Remove the option to disable `llvm-version-check`
We don't support old versions of LLVM; there's no reason to have an easy way to force bootstrap to use them anyway. If someone really needs to use an unsupported version, they can modify bootstrap to change the version range.
r? ``@cuviper`` on whether we want to do this or not, since you maintain rust on Fedora and touched this config last.
We don't support old versions of LLVM; there's no reason to have an easy
way to force bootstrap to use them anyway. If someone really needs to
use an unsupported version, they can modify bootstrap to change the
version range.
Remove `llvm.skip-rebuild` option
This was added to in 2019 to speed up rebuild times when LLVM was modified. Now that download-ci-llvm exists, I don't think it makes sense to support an unsound option like this that can lead to miscompiles; and the code cleanup is nice too.
r? `@Mark-Simulacrum` cc `@varkor` #65612
This was added to in 2019 to speed up rebuild times when LLVM was
modified. Now that download-ci-llvm exists, I don't think it makes sense
to support an unsound option like this that can lead to miscompiles; and
the code cleanup is nice too.
Add sanitizer support for modern iOS platforms
asan and tsan generally support iOS, but that previously wasn't configured in rust. This only adds support for the simulator architectures, and arm64 device architecture, not the older 32 bit architectures.
asan and tsan generally support iOS, but that previously wasn't
configured in rust. This only adds support for the simulator
architectures, and arm64 device architecture, not the older 32 bit
architectures.
Include sanitizers supported by LLVM on s390x (asan, lsan, msan, tsan)
in the target definition, as well as in the compiletest supported list.
Build sanitizer runtime for the target. Enable sanitizers in the CI.
Handle non-existent upstream master branches in `x fmt`
People who do have a remote for `rust-lang/rust` but don't have the master branch checked out there used to get this error when running `x fmt`:
> fatal: ambiguous argument 'rust/master': unknown revision or path not in the working tree.
> Use '--' to separate paths from revisions, like this:
> 'git <command> [<revision>...] -- [<file>...]'
> rust/master
Which is not exactly helpful.
Now, we fall back to `origin/master` (hoping that at least that remote exists) for that case. If there is still some error, we just fall back to `x fmt .` and print a warning.
r? `@jyn514`
Historically, Rust's Fuchsia targets have been labeled x86_64-fuchsia
and aarch64-fuchsia. However, they should technically contain vendor
information. This CL changes Fuchsia's target triples to include the
"unknown" vendor since Clang now does normalization and handles all
triple spellings.
This was previously attempted in #90510, which was closed due to
inactivity.
Make LLD build forward-compatible with LLVM 16
Switch to using the cmake module instead of llvm-config. I believe this also removes the need for llvm-config-wrapper.
LLVM_CONFIG_PATH is no longer supported as of LLVM 16, switch to
using the cmake module instead.
We separately return the llvm-config and cmake directory paths,
because llvm-config always refers to the host binary, while
the cmake directory is for the target triple.
LLVM does this itself since 606cb8548a1b7763e0c8489c5efe66803a7ede72,
and 14 is no longer the correct standard when building lld 16,
causing build failures.
On stable, our artifacts are uploaded with the raw version number (e.g.,
1.65.0), not the channel. This adjusts our detection logic to use the
version number from src/version when we detect the stable channel.
Use BOLT in CI to optimize LLVM
This PR adds an optimization step in the Linux `dist` CI pipeline that uses [BOLT](https://github.com/llvm/llvm-project/tree/main/bolt) to optimize the `libLLVM.so` library built by boostrap.
Steps:
- [x] Use LLVM 15 as a bootstrap compiler and use it to build BOLT
- [x] Compile LLVM with support for relocations (`-DCMAKE_SHARED_LINKER_FLAGS="-Wl,-q"`)
- [x] Gather profile data using instrumented LLVM
- [x] Apply profile to LLVM that has already been PGOfied
- [x] Run with BOLT profiling on more benchmarks
- [x] Decide on the order of optimization (PGO -> BOLT?)
- [x] Decide how we should get `bolt` (currently we use the host `bolt`)
- [x] Clean up
The latest perf results can be found [here](https://github.com/rust-lang/rust/pull/94381#issuecomment-1258269440). The current CI build time with BOLT applied is around 1h 55 minutes.
fix: use git-commit-info for version information
Fixes#33286.
Fixes#86587.
This PR changes the current `git-commit-hash` file that `./x.py` dist puts in the `rustc-{version}-src.tar.{x,g}z` to contain the hash, the short hash, and the commit date from which the tarball was created, assuming git was available when it was. It uses this for reading the version so that rustc has all the appropriate metadata.
# Testing
Testing this is kind of a pain. I did it with something like
```sh
./x.py dist # ensure that `ignore-git` is `false` in config.toml
cp ./build/dist/rustc-1.65.0-dev-src.tar.gz ../rustc-1.65.0-dev-src.tar.gz
cd .. && tar -xzf rustc-1.65.0-dev-src && cd rustc-1.65.0-dev-src
./x.py build
```
Then, the output of `rustc -vV` with the stage1 compiler should have the `commit-hash` and `commit-date` fields filled, rather than be `unknown`. To be completely sure, you can use `rustc --sysroot` with the stdlib that the original `./x.py dist` made, which will require that the metadata matches.
Fix a bunch of typo
This PR will fix some typos detected by [typos].
I only picked the ones I was sure were spelling errors to fix, mostly in
the comments.
[typos]: https://github.com/crate-ci/typos
This PR will fix some typos detected by [typos].
I only picked the ones I was sure were spelling errors to fix, mostly in
the comments.
[typos]: https://github.com/crate-ci/typos
See comment added for details on the test builder restriction. This is primarily
intended for macOS CI, but is likely to be a slight win on other builders too.
This verifies if the HEAD sha matches with the detected LLVM SHA, and if not,
permits usage of the detected LLVM. Otherwise, we fallback on regular
non-downloaded LLVM (currently still cached with sccache, though that's still
10+ minutes on macOS).
When building LLVM/LLD as part of a build that asks LLVM to generate profiles, e.g. when
doing PGO, cmake or clang-cl don't automatically link clang's profiler runtime in,
causing undefined reference errors at link-time.
We do that manually, by adding clang's resource library folder to the library search path:
- for LLVM itself, by extending the linker args that `rustc_llvm`'s build script
uses, to avoid the linker errors when linking `rustc_driver`.
- for LLD, by extending cmake's linker flags during the LLD build step.
