Before this PR, the compiler would emit the debuginfo name `slice$<T>`
for all kinds of slices, regardless of whether they are behind a
reference or not and regardless of the kind of reference. As a
consequence, the types `Foo<&[T]>`, `Foo<[T]>`, and `Foo<&mut [T]>`
would end up with the same type name `Foo<slice$<T> >` in debuginfo,
making it impossible to disambiguate between them by name. Similarly,
`&str` would get the name `str` in debuginfo, so the debuginfo name for
`Foo<str>` and `Foo<&str>` would be the same. In contrast,
`*const [bool]` and `*mut [bool]` would be `ptr_const$<slice$<bool> >`
and `ptr_mut$<slice$<bool> >`, i.e. the encoding does not lose
information about the type.
This PR removes all special handling for slices and `str`. The types
`&[bool]`, `&mut [bool]`, and `&str` thus get the names
`ref$<slice2$<bool> >`, `ref_mut$<slice2$<bool> >`, and
`ref$<str$>` respectively -- as one would expect.
`codegen_switchint_terminator` already uses `br` instead of `switch`
when there is one normal target plus the `otherwise` target. But there's
another common case with two normal targets and an `otherwise` target
that points to an empty unreachable BB. This comes up a lot when
switching on the tags of enums that use niches.
The pattern looks like this:
```
bb1: ; preds = %bb6
%3 = load i8, ptr %_2, align 1, !range !9, !noundef !4
%4 = sub i8 %3, 2
%5 = icmp eq i8 %4, 0
%_6 = select i1 %5, i64 0, i64 1
switch i64 %_6, label %bb3 [
i64 0, label %bb4
i64 1, label %bb2
]
bb3: ; preds = %bb1
unreachable
```
This commit adds code to convert the `switch` to a `br`:
```
bb1: ; preds = %bb6
%3 = load i8, ptr %_2, align 1, !range !9, !noundef !4
%4 = sub i8 %3, 2
%5 = icmp eq i8 %4, 0
%_6 = select i1 %5, i64 0, i64 1
%6 = icmp eq i64 %_6, 0
br i1 %6, label %bb4, label %bb2
bb3: ; No predecessors!
unreachable
```
This has a surprisingly large effect on compile times, with reductions
of 5% on debug builds of some crates. The reduction is all due to LLVM
taking less time. Maybe LLVM is just much better at handling `br` than
`switch`.
The resulting code is still suboptimal.
- The `icmp`, `select`, `icmp` sequence is silly, converting an `i1` to an `i64`
and back to an `i1`. But with the current code structure it's hard to avoid,
and LLVM will easily clean it up, in opt builds at least.
- `bb3` is usually now truly dead code (though not always, so it can't
be removed universally).
`TerminatorCodegenHelper` has three methods `llblock`, `llbb`, and
`lltarget`. They're all similar, but the names given no indication of
the differences.
This commit renames `lltarget` as `llbb_with_landing_pad`, and `llblock`
as `llbb_with_cleanup`. These aren't fantastic names, but at least it's
now clear that `llbb` is the lowest-level of the three and the other two
wrap it.
Ensure:
- builders always have a `bx` suffix;
- backend basic blocks always have an `llbb` suffix,
- paired builders and basic blocks have consistent prefixes.
Remove byte swap of valtree hash on big endian
This addresses problem reported in #103183. The code was originally introduced in e14b34c386. (see https://github.com/rust-lang/rust/pull/96591)
On big-endian environment, this operation sequence actually put the other half from 128-bit result, thus we got different hash result on LE and BE.
Get rid of native_library projection queries
They don't seem particularly useful as I don't expect native libraries to change frequently.
Maybe they do provide significant value of keeping incremental compilation green though, I'm not sure.
linker: Fix weak lang item linking with combination windows-gnu + LLD + LTO
In https://github.com/rust-lang/rust/pull/100404 this logic was originally disabled for MSVC due to issues with LTO, but the same issues appear on windows-gnu with LLD because that LLD uses the same underlying logic as MSVC LLD, just with re-syntaxed command line options.
So this PR just disables it for LTO builds in general.
The illumos linker does not support --strip-debug
When building and testing rust 1.64.0 on illumos, we saw a large number of failing tests associated with:
```
= note: ld: fatal: unrecognized option '--strip-debug'
ld: fatal: use the -z help option for usage information
collect2: error: ld returned 1 exit status
```
The illumos linker does not support the `--strip-debug` option (although it does support `--strip-all`).
Make `dyn*` casts into a coercion, allow `dyn*` upcasting
I know that `dyn*` is likely not going to be a feature exposed to surface Rust, but this makes it slightly more ergonomic to write tests for these types anyways. ... and this was just fun to implement anyways.
1. Make `dyn*` into a coercion instead of a cast
2. Enable `dyn*` upcasting since we basically get it for free
3. Simplify some of the cast checking code since we're using the coercion path now
r? `@eholk` but feel free to reassign
cc `@nikomatsakis` and `@tmandry` who might care about making `dyn*` casts into a coercion
Support casting boxes to dyn*
Boxes have a pointer type at codegen time which LLVM does not allow to be transparently converted to an integer. Work around this by inserting a `ptrtoint` instruction if the argument is a pointer.
r? ``@compiler-errors``
Fixes#102427
translation: eager translation
Part of #100717. See [Zulip thread](https://rust-lang.zulipchat.com/#narrow/stream/336883-i18n/topic/.23100717.20lists!/near/295010720) for additional context.
- **Store diagnostic arguments in a `HashMap`**: Eager translation will enable subdiagnostics to be translated multiple times with different arguments - this requires the ability to replace the value of one argument with a new value, which is better suited to a `HashMap` than the previous storage, a `Vec`.
- **Add `AddToDiagnostic::add_to_diagnostic_with`**: `AddToDiagnostic::add_to_diagnostic_with` is similar to the previous `AddToDiagnostic::add_to_diagnostic` but takes a function that can be used by the caller to modify diagnostic messages originating from the subdiagnostic (such as performing translation eagerly). `add_to_diagnostic` now just calls `add_to_diagnostic_with` with an empty closure.
- **Add `DiagnosticMessage::Eager`**: Add variant of `DiagnosticMessage` for eagerly translated messages
(messages in the target language which don't need translated by the emitter during emission). Also adds `eager_subdiagnostic` function which is intended to be invoked by the diagnostic derive for subdiagnostic fields which are marked as needing eager translation.
- **Support `#[subdiagnostic(eager)]`**: Add support for `eager` argument to the `subdiagnostic` attribute which generates a call to `eager_subdiagnostic`.
- **Finish migrating `rustc_query_system`**: Using eager translation, migrate the remaining repeated cycle stack diagnostic.
- **Split formatting initialization and use in diagnostic derives**: Diagnostic derives have previously had to take special care when ordering the generated code so that fields were not used after a move.
This is unlikely for most fields because a field is either annotated with a subdiagnostic attribute and is thus likely a `Span` and copiable, or is a argument, in which case it is only used once by `set_arg`
anyway.
However, format strings for code in suggestions can result in fields being used after being moved if not ordered carefully. As a result, the derive currently puts `set_arg` calls last (just before emission), such as:
let diag = { /* create diagnostic */ };
diag.span_suggestion_with_style(
span,
fluent::crate::slug,
format!("{}", __binding_0),
Applicability::Unknown,
SuggestionStyle::ShowAlways
);
/* + other subdiagnostic additions */
diag.set_arg("foo", __binding_0);
/* + other `set_arg` calls */
diag.emit();
For eager translation, this doesn't work, as the message being translated eagerly can assume that all arguments are available - so arguments _must_ be set first.
Format strings for suggestion code are now separated into two parts - an initialization line that performs the formatting into a variable, and a usage in the subdiagnostic addition.
By separating these parts, the initialization can happen before arguments are set, preserving the desired order so that code compiles, while still enabling arguments to be set before subdiagnostics are added.
let diag = { /* create diagnostic */ };
let __code_0 = format!("{}", __binding_0);
/* + other formatting */
diag.set_arg("foo", __binding_0);
/* + other `set_arg` calls */
diag.span_suggestion_with_style(
span,
fluent::crate::slug,
__code_0,
Applicability::Unknown,
SuggestionStyle::ShowAlways
);
/* + other subdiagnostic additions */
diag.emit();
- **Remove field ordering logic in diagnostic derive:** Following the approach taken in earlier commits to separate formatting initialization from use in the subdiagnostic derive, simplify the diagnostic derive by removing the field-ordering logic that previously solved this problem.
r? ```@compiler-errors```
Migrate `codegen_ssa` to diagnostics structs - [Part 1]
Initial migration of `codegen_ssa`. Going to split this crate migration in at least two PRs in order to avoid a huge PR and to quick off some questions around:
1. Translating messages from "external" crates.
2. Interfacing with OS messages.
3. Adding UI tests while migrating diagnostics.
_See comments below._
Eager translation will enable subdiagnostics to be translated multiple
times with different arguments - this requires the ability to replace
the value of one argument with a new value, which is better suited to a
`HashMap` than the previous storage, a `Vec`.
Signed-off-by: David Wood <david.wood@huawei.com>
Remove `mir::CastKind::Misc`
As discussed in #97649 `mir::CastKind::Misc` is not clear, this PR addresses that by creating a new enum variant for every valid cast.
r? ````@oli-obk````
- UPDATE - revert migration of logs
- UPDATE - use derive on LinkRlibError enum
- [Gardening] UPDATE - alphabetically sort fluent_messages
- UPDATE - use PathBuf and unify both AddNativeLibrary to use Display (which is what PathBuf uses when conforming to IntoDiagnosticArg)
- UPDATE - fluent messages sort after rebase
Remove `-Ztime`
Because it has a lot of overlap with `-Ztime-passes` but is generally less useful. Plus some related cleanups.
Best reviewed one commit at a time.
r? `@davidtwco`
The compiler currently has `-Ztime` and `-Ztime-passes`. I've used
`-Ztime-passes` for years but only recently learned about `-Ztime`.
What's the difference? Let's look at the `-Zhelp` output:
```
-Z time=val -- measure time of rustc processes (default: no)
-Z time-passes=val -- measure time of each rustc pass (default: no)
```
The `-Ztime-passes` description is clear, but the `-Ztime` one is less so.
Sounds like it measures the time for the entire process?
No. The real difference is that `-Ztime-passes` prints out info about passes,
and `-Ztime` does the same, but only for a subset of those passes. More
specifically, there is a distinction in the profiling code between a "verbose
generic activity" and an "extra verbose generic activity". `-Ztime-passes`
prints both kinds, while `-Ztime` only prints the first one. (It took me
a close reading of the source code to determine this difference.)
In practice this distinction has low value. Perhaps in the past the "extra
verbose" output was more voluminous, but now that we only print stats for a
pass if it exceeds 5ms or alters the RSS, `-Ztime-passes` is less spammy. Also,
a lot of the "extra verbose" cases are for individual lint passes, and you need
to also use `-Zno-interleave-lints` to see those anyway.
Therefore, this commit removes `-Ztime` and the associated machinery. One thing
to note is that the existing "extra verbose" activities all have an extra
string argument, so the commit adds the ability to accept an extra argument to
the "verbose" activities.
Only export `__tls_*` on wasm32-unknown-unknown.
From talking with `@abrown,` we aren't planning to have hosts call these `__tls_*` functions; instead, TLS initialization will be handled transparently within libc. Consequently, these functions don't need to be exported.
Leave them exported on wasm32-unknown-unknown though, as wasm-bindgen does call them.
From talking with @abrown, we aren't planning to have hosts call these
`__tls_*` functions; instead, TLS initialization will be handled
transparently within libc. Consequently, these functions don't need to
be exported.
Leave them exported on wasm32-unknown-unknown though, as wasm-bindgen
does call them.
Don't export `__wasm_init_memory` on WebAssembly.
Since #72889, the Rust wasm target doesn't use --passive-segments, so remove the `--export=__wasm_init_memory`.
As documented in the [tool-conventions Linking convention], `__wasm_init_memory` is not intended to be exported.
[tool-conventions Linking convention]: 7c064f3048/Linking.md (shared-memory-and-passive-segments)
Since #72889, the Rust wasm target doesn't use --passive-segments, so
remove the `--export=__wasm_init_memory`.
As documented in the [tool-conventions Linking convention],
`__wasm_init_memory` is not intended to be exported.
[tool-conventions Linking convention]: 7c064f3048/Linking.md (shared-memory-and-passive-segments)
`__heap_base` and `__data_end` are exported for use by wasm-bindgen, which
uses the wasm32-unknown-unknown target. On wasm32-wasi, as a step toward
implementing the Canonical ABI, and as an aid to building speicalized WASI
API polyfill wrappers, don't export `__heap_base` and `__data_end` on
wasm32-wasi.
Remove support for legacy PM
This removes support for optimizing with LLVM's legacy pass manager, as well as the unstable `-Znew-llvm-pass-manager` option. We have been defaulting to the new PM since LLVM 13 (except for s390x that waited for 14), and LLVM 15 removed support altogether. The only place we still use the legacy PM is for writing the output file, just like `llc` does.
cc #74705
r? ``@nikic``
This implements the `Linker::{link_whole_staticlib,link_whole_rlib}`
methods for the `WasmLd` linker used on wasm targets. Previously these
methods were noops since I think historically `wasm-ld` did not have
support for `--whole-archive` but nowadays it does, so the flags are
passed through.
Update rustc's information on Android's sanitizers
This patch updates sanitizer support definitions for Android inside the compiler. It also adjusts the logic to make sure no pre-built sanitizer runtime libraries are emitted as these are instead provided dynamically on Android targets.
This patch updates sanitizier support definitions for Android inside the
compiler. It also adjusts the logic to make sure no pre-built sanitizer
runtime libraries are emitted as these are instead provided dynamically
on Android targets.
On later stages, the feature is already stable.
Result of running:
rg -l "feature.let_else" compiler/ src/librustdoc/ library/ | xargs sed -s -i "s#\\[feature.let_else#\\[cfg_attr\\(bootstrap, feature\\(let_else\\)#"
Initial implementation of dyn*
This PR adds extremely basic and incomplete support for [dyn*](https://smallcultfollowing.com/babysteps//blog/2022/03/29/dyn-can-we-make-dyn-sized/). The goal is to get something in tree behind a flag to make collaboration easier, and also to make sure the implementation so far is not unreasonable. This PR does quite a few things:
* Introduce `dyn_star` feature flag
* Adds parsing for `dyn* Trait` types
* Defines `dyn* Trait` as a sized type
* Adds support for explicit casts, like `42usize as dyn* Debug`
* Including const evaluation of such casts
* Adds codegen for drop glue so things are cleaned up properly when a `dyn* Trait` object goes out of scope
* Adds codegen for method calls, at least for methods that take `&self`
Quite a bit is still missing, but this gives us a starting point. Note that this is never intended to become stable surface syntax for Rust, but rather `dyn*` is planned to be used as an implementation detail for async functions in dyn traits.
Joint work with `@nikomatsakis` and `@compiler-errors.`
r? `@bjorn3`
Add support for MIPS VZ ISA extension
[Link to relevant LLVM line where virt extension is specified](83fab8cee9/llvm/lib/Target/Mips/Mips.td (L172-L173))
This has been tested on mips-unknown-linux-musl with a target-cpu that is >= MIPS32 5 and `target-features=+virt`. The example was checked in a disassembler to ensure the correct assembly sequence was being generated using the virtualization instructions.
Needed additional work:
* MIPS is missing from [the Rust reference CPU feature lists](https://doc.rust-lang.org/reference/attributes/codegen.html#available-features)
Example docs for later:
```md
#### `mips` or `mips64`
This platform requires that `#[target_feature]` is only applied to [`unsafe`
functions][unsafe function]. This target's feature support is currently unstable
and must be enabled by `#![feature(mips_target_feature)]` ([Issue #44839])
[Issue #44839]: https://github.com/rust-lang/rust/issues/44839
Further documentation on these features can be found in the [MIPS Instruction Set
Reference Manual], or elsewhere on [mips.com].
[MIPS Instruction Set Reference Manual]: https://s3-eu-west-1.amazonaws.com/downloads-mips/documents/MD00086-2B-MIPS32BIS-AFP-6.06.pdf
[developer.arm.com]: https://www.mips.com/products/architectures/ase/
Feature | Implicitly Enables | Description
---------------|--------------------|-------------------
`fp64` | | 64-bit Floating Point
`msa` | | "MIPS SIMD Architecture"
`virt` | | Virtualization instructions (VZ ASE)
```
If the above is good I can also submit a PR for that if there's interest in documenting it while it's still unstable. Otherwise that can be dropped, I just wrote it before realizing it was possibly not a good idea.
Relevant to #44839
change rlib format to distinguish native dependencies
Another one method to solve problem mentioned in #99429.
Changed .rlib format, it contains all bundled native libraries as archieves.
At link time rlib is unpacked and native dependencies linked separately.
New behavior hidden under separate_native_rlib_dependencies flag.
ssa: implement `#[collapse_debuginfo]`
cc #39153rust-lang/compiler-team#386
Debuginfo line information for macro invocations are collapsed by default - line information are replaced by the line of the outermost expansion site. Using `-Zdebug-macros` disables this behaviour.
When the `collapse_debuginfo` feature is enabled, the default behaviour is reversed so that debuginfo is not collapsed by default. In addition, the `#[collapse_debuginfo]` attribute is available and can be applied to macro definitions which will then have their line information collapsed.
r? rust-lang/wg-debugging
