Fix: On wasm targets, call `panic_in_cleanup` if panic occurs in cleanup
Previously this was not correctly implemented. Each funclet may need its own terminate block, so this changes the `terminate_block` into a `terminate_blocks` `IndexVec` which can have a terminate_block for each funclet. We key on the first basic block of the funclet -- in particular, this is the start block for the old case of the top level terminate function.
I also fixed the `terminate` handler to not be invoked when a foreign exception is raised, mimicking the behavior from msvc. On wasm, in order to avoid generating a `catch_all` we need to call `llvm.wasm.get.exception` and `llvm.wasm.get.ehselector`.
Fix attribute parser and kind names.
For the attribute `FooBar` the parser is generally called `FooBarParser` and the kind is called `AttributeKind::FooBar`. This commit renames some cases that don't match that pattern. The most common cases:
- Adding `Rustc` to the front of the parser name for a `rustc_*` attribute.
- Adding `Parser` to the end of a parser name.
- Slight word variations, e.g. `Deprecation` instead of `Deprecated`, `Pointer` instead of `Ptr`, `Stability` instead of `Stable`.
r? @JonathanBrouwer
Implement debuginfo for unsafe binder types
Fixes: rust-lang/rust#139462
This treats an unsafe binder like a struct with a single field. This way we'd have the binder's distinct type name while keeping the wrapped value accessible.
Tracking:
- https://github.com/rust-lang/rust/issues/130516
For the attribute `FooBar` the parser is generally called `FooBarParser`
and the kind is called `AttributeKind::FooBar`. This commit renames some
cases that don't match that pattern. The most common cases:
- Adding `Rustc` to the front of the parser name for a `rustc_*`
attribute.
- Adding `Parser` to the end of a parser name.
- Slight word variations, e.g. `Deprecation` instead of `Deprecated`,
`Pointer` instead of `Ptr`, `Stability` instead of `Stable`.
Previously this was not correctly implemented. Each funclet may need its own terminate
block, so this changes the `terminate_block` into a `terminate_blocks` `IndexVec` which
can have a terminate_block for each funclet. We key on the first basic block of the
funclet -- in particular, this is the start block for the old case of the top level
terminate function.
Rather than using a catchswitch/catchpad pair, I used a cleanuppad. The reason for the
pair is to avoid catching foreign exceptions on MSVC. On wasm, it seems that the
catchswitch/catchpad pair is optimized back into a single cleanuppad and a catch_all
instruction is emitted which will catch foreign exceptions. Because the new logic is
only used on wasm, it seemed better to take the simpler approach seeing as they do the
same thing.
Simplify ThinLTO handling
This reduces the amount of complexity around ThinLTO module buffers. It removes duplication between `ModuleBuffer` and `ThinBuffer` (the latter was also used for fat LTO in some cases), clarifies when and where the ThinLTO summary is used (only for `--emit thin-link-bitcode`, ThinLTO performed by rustc rebuilds the summary every time). This also enables a follow up cleanup to reduce code duplication between red, green and imported codegen units.
Part of https://github.com/rust-lang/compiler-team/issues/908
Start migration for `LintDiagnostic` items by adding API and migrating `LinkerOutput` lint
This is more or less the same approach as https://github.com/rust-lang/rust/pull/152811, but in a much smaller size to make it reviewable. A lot of PRs will follow though. :)
This PR creates the equivalent of `lint_level` working with `Diagnostic` and add new methods on `MultiSpan` to make it work as well (in particular because we need to copy messages/spans from one context to another).
r? @JonathanBrouwer
Tighten the `!range` bounds on alignments in vtables
Right now we're only telling LLVM that they're non-zero, but alignments must be powers of two so can't be more than `isize::MAX+1`. And we actually never emit anything beyond LLVM's limit of 2²⁹, so outside of 16-bit targets the limit is that.
(Pulled out from rust-lang/rust#152867 which is starting to have too much in it.)
As far as I can tell it was introduced to allow fat LTO with
-Clinker-plugin-lto. Later a change was made to automatically disable
ThinLTO summary generation when -Clinker-plugin-lto -Clto=fat is used,
so we can safely remove it.
The next commit will bring back `enum DepKind` and there would be a
variant `DepKind::dep_kind`. This makes it impossible to have a variable
named `dep_kind`, because the `bindings_with_variant_name` lint is
overzealous. For this code:
```
let dep_kind = DepKind::dep_kind;
```
the lint will give this error:
```
pattern binding `dep_kind` is named the same as one of the variants of the type `DepKind`
```
This is arguably a bug in the lint. To work around it, this commit
renames the `dep_kind` query as `crate_dep_kind`. It is a better name
anyway given that `DepKind` and `CrateDepKind` are different things.
Do no add -no-pie on Windows
Windows binaries are always position independent and Clang warns when trying to enable or disable that:
```
❯ clang hello.c -pie
clang: warning: argument unused during compilation: '-pie' [-Wunused-command-line-argument]
❯ clang hello.c -no-pie
clang: warning: argument unused during compilation: '-no-pie' [-Wunused-command-line-argument]
```
https://github.com/rust-lang/rust/pull/149937 will turn these warnings into build errors:
```
❯ cargo rustc -- -D linker-messages
Compiling hello v0.1.0 (E:\tmp\hello)
error: linker stderr: x86_64-w64-mingw32-clang: argument unused during compilation: '-nolibc' [-Wunused-command-line-argument]␍
x86_64-w64-mingw32-clang: argument unused during compilation: '-no-pie' [-Wunused-command-line-argument]␍
|
= note: requested on the command line with `-D linker-messages`
error: could not compile `hello` (bin "hello") due to 1 previous error
```
Windows binaries are always position independent
and Clang warns when trying to enable or disable
that:
```
❯ clang hello.c -pie
clang: warning: argument unused during compilation: '-pie' [-Wunused-command-line-argument]
❯ clang hello.c -no-pie
clang: warning: argument unused during compilation: '-no-pie' [-Wunused-command-line-argument]
```
It was just a dummy implementation to workarround the fact that thin
local lto is the default in rustc. By adding a thin_lto_supported thin
local lto can be automatically disabled for cg_gcc, removing the need
for this dummy implementation. This makes improvements to the LTO
handling on the cg_ssa side a lot easier.
Set hidden visibility on naked functions in compiler-builtins
88b46460fa made builtin functions hidden, but it doesn't apply to naked functions, which are generated through a different code path.
This was discovered in rust-lang/rust#151486 where aarch64 outline atomics showed up in shared objects, overriding the symbols from compiler-rt.
Jonathan Brouwer