Separate L4Re from Linux code, add aarch64 and enable tests by farao · Pull Request #4479 · rust-lang/libc

farao · 2025-06-09T17:11:05Z

The L4Re code was previously attached to the Linux code which was not correct in many ways. This commit separates the L4Re code and enables the libc-tests and includes the fixes for the failing tests. Aarch64 is added as a second supported architecture (more to come).

Sources

L4Re-adapted version of uclibc: https://github.com/kernkonzept/l4re-core/tree/master/uclibc/lib.

Checklist

Relevant tests in libc-test/semver have been updated
No placeholder or unstable values like *LAST or *MAX are
included (see #3131)
Tested locally (cd libc-test && cargo test --target mytarget);
especially relevant for platforms that may not be checked in CI

rustbot · 2025-06-09T17:11:12Z

Some changes occurred in the Android module

cc @maurer

farao · 2025-06-13T08:50:10Z

Hi @tgross35, this is the refactoring of the recent L4Re PR: #4383 (which is kept around just in case).

I think the failure of the freebsd nightly checks is not my fault. The rest succeeds now.

I did most of what you requested in your comment on the old PR, I just kept the linux/mod.rs file around since there is a massive part of code that is linux-only and is not supported by L4Re and it did not seem to make sense to put it all in shared.rs.

Also, I put some more code that's shared from the sub modules (emscripten, android, linux, l4re) up in linux_like/mod.rs. In theory, I think there's potential to do that with more code, that's just the one that I would have put in shared.rs but realized that it could go into linux_like/mod.rs instead.

tgross35

Sorry this has taken a while to get to, but thank you for all the changes here! The shape of this one looks much better. I have a handful of small comments but will need to take a deeper look at the big refactor portions again.

ci/run-docker.sh

libc-test/build.rs

src/unix/linux_like/l4re/uclibc/mod.rs

src/unix/linux_like/l4re/uclibc/x86_64/mod.rs

src/unix/linux_like/shared.rs

tgross35 · 2025-07-07T17:20:28Z

@rustbot author, for the above review and a rebase

rustbot · 2025-07-07T17:20:33Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

farao · 2025-11-19T14:31:39Z

Yes, getting close... :)

Regarding L4Re and Linux: Generally, they are very different OSes, L4Re is a microkernel-based OS and has its very own API but for compatibility purposes we aim to implement not only the C standard library but also as much as sensibly possible from the posix standard. We do not implement linux specific features, I guess the larger overlap is rather due to a choice of a similar set of feature test macros.

src/unix/linux_like/emscripten/mod.rs

src/unix/linux_like/android/mod.rs

tgross35 · 2025-11-20T01:09:46Z

src/unix/linux_like/linux_l4re_shared.rs

+    #[cfg_attr(
+        any(
+            target_pointer_width = "32",
+            target_arch = "x86_64",
+            target_arch = "powerpc64",
+            target_arch = "mips64",
+            target_arch = "mips64r6",
+            target_arch = "s390x",
+            target_arch = "sparc64",
+            target_arch = "aarch64",
+            target_arch = "riscv64",
+            target_arch = "riscv32",
+            target_arch = "loongarch64",
+            target_os = "emscripten"
+        ),
+        repr(align(4))
+    )]
+    #[cfg_attr(
+        not(any(
+            target_pointer_width = "32",
+            target_arch = "x86_64",
+            target_arch = "powerpc64",
+            target_arch = "mips64",
+            target_arch = "mips64r6",
+            target_arch = "s390x",
+            target_arch = "sparc64",
+            target_arch = "aarch64",
+            target_arch = "riscv64",
+            target_arch = "riscv32",
+            target_arch = "loongarch64",
+            target_os = "emscripten"
+        )),
+        repr(align(8))
+    )]
+    pub struct pthread_mutexattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_MUTEXATTR_T],
+    }


Don't try to share these pthread_* types between linux and l4re, they're a mess and most of the config isn't relevant for l4re. Just redefine them in the l4re module with something that matches the source headers.

src/unix/linux_like/linux_l4re_shared.rs

tgross35 · 2025-11-20T04:27:55Z

src/unix/linux_like/linux_l4re_shared.rs

+    #[cfg(not(target_env = "uclibc"))]
+    pub fn pthread_setschedprio(native: crate::pthread_t, priority: c_int) -> c_int;


Is this meant to be unsupported on linux-uclibc also? It seems to be available at https://github.com/kraj/uClibc/blob/ca1c74d67dd115d059a875150e10b8560a9c35a8/libpthread/nptl/sysdeps/pthread/pthread.h#L427-L429

Similar for a few other API that is disabled on uclibc, e.g. pthread_mutexattr_getprotocol

Oh, true, sorry, I think that was a mistake. I see it's fixed now that they are in "new"

This is imperfect but should get things closer to what is available on l4re's uclibc fork. Based on work by Marius Melzer in [rust-lang#4479]. [rust-lang#4479: rust-lang#4479

This is imperfect but should get things closer to what is available on l4re's uclibc fork. Based on work by Marius Melzer [rust-lang#4479]. [rust-lang#4479]: rust-lang#4479

This is imperfect but should get things closer to what is available on l4re's uclibc fork. Based on work by Marius Melzer [#4479]. [#4479]: #4479

farao · 2025-11-20T12:40:10Z

@rustbot ready

src/unix/linux_like/linux/gnu/mod.rs

src/unix/linux_like/linux/mod.rs

src/unix/linux_like/linux_l4re_shared.rs

src/unix/linux_like/linux/mod.rs

src/unix/linux_like/linux_l4re_shared.rs

src/unix/linux_like/linux/mod.rs

tgross35 · 2025-11-20T14:29:39Z

src/unix/linux_like/l4re/mod.rs

+use crate::prelude::*;
+
+pub type l4_umword_t = c_ulong; // Unsigned machine word.
+pub type pthread_t = *mut c_void;


Where does this come from? From https://github.com/search?q=repo%3Akernkonzept%2Fl4re-core+%2Ftypedef.*pthread_t%3B%2F&type=code, it seems like they are all c_ulong, like upstream uclibc. (Not that it really matters too much.)

I will try to investigate (since changing it brings me a warning). I'd like to leave it for another PR if that's ok.

tgross35 · 2025-11-20T14:32:42Z

src/unix/linux_like/l4re/uclibc/mod.rs

+    #[cfg_attr(any(target_pointer_width = "32",), repr(align(4)))]
+    #[cfg_attr(not(any(target_pointer_width = "32",)), repr(align(8)))]
+    pub struct pthread_mutexattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_MUTEXATTR_T],
+    }
+
+    #[cfg_attr(any(target_pointer_width = "32"), repr(align(4)))]
+    #[cfg_attr(all(target_pointer_width = "64"), repr(align(8)))]
+    pub struct pthread_rwlockattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_RWLOCKATTR_T],
+    }
+
+    #[repr(align(4))]
+    pub struct pthread_condattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_CONDATTR_T],
+    }
+
+    #[repr(align(4))]
+    pub struct pthread_barrierattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_BARRIERATTR_T],
+    }


I'm looking at the definitions here https://github.com/kernkonzept/l4re-core/blob/e1bf5beeb2933af404b03ca1571e82f33c3385c2/libc/uclibc-ng/libpthread/include/bits/pthreadtypes.h and they don't seem to match up at all - is that the wrong place?

No, the definitions are different. I thought it might have been a concious design decision in the linux part of the code to commonize the pthread structs by just putting an array of bytes into them so I kept this (especially because some time in the near future we might support other libcs). But I'm also happy to choose the actual struct definitions if you prefer this?

Oh no it wasn't a deviant decision, that's actually how they're defined in the linux libs https://github.com/bminor/glibc/blob/f9e61cd446d45016e20b6fe85ab87364ebdbec1b/sysdeps/nptl/bits/pthreadtypes.h#L67-L72. I would indeed prefer to match the struct definitions (with all fields private). It's much easier to make sure the fields line up than to add manually compute the size/align, and I don't know that the sizes are likely to always stay the same across platforms anyway.

farao · 2025-11-21T08:49:56Z

@rustbot ready

tgross35

Still some leftovers in the pthread definitions, and I would like the uclibc pthread defs to match source rather than using the array. But after this, LGTM!

tgross35 · 2025-11-30T06:45:23Z

src/unix/linux_like/l4re/uclibc/mod.rs

+    #[cfg_attr(any(target_pointer_width = "32",), repr(align(4)))]
+    #[cfg_attr(not(any(target_pointer_width = "32",)), repr(align(8)))]
+    pub struct pthread_mutexattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_MUTEXATTR_T],
+    }
+
+    #[cfg_attr(any(target_pointer_width = "32"), repr(align(4)))]
+    #[cfg_attr(all(target_pointer_width = "64"), repr(align(8)))]
+    pub struct pthread_rwlockattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_RWLOCKATTR_T],
+    }
+
+    #[repr(align(4))]
+    pub struct pthread_condattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_CONDATTR_T],
+    }
+
+    #[repr(align(4))]
+    pub struct pthread_barrierattr_t {
+        #[doc(hidden)]
+        size: [u8; crate::__SIZEOF_PTHREAD_BARRIERATTR_T],
+    }


Oh no it wasn't a deviant decision, that's actually how they're defined in the linux libs https://github.com/bminor/glibc/blob/f9e61cd446d45016e20b6fe85ab87364ebdbec1b/sysdeps/nptl/bits/pthreadtypes.h#L67-L72. I would indeed prefer to match the struct definitions (with all fields private). It's much easier to make sure the fields line up than to add manually compute the size/align, and I don't know that the sizes are likely to always stay the same across platforms anyway.

tgross35 · 2025-11-30T06:58:09Z

src/unix/linux_like/linux/mod.rs

-        size: [u8; crate::__SIZEOF_PTHREAD_COND_T],
+        pub size: [u8; crate::__SIZEOF_PTHREAD_COND_T],


tgross35 · 2025-11-30T06:58:39Z

src/unix/linux_like/linux/mod.rs

    #[cfg_attr(
        all(
-            any(target_env = "musl", target_env = "ohos"),
+            any(target_env = "musl", target_env = "ohos", target_os = "emscripten"),
            target_pointer_width = "32"
        ),
        repr(align(4))
    )]
    #[cfg_attr(
        all(
-            any(target_env = "musl", target_env = "ohos"),
+            any(target_env = "musl", target_env = "ohos", target_os = "emscripten"),
            target_pointer_width = "64"
        ),
        repr(align(8))
    )]
    #[cfg_attr(
        all(
-            not(any(target_env = "musl", target_env = "ohos")),
+            not(any(target_env = "musl", target_env = "ohos", target_os = "emscripten")),
            target_arch = "x86"
        ),
        repr(align(4))
    )]
    #[cfg_attr(
        all(
-            not(any(target_env = "musl", target_env = "ohos")),
+            not(any(target_env = "musl", target_env = "ohos", target_os = "emscripten")),
            not(target_arch = "x86")
        ),
        repr(align(8))
    )]
    pub struct pthread_cond_t {


Could you please make sure to self-review these pthread diffs? Again this has some changes for emscripten that aren't relevant.

tgross35 · 2025-11-30T06:59:29Z

src/unix/linux_like/linux/mod.rs

                target_arch = "powerpc",
                target_arch = "sparc",
                target_arch = "x86_64",
-                target_arch = "x86"
+                target_arch = "x86",
+                target_os = "emscripten",
            )
        ),
        repr(align(4))


Ditto, drop this

tgross35 · 2025-11-30T06:59:35Z

src/unix/linux_like/linux/mod.rs

                target_arch = "powerpc",
                target_arch = "sparc",
                target_arch = "x86_64",
-                target_arch = "x86"
+                target_arch = "x86",
+                target_os = "emscripten",
            ))
        ),
        repr(align(8))
    )]
    pub struct pthread_mutex_t {
        #[doc(hidden)]
-        size: [u8; crate::__SIZEOF_PTHREAD_MUTEX_T],
+        pub size: [c_char; crate::__SIZEOF_PTHREAD_MUTEX_T],
    }

    #[cfg_attr(
        all(
            target_pointer_width = "32",
            any(
+                target_os = "emscripten",
                target_arch = "mips",
                target_arch = "mips32r6",
                target_arch = "arm",


tgross35 · 2025-11-30T06:59:43Z

src/unix/linux_like/linux/mod.rs

        any(
            target_pointer_width = "64",
            not(any(
+                target_os = "emscripten",
                target_arch = "mips",
                target_arch = "mips32r6",
                target_arch = "arm",


tgross35 · 2025-11-30T06:59:49Z

src/unix/linux_like/linux/mod.rs

        repr(align(8))
    )]
    pub struct pthread_rwlock_t {
-        size: [u8; crate::__SIZEOF_PTHREAD_RWLOCK_T],
+        pub size: [u8; crate::__SIZEOF_PTHREAD_RWLOCK_T],
    }

    #[cfg_attr(


The L4Re code was previously attached to the Linux code which was not correct in many ways. This commit separates the L4Re code and enables the libc-tests and includes the fixes for the failing tests.

rustbot · 2025-12-04T16:24:49Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

farao · 2025-12-04T17:01:58Z

I replaced the pthread_* definitions with the uclibc (linuxthreads) ones in use by l4re. I removed the remaining emscripten lines after the last refactoring, sorry for missing those.

@rustbot ready

tgross35 · 2025-12-04T18:49:15Z

src/unix/linux_like/l4re/uclibc/mod.rs

+    pub struct _pthread_fastlock {
+        pub __status: c_long,
+        pub __spinlock: c_int,
+    }
+
+    pub struct pthread_cond_t {
+        pub __c_lock: _pthread_fastlock,
+        pub __c_waiting: _pthread_descr,
+        pub __padding: [u8; PTHREAD_COND_PADDING_SIZE],
+        pub __align: __pthread_cond_align_t,
+    }
+
+    pub struct pthread_condattr_t {
+        pub __dummy: c_int,
+    }
+
+    pub struct pthread_mutex_t {
+        pub __m_reserved: c_int,
+        pub __m_count: c_int,
+        pub __m_owner: _pthread_descr,
+        pub __m_kind: c_int,
+        pub __m_lock: _pthread_fastlock,
+    }
+
+    pub struct pthread_mutexattr_t {
+        pub __mutexkind: c_int,
+    }
+
+    pub struct pthread_rwlock_t {
+        pub __rw_lock: _pthread_fastlock,
+        pub __rw_readers: c_int,
+        pub __rw_writer: _pthread_descr,
+        pub __rw_read_waiting: _pthread_descr,
+        pub __rw_write_waiting: _pthread_descr,
+        pub __rw_kind: c_int,
+        pub __rw_pshared: c_int,
+    }
+
+    pub struct pthread_rwlockattr_t {
+        pub __lockkind: c_int,
+        pub __pshared: c_int,
+    }
+
+    pub struct pthread_barrier_t {
+        pub __ba_lock: _pthread_fastlock,
+        pub __ba_required: c_int,
+        pub __ba_present: c_int,
+        pub __ba_waiting: _pthread_descr,
+    }
+
+    pub struct pthread_barrierattr_t {
+        pub __pshared: c_int,
+    }


The types are intended to be opaque, so please make the fields private. (that's generally the rule for __-prefixed types/fields anyway).

Last thing, after that I'm happy to merge!

tgross35

You know what, please just send a followup making those fields private. Let's get this in.

I'm sorry this needed such a huge diff and a difficult process, but thank you for the persistence here! I'll be doing a new release in the next couple of days and will include this.

farao · 2025-12-05T10:09:53Z

Nice, thanks a lot for baring with me for this large change(s)!

rustbot added ctest Issues relating to the ctest crate O-android O-gnu O-linux O-linux-like O-musl O-unix O-x86 S-waiting-on-review labels Jun 9, 2025

farao force-pushed the revert branch 11 times, most recently from 92284e5 to 40a082a Compare June 12, 2025 16:33

farao mentioned this pull request Jun 13, 2025

Separate L4Re code, add aarch64 and enable tests #4383

Closed

3 tasks

farao force-pushed the revert branch 2 times, most recently from 26e9993 to 5b60e70 Compare June 23, 2025 09:15

tgross35 reviewed Jun 26, 2025

View reviewed changes

rustbot added S-waiting-on-author and removed S-waiting-on-review labels Jul 7, 2025

farao force-pushed the revert branch from 929c84c to 0875e06 Compare November 19, 2025 15:49

This comment has been minimized.

Sign in to view

farao force-pushed the revert branch 2 times, most recently from 2b40eb1 to f438c6d Compare November 19, 2025 16:15

tgross35 reviewed Nov 20, 2025

View reviewed changes

github-merge-queue bot pushed a commit that referenced this pull request Nov 20, 2025

l4re: Update available pthread API

34d8ce7

This is imperfect but should get things closer to what is available on l4re's uclibc fork. Based on work by Marius Melzer [#4479]. [#4479]: #4479

This comment has been minimized.

Sign in to view

tgross35 requested changes Nov 20, 2025

View reviewed changes

maurer approved these changes Nov 24, 2025

View reviewed changes

tgross35 requested changes Nov 30, 2025

View reviewed changes

farao added 5 commits December 4, 2025 17:15

ci: fix docker mount volumes SE-Linux enabled systems

721193c

Separate L4Re from Linux code and enable tests

93d0936

The L4Re code was previously attached to the Linux code which was not correct in many ways. This commit separates the L4Re code and enables the libc-tests and includes the fixes for the failing tests.

Add uclibc aarch64 support for L4Re

3efdce1

l4re: pthread fixes in new

c70208a

linux_like: communize some structs

19add0d

tgross35 reviewed Dec 4, 2025

View reviewed changes

tgross35 approved these changes Dec 4, 2025

View reviewed changes

xbjfk mentioned this pull request Dec 5, 2025

musl: 64-bit time support #4463

Merged

3 tasks

tgross35 mentioned this pull request Dec 28, 2025

[0.2] Backports #4897

Merged

		#[cfg(not(target_env = "uclibc"))]
		pub fn pthread_setschedprio(native: crate::pthread_t, priority: c_int) -> c_int;

		size: [u8; crate::__SIZEOF_PTHREAD_COND_T],
		pub size: [u8; crate::__SIZEOF_PTHREAD_COND_T],

Conversation

farao commented Jun 9, 2025

Sources

Checklist

Uh oh!

rustbot commented Jun 9, 2025

Uh oh!

farao commented Jun 13, 2025

Uh oh!

tgross35 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tgross35 commented Jul 7, 2025

Uh oh!

rustbot commented Jul 7, 2025

Uh oh!

farao commented Nov 19, 2025

Uh oh!

This comment has been minimized.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

farao commented Nov 20, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

farao Nov 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

farao commented Nov 21, 2025

Uh oh!

tgross35 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

farao Nov 21, 2025 •

edited

Loading