Subtree cg_gcc sync (2026-07-24) by GuillaumeGomez · Pull Request #159844 · rust-lang/rust

GuillaumeGomez · 2026-07-24T14:52:50Z

r? ghost

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.

…update_cg_gcc_2026-04-29

…_2026-04-29, r=antoyo GCC backend subtree sync r? ghost

This way they also apply to the allocator shim.

Submission to the Apple App Store for iOS no longer requires embedding bitcode, but even back when it did, it needed LLVM bitcode, so GCC wouldn't work anyway.

This is necessary to fix incremental LTO in cg_gcc as well as to do some LTO refactorings I want to do. The actual fix for cg_gcc will be done on the cg_gcc repo to test it in CI.

Move most flags from module_codegen to new_context

`rustc_error_messages` currently depends on `rustc_ast`/`rustc_ast_pretty`. This is odd, because `rustc_error_messages` feels like a very low-level module but `rustc_ast`/`rustc_ast_pretty` do not. The reason is that a few AST types impl `IntoDiagArg` via pretty-printing. `rustc_error_messages` can define `IntoDiagArg` and then impl it for the AST types. But if we invert the dependency we hit a problem with the orphan rule: `rustc_ast` must impl `IntoDiagArg` for the AST types, but that requires calling pretty-printing code which is in `rustc_ast_pretty`, a downstream crate. This commit avoids this problem by just removing the `IntoDiagArg` impls for these AST types. There aren't that many of them, and we can just use `String` in the relevant error structs and use the pretty printer in the downstream crates that construct the error structs. There are plenty of existing examples where `String` is used in error structs. There is now no dependency between `rustc_ast*` and `rustc_error_messages`.

…and, r=bjorn3 Fix: On wasm targets, call `panic_in_cleanup` if panic occurs in cleanup Relies on rust-lang/llvm-project#194. Reland of rust-lang#151771. 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. - [ ] Add test for rust-lang#153948

Pass Session to optimize_and_codegen_fat_lto This is necessary to fix incremental LTO in cg_gcc as well as to do some LTO refactorings I want to do. The actual fix for cg_gcc will be done on the cg_gcc repo to test it in CI.

…henkov Add rlib digest to identify Rust object files This adds a metadata entry to `rlib` archives that lists which members are Rust object files instead of relying on the filename heuristic in `looks_like_rust_object.file`. I also added a fallback to the old behavior for `rlibs` built by older compilers. Part of rust-lang#138243.

Rustup to rustc 1.97.0-nightly (e95e732 2026-05-05)

Handle all modules being serialized during LTO

CrateInfo is only necessary during linking and non-local LTO.

While it was previously defined in Session, it is only ever used with OutputFilenames methods.

@RalfJung

…lfJung,scottmcm,saethlin change the type of the argument of `drop_in_place` lang item to `&mut _` We used to special case `core::ptr::drop_in_place` when computing LLVM argument attributes with this hack: https://github.com/rust-lang/rust/blob/db5e2dc248fe5bb26f70d7baec46a3bca9fa3e1d/compiler/rustc_ty_utils/src/abi.rs#L383-L392 This is because even though `drop_in_place` takes a `*mut T` it is semantically a `&mut T` (remember how `&mut Self` is passed to `Drop::drop`). This is apparently relevant for perf. This PR replaces this hack with a simpler solution -- it makes `drop_in_place` a thin wrapper around newly added `core::ptr::drop_glue`, which is the actual lang item and takes a `&mut T`: https://github.com/rust-lang/rust/blob/d2563d5003bbecff1efc40c1f5673ceec603825b/library/core/src/ptr/mod.rs#L810-L833 ------ The rest of the PR is blessing tests and cleaning up things which are not necessary after this change. One thing that is a bit awkward is that now that `drop_glue` is the actual lang item, a lot of the comments referring to `drop_in_place` are outdated. Should I try fixing that? I've also changed `async_drop_in_place` to take a `&mut T`, and it simplified the code handling it a bit. (since it's unstable we don't need to introduce a wrapper) ------- cc @RalfJung Closes rust-lang#154274

Move CrateInfo computation after codegen_crate CrateInfo is only necessary during linking and non-local LTO. Part of rust-lang/compiler-team#908

…-obk Move invocation_temp into OutputFilenames While it was previously defined in Session, it is only ever used with OutputFilenames methods.

@antoyo

…24, r=GuillaumeGomez Subtree cg_gcc sync (2026-07-24) cc @antoyo @bjorn3 r? ghost

JonathanBrouwer · 2026-08-03T13:39:50Z

@bors yield
Scheduling

rust-bors · 2026-08-03T13:39:54Z

Auto build was cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/30818938172

The next pull request likely to be tested is #159789.

@antoyo

…cc_2026-07-24, r=GuillaumeGomez Subtree cg_gcc sync (2026-07-24) cc @antoyo @bjorn3 r? ghost

JonathanBrouwer · 2026-08-03T13:41:16Z

@bors try jobs=dist-i686-linux,dist-aarch64-linux
Debugging stuck CI builds

…24, r=<try> Subtree cg_gcc sync (2026-07-24) try-job: dist-i686-linux try-job: dist-aarch64-linux

Kobzol · 2026-08-03T13:49:37Z

It wasn't stuck, this PR invalidated the Docker image, so it would take much longer to build.

JonathanBrouwer · 2026-08-03T13:50:37Z

Oh sorry :c
The LLVM build hadn't produced any output for more than an hour tho, which I found to be suspicious, but that might just be github not showing the output live correctly then?

JonathanBrouwer · 2026-08-03T13:53:59Z

@bors rollup=iffy Invalidates docker image causing long builds

…uwer Rollup of 21 pull requests Successful merges: - #159844 (Subtree cg_gcc sync (2026-07-24)) - #159326 (Deny multiple EII impls on a single item) - #159535 (Optimize slice::contains for bytewise types) - #159595 (Promote loongarch32-unknown-none* to Tier 2) - #160007 (allow `-Ldependency` search paths for panic runtimes) - #160320 (point at trait definition when it is used as a derive macro) - #160369 (When suggesting method names, prefer *exact* doc aliases over similar names) - #160406 (`DepKind` cleanups) - #160424 (Use `thread::available_parallelism` as the default limit for backend parallelism) - #159014 ([rustdoc] Do not take `doc(cfg())` into account when filtering doctests) - #159303 (Fix ICE for direct inline const generic defaults) - #159977 (Add regression test for bool indexing codegen) - #160165 (reject `...` without pattern post-expansion) - #160295 (Fix rustdoc ICE when checking if a generic arg can be elided) - #160305 (Linkify C-SKY targets in `platform-support.md`) - #160314 (fix borrowck ICE for consts with fn pointer type) - #160322 (ElaborateBoxDeref: remove unnecessary projection) - #160338 (Add regression test for supertrait associated type normalization through dyn) - #160340 (Add regression test for unused_parens on contract clauses) - #160371 (Add doc aliases for transpositions `read_exact_buf` and `read_exact_buf_at`) - #160384 (Add PR body notes for Cargo lock file maintenance)

Kobzol · 2026-08-03T16:15:41Z

Oh sorry :c The LLVM build hadn't produced any output for more than an hour tho, which I found to be suspicious, but that might just be github not showing the output live correctly then?

Don't worry, the Docker cache was saved, so the next time should be faster :) It is possible that only the logs got delayed, but maybe the job was indeed also stuck, hard to tell.

rust-bors · 2026-08-03T17:10:07Z

☀️ Try build successful (CI)
Build commit: f671346 (f671346cb1c4ed925f370cf9dcedc537b639ff26)
Base parent: 7c329d6 (7c329d6c76e11ca40c5673818ab0439c1be8962c)

rust-bors · 2026-08-03T18:57:46Z

☀️ Test successful - CI
Approved by: GuillaumeGomez
Duration: 3h 16m 27s
Pushing 5048696 to main...

github-actions · 2026-08-03T19:00:20Z

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 7c329d6 (parent) -> 5048696 (this PR)

Test differences

Show 3 test diffs

3 doctest diffs were found. These are ignored, as they are noisy.

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard 504869653f510b279c542e65ccd1ea9710c119ba --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-x86_64-illumos: 1h 56m -> 1h 15m (-35.2%)
x86_64-gnu-nopt: 2h 20m -> 1h 31m (-34.8%)
dist-various-2: 44m 46s -> 30m 23s (-32.2%)
i686-msvc-1: 2h 21m -> 3h 4m (+30.5%)
x86_64-gnu-next-trait-solver-polonius: 41m 23s -> 52m 52s (+27.7%)
x86_64-gnu-distcheck: 2h 17m -> 1h 40m (-26.8%)
x86_64-gnu-gcc-core-tests: 11m 17s -> 14m 15s (+26.3%)
x86_64-gnu-llvm-21-1: 55m 54s -> 41m 47s (-25.2%)
dist-loongarch64-linux: 1h 29m -> 1h 52m (+25.1%)
x86_64-msvc-ext2: 1h 7m -> 1h 23m (+24.4%)

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-timer · 2026-08-03T19:44:27Z

Finished benchmarking commit (5048696): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This perf run didn't have relevant results for this metric.

Max RSS (memory usage)

Results (primary -0.9%, secondary -1.8%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.5%	[0.4%, 0.5%]	3
Regressions ❌ (secondary)	0.5%	[0.4%, 0.5%]	2
Improvements ✅ (primary)	-4.9%	[-4.9%, -4.9%]	1
Improvements ✅ (secondary)	-4.1%	[-5.5%, -2.6%]	2
All ❌✅ (primary)	-0.9%	[-4.9%, 0.5%]	4

Cycles

Results (primary -0.3%, secondary 0.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.9%	[0.4%, 1.4%]	5
Regressions ❌ (secondary)	1.4%	[0.6%, 2.2%]	5
Improvements ✅ (primary)	-1.1%	[-2.1%, -0.6%]	7
Improvements ✅ (secondary)	-0.9%	[-1.5%, -0.4%]	6
All ❌✅ (primary)	-0.3%	[-2.1%, 1.4%]	12

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 491.018s -> 491.182s (0.03%)
Artifact size: 390.29 MiB -> 390.23 MiB (-0.02%)

…2026-07-24, r=GuillaumeGomez" This reverts commit 5048696, reversing changes made to 7c329d6.

@JonathanBrouwer

Revert #159844 r? @JonathanBrouwer

RalfJung · 2026-08-03T21:57:40Z

Seems like this broke PR CI, x86_64-gnu-gcc now fails everywhere. Can you either quickly fix that or revert?

Also how is it even possible that this was allowed to land if it does not pass CI?

antoyo · 2026-08-03T21:58:49Z

Can you either quickly fix that or revert?

The revert PR is up.

GuillaumeGomez · 2026-08-03T22:00:24Z

Seems like this broke PR CI, x86_64-gnu-gcc now fails everywhere. Can you either quickly fix that or revert?

Also how is it even possible that this was allowed to land if it does not pass CI?

I was very surprised by that too. Asked on zulip.

hoodmane and others added 30 commits April 28, 2026 09:56

simd_reduce_min/max: remove float support

062e9f3

Merge commit 'd189e9f23c4c971546cb59bf43ab4df0e5552770' into subtree-…

e001838

…update_cg_gcc_2026-04-29

Fix tidy errors in cg_gcc

f81e070

Auto merge of rust-lang#155978 - GuillaumeGomez:subtree-update_cg_gcc…

efbbd33

…_2026-04-29, r=antoyo GCC backend subtree sync r? ghost

Move most flags from module_codegen to new_context

68fc1e2

This way they also apply to the allocator shim.

Remove a fixme

6b0348d

Submission to the Apple App Store for iOS no longer requires embedding bitcode, but even back when it did, it needed LLVM bitcode, so GCC wouldn't work anyway.

Update comment

1ab95d4

Pass Session to optimize_and_codegen_fat_lto

f992a3d

This is necessary to fix incremental LTO in cg_gcc as well as to do some LTO refactorings I want to do. The actual fix for cg_gcc will be done on the cg_gcc repo to test it in CI.

Merge pull request rust-lang#868 from bjorn3/misc_changes

7be6ee1

Move most flags from module_codegen to new_context

Fix asm pointer inputs (rust-lang#874)

6ecabe8

Add rlib digest to identify Rust object files

c32743a

rename drop_in_place lang item to drop_glue

43cdceb

Merge branch 'master'

ea8fd81

Rustup to rustc 1.97.0-nightly (e95e732 2026-05-05)

fbf722a

Stop cspell from complaining about rmeta

b69d074

Add failing run-make test

50056a8

Merge pull request rust-lang#884 from bjorn3/rustup

b5a73af

Rustup to rustc 1.97.0-nightly (e95e732 2026-05-05)

Handle all modules being serialized during LTO

da9e1aa

Merge pull request rust-lang#880 from bjorn3/lto_all_serialized

b4f3025

Handle all modules being serialized during LTO

Move CrateInfo computation after codegen_crate

d2a379f

CrateInfo is only necessary during linking and non-local LTO.

Move invocation_temp into OutputFilenames

3ce0567

While it was previously defined in Session, it is only ever used with OutputFilenames methods.

Rollup merge of rust-lang#156243 - bjorn3:lto_refactors18, r=mu001999

2cfff08

Move CrateInfo computation after codegen_crate CrateInfo is only necessary during linking and non-local LTO. Part of rust-lang/compiler-team#908

Rollup merge of rust-lang#156245 - bjorn3:move_invocation_temp, r=oli…

31a27ca

…-obk Move invocation_temp into OutputFilenames While it was previously defined in Session, it is only ever used with OutputFilenames methods.

Update a bunch of dependencies to remove windows-targets

86bb28c

rust-bors Bot pushed a commit that referenced this pull request Aug 3, 2026

Auto merge of #159844 - GuillaumeGomez:subtree-update_cg_gcc_2026-07-…

076db75

…24, r=GuillaumeGomez Subtree cg_gcc sync (2026-07-24) cc @antoyo @bjorn3 r? ghost

JonathanBrouwer added a commit to JonathanBrouwer/rust that referenced this pull request Aug 3, 2026

Rollup merge of rust-lang#159844 - GuillaumeGomez:subtree-update_cg_g…

ac65895

…cc_2026-07-24, r=GuillaumeGomez Subtree cg_gcc sync (2026-07-24) cc @antoyo @bjorn3 r? ghost

JonathanBrouwer mentioned this pull request Aug 3, 2026

Rollup of 21 pull requests #160437

Closed

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 3, 2026

Auto merge of #159844 - GuillaumeGomez:subtree-update_cg_gcc_2026-07-…

f671346

…24, r=<try> Subtree cg_gcc sync (2026-07-24) try-job: dist-i686-linux try-job: dist-aarch64-linux

This comment has been minimized.

Sign in to view

rust-bors Bot added merged-by-bors This PR was explicitly merged by bors. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Aug 3, 2026

rust-bors Bot merged commit 5048696 into rust-lang:main Aug 3, 2026
15 checks passed

rustbot added this to the 1.99.0 milestone Aug 3, 2026

This was referenced Aug 3, 2026

Refactor the way cg_ssa handles indirect returns (returns via sret) — Take 2 #160023

Open

Mangle personality symbol #148413

Open

GuillaumeGomez deleted the subtree-update_cg_gcc_2026-07-24 branch August 3, 2026 19:10

rust-bors Bot pushed a commit that referenced this pull request Aug 3, 2026

Revert "Auto merge of #159844 - GuillaumeGomez:subtree-update_cg_gcc_…

bf57193

…2026-07-24, r=GuillaumeGomez" This reverts commit 5048696, reversing changes made to 7c329d6.

rust-bors Bot pushed a commit that referenced this pull request Aug 3, 2026

Auto merge of #160468 - GuillaumeGomez:revert-159844, r=JonathanBrouwer

c9ff496

Revert #159844 r? @JonathanBrouwer

Uh oh!

Uh oh!

Conversation

GuillaumeGomez commented Jul 24, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JonathanBrouwer commented Aug 3, 2026

Uh oh!

rust-bors Bot commented Aug 3, 2026

Uh oh!

JonathanBrouwer commented Aug 3, 2026

Uh oh!

This comment has been minimized.

Kobzol commented Aug 3, 2026

Uh oh!

JonathanBrouwer commented Aug 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JonathanBrouwer commented Aug 3, 2026

Uh oh!

This comment has been minimized.

Kobzol commented Aug 3, 2026

Uh oh!

rust-bors Bot commented Aug 3, 2026

Uh oh!

rust-bors Bot commented Aug 3, 2026

Uh oh!

Uh oh!

github-actions Bot commented Aug 3, 2026

Test differences

Job duration changes

Uh oh!

rust-timer commented Aug 3, 2026

Overall result: no relevant changes - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

RalfJung commented Aug 3, 2026

Uh oh!

antoyo commented Aug 3, 2026

Uh oh!

GuillaumeGomez commented Aug 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

20 participants

GuillaumeGomez commented Jul 24, 2026 •

edited by rustbot

Loading

JonathanBrouwer commented Aug 3, 2026 •

edited

Loading