Merge `compiler-builtins` as a Josh subtree #141229

tgross35 · 2025-05-18T16:07:36Z

Use the Josh 1 utility to add compiler-builtins as a subtree, which
will allow us to stop using crates.io for updates. This is intended to
help resolve some problems when unstable features change and require
code changes in compiler-builtins, which sometimes gets trapped in a
bootstrap cycle.

This was done using josh-filter built from the r24.10.04 tag:

git fetch https://github.com/rust-lang/compiler-builtins.git 233434412fe7eced8f1ddbfeddabef1d55e493bd
josh-filter ":prefix=library/compiler-builtins" FETCH_HEAD
git merge --allow-unrelated FILTERED_HEAD

The HEAD in the compiler-builtins repository is 233434412f ("fix an if
statement that can be collapsed").

Add `fminf16`, `fmaxf16`, `fminf128`, and `fmaxf128`

This can replace `fmod` and `fmodf`. As part of this change I was able to replace some of the `while` loops with `leading_zeros`.

Add `fmodf16`

With the new routines, some of our tests are running close to their timeouts. Increase the timeout for test jobs, and set a short timeout for all other jobs that did not have one.

Certain functions (`fmodf128`) are significantly slower than others, to the point that running the default number of tests adds tens of minutes to PR CI and extensive test time increases to ~1day. It does not make sense to do this by default; so, introduce `EXTREMELY_SLOW_TESTS` to test configuration that allows setting specific tests that need to have a reduced iteration count.

This function is significantly slower than all others so includes an override in `EXTREMELY_SLOW_TESTS`. Without it, PR CI takes ~1hour and the extensive tests in CI take ~1day.

Add `fmodf128`

A few new functions were added but this list did not get updated. Do so here.

In particular, this includes updates to Rug that we can make use of [1], [2], [3], [4]. [1]: https://gitlab.com/tspiteri/rug/-/issues/78 [2]: https://gitlab.com/tspiteri/rug/-/issues/80 [3]: https://gitlab.com/tspiteri/rug/-/issues/76 [4]: https://gitlab.com/tspiteri/rug/-/issues/73

Since Rug 1.27.0, `az` is reexported. This means we no longer need to track it as a separate dependency.

Rug 1.27.0 exposes `frexp`. Make use of it for our tests.

Rug 1.27.0 exposes `remquo`; make use of it for our tests. Removing our workaround also allows removing the direct dependency on `gmp-mpfr-sys`

Upgrade all dependencies to the latest version

The Cargo feature `checked` was added in 410b0633a6b9 ("Overhaul tests") and later removed in e4ac1399062c ("swap stable to be unstable, checked is now debug_assertions"). However, there are a few remaining uses of `feature = "checked"` that did not get removed. Clean these up here.

Currently the default release profile enables LTO and single CGU builds, which is very slow to build. Most tests are better run with optimizations enabled since it allows testing a much larger number of inputs, so it is inconvenient that building can sometimes take significantly longer than the tests. Remedy this by doing the following: * Move the existing `release` profile to `release-opt`. * With the above, the default `release` profile is untouched (16 CGUs and thin local LTO). * `release-checked` inherits `release`, so no LTO or single CGU. This means that the simple `cargo test --release` becomes much faster for local development. We are able to enable the other profiles as needed in CI. Tests should ideally still be run with `--profile release-checked` to ensure there are no debug assetions or unexpected wrapping math hit. `no-panic` still needs a single CGU, so must be run with `--profile release-opt`. Since it is not possible to detect CGU or profilel configuration from within build scripts, the `ENSURE_NO_PANIC` environment variable must now always be set.

There seems to be a case of unsoundness with the `i586` version of `atan2`. For the following test: assert_eq!(atan2(2.0, -1.0), atan(2.0 / -1.0) + PI);atan2(2.0, -1.0) The output is optimization-dependent. The new `release-checked` profile produces the following failure: thread 'math::atan2::sanity_check' panicked at src/math/atan2.rs:123:5: assertion `left == right` failed left: 2.0344439357957027 right: 2.0344439357957027 Similarly, `sin::test_near_pi` fails with the following: thread 'math::sin::test_near_pi' panicked at src/math/sin.rs:91:5: assertion `left == right` failed left: 6.273720864039203e-7 right: 6.273720864039205e-7 Mark the tests ignored on `i586` for now.

`#![feature(start)]` was removed in [1], but we make use of it in the intrinsics example. Replace use of this feature with `#[no_mangle]` applied to `#[main]`. We don't actually run this example so it is not a problem if this is not entirely accurate. Currently the example does not run to completion, instead invoking `rust_begin_unwind`. [1]: rust-lang#134299

…ates Rework the available Cargo profiles

Rather than keeping a script that downloads the tarball, we can just add musl as a submodule and let git handle the synchronizatoin. Do so here.

0.17.10 introduced a change that removes `Sync` from `ProgressStyle`, which makes it more difficult to share in a callback. Pin the dependency for now until we see if `indicatif` will change this back or if we need to find a workaround.

llvm/llvm-project#116706 added Windows support to cpu_model. Compiling for UEFI also goes through that code path, because we treat it as a windows target. However, including windows.h is not actually going to work (and the used API would not be available in an UEFI environment). Disable building of cpu_model on UEFI to fix this.

This reverts commit 1dacdabdb6186f97144c50f8952575576deb3730.

This isn't very useful for constants since the trait constants are available, but does enable roundtripping via hex float syntax.

Add support for printing hex float syntax

Simplify the SPDX string to the user-facing version to make it easier for users and tooling to understand. Contributions must still be `MIT OR Apache-2.0`. [ add commit body with context - Trevor ]

`EXP_MAX` sounds like it would be the maximum value representable by that float type's exponent, rather than the maximum unsigned value of its bits. Clarify this by renaming to `EXP_SAT`, the "saturated" exponent representation.

bors · 2025-06-02T23:55:46Z

💔 Test failed - checks-actions

compiler-builtins has a symlink to the `libm` source directory so the two crates can share files but still act as two separate crates. This causes problems with some sysroot-related tooling, however, since directory symlinks seem to not be supported. The reason this was a symlink in the first place is that there isn't an easy for Cargo to publish two crates that share source (building works fine but publishing rejects `include`d files from parent directories, as well as nested package roots). However, after the switch to a subtree, we no longer need to publish compiler-builtins; this means that we can eliminate the link and just use `#[path]`. Similarly, the LICENSE file was symlinked so it could live in the repository root but be included in the package. This is also removed as it caused problems with the dist job (error from bootstrap's `tarball.rs`, "generated a symlink in a tarball"). If we need to publish compiler-builtins again for any reason, it would be easy to revert these changes in a preprocess step.

tgross35 · 2025-06-03T00:04:49Z

It seems the dist jobs very much do not like links. I removed the LICENSE symlink.

@bors r=Kobzol

bors · 2025-06-03T00:04:52Z

📌 Commit aff21f6 has been approved by Kobzol

It is now in the queue for this repository.

Kobzol · 2025-06-03T05:18:44Z

Hmm, for these simple cases we should teach bootstrap to just unwrap the symlinks.

RalfJung · 2025-06-03T05:43:04Z

The compiler-builtins repo still exists so it still needs a license, no? The license doesn't usually change so duplicating instead of symlinking seems fine.

tgross35 · 2025-06-03T05:48:20Z

It still exists at the repo root. It was just symlinked to within the compiler-builtins crate so it would get included in the package (Cargo unfortunately doesn't like ../LICENSE.txt) which shouldn't matter anymore if we're not publishing it.

tgross35 · 2025-06-03T17:02:38Z

@bors p=1 subtree + followup work

bors · 2025-06-03T19:52:17Z

⌛ Testing commit aff21f6 with merge 59aa1e8...

bors · 2025-06-03T23:14:41Z

☀️ Test successful - checks-actions
Approved by: Kobzol
Pushing 59aa1e8 to master...

github-actions · 2025-06-03T23:18:03Z

What is this?

This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.

Comparing a124fb3 (parent) -> 59aa1e8 (this PR)

Test differences

No test diffs found

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard 59aa1e873028948faaf8b97e5e02d4db340ad7b1 --output-dir test-dashboard

And then open test-dashboard/index.html in your browser to see an overview of all executed tests.

Job duration changes

dist-aarch64-linux: 5443.9s -> 8079.9s (48.4%)
x86_64-apple-1: 9194.8s -> 6891.9s (-25.0%)
mingw-check-1: 1956.1s -> 1591.0s (-18.7%)
mingw-check-2: 2231.4s -> 1839.8s (-17.6%)
dist-ohos-aarch64: 5003.1s -> 4213.6s (-15.8%)
aarch64-apple: 5625.5s -> 4860.9s (-13.6%)
i686-gnu-2: 6023.6s -> 5276.7s (-12.4%)
i686-gnu-1: 8060.1s -> 7165.5s (-11.1%)
x86_64-gnu-debug: 6002.3s -> 5360.4s (-10.7%)
i686-gnu-nopt-1: 7977.5s -> 7134.5s (-10.6%)

How to interpret the job duration changes?

Job durations can vary a lot, based on the actual runner instance
that executed the job, system noise, invalidated caches, etc. The table above is provided
mostly for t-infra members, for simpler debugging of potential CI slow-downs.

rust-log-analyzer · 2025-06-04T00:00:50Z

A job failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

Version: 4.2.2
Digest: sha256:ccb2698953eaebd21c7bf6268a94f9c26518a7e38e27e0b83c1fe1ad049819b1
Source commit SHA: 11bd71901bbe5b1630ceea73d27597364c9af683
##[endgroup]
Complete job name: DockerHub mirror
##[group]Run actions/checkout@v4
with:
  persist-credentials: false
  repository: rust-lang/rust
  token: ***

Kobzol · 2025-06-04T06:17:11Z

Oh crap, the compiler-builtins repository has its own submodule? 🥶 I wonder what problems that will cause.. 😆 The above CI error is the first one.

tgross35 · 2025-06-04T06:24:25Z

It's only used for testing, so never needs to be checked out by default here. I can switch it to a download script though, seeing as any other workaround is probably more difficult.

rust-timer · 2025-06-04T09:27:09Z

Finished benchmarking commit (59aa1e8): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results (primary 1.3%, secondary 3.2%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.3%	[1.3%, 1.3%]	1
Regressions ❌ (secondary)	5.3%	[4.7%, 5.9%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.2%	[-1.2%, -1.2%]	1
All ❌✅ (primary)	1.3%	[1.3%, 1.3%]	1

Cycles

Results (primary 1.8%, secondary -3.0%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.8%	[1.8%, 1.8%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.0%	[-3.0%, -3.0%]	1
All ❌✅ (primary)	1.8%	[1.8%, 1.8%]	1

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 745.424s -> 745.508s (0.01%)
Artifact size: 372.33 MiB -> 372.31 MiB (-0.00%)

durin42 · 2025-06-04T15:20:44Z

Oh crap, the compiler-builtins repository has its own submodule? 🥶 I wonder what problems that will cause.. 😆 The above CI error is the first one.

Anecdotally, I can't do anything terribly interesting with git submodule now, which has broken a CI job we run to keep Rust HEAD green against LLVM HEAD. Probably the submodule reference needs to get pruned on the way in...

lambdageek · 2025-06-04T15:22:11Z

This broke my workflow, too. Would just adding the submodule to the .gitmodules in the repo root fix it (it fixes the issue for me locally, but I'm not sure if it's the right approach) #142029

tgross35 · 2025-06-04T19:14:56Z

Sorry this broke some things; this will be removed in #142036

tgross35 and others added 30 commits January 23, 2025 22:30

Merge pull request rust-lang/libm#466 from tgross35/generic-fmin-fmax

df5ac49

Add `fminf16`, `fmaxf16`, `fminf128`, and `fmaxf128`

Add a generic version of fmod

08eda86

This can replace `fmod` and `fmodf`. As part of this change I was able to replace some of the `while` loops with `leading_zeros`.

Add fmodf16 using the generic implementation

67218cb

Merge pull request rust-lang/libm#469 from tgross35/generic-mod

ddffab3

Add `fmodf16`

Increase or set CI timeouts

9d38c93

With the new routines, some of our tests are running close to their timeouts. Increase the timeout for test jobs, and set a short timeout for all other jobs that did not have one.

Add fmodf128

71200bc

This function is significantly slower than all others so includes an override in `EXTREMELY_SLOW_TESTS`. Without it, PR CI takes ~1hour and the extensive tests in CI take ~1day.

Merge pull request rust-lang/libm#470 from tgross35/f128-fmod

3524632

Add `fmodf128`

Enable missing icount benchmarks

173a48c

A few new functions were added but this list did not get updated. Do so here.

Use az exported from Rug

62f0fd4

Since Rug 1.27.0, `az` is reexported. This means we no longer need to track it as a separate dependency.

Use frexp from Rug

edd1289

Rug 1.27.0 exposes `frexp`. Make use of it for our tests.

Use remquo from Rug

b9bd107

Rug 1.27.0 exposes `remquo`; make use of it for our tests. Removing our workaround also allows removing the direct dependency on `gmp-mpfr-sys`

Merge pull request rust-lang/libm#477 from tgross35/upgrade-dependencies

9c6548e

Upgrade all dependencies to the latest version

Merge pull request rust-lang/libm#479 from tgross35/cargo-profile-upd…

98fd72b

…ates Rework the available Cargo profiles

Add a version to some FIXMEs that will be resolved in LLVM 20

4e1b885

Switch musl from a script download to a submodule

259e544

Rather than keeping a script that downloads the tarball, we can just add musl as a submodule and let git handle the synchronizatoin. Do so here.

Temporarily pin indicatif to 0.17.9

176d22d

0.17.10 introduced a change that removes `Sync` from `ProgressStyle`, which makes it more difficult to share in a callback. Pin the dependency for now until we see if `indicatif` will change this back or if we need to find a workaround.

Revert "Temporarily pin indicatif to 0.17.9"

396725e

This reverts commit 1dacdabdb6186f97144c50f8952575576deb3730.

Support parsing NaN and infinities from the hf* functions

99c1080

This isn't very useful for constants since the trait constants are available, but does enable roundtripping via hex float syntax.

Introduce a wrapper type for IEEE hex float formatting

f767b58

Util: also print the hex float format for outputs

e283c25

Merge pull request rust-lang/libm#483 from tgross35/hex-print

9d62bd0

Add support for printing hex float syntax

Specify license as just MIT

7d67414

Simplify the SPDX string to the user-facing version to make it easier for users and tooling to understand. Contributions must still be `MIT OR Apache-2.0`. [ add commit body with context - Trevor ]

Rename EXP_MAX to EXP_SAT

8db5ff7

`EXP_MAX` sounds like it would be the maximum value representable by that float type's exponent, rather than the maximum unsigned value of its bits. Clarify this by renaming to `EXP_SAT`, the "saturated" exponent representation.

This comment has been minimized.

Sign in to view

bors added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Jun 2, 2025

tgross35 force-pushed the builtins-josh-subtree branch from e8a36fb to aff21f6 Compare June 3, 2025 00:02

This comment was marked as off-topic.

Sign in to view

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jun 3, 2025

bors mentioned this pull request Jun 3, 2025

Remove i128 and u128 from improper_ctypes_definitions #137306

Merged

bors added the merged-by-bors This PR was explicitly merged by bors. label Jun 3, 2025

bors merged commit 59aa1e8 into rust-lang:master Jun 3, 2025
11 checks passed

rustbot added this to the 1.89.0 milestone Jun 3, 2025

tgross35 deleted the builtins-josh-subtree branch June 3, 2025 23:18

Merge compiler-builtins as a Josh subtree #141229

Merge compiler-builtins as a Josh subtree #141229

Uh oh!

Conversation

tgross35 commented May 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

bors commented Jun 2, 2025

Uh oh!

This comment was marked as off-topic.

tgross35 commented Jun 3, 2025

Uh oh!

bors commented Jun 3, 2025

Uh oh!

Kobzol commented Jun 3, 2025

Uh oh!

RalfJung commented Jun 3, 2025 via email

Uh oh!

tgross35 commented Jun 3, 2025

Uh oh!

tgross35 commented Jun 3, 2025

Uh oh!

bors commented Jun 3, 2025

Uh oh!

bors commented Jun 3, 2025

Uh oh!

Uh oh!

github-actions bot commented Jun 3, 2025

Test differences

Job duration changes

Uh oh!

rust-log-analyzer commented Jun 4, 2025

Uh oh!

Kobzol commented Jun 4, 2025

Uh oh!

tgross35 commented Jun 4, 2025

Uh oh!

rust-timer commented Jun 4, 2025

Overall result: no relevant changes - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

durin42 commented Jun 4, 2025

Uh oh!

lambdageek commented Jun 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tgross35 commented Jun 4, 2025

Uh oh!

Uh oh!

Merge `compiler-builtins` as a Josh subtree #141229

Merge `compiler-builtins` as a Josh subtree #141229

tgross35 commented May 18, 2025 •

edited

Loading

lambdageek commented Jun 4, 2025 •

edited

Loading