test that target features given in the target spec are enabled

[RalfJung]

This currently fails... not sure if that is a regression, but it might be the same as rust-lang/rust#149314.
@bjorn3 do you think your PR will fix this?

[rustbot]

Thank you for contributing to Miri! A reviewer will take a look at your PR, typically within a week or two.
Please remember to not force-push to the PR branch except when you need to rebase due to a conflict or when the reviewer asks you for it.

[RalfJung]

I could just test this, I still have your PR checked out after all... no, turns out it does not fix this.

$ ./x run miri --target armv7-unknown-linux-gnueabihf -- ../miri/tests/pass/target-spec-implies-target-feature.rs
thread 'main' (1000) panicked at ../miri/tests/pass/target-spec-implies-target-feature.rs:6:5:
assertion failed: cfg!(target_feature = "vfp3")

[RalfJung]

However it also already fails before #4648. So we probably need a different test.

[bjorn3]

My PR should fix this for v7 and thumb2. vfp3 is not a target feature recognized by rustc.

warning: unstable feature specified for `-Ctarget-feature`: `vfp3`
  |
  = note: this feature is not stably supported; its behavior can change in the future

[RalfJung]

vfp3 is missing due to rust-lang/rust#149399.

[RalfJung]

FWIW I also tried std::arch::is_arm_feature_detected and that still fails even when I set the LLVM backend. Not sure what is up with that.

error: "v7" feature cannot be detected at run-time
 --> ../miri/tests/pass/target-spec-implies-target-feature.rs:7:13
  |
7 |     assert!(std::arch::is_arm_feature_detected!("v7"));
  |             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |

The macro docs make it sound like if the feature is available at compile time, the macro will work. Maybe I am misunderstanding the comment.

EDIT: I just misunderstood the code, these never work.

[RalfJung]

My PR should fix this for v7 and thumb2

Indeed it does, I confirmed that locally. Great!

[RalfJung]

I integrated this test into #4730.