ICE in smol-potat 1.1.2 (data race detected in EnvVars::cleanup)

[saethlin]

#2020 is back. This one is a bit spooky though, the combination of RUST_BACKTRACE=0 and -Zmiri-panic-on-unsupported is required. But we don't see a panic. So perhaps the panic is being caught and we're exiting the interpreter in a strange state?

╭ ➜ ben@archlinux:/tmp/smol-potat-1.1.2
╰ ➤ RUST_BACKTRACE=0 MIRIFLAGS="-Zmiri-disable-isolation -Zmiri-panic-on-unsupported" cargo +miri miri test
Preparing a sysroot for Miri (target: x86_64-unknown-linux-gnu)... done
    Finished test [unoptimized + debuginfo] target(s) in 0.03s
     Running unittests src/lib.rs (target/miri/x86_64-unknown-linux-gnu/debug/deps/smol_potat-1908872ee5450113)

running 0 tests

test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

     Running tests/hygiene.rs (target/miri/x86_64-unknown-linux-gnu/debug/deps/hygiene-42b8ceeee13ed168)
warning: unused import: `::std::panic`
 --> tests/hygiene.rs:3:5
  |
3 | use ::std::panic;
  |     ^^^^^^^^^^^^
  |
  = note: `#[warn(unused_imports)]` on by default


running 1 test
test test ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

thread 'rustc' panicked at 'called `Result::unwrap()` on an `Err` value: InterpErrorInfo(InterpErrorInfoInner { kind: Data race detected between Deallocate on thread `main` and Read on thread `async-io` at alloc65+0xf, backtrace: None })', src/eval.rs:378:36
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

error: internal compiler error: unexpected panic

note: the compiler unexpectedly panicked. this is a bug.

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: rustc 1.65.0-nightly (e1b28cd2f 2022-08-19) running on x86_64-unknown-linux-gnu

note: compiler flags: -C embed-bitcode=no -C debuginfo=2 -C linker=clang -C incremental -C link-arg=-fuse-ld=mold -Z miri-disable-isolation -Z miri-panic-on-unsupported

note: some of the compiler flags provided by cargo are hidden

query stack during panic:
end of query stack
warning: 1 warning emitted

error: test failed, to rerun pass '--test hygiene'

╭ ➜ ben@archlinux:/tmp/futures-concurrency-3.0.2
╰ ➤ RUST_BACKTRACE=0 MIRIFLAGS="-Zmiri-panic-on-unsupported -Zmiri-disable-isolation" cargo +miri miri test
Preparing a sysroot for Miri (target: x86_64-unknown-linux-gnu)... done
    Finished test [unoptimized + debuginfo] target(s) in 0.01s
     Running unittests src/lib.rs (target/miri/x86_64-unknown-linux-gnu/debug/deps/futures_concurrency-1619211f71bea4c9)

running 5 tests
test merge::tuple::tests::merge_tuple_4 ... ok
test race::array::test::no_fairness ... ok
test race::tuple::test::no_fairness ... ok
test race::tuple::test::thruple ... ok
test race::vec::test::no_fairness ... ok

test result: ok. 5 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

thread 'rustc' panicked at 'called `Result::unwrap()` on an `Err` value: InterpErrorInfo(InterpErrorInfoInner { kind: Data race detected between Deallocate on thread `main` and Read on thread `async-io` at alloc151+0xf, backtrace: None })', src/eval.rs:381:36
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

error: internal compiler error: unexpected panic

note: the compiler unexpectedly panicked. this is a bug.

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: rustc 1.65.0-nightly (e1b28cd2f 2022-08-19) running on x86_64-unknown-linux-gnu

note: compiler flags: -C embed-bitcode=no -C debuginfo=2 -C linker=clang -C incremental -C link-arg=-fuse-ld=mold -Z miri-panic-on-unsupported -Z miri-disable-isolation

note: some of the compiler flags provided by cargo are hidden

query stack during panic:
end of query stack
error: test failed, to rerun pass '--lib'

miri 0.1.0 (d5853bc 2022-08-25)

[RalfJung]

That's wild. I have removed the "disable data race detection during env cleanup" in favor of "just don't do env cleanup unless all threads have terminated", so I have no idea how there can be a race here...

Interesting observation: with RUST_BACKTRACE=1, it instead errors saying

error: the main thread terminated without waiting for all remaining threads