tail calls: fix copying non-scalar arguments to callee by RalfJung · Pull Request #150473 · rust-lang/rust

RalfJung · 2025-12-28T23:57:03Z

Alternative to #144933: when invoking a tail call with a non-scalar argument, we need to delay freeing the caller's local variables until after the callee is initialized, so that we can copy things from the caller to the callee.

Fixes #144820... but as the FIXMEs in the code show, it's not clear to me whether these are the right semantics.
r? @WaffleLapkin

rustbot · 2025-12-28T23:57:05Z

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

The Miri subtree was changed

cc @rust-lang/miri

bors · 2026-01-01T19:04:16Z

☔ The latest upstream changes (presumably #150565) made this pull request unmergeable. Please resolve the merge conflicts.

folkertdev · 2026-01-11T14:51:21Z

I looked into this some at #144855 (comment). In summary, I believe that a tail call that uses an argument that is PassMode::Indirect { on_stack: false } fundamentally cannot work. That value lives on the caller's stack, and the callee has no space where that argument can be stored. The argument stack value becomes dead upon the call.

Supporting PassMode::Indirect { on_stack: true } should be possible once LLVM supports it on more targets. As of LLVM 21 at least aarch64 and s390x support that, riscv and x86 don't yet.

RalfJung · 2026-01-11T18:15:56Z

Thanks, I had not seen that issue! What a mess. Whatever rules we come up with should be portable to all targets...

Anyway, that's mostly orthogonal to what we do in the interpreter. Fixing the order in which we copy arguments and deallocate the old locals still makes sense IMO.

RalfJung · 2026-01-31T14:14:48Z

@WaffleLapkin just a friendly ping -- are you okay with reviewing this or should I find another reviewer?

WaffleLapkin

Sorry for such a delayed review. This looks good to me; r=me after resolving conflicts.

View changes since this review

WaffleLapkin · 2026-02-17T14:22:24Z

src/tools/miri/tests/fail/tail_calls/dangling-local-var.stderr

 help: ALLOC was deallocated here:
  --> tests/fail/tail_calls/dangling-local-var.rs:LL:CC
   |
-LL |     f(std::ptr::null());
-   |     ^^^^^^^^^^^^^^^^^^^
+LL |     let _val = unsafe { *x };
+   |         ^^^^


arguably neither the old nor the new diagnostic is correct...

Fair, though the new one is really strange... good point.

The old one kind of makes sense -- doing a tail call deallocates the locals of the old stack frame. (Though it'd be better to point at the become.)

Oh wait, it used to point into main? Yeah that's nonsense.^^

rustbot · 2026-02-17T21:18:09Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

RalfJung · 2026-02-17T21:31:46Z

@bors r=WaffleLapkin

rust-bors · 2026-02-17T21:31:49Z

📌 Commit ad5108e has been approved by WaffleLapkin

It is now in the queue for this repository.

@WaffleLapkin

…ffleLapkin tail calls: fix copying non-scalar arguments to callee Alternative to rust-lang#144933: when invoking a tail call with a non-scalar argument, we need to delay freeing the caller's local variables until after the callee is initialized, so that we can copy things from the caller to the callee. Fixes rust-lang#144820... but as the FIXMEs in the code show, it's not clear to me whether these are the right semantics. r? @WaffleLapkin

@WaffleLapkin

…ffleLapkin tail calls: fix copying non-scalar arguments to callee Alternative to rust-lang#144933: when invoking a tail call with a non-scalar argument, we need to delay freeing the caller's local variables until after the callee is initialized, so that we can copy things from the caller to the callee. Fixes rust-lang#144820... but as the FIXMEs in the code show, it's not clear to me whether these are the right semantics. r? @WaffleLapkin

@WaffleLapkin

…ffleLapkin tail calls: fix copying non-scalar arguments to callee Alternative to rust-lang#144933: when invoking a tail call with a non-scalar argument, we need to delay freeing the caller's local variables until after the callee is initialized, so that we can copy things from the caller to the callee. Fixes rust-lang#144820... but as the FIXMEs in the code show, it's not clear to me whether these are the right semantics. r? @WaffleLapkin

Rollup of 19 pull requests Successful merges: - #145399 (Unify wording of resolve error) - #150473 (tail calls: fix copying non-scalar arguments to callee) - #152637 (Add a note about elided lifetime) - #152657 (std: move `exit` out of PAL) - #152729 (compiler: Don't mark `SingleUseConsts` MIR pass as "required for soundness") - #152753 (remove the explicit error for old `rental` versions) - #152758 (Remove ShallowInitBox.) - #151530 (Fix invalid `mut T` suggestion for `&mut T` in missing lifetime error) - #152179 (Add documentation note about signed overflow direction) - #152474 (Implement opt-bisect-limit for MIR) - #152509 (tests/ui/test-attrs: add annotations for reference rules) - #152672 (std: use libc version of `_NSGetArgc`/`_NSGetArgv`) - #152711 (resolve: Disable an assert that no longer holds) - #152732 (add regression test for 147958) - #152745 (Fix ICE in `suggest_param_env_shadowing` with incompatible args) - #152749 (make `rustc_allow_const_fn_unstable` an actual `rustc_attrs` attribute) - #152756 (Miri: recursive validity: also recurse into Boxes) - #152770 (carryless_mul: mention the base) - #152778 (Update tracking issue number for final_associated_functions)

Rollup of 20 pull requests Successful merges: - #145399 (Unify wording of resolve error) - #150473 (tail calls: fix copying non-scalar arguments to callee) - #152637 (Add a note about elided lifetime) - #152729 (compiler: Don't mark `SingleUseConsts` MIR pass as "required for soundness") - #152751 (Rename dep node "fingerprints" to distinguish key and value hashes) - #152753 (remove the explicit error for old `rental` versions) - #152758 (Remove ShallowInitBox.) - #151530 (Fix invalid `mut T` suggestion for `&mut T` in missing lifetime error) - #152179 (Add documentation note about signed overflow direction) - #152474 (Implement opt-bisect-limit for MIR) - #152509 (tests/ui/test-attrs: add annotations for reference rules) - #152672 (std: use libc version of `_NSGetArgc`/`_NSGetArgv`) - #152711 (resolve: Disable an assert that no longer holds) - #152725 (Rework explanation of CLI lint level flags) - #152732 (add regression test for 147958) - #152745 (Fix ICE in `suggest_param_env_shadowing` with incompatible args) - #152749 (make `rustc_allow_const_fn_unstable` an actual `rustc_attrs` attribute) - #152756 (Miri: recursive validity: also recurse into Boxes) - #152770 (carryless_mul: mention the base) - #152778 (Update tracking issue number for final_associated_functions)

@WaffleLapkin

Rollup merge of #150473 - RalfJung:interpret-tail-call, r=WaffleLapkin tail calls: fix copying non-scalar arguments to callee Alternative to #144933: when invoking a tail call with a non-scalar argument, we need to delay freeing the caller's local variables until after the callee is initialized, so that we can copy things from the caller to the callee. Fixes #144820... but as the FIXMEs in the code show, it's not clear to me whether these are the right semantics. r? @WaffleLapkin

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 Dec 28, 2025

rustbot assigned WaffleLapkin Dec 28, 2025

This comment has been minimized.

Sign in to view

RalfJung force-pushed the interpret-tail-call branch from 3604bf5 to 0d2a400 Compare December 29, 2025 08:49

RalfJung force-pushed the interpret-tail-call branch from 0d2a400 to 1bccfe8 Compare January 2, 2026 09:54

This comment has been minimized.

Sign in to view

RalfJung force-pushed the interpret-tail-call branch from 1bccfe8 to b3bdf0f Compare January 11, 2026 18:12

This comment has been minimized.

Sign in to view

WaffleLapkin approved these changes Feb 17, 2026

View reviewed changes

WaffleLapkin 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 Feb 17, 2026

tail calls: fix copying non-scalar arguments to callee

ad5108e

RalfJung force-pushed the interpret-tail-call branch from b3bdf0f to ad5108e Compare February 17, 2026 21:18

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 17, 2026

jhpratt mentioned this pull request Feb 18, 2026

Rollup of 12 pull requests #152776

Closed

jhpratt mentioned this pull request Feb 18, 2026

Rollup of 20 pull requests #152781

Closed

jhpratt mentioned this pull request Feb 18, 2026

Rollup of 19 pull requests #152782

Closed

Zalathar mentioned this pull request Feb 18, 2026

Rollup of 20 pull requests #152785

Merged

rust-bors bot merged commit b1c72fb into rust-lang:main Feb 18, 2026
11 checks passed

RalfJung deleted the interpret-tail-call branch February 18, 2026 17:21

Uh oh!

Conversation

RalfJung commented Dec 28, 2025

Uh oh!

rustbot commented Dec 28, 2025

Uh oh!

This comment has been minimized.

bors commented Jan 1, 2026

Uh oh!

This comment has been minimized.

folkertdev commented Jan 11, 2026

Uh oh!

This comment has been minimized.

RalfJung commented Jan 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

RalfJung commented Jan 31, 2026

Uh oh!

This comment has been minimized.

WaffleLapkin 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!

WaffleLapkin Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

RalfJung Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

RalfJung Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

rustbot commented Feb 17, 2026

Uh oh!

RalfJung commented Feb 17, 2026

Uh oh!

rust-bors bot commented Feb 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Comments

RalfJung commented Jan 11, 2026 •

edited

Loading

WaffleLapkin left a comment •

edited by rustbot

Loading