Implement support for `become` and explicit tail call codegen for the LLVM backend #144232

xacrimon · 2025-07-20T22:46:59Z

This PR implements codegen of explicit tail calls via become in rustc_codegen_ssa and support within the LLVM backend. Completes a task on (#112788). This PR implements all the necessary bits to make explicit tail calls usable, other backends have received stubs for now and will ICE if you use become on them. I suspect there is some bikeshedding to be done on how we should go about implementing this for other backends, but it should be relatively straightforward for GCC after this is merged.

During development I also put together a POC bytecode VM based on tail call dispatch to test these changes out and analyze the codegen to make sure it generates expected assembly. That is available here.

rustbot · 2025-07-20T22:47:03Z

r? @oli-obk

rustbot has assigned @oli-obk.
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

rustbot · 2025-07-20T22:47:05Z

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo, @GuillaumeGomez

Some changes occurred in compiler/rustc_codegen_ssa

cc @WaffleLapkin

jeffparsons · 2025-07-21T03:41:46Z

I suspect there is quite a bit of bikeshedding on how we should go about implementing this for WASM

I thought LLVM already supported the Wasm tail-call instructions — what's the missing piece that needs bikeshedding?

oli-obk · 2025-07-21T06:56:48Z

@rustbot author

it seems a tailcall test is still hitting a stack overflow, do you know why?

rustbot · 2025-07-21T06:56:52Z

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

compiler/rustc_codegen_ssa/src/mir/block.rs

xacrimon · 2025-07-21T15:55:42Z

@rustbot author

it seems a tailcall test is still hitting a stack overflow, do you know why?

This was me mistakenly marking the rest @run-fail instead of @run-crash, that was intentional and it has now been fixed. That one is marked no-tce and doesn't use become. There is another one called tce instead of no-tce which uses become and is set to @run-pass.

xacrimon · 2025-07-21T16:36:03Z

I suspect there is quite a bit of bikeshedding on how we should go about implementing this for WASM

I thought LLVM already supported the Wasm tail-call instructions — what's the missing piece that needs bikeshedding?

Ah, my knowledge is outdated then. I've edited my comment, though point still stands for backends that aren't llvm/gcc.

xacrimon · 2025-07-21T16:38:28Z

@rustbot ready

xacrimon · 2025-07-21T19:31:44Z

@WaffleLapkin I'm going to go ahead and guess this is something you want to have a look at.

oli-obk · 2025-07-22T07:42:20Z

@bors r+

Thanks! Glad we finally support it!

bors · 2025-07-22T07:42:23Z

📌 Commit ecf34e1 has been approved by oli-obk

It is now in the queue for this repository.

xacrimon · 2025-07-24T16:05:07Z

@rustbot ready

Alright, that makes sense. It's been removed.

compiler/rustc_codegen_gcc/src/builder.rs

compiler/rustc_codegen_llvm/src/llvm/ffi.rs

compiler/rustc_llvm/llvm-wrapper/RustWrapper.cpp

tests/ui/explicit-tail-calls/recursion-tce.rs

compiler/rustc_codegen_ssa/src/mir/block.rs

WaffleLapkin · 2025-07-25T09:18:28Z

compiler/rustc_codegen_ssa/src/mir/block.rs

+            if tail && matches!(fn_abi.args[i].mode, PassMode::Indirect { .. }) {
+                span_bug!(
+                    fn_span,
+                    "arguments using PassMode::Indirect are currently not supported for tail calls"


#144232 (comment)

Also #144293

compiler/rustc_codegen_llvm/src/builder.rs

lqd · 2025-07-25T10:32:43Z

Could you also clean up the git history, to avoid the back-and-forths and simple fixups of later commits? Thanks.

oli-obk · 2025-07-25T19:05:38Z

r? WaffleLapkin

rustbot · 2025-07-25T19:05:42Z

WaffleLapkin is not on the review rotation at the moment.
They may take a while to respond.

xacrimon · 2025-07-25T19:22:08Z

Could you also clean up the git history, to avoid the back-and-forths and simple fixups of later commits? Thanks.

Done.

…nd the LLVM codegen backend.

WaffleLapkin · 2025-07-30T14:48:53Z

Thanks!
@bors r+

bors · 2025-07-30T14:48:56Z

📌 Commit a448837 has been approved by WaffleLapkin

It is now in the queue for this repository.

…fleLapkin Implement support for `become` and explicit tail call codegen for the LLVM backend This PR implements codegen of explicit tail calls via `become` in `rustc_codegen_ssa` and support within the LLVM backend. Completes a task on (rust-lang#112788). This PR implements all the necessary bits to make explicit tail calls usable, other backends have received stubs for now and will ICE if you use `become` on them. I suspect there is some bikeshedding to be done on how we should go about implementing this for other backends, but it should be relatively straightforward for GCC after this is merged. During development I also put together a POC bytecode VM based on tail call dispatch to test these changes out and analyze the codegen to make sure it generates expected assembly. That is available [here](https://github.com/xacrimon/tcvm).

Rollup of 6 pull requests Successful merges: - #135975 (Implement `push_mut`) - #143672 (Fix Box allocator drop elaboration) - #144232 (Implement support for `become` and explicit tail call codegen for the LLVM backend) - #144663 (coverage: Re-land "Enlarge empty spans during MIR instrumentation") - #144683 (Simplify library dependencies on `compiler-builtins`) - #144685 (Only extract lang items once in codegen_fn_attrs) r? `@ghost` `@rustbot` modify labels: rollup

Rollup of 6 pull requests Successful merges: - #135975 (Implement `push_mut`) - #143672 (Fix Box allocator drop elaboration) - #144232 (Implement support for `become` and explicit tail call codegen for the LLVM backend) - #144663 (coverage: Re-land "Enlarge empty spans during MIR instrumentation") - #144685 (Only extract lang items once in codegen_fn_attrs) - #144717 (Move `rustc_middle::parameterized`) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of #144232 - xacrimon:explicit-tail-call, r=WaffleLapkin Implement support for `become` and explicit tail call codegen for the LLVM backend This PR implements codegen of explicit tail calls via `become` in `rustc_codegen_ssa` and support within the LLVM backend. Completes a task on (#112788). This PR implements all the necessary bits to make explicit tail calls usable, other backends have received stubs for now and will ICE if you use `become` on them. I suspect there is some bikeshedding to be done on how we should go about implementing this for other backends, but it should be relatively straightforward for GCC after this is merged. During development I also put together a POC bytecode VM based on tail call dispatch to test these changes out and analyze the codegen to make sure it generates expected assembly. That is available [here](https://github.com/xacrimon/tcvm).

rustbot assigned oli-obk Jul 20, 2025

xacrimon force-pushed the explicit-tail-call branch from 6a6d3c2 to f912c90 Compare July 20, 2025 22:49

This comment has been minimized.

Sign in to view

xacrimon force-pushed the explicit-tail-call branch 2 times, most recently from b36cf15 to ac87434 Compare July 20, 2025 23:10

This comment has been minimized.

Sign in to view

xacrimon force-pushed the explicit-tail-call branch from 138b2a7 to 3fac279 Compare July 21, 2025 00:14

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jul 21, 2025

rustbot added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Jul 21, 2025

tmiasko reviewed Jul 21, 2025

View reviewed changes

compiler/rustc_codegen_ssa/src/mir/block.rs Outdated Show resolved Hide resolved

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 21, 2025

xacrimon requested review from bjorn3 and tmiasko July 21, 2025 18:37

xacrimon force-pushed the explicit-tail-call branch from 4028c67 to 58c6633 Compare July 24, 2025 16:05

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 24, 2025

WaffleLapkin requested changes Jul 25, 2025

View reviewed changes

rustbot assigned WaffleLapkin Jul 25, 2025

rustbot unassigned oli-obk Jul 25, 2025

xacrimon requested a review from WaffleLapkin July 25, 2025 19:21

xacrimon force-pushed the explicit-tail-call branch from 12c0f0d to ea59690 Compare July 25, 2025 19:22

xacrimon force-pushed the explicit-tail-call branch from ea59690 to 3a0a2da Compare July 25, 2025 19:29

Implement support for explicit tail calls in the MIR block builders a…

a448837

…nd the LLVM codegen backend.

xacrimon force-pushed the explicit-tail-call branch from 3a0a2da to a448837 Compare July 25, 2025 23:02

Borgerr mentioned this pull request Jul 29, 2025

Additional tce tests #144650

Open

WaffleLapkin approved these changes Jul 30, 2025

View reviewed changes

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 Jul 30, 2025

Zalathar mentioned this pull request Jul 31, 2025

Rollup of 6 pull requests #144709

Closed

Zalathar mentioned this pull request Jul 31, 2025

Rollup of 6 pull requests #144718

Merged

bors merged commit 8628b78 into rust-lang:master Jul 31, 2025
10 checks passed

rustbot added this to the 1.90.0 milestone Jul 31, 2025

Implement support for become and explicit tail call codegen for the LLVM backend #144232

Implement support for become and explicit tail call codegen for the LLVM backend #144232

Uh oh!

Conversation

xacrimon commented Jul 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jul 20, 2025

Uh oh!

rustbot commented Jul 20, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

jeffparsons commented Jul 21, 2025

Uh oh!

oli-obk commented Jul 21, 2025

Uh oh!

rustbot commented Jul 21, 2025

Uh oh!

Uh oh!

xacrimon commented Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

xacrimon commented Jul 21, 2025

Uh oh!

xacrimon commented Jul 21, 2025

Uh oh!

xacrimon commented Jul 21, 2025

Uh oh!

oli-obk commented Jul 22, 2025

Uh oh!

bors commented Jul 22, 2025

Uh oh!

xacrimon commented Jul 24, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

WaffleLapkin Jul 25, 2025

Choose a reason for hiding this comment

Uh oh!

xacrimon Jul 25, 2025

Choose a reason for hiding this comment

Uh oh!

xacrimon Jul 25, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

lqd commented Jul 25, 2025

Uh oh!

oli-obk commented Jul 25, 2025

Uh oh!

rustbot commented Jul 25, 2025

Uh oh!

xacrimon commented Jul 25, 2025

Uh oh!

WaffleLapkin commented Jul 30, 2025

Uh oh!

bors commented Jul 30, 2025

Uh oh!

Uh oh!

Uh oh!

Implement support for `become` and explicit tail call codegen for the LLVM backend #144232

Implement support for `become` and explicit tail call codegen for the LLVM backend #144232

xacrimon commented Jul 20, 2025 •

edited

Loading

xacrimon commented Jul 21, 2025 •

edited

Loading