ICE: malformed malformed repr(align(N))

[matthiaskrgr]

Code

trait MyTrait {
    #[repr(align)]
    fn myfun();
}

Meta

rustc --version --verbose:

rustc 1.84.0-nightly (759e07f06 2024-10-30)
binary: rustc
commit-hash: 759e07f063fb8e6306ff1bdaeb70af56a878b415
commit-date: 2024-10-30
host: x86_64-unknown-linux-gnu
release: 1.84.0-nightly
LLVM version: 19.1.1

Error output

<output>

Backtrace

warning: trait `MyTrait` is never used
 --> a.rs:1:7
  |
1 | trait MyTrait {
  |       ^^^^^^^
  |
  = note: `#[warn(dead_code)]` on by default

warning: 1 warning emitted

note: no errors encountered even though delayed bugs were created

note: those delayed bugs will now be shown as internal compiler errors

error: internal compiler error: malformed repr(align(N))
 --> a.rs:2:12
  |
2 |     #[repr(align)]
  |            ^^^^^
  |
note: delayed at compiler/rustc_passes/src/check_attr.rs:1932:28
         0: <rustc_errors::DiagCtxtInner>::emit_diagnostic
         1: <rustc_errors::DiagCtxtHandle>::emit_diagnostic
         2: <rustc_span::ErrorGuaranteed as rustc_errors::diagnostic::EmissionGuarantee>::emit_producing_guarantee
         3: <rustc_errors::DiagCtxtHandle>::span_delayed_bug::<rustc_span::span_encoding::Span, &str>
         4: <rustc_passes::check_attr::CheckAttrVisitor>::check_attributes
         5: <rustc_passes::check_attr::CheckAttrVisitor as rustc_hir::intravisit::Visitor>::visit_trait_item
         6: rustc_passes::check_attr::check_mod_attrs
         7: rustc_query_impl::plumbing::__rust_begin_short_backtrace::<rustc_query_impl::query_impl::check_mod_attrs::dynamic_query::{closure#2}::{closure#0}, rustc_middle::query::erase::Erased<[u8; 0]>>
         8: rustc_query_system::query::plumbing::try_execute_query::<rustc_query_impl::DynamicConfig<rustc_query_system::query::caches::DefaultCache<rustc_span::def_id::LocalModDefId, rustc_middle::query::erase::Erased<[u8; 0]>>, false, false, false>, rustc_query_impl::plumbing::QueryCtxt, false>
         9: rustc_query_impl::query_impl::check_mod_attrs::get_query_non_incr::__rust_end_short_backtrace
        10: <rustc_middle::hir::map::Map>::par_for_each_module::<rustc_interface::passes::run_required_analyses::{closure#0}::{closure#0}::{closure#1}::{closure#0}>::{closure#0}
        11: rustc_interface::passes::run_required_analyses
        12: rustc_interface::passes::analysis
        13: rustc_query_impl::plumbing::__rust_begin_short_backtrace::<rustc_query_impl::query_impl::analysis::dynamic_query::{closure#2}::{closure#0}, rustc_middle::query::erase::Erased<[u8; 1]>>
        14: rustc_query_system::query::plumbing::try_execute_query::<rustc_query_impl::DynamicConfig<rustc_query_system::query::caches::SingleCache<rustc_middle::query::erase::Erased<[u8; 1]>>, false, false, false>, rustc_query_impl::plumbing::QueryCtxt, false>
        15: rustc_query_impl::query_impl::analysis::get_query_non_incr::__rust_end_short_backtrace
        16: rustc_interface::interface::run_compiler::<core::result::Result<(), rustc_span::ErrorGuaranteed>, rustc_driver_impl::run_compiler::{closure#0}>::{closure#1}
        17: std::sys::backtrace::__rust_begin_short_backtrace::<rustc_interface::util::run_in_thread_with_globals<rustc_interface::util::run_in_thread_pool_with_globals<rustc_interface::interface::run_compiler<core::result::Result<(), rustc_span::ErrorGuaranteed>, rustc_driver_impl::run_compiler::{closure#0}>::{closure#1}, core::result::Result<(), rustc_span::ErrorGuaranteed>>::{closure#0}, core::result::Result<(), rustc_span::ErrorGuaranteed>>::{closure#0}::{closure#0}, core::result::Result<(), rustc_span::ErrorGuaranteed>>
        18: <<std::thread::Builder>::spawn_unchecked_<rustc_interface::util::run_in_thread_with_globals<rustc_interface::util::run_in_thread_pool_with_globals<rustc_interface::interface::run_compiler<core::result::Result<(), rustc_span::ErrorGuaranteed>, rustc_driver_impl::run_compiler::{closure#0}>::{closure#1}, core::result::Result<(), rustc_span::ErrorGuaranteed>>::{closure#0}, core::result::Result<(), rustc_span::ErrorGuaranteed>>::{closure#0}::{closure#0}, core::result::Result<(), rustc_span::ErrorGuaranteed>>::{closure#1} as core::ops::function::FnOnce<()>>::call_once::{shim:vtable#0}
        19: std::sys::pal::unix::thread::Thread::new::thread_start
        20: <unknown>
        21: <unknown>
 --> a.rs:2:12
  |
2 |     #[repr(align)]
  |            ^^^^^

note: we would appreciate a bug report: https://github.com/rust-lang/rust/issues/new?labels=C-bug%2C+I-ICE%2C+T-compiler&template=ice.md

note: please make sure that you have updated to the latest nightly

note: please attach the file at `/tmp/im/rustc-ice-2024-10-31T07_01_17-3418840.txt` to your bug report

note: compiler flags: --crate-type lib

query stack during panic:
end of query stack

[jieyouxu]

Without performing an exact bisection, I think this is #131633.

[jieyouxu]

Actually, #131633 simply exposed a pre-existing bug: this doesn't even error on stable or beta, lmao. Actually there's like at least 2 separate bugs here:

1. I don't think it's allowed to use #[repr()] on a method, first of all. EDIT: based on Tracking Issue for #[repr(align(...))] on function items (fn_align) #82232, that should at least be feature-gated under fn_align.
   • To be pedantic, #[repr()] should additionally emit a warning about empty repr() doing, well, nothing. There should be already UI tests for that on some other targets, just need to also be the case for repr() on a trait's fn.
2. align requires a value.

[workingjubilee]

cargo bisect-rustc does pin it down to that:

• 66701c4
• rust-lang-ci@69dcce6

[jieyouxu]

Yes, #131633, because we should've emitted an error already by that point but didn't.

[workingjubilee]

...I guess this is technically a regression?

[jieyouxu]

The previous behavior is also unintentionally accepted code, I think.

[jieyouxu]

This is actually a deficiency in how attributes are currently handled, pending a work cf. rust-lang/compiler-team#796. This should not be spot-fixed, how this is even handled in the first place needs to be fixed (allow-list rather than deny-list), and preferably don't mix AST with HIR.

cf. https://rust-lang.zulipchat.com/#narrow/channel/233931-t-compiler.2Fmajor-changes/topic/Attribute.20handling.20reworks.20compiler-team.23796/near/479823942

[workingjubilee]

expecting this to be resolved either by #131808 or a followup

[jdonszelmann]

@rustbot claim

[jieyouxu]

Marking as P-low because it's accepting code that we shouldn't be accepting, the ICE is just "don't write malformed code" but yeah.

[asquared31415]

woo i uncovered a bug! i knew attribute handling was weird, but i'm shocked that it's possible to hit this code path without trying to parse the repr attribute anywhere else that would error.

[jdonszelmann]

@rustbot label +A-repr +A-align

[fmease]

(due diligence, cc #132693 (comment))