each platform has its own linker arguments. So that it's stable across machines, pass BUILD_ROOT as an argument through compiletest through to run-make tests.
r? `@bjorn3`
try-job: aarch64-apple
|
|
- Don't show environment variables. Seeing PATH is almost never useful, and it can be extremely long.
- For .rlibs in the sysroot, replace crate hashes with a `"-*"` string. This will expand to the full crate name when pasted into the shell.
- Move `.rlib` to outside the glob.
- Abbreviate the sysroot path to `<sysroot>` wherever it appears in the arguments.
This also adds an example of the linker output as a run-make test. Currently it only runs on x86_64-unknown-linux-gnu, because each platform has its own linker arguments. So that it's stable across machines, pass BUILD_ROOT as an argument through compiletest through to run-make tests.
- Only use linker-flavor=gnu-cc if we're actually going to compare the output. It doesn't exist on MacOS.
|
|
- Simplify some of the language
- Minor grammar fixes
- Don't imply that pipes *only* work across multiple processes; instead,
*suggest* that they're typically used across two or more separate
processes.
- Specify that portable applications cannot use multiple readers or
multiple writers for messages larger than a byte, due to potential
interleaving.
- Remove no-longer-referenced footnote URLs.
|
|
show linker output even if the linker succeeds
Show stderr and stderr by default, controlled by a new `linker_messages` lint.
fixes https://github.com/rust-lang/rust/issues/83436. fixes https://github.com/rust-lang/rust/issues/38206. cc https://rust-lang.zulipchat.com/#narrow/stream/233931-t-compiler.2Fmajor-changes/topic/uplift.20some.20-Zverbose.20calls.20and.20rename.20to.E2.80.A6.20compiler-team.23706/near/408986134
<!-- try-job: dist-x86_64-msvc -->
try-job: aarch64-apple
r? `@bjorn3`
|
|
|
|
Reword resolve errors caused by likely missing crate in dep tree
Reword label and add `help`:
```
error[E0432]: unresolved import `some_novel_crate`
--> f704.rs:1:5
|
1 | use some_novel_crate::Type;
| ^^^^^^^^^^^^^^^^ use of unresolved module or unlinked crate `some_novel_crate`
|
= help: if you wanted to use a crate named `some_novel_crate`, use `cargo add some_novel_crate` to add it to your `Cargo.toml`
```
Fix #133137.
|
|
|
|
|
|
Rollup of 9 pull requests
Successful merges:
- #135971 (Properly report error when object type param default references self)
- #135977 (Fix `FormattingOptions` instantiation with `Default`)
- #135985 (Rename test to `unresolvable-upvar-issue-87987.rs` and add some notes)
- #135991 (Fix set_name in thread mod for NuttX)
- #136009 (bootstrap: Handle bootstrap lockfile race condition better)
- #136018 (Use short ty string for move errors)
- #136027 (Skip suggestions in `derive`d code)
- #136029 (Bootstrap: Don't move ownership of job object)
- #136034 (fix(bootstrap): deserialize null as `f64::NAN`)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
|
|
|
|
escape the regular raw MIR into its own section
|
|
|
|
for cases where we want to dump the MIR to a given writer instead of a
new file as the default does.
this will be used when dumping the MIR to a buffer to process
differently, e.g. post-process to escape for an HTML dump.
|
|
fix(bootstrap): deserialize null as `f64::NAN`
When doing optimized build through opt-dist,
I've often run into errors like `invalid type: null, expected f64`.
This is likely because some f64 fields might actually bet set null.
Unfortunately, serde_json doesn't handle null <-> NaN well.
This commit addresses it by having a custom deserialize method, so null is always be deserialized to `f64:NAN`.
See:
* https://rust-lang.zulipchat.com/#narrow/channel/242791-t-infra/topic/opt-dist.3A.20.60invalid.20type.3A.20null.2C.20expect.20f64.60.20failure
* https://github.com/serde-rs/json/issues/202
r? `@Kobzol`
|
|
Bootstrap: Don't move ownership of job object
I've been thinking about this since the last time I looked at bootstrap's use of job objects. We currently pass ownership of the job object to Python. I feel this is unneeded complexity.
The rationale given (in a comment) is that it helps with `ctrl-c` on `x.py`. But using `ctrl-c` when running `x.py` will also cause `bootstrap.exe` to immediately exit so I don't find that convincing.
|
|
Skip suggestions in `derive`d code
Do not suggest
```
help: use parentheses to call these
|
5 | (callback: Rc<dyn Fn()>)(),
| + +++
```
Skip all "call function for this binop" suggestions when in a derive context.
Fix #135989.
|
|
Use short ty string for move errors
```
error[E0382]: use of moved value: `x`
--> bay.rs:14:14
|
12 | fn foo(x: D) {
| - move occurs because `x` has type `(((..., ..., ..., ...), ..., ..., ...), ..., ..., ...)`, which does not implement the `Copy` trait
13 | let _a = x;
| - value moved here
14 | let _b = x; //~ ERROR use of moved value
| ^ value used here after move
|
= note: the full type name has been written to 'bay.long-type-14349227078439097973.txt'
= note: consider using `--verbose` to print the full type name to the console
help: consider cloning the value if the performance cost is acceptable
|
13 | let _a = x.clone();
| ++++++++
```
Address 4th case in #135919.
|
