Skip to content

libc: uncomment items that were commented due to old ctest #4682

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 6, 2025
+4 −74
Merged

libc: uncomment items that were commented due to old ctest #4682

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 0 additions & 3 deletions src/new/linux_uapi/linux/can.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,9 +1,6 @@
//! Header: `uapi/linux/can.h`

// FIXME(ctest): we shouldn't have to specify the path but garando doesn't find modules otherwise
#[path = "can/j1939.rs"]
pub(crate) mod j1939;
#[path = "can/raw.rs"]
pub(crate) mod raw;

pub use j1939::*;
Expand Down
64 changes: 0 additions & 64 deletions src/primitives.rs
View file
Open in desktop
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry for the confusion; to clarify a bit, for #4682 (comment) I only meant that the quoted bit (long explanation + checks) can be deleted. __u?int128(_t)? need to stay around on the same platforms, so the cfg_if block needs to stay.

This should let you drop the build.rs changes

Original file line number Diff line number Diff line change
Expand Up @@ -99,35 +99,6 @@ cfg_if! {
target_os = "watchos"
))
))] {
// This introduces partial support for FFI with __int128 and
// equivalent types on platforms where Rust's definition is validated
// to match the standard C ABI of that platform.
//
// Rust does not guarantee u128/i128 are sound for FFI, and its
// definitions are in fact known to be incompatible. [0]
//
// However these problems aren't fundamental, and are just platform
// inconsistencies. Specifically at the time of this writing:
//
// * For x64 SysV ABIs (everything but Windows), the types are underaligned.
// * For all Windows ABIs, Microsoft doesn't actually officially define __int128,
// and as a result different implementations don't actually agree on its ABI.
//
// But on the other major aarch64 platforms (android, linux, ios, macos) we have
// validated that rustc has the right ABI for these types. This is important because
// aarch64 uses these types in some fundamental OS types like user_fpsimd_struct,
// which represents saved simd registers.
//
// Any API which uses these types will need to `#[ignore(improper_ctypes)]`
// until the upstream rust issue is resolved, but this at least lets us make
// progress on platforms where this type is important.
//
// The list of supported architectures and OSes is intentionally very restricted,
// as careful work needs to be done to verify that a particular platform
// has a conformant ABI.
//
// [0]: https://github.com/rust-lang/rust/issues/54341

/// C `__int128` (a GCC extension that's part of many ABIs)
pub type __int128 = i128;
/// C `unsigned __int128` (a GCC extension that's part of many ABIs)
Expand All @@ -136,41 +107,6 @@ cfg_if! {
pub type __int128_t = i128;
/// C __uint128_t (alternate name for [__uint128][])
pub type __uint128_t = u128;

// NOTE: if you add more platforms to here, you may need to cfg
// these consts. They should always match the platform's values
// for `sizeof(__int128)` and `_Alignof(__int128)`.
const _SIZE_128: usize = 16;
const _ALIGN_128: usize = 16;

// FIXME(ctest): ctest doesn't handle `_` as an identifier so these tests are temporarily
// disabled.
// macro_rules! static_assert_eq {
// ($a:expr, $b:expr) => {
// const _: [(); $a] = [(); $b];
// };
// }
//
// // Since Rust doesn't officially guarantee that these types
// // have compatible ABIs, we const assert that these values have the
// // known size/align of the target platform's libc. If rustc ever
// // tries to regress things, it will cause a compilation error.
// //
// // This isn't a bullet-proof solution because e.g. it doesn't
// // catch the fact that llvm and gcc disagree on how x64 __int128
// // is actually *passed* on the stack (clang underaligns it for
// // the same reason that rustc *never* properly aligns it).
// static_assert_eq!(size_of::<__int128>(), _SIZE_128);
// static_assert_eq!(align_of::<__int128>(), _ALIGN_128);

// static_assert_eq!(size_of::<__uint128>(), _SIZE_128);
// static_assert_eq!(align_of::<__uint128>(), _ALIGN_128);

// static_assert_eq!(size_of::<__int128_t>(), _SIZE_128);
// static_assert_eq!(align_of::<__int128_t>(), _ALIGN_128);

// static_assert_eq!(size_of::<__uint128_t>(), _SIZE_128);
// static_assert_eq!(align_of::<__uint128_t>(), _ALIGN_128);
} else if #[cfg(all(
target_arch = "aarch64",
any(
Expand Down
11 changes: 4 additions & 7 deletions src/unix/linux_like/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1783,13 +1783,10 @@ cfg_if! {

/// Build an ioctl number, analogous to the C macro of the same name.
const fn _IOC(dir: u32, ty: u32, nr: u32, size: usize) -> Ioctl {
// FIXME(ctest) the `garando_syntax` crate (used by ctest in the CI test suite)
// cannot currently parse these `debug_assert!`s
//
// debug_assert!(dir <= _IOC_DIRMASK);
// debug_assert!(ty <= _IOC_TYPEMASK);
// debug_assert!(nr <= _IOC_NRMASK);
// debug_assert!(size <= (_IOC_SIZEMASK as usize));
core::debug_assert!(dir <= _IOC_DIRMASK);
core::debug_assert!(ty <= _IOC_TYPEMASK);
core::debug_assert!(nr <= _IOC_NRMASK);
core::debug_assert!(size <= (_IOC_SIZEMASK as usize));

((dir << _IOC_DIRSHIFT)
| (ty << _IOC_TYPESHIFT)
Expand Down