add `extern "tail"` calling convention by folkertdev · Pull Request #157016 · rust-lang/rust

folkertdev · 2026-05-27T10:26:07Z

View all comments

Maps to LLVM's tailcc and hence inherits all its problems.

This calling convention at least compiles on:

x86_64
aarch64

It also compiles on

s390x
loongarch64

But I've been told by target maintainers that this is kind of unintentional, and it's not tested in LLVM.

For most backends LLVM does not have support and fails loudly.

folkertdev

This PR does not yet lift the same-signature restriction for this calling convention. Also tailcc is currently kind of broken on x86_64 due to llvm/llvm-project#199224 (which I caused, whoops). There is an open PR for a fix that should make it into LLVM 23.

cc @scottmcm as our T-lang contact

View changes since this review

folkertdev · 2026-06-02T13:40:32Z

 RustCold =><= "rust-cold",
 RustInvalid =><= "rust-invalid",
 RustPreserveNone =><= "rust-preserve-none",
+ RustTail =><= "tail",


I'm not entirely sure what the most consistent naming scheme is here, but we can change it later. Waffle preferred the shorter "tail" over "rust-tail" because this is meant to be a user-facing calling convention eventually.

rustbot · 2026-06-02T13:45:06Z

The GCC codegen subtree was changed

cc @antoyo, @GuillaumeGomez

This PR changes rustc_public

cc @oli-obk, @celinval, @ouz-a, @makai410

The Cranelift subtree was changed

cc @bjorn3

These commits modify compiler targets.
(See the Target Tier Policy.)

WaffleLapkin · 2026-06-04T11:30:03Z

Pre-existing, but I'm confused why names in abi.rs and ty.rs are different...
View changes since the review

Not sure, that's kind of up to the rustc_public maintainers.

…uwer Rollup of 10 pull requests Successful merges: - #157016 (add `extern "tail"` calling convention) - #157264 (diagnostics: Fix ICE building a trait ref in method suggestions) - #157488 (compiletest: inject `#![windows_subsystem = "windows"]` to debuginfo tests on Windows) - #157509 (remove solaris implementation for File::lock, it has the wrong semantics) - #157521 (Rename `SyncView::{as_pin => as_pin_ref}`) - #156136 (Move tests box) - #157289 (Add infallible primitive type lookups to template arg resolver) - #157365 (Revert "LLVM 23: Run AssignGUIDPass in some places") - #157494 (Convert `QueryRegionConstraint` into a struct) - #157526 (std tests: skip a slow test on Miri)

…eLapkin add `extern "tail"` calling convention Maps to LLVM's `tailcc` and hence inherits all its problems. This calling convention at least compiles on: - `x86_64` - `aarch64` It also compiles on - `s390x` - `loongarch64` But I've been told by target maintainers that this is kind of unintentional, and it's not tested in LLVM. For most backends LLVM does not have support and fails loudly.

…uwer Rollup of 17 pull requests Successful merges: - #157251 (`rust-analyzer` subtree update) - #157533 (Subtree sync for rustc_codegen_cranelift) - #154742 (Add APIs for case folding to the standard library) - #155144 (mir_build: Add an extra intermediate step in MIR building for patterns ) - #157016 (add `extern "tail"` calling convention) - #157264 (diagnostics: Fix ICE building a trait ref in method suggestions) - #157386 (Parse deprecated note links separately in rustc_resolve) - #157483 (fix windows-gnu TLS leak) - #157488 (compiletest: inject `#![windows_subsystem = "windows"]` to debuginfo tests on Windows) - #157509 (remove solaris implementation for File::lock, it has the wrong semantics) - #157521 (Rename `SyncView::{as_pin => as_pin_ref}`) - #156136 (Move tests box) - #157365 (Revert "LLVM 23: Run AssignGUIDPass in some places") - #157471 (Debug assert that parsed attributes are in the `BUILTIN_ATTRIBUTE_MAP`) - #157485 (Rename `errors.rs` file to `diagnostics.rs` (1/N)) - #157494 (Convert `QueryRegionConstraint` into a struct) - #157526 (std tests: skip a slow test on Miri) Failed merges: - #155527 (Replace printables table with `unicode_data.rs` tables)

…eLapkin add `extern "tail"` calling convention Maps to LLVM's `tailcc` and hence inherits all its problems. This calling convention at least compiles on: - `x86_64` - `aarch64` It also compiles on - `s390x` - `loongarch64` But I've been told by target maintainers that this is kind of unintentional, and it's not tested in LLVM. For most backends LLVM does not have support and fails loudly.

Rollup of 25 pull requests Successful merges: - #157251 (`rust-analyzer` subtree update) - #157533 (Subtree sync for rustc_codegen_cranelift) - #154742 (Add APIs for case folding to the standard library) - #155144 (mir_build: Add an extra intermediate step in MIR building for patterns ) - #156222 (Stabilize `Result::map_or_default` and `Option::map_or_default`) - #157016 (add `extern "tail"` calling convention) - #157264 (diagnostics: Fix ICE building a trait ref in method suggestions) - #157386 (Parse deprecated note links separately in rustc_resolve) - #157483 (fix windows-gnu TLS leak) - #157488 (compiletest: inject `#![windows_subsystem = "windows"]` to debuginfo tests on Windows) - #157509 (remove solaris implementation for File::lock, it has the wrong semantics) - #157521 (Rename `SyncView::{as_pin => as_pin_ref}`) - #156136 (Move tests box) - #156573 (Add unwinder_private_data_size for wasm64 target) - #156783 (docs: make `Rc::into_raw` clickable in `Rc::increment_strong_count` doc) - #156840 (Stabilize `PathBuf::into_string`) - #156936 (Remove FIXME about impl PinCoerceUnsized for UnsafePinned<T>) - #157365 (Revert "LLVM 23: Run AssignGUIDPass in some places") - #157380 (clarify compiler_fence (and fence) docs) - #157471 (Debug assert that parsed attributes are in the `BUILTIN_ATTRIBUTE_MAP`) - #157485 (Rename `errors.rs` file to `diagnostics.rs` (1/N)) - #157494 (Convert `QueryRegionConstraint` into a struct) - #157526 (std tests: skip a slow test on Miri) - #157531 (ci: bump x86_64-gnu base image to 26.04) - #157556 (Add `BTree::append()` change to 1.96.0 relnotes) Failed merges: - #155527 (Replace printables table with `unicode_data.rs` tables)

Rollup merge of #157016 - folkertdev:add-tailcc-abi, r=WaffleLapkin add `extern "tail"` calling convention Maps to LLVM's `tailcc` and hence inherits all its problems. This calling convention at least compiles on: - `x86_64` - `aarch64` It also compiles on - `s390x` - `loongarch64` But I've been told by target maintainers that this is kind of unintentional, and it's not tested in LLVM. For most backends LLVM does not have support and fails loudly.

Rollup of 25 pull requests Successful merges: - rust-lang/rust#157251 (`rust-analyzer` subtree update) - rust-lang/rust#157533 (Subtree sync for rustc_codegen_cranelift) - rust-lang/rust#154742 (Add APIs for case folding to the standard library) - rust-lang/rust#155144 (mir_build: Add an extra intermediate step in MIR building for patterns ) - rust-lang/rust#156222 (Stabilize `Result::map_or_default` and `Option::map_or_default`) - rust-lang/rust#157016 (add `extern "tail"` calling convention) - rust-lang/rust#157264 (diagnostics: Fix ICE building a trait ref in method suggestions) - rust-lang/rust#157386 (Parse deprecated note links separately in rustc_resolve) - rust-lang/rust#157483 (fix windows-gnu TLS leak) - rust-lang/rust#157488 (compiletest: inject `#![windows_subsystem = "windows"]` to debuginfo tests on Windows) - rust-lang/rust#157509 (remove solaris implementation for File::lock, it has the wrong semantics) - rust-lang/rust#157521 (Rename `SyncView::{as_pin => as_pin_ref}`) - rust-lang/rust#156136 (Move tests box) - rust-lang/rust#156573 (Add unwinder_private_data_size for wasm64 target) - rust-lang/rust#156783 (docs: make `Rc::into_raw` clickable in `Rc::increment_strong_count` doc) - rust-lang/rust#156840 (Stabilize `PathBuf::into_string`) - rust-lang/rust#156936 (Remove FIXME about impl PinCoerceUnsized for UnsafePinned<T>) - rust-lang/rust#157365 (Revert "LLVM 23: Run AssignGUIDPass in some places") - rust-lang/rust#157380 (clarify compiler_fence (and fence) docs) - rust-lang/rust#157471 (Debug assert that parsed attributes are in the `BUILTIN_ATTRIBUTE_MAP`) - rust-lang/rust#157485 (Rename `errors.rs` file to `diagnostics.rs` (1/N)) - rust-lang/rust#157494 (Convert `QueryRegionConstraint` into a struct) - rust-lang/rust#157526 (std tests: skip a slow test on Miri) - rust-lang/rust#157531 (ci: bump x86_64-gnu base image to 26.04) - rust-lang/rust#157556 (Add `BTree::append()` change to 1.96.0 relnotes) Failed merges: - rust-lang/rust#155527 (Replace printables table with `unicode_data.rs` tables)

@bors

mention in the `extern "tail"` error that it's supported on x86 rust-lang#157016 (comment) r? folkertdev @bors rollup

@bors

mention in the `extern "tail"` error that it's supported on x86 rust-lang#157016 (comment) r? folkertdev @bors rollup

@bors

Rollup merge of #157919 - WaffleLapkin:x86tc, r=folkertdev mention in the `extern "tail"` error that it's supported on x86 #157016 (comment) r? folkertdev @bors rollup

…eLapkin add `extern "tail"` calling convention Maps to LLVM's `tailcc` and hence inherits all its problems. This calling convention at least compiles on: - `x86_64` - `aarch64` It also compiles on - `s390x` - `loongarch64` But I've been told by target maintainers that this is kind of unintentional, and it's not tested in LLVM. For most backends LLVM does not have support and fails loudly.

This comment has been minimized.
Sign in to view

bjorn3 reviewed May 27, 2026
View reviewed changes

Comment threadcompiler/rustc_codegen_cranelift/src/abi/mod.rs Outdated

folkertdev force-pushed the add-tailcc-abi branch from b188b60 to 4d12202Compare May 27, 2026 15:22

bjorn3 reviewed May 27, 2026
View reviewed changes

Comment threadcompiler/rustc_codegen_cranelift/src/abi/mod.rs Outdated

This comment has been minimized.
Sign in to view

folkertdev force-pushed the add-tailcc-abi branch from 4d12202 to c3e1d05Compare May 28, 2026 16:57

This comment has been minimized.
Sign in to view

folkertdev force-pushed the add-tailcc-abi branch from c3e1d05 to b4be083Compare May 29, 2026 22:22

This comment has been minimized.
Sign in to view

folkertdev force-pushed the add-tailcc-abi branch from b4be083 to de85385Compare May 29, 2026 23:09

This comment has been minimized.
Sign in to view

folkertdev force-pushed the add-tailcc-abi branch from de85385 to 35e3d57Compare May 30, 2026 09:25

WaffleLapkin self-assigned this Jun 1, 2026

folkertdev force-pushed the add-tailcc-abi branch from 35e3d57 to 433e4c2Compare June 2, 2026 13:26

folkertdev commented Jun 2, 2026
View reviewed changes

folkertdev marked this pull request as ready for review June 2, 2026 13:45

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 Jun 2, 2026

folkertdev changed the title ~~add extern "rust-tail" calling convention~~add extern "tail" calling conventionJun 2, 2026

antoyo reviewed Jun 2, 2026
View reviewed changes

Comment threadcompiler/rustc_codegen_gcc/src/abi.rs Outdated

This comment has been minimized.
Sign in to view

folkertdev force-pushed the add-tailcc-abi branch from 433e4c2 to 7d8191aCompare June 2, 2026 14:07

This comment has been minimized.
Sign in to view

folkertdev force-pushed the add-tailcc-abi branch from 7d8191a to c19423cCompare June 2, 2026 15:01

WaffleLapkin mentioned this pull request Jun 4, 2026
Tracking Issue for extern "tail" calling convention #157427
Open
7 tasks

WaffleLapkin requested changes Jun 4, 2026
View reviewed changes

JonathanBrouwer mentioned this pull request Jun 6, 2026
Rollup of 10 pull requests #157528
Closed

JonathanBrouwer mentioned this pull request Jun 6, 2026
Rollup of 10 pull requests #157530
Closed

JonathanBrouwer mentioned this pull request Jun 6, 2026
Rollup of 17 pull requests #157536
Closed

JonathanBrouwer mentioned this pull request Jun 6, 2026
Rollup of 17 pull requests #157541
Closed

jhpratt mentioned this pull request Jun 7, 2026
Rollup of 21 pull requests #157554
Closed

This was referenced Jun 7, 2026
Rollup of 22 pull requests #157555
Closed
Rollup of 25 pull requests #157558
Merged

rust-borsBot merged commit d66ac24 into rust-lang:mainJun 7, 2026
13 of 14 checks passed

rustbot added this to the 1.98.0 milestone Jun 7, 2026

RalfJung reviewed Jun 11, 2026
View reviewed changes

Comment threadcompiler/rustc_codegen_llvm/src/abi.rs

folkertdev mentioned this pull request Jun 15, 2026
Explicit tail calls & loop_matchrust-lang/goals#634
Open

WaffleLapkin mentioned this pull request Jun 15, 2026
mention in the extern "tail" error that it's supported on x86 #157919
Merged

GuillaumeGomez added a commit to GuillaumeGomez/rust that referenced this pull request Jun 15, 2026
Rollup merge ofrust-lang#157919- WaffleLapkin:x86tc, r=folkertdev
89167ef
mention in the `extern "tail"` error that it's supported on x86 rust-lang#157016 (comment) r? folkertdev @bors rollup

jhpratt added a commit to jhpratt/rust that referenced this pull request Jun 15, 2026
Rollup merge ofrust-lang#157919- WaffleLapkin:x86tc, r=folkertdev
ad55d52
mention in the `extern "tail"` error that it's supported on x86 rust-lang#157016 (comment) r? folkertdev @bors rollup

rust-timer added a commit that referenced this pull request Jun 16, 2026
Unrolled build for#157919
c746055
Rollup merge of #157919 - WaffleLapkin:x86tc, r=folkertdev mention in the `extern "tail"` error that it's supported on x86 #157016 (comment) r? folkertdev @bors rollup

rust-timer mentioned this pull request Jul 22, 2026
Rollup of 11 pull requests #159728
Merged

Navigation Menu

Uh oh!

Uh oh!

add `extern "tail"` calling convention - #157016

add `extern "tail"` calling convention#157016
rust-bors[bot] merged 1 commit into
rust-lang:mainfrom
folkertdev:add-tailcc-abi

folkertdev commented May 27, 2026•
edited by rustbot
Loading

Uh oh!

This comment has been minimized.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

folkertdev left a comment •
edited by rustbot
Loading

Uh oh!

folkertdevJun 2, 2026

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

Uh oh!

Uh oh!

WaffleLapkinJun 4, 2026•
edited by rustbot
Loading

Uh oh!

folkertdevJun 4, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

Uh oh!

Uh oh!

Conversation

folkertdev commented May 27, 2026• edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

folkertdev 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!

folkertdevJun 2, 2026

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

Uh oh!

Uh oh!

WaffleLapkinJun 4, 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!

folkertdevJun 4, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

folkertdev commented May 27, 2026•
edited by rustbot
Loading

folkertdev left a comment •
edited by rustbot
Loading

WaffleLapkinJun 4, 2026•
edited by rustbot
Loading