codegen: handle OperandValue::Uninit in codegen_return_terminator by glandium · Pull Request #159825 · rust-lang/rust

glandium · 2026-07-24T10:17:16Z

#157797 added OperandValue::Uninit to skip stores for entirely-uninit constants, but missed the PassMode::Direct | PassMode::Pair branch of codegen_return_terminator, which called immediate_or_packed_pair unconditionally. A function directly returning an all-uninit value (e.g. MaybeUninit::uninit()) would hit that branch and ICE in OperandRef::immediate, as reported against tokio and reduced by tmiasko.

rustbot · 2026-07-24T10:17:23Z

r? @tiif

rustbot has assigned @tiif.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 74 candidates
Random selection from 18 candidates

glandium · 2026-07-24T10:17:44Z

r? @oli-obk

oli-obk · 2026-07-25T08:26:31Z

                let op = self.codegen_consume(bx, mir::Place::return_place().as_ref());
                if let Ref(place_val) = op.val {
                    bx.load_from_place(bx.backend_type(op.layout), place_val)
+                } else if let Uninit = op.val {


Use an exhaustive match instead of an if let chain

View changes since the review

oli-obk

r=me with nit

View changes since this review

rustbot · 2026-07-25T08:26:54Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

When adding `OperandValue::Uninit` to skip stores for entirely-uninit constants, we missed the `PassMode::Direct | PassMode::Pair` branch of `codegen_return_terminator`, which called `immediate_or_packed_pair` unconditionally. A function directly returning an all-uninit value (e.g. `MaybeUninit::uninit()`) would hit that branch and ICE in `OperandRef::immediate`, as reported against tokio and reduced by tmiasko.

glandium · 2026-07-25T15:08:55Z

@rustbot ready

oli-obk · 2026-07-25T15:16:08Z

@bors r+ rollup

rust-bors · 2026-07-25T15:16:11Z

📌 Commit 9e9467a has been approved by oli-obk

It is now in the queue for this repository.

Mark-Simulacrum · 2026-07-25T17:43:47Z

@bors p=1

(Since we're getting a non-trivial number of reports of this.)

codegen: handle OperandValue::Uninit in codegen_return_terminator Fixes rust-lang#159815. rust-lang#157797 added `OperandValue::Uninit` to skip stores for entirely-uninit constants, but missed the `PassMode::Direct | PassMode::Pair` branch of `codegen_return_terminator`, which called `immediate_or_packed_pair` unconditionally. A function directly returning an all-uninit value (e.g. `MaybeUninit::uninit()`) would hit that branch and ICE in `OperandRef::immediate`, as reported against tokio and reduced by tmiasko.

codegen: handle OperandValue::Uninit in codegen_return_terminator Fixes #159815. #157797 added `OperandValue::Uninit` to skip stores for entirely-uninit constants, but missed the `PassMode::Direct | PassMode::Pair` branch of `codegen_return_terminator`, which called `immediate_or_packed_pair` unconditionally. A function directly returning an all-uninit value (e.g. `MaybeUninit::uninit()`) would hit that branch and ICE in `OperandRef::immediate`, as reported against tokio and reduced by tmiasko.

JonathanBrouwer · 2026-07-25T20:40:16Z

@bors yield

rust-bors · 2026-07-25T20:40:19Z

Auto build was cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/30173802025

The next pull request likely to be tested is #159825.

codegen: handle OperandValue::Uninit in codegen_return_terminator Fixes #159815. #157797 added `OperandValue::Uninit` to skip stores for entirely-uninit constants, but missed the `PassMode::Direct | PassMode::Pair` branch of `codegen_return_terminator`, which called `immediate_or_packed_pair` unconditionally. A function directly returning an all-uninit value (e.g. `MaybeUninit::uninit()`) would hit that branch and ICE in `OperandRef::immediate`, as reported against tokio and reduced by tmiasko.

JonathanBrouwer · 2026-07-25T20:40:56Z

@bors yield

rust-bors · 2026-07-25T20:41:00Z

Auto build was cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/30174049598

The next pull request likely to be tested is #159920.

codegen: handle OperandValue::Uninit in codegen_return_terminator Fixes #159815. #157797 added `OperandValue::Uninit` to skip stores for entirely-uninit constants, but missed the `PassMode::Direct | PassMode::Pair` branch of `codegen_return_terminator`, which called `immediate_or_packed_pair` unconditionally. A function directly returning an all-uninit value (e.g. `MaybeUninit::uninit()`) would hit that branch and ICE in `OperandRef::immediate`, as reported against tokio and reduced by tmiasko.

JonathanBrouwer · 2026-07-25T20:50:48Z

@bors yield

rust-bors · 2026-07-25T20:50:51Z

Auto build was cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/30174282729

The next pull request likely to be tested is #159920.

codegen: handle OperandValue::Uninit in codegen_return_terminator Fixes rust-lang#159815. rust-lang#157797 added `OperandValue::Uninit` to skip stores for entirely-uninit constants, but missed the `PassMode::Direct | PassMode::Pair` branch of `codegen_return_terminator`, which called `immediate_or_packed_pair` unconditionally. A function directly returning an all-uninit value (e.g. `MaybeUninit::uninit()`) would hit that branch and ICE in `OperandRef::immediate`, as reported against tokio and reduced by tmiasko.

…uwer Rollup of 25 pull requests Successful merges: - #159825 (codegen: handle OperandValue::Uninit in codegen_return_terminator) - #138618 (Support using const pointers in asm `const` operand) - #157962 (Lower paths to functions in const args as ConstKind::Error) - #158404 (trait_solver: normalize next-gen region constraints) - #158709 (rustdoc: warn on improperly interleaved HTML/MD) - #159174 (Fix implicit_provenance_casts warnings on Xous) - #159179 (enable `unreachable_cfg_select_predicates` lint as part of `unused` lint group) - #159518 (iter: extend step_by specialization to cover StepBy<RangeIter<{integer}>>) - #159673 (bootstrap: forward -fdebug-prefix-map when using cc) - #159700 (Split non-local `semicolon_in_expressions_from_macros` into a separate lint) - #159720 (document #[global_allocator] constraints) - #159732 (optimization: don't look for diagnostic/canonical items without rustc_attrs enabled) - #159738 (implement `CovariantUnsafeCell`) - #159740 (reuse regular exported_non_generic_symbols logic in Miri) - #159780 (check `extern "custom"` function pointers) - #159786 (rustdoc-js: ignore editor temp files in test folder discovery) - #159819 (std::sync::poison: disable auto_cfg on PoisonError::new) - #155388 (stepping into where-clauses during normalization may be productive) - #155914 (when bailing on ambiguity, don't force other results to ambig) - #159204 (Add support to caller_location to rustc_public) - #159439 (Fix(lib/fs/win): Fall back on Win32 delete for `Dir::remove_file`) - #159676 (Update wasm-component-ld to 0.5.27) - #159695 (proc_macro: Fix cfg_attr inner attrs in file modules) - #159730 (allow accessing the contents of UnsafeCell without going through get) - #159809 (Avoid `#[target_features]`)

JonathanBrouwer · 2026-07-25T21:56:32Z

@bors p=101
Scheduling

rust-bors · 2026-07-26T06:29:11Z

☀️ Test successful - CI
Approved by: oli-obk
Duration: 3h 11m 36s
Pushing fdbd8ee to main...

github-actions · 2026-07-26T06:32:13Z

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 008fa22 (parent) -> fdbd8ee (this PR)

Test differences

Show 6 test diffs

Stage 1

[codegen] tests/codegen-llvm/uninit-return-value.rs: [missing] -> pass (J0)

Stage 2

[codegen] tests/codegen-llvm/uninit-return-value.rs: [missing] -> pass (J1)

Additionally, 4 doctest diffs were found. These are ignored, as they are noisy.

Job group index

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard fdbd8ee70d9b68d98be873e1095d74cf9441f0c8 --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

x86_64-gnu-gcc-core-tests: 8m 10s -> 13m 54s (+70.2%)
optional-x86_64-gnu-parallel-frontend: 1h 34m -> 2h 18m (+46.9%)
x86_64-gnu-llvm-22-1: 51m 36s -> 1h 14m (+44.4%)
x86_64-gnu-next-trait-solver-polonius: 34m 12s -> 46m 33s (+36.1%)
x86_64-msvc-ext1: 1h 38m -> 2h 11m (+33.9%)
x86_64-gnu: 2h 19m -> 1h 37m (-30.4%)
dist-armhf-linux: 1h 10m -> 1h 31m (+30.0%)
dist-x86_64-llvm-mingw: 2h 7m -> 1h 29m (-29.7%)
dist-x86_64-msvc: 2h 23m -> 1h 43m (-28.0%)
dist-i686-msvc: 2h 2m -> 1h 30m (-26.4%)

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 · 2026-07-26T07:09:52Z

Finished benchmarking commit (fdbd8ee): 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 1.9%, secondary -2.2%)

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

	mean	range	count
Regressions ❌ (primary)	1.9%	[1.9%, 1.9%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.2%	[-2.2%, -2.2%]	1
All ❌✅ (primary)	1.9%	[1.9%, 1.9%]	1

Cycles

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

Binary size

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

Bootstrap: 489.346s -> 487.865s (-0.30%)
Artifact size: 387.69 MiB -> 387.69 MiB (0.00%)

The ICE was already reported as rust-lang/rust#159815 and fixed by rust-lang/rust#159825, merged 2026-07-26 06:29 UTC — after the nightly that broke us was cut, so 2026-07-27 is the first nightly containing it. Record that in the pin comment, along with what to check before unpinning, so the pin does not outlive its cause. rust-lang/rust#159867 is a still-open ICE in the same `OperandValue::Uninit` family, so the unpin wants verifying rather than assuming. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

… build (#984) `nix build .#packages.x86_64-linux.bedwars` died in the compiler: ``` error: internal compiler error: compiler/rustc_codegen_ssa/src/mir/operand.rs:291:18: not immediate: OperandRef(Uninit @ TyAndLayout { ty: unsafe_cell::UnsafeCell<MaybeUninit<runtime::task::Notified<Arc<multi_thread::handle::Handle>>>>, ... backend_repr: Scalar(union pointer) ... }) ``` ## What the differentials said The upstream issue carries an 8-line reproduction, which turns a 40-minute `nix build` into a sub-second `rustc` call, so all of these ran against that rather than the workspace. - **Optimisation level.** ICEs at 2 and 3, clean at 0 and 1. So `Tests (ubuntu-latest)` being green proves nothing: it is a debug build and cargoUnit compiles at `-C opt-level=3`. - **Architecture.** Not x86. It ICEs identically on aarch64-apple-darwin, same assertion, same line. The bug is in target-independent `codegen_return_terminator`. - **Nightly.** Clean on 07-23, ICE on 07-24 through 07-26, clean again on 07-27. - **A newer tokio.** Does not exist. 1.53.1 is the newest published. ## The fix Bump the pin forward one day, to the nightly that carries [rust-lang/rust#159825](rust-lang/rust#159825), which fixes [#159815](rust-lang/rust#159815). Root cause upstream: #157797 added `OperandValue::Uninit` and missed the `PassMode::Direct | PassMode::Pair` branch. Forward rather than back, deliberately. Rolling back to 07-23 also works, but it is a knob someone has to remember to remove and it strands the workspace three days behind. Bumping forward becomes an ordinary pin the moment the fix reaches beta. Rejected: `opt-level=1` for tokio only, which permanently deoptimises the async runtime of a game server to work around a compiler bug that is already fixed. ## Verified The whole workspace moves with a toolchain pin, so every gate was re-run. ``` cargo check --workspace --all-targets rc=0 nix run .#lint rc=0 nix run .#test rc=0 447 passed, 1 skipped nix run .#e2e -- --seconds 12 rc=0 3729 LevelChunkWithLight nix build x86_64-linux bedwars + proxy rc=0 ``` ``` /nix/store/rabbfr795mbs20n0qwdmjvlfa3ibprv3-bedwars-0.1.0 /nix/store/qp3ykmwsy35df2hrblaaz77frs6xbjyb-hyperion-proxy-0.1.0 ``` Zero occurrences of `internal compiler error` in the full build log. (sent by an AI agent via Claude Code, model claude-opus-4-6)

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jul 24, 2026

rustbot assigned tiif Jul 24, 2026

This comment has been minimized.

Sign in to view

rustbot assigned oli-obk and unassigned tiif Jul 24, 2026

glandium force-pushed the issue159815 branch from eafe934 to 62c243b Compare July 24, 2026 10:35

This comment has been minimized.

Sign in to view

glandium force-pushed the issue159815 branch from 62c243b to 415030f Compare July 24, 2026 10:43

tmiasko reviewed Jul 24, 2026

View reviewed changes

Comment thread tests/codegen-llvm/uninit-return-value.rs Outdated

glandium mentioned this pull request Jul 25, 2026

[ICE]: operand.rs:114:32: OperandValue cannot be a pointer: Uninit #159867

Open

oli-obk reviewed Jul 25, 2026

View reviewed changes

oli-obk requested changes Jul 25, 2026

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 25, 2026

glandium force-pushed the issue159815 branch from 415030f to 9e9467a Compare July 25, 2026 15:08

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 25, 2026

rust-bors Bot 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 Jul 25, 2026

JonathanBrouwer mentioned this pull request Jul 25, 2026

Rollup of 15 pull requests #159925

Closed

This comment has been minimized.

Sign in to view

JonathanBrouwer mentioned this pull request Jul 25, 2026

Rollup of 25 pull requests #159926

Closed

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 Jul 26, 2026

rust-bors Bot merged commit fdbd8ee into rust-lang:main Jul 26, 2026
14 checks passed

rustbot added this to the 1.99.0 milestone Jul 26, 2026

andrewgazelka mentioned this pull request Jul 27, 2026

fix: bump the nightly past a codegen ICE that broke the Linux release build hyperion-mc/hyperion#984

Merged

Uh oh!

Uh oh!

Conversation

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

Uh oh!

Uh oh!

rustbot commented Jul 24, 2026

Uh oh!

This comment has been minimized.

glandium commented Jul 24, 2026

Uh oh!

This comment has been minimized.

Uh oh!

oli-obk Jul 25, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oli-obk left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jul 25, 2026

Uh oh!

glandium commented Jul 25, 2026

Uh oh!

oli-obk commented Jul 25, 2026

Uh oh!

rust-bors Bot commented Jul 25, 2026

Uh oh!

Mark-Simulacrum commented Jul 25, 2026

Uh oh!

This comment has been minimized.

JonathanBrouwer commented Jul 25, 2026

Uh oh!

rust-bors Bot commented Jul 25, 2026

Uh oh!

This comment has been minimized.

JonathanBrouwer commented Jul 25, 2026

Uh oh!

rust-bors Bot commented Jul 25, 2026

Uh oh!

This comment has been minimized.

JonathanBrouwer commented Jul 25, 2026

Uh oh!

rust-bors Bot commented Jul 25, 2026

Uh oh!

JonathanBrouwer commented Jul 25, 2026

Uh oh!

This comment has been minimized.

rust-bors Bot commented Jul 26, 2026

Uh oh!

Uh oh!

github-actions Bot commented Jul 26, 2026

Test differences

Stage 1

Stage 2

Job duration changes

Uh oh!

rust-timer commented Jul 26, 2026

Overall result: no relevant changes - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

glandium commented Jul 24, 2026 •

edited by rustbot

Loading

oli-obk Jul 25, 2026 •

edited by rustbot

Loading

oli-obk left a comment •

edited by rustbot

Loading