Add __isPlatformVersionAtLeast and __isOSVersionAtLeast symbols

[madsmtm]

Motivation

When Objective-C code uses @available(...), Clang inserts a call to __isPlatformVersionAtLeast (__isOSVersionAtLeast in older Clang versions). These symbols not being available sometimes ends up causing linker errors. See the new test tests/run-make/apple-c-available-links for a minimal reproducer.

The workaround is to link libclang_rt.osx.a, see e.g. alexcrichton/curl-rust#279. But that's very difficult for users to figure out (and the backreferences to that issue indicates that people are still running into this in their own projects every so often).

For another recent example, this is preventing rustc from using LLVM assertions on macOS, see #62592 (comment) and #134275 (comment).

It is also a blocker for setting the correct minimum OS version in cc-rs, since fixing this in cc-rs might end up introducing linker errors in places where we weren't before (by default, if using e.g. @available(macos 10.15, *), the symbol usually happens to be left out, since clang defaults to compiling for the host macOS version, and thus things seem to work - but the availability check actually compiles down to nothing, which is a huge correctness footgun for running on older OSes).

(My super secret evil agenda is also to expose some variant of @available in Rust's std after rust-lang/rfcs#3750 progresses further, will probably file an ACP for this later. But I believe this PR has value regardless of those future plans, since we'd be making C/Objective-C/Swift interop easier).

Solution

Implement __isPlatformVersionAtLeast and __isOSVersionAtLeast as part of the "public ABI" that std exposes.

This is insta-stable, in the same sense that additions to compiler-builtins are insta-stable, though the availability of these symbols can probably be considered a "quality of implementation" detail rather than a stable promise.

I originally proposed to implement this in compiler-builtins, see rust-lang/compiler-builtins#794, but we discussed moving it to std instead (Zulip thread), which makes the implementation substantially simpler, and we avoid gnarly issues with requiring the user to link libSystem.dylib (since std unconditionally does that).

Note that this does not solve the linker errors for (pure) #![no_std] users, but that's probably fine, if you are using @available to test the OS version on Apple platforms, you're likely also using std (and it is still possible to work around by linking libclang_rt.*.a).

A thing to note about the implementation, I've choosen to stray a bit from LLVM's upstream implementation, and not use _availability_version_check since it has problems when compiling with an older SDK. Instead, we use sysctl kern.osproductversion when available to still avoid the costly PList lookup in most cases, but still with a fall back to the PList lookup when that is not available (with the PList fallback being is similar to LLVM's implementation).

Testing

Apple has a lot of different "modes" that they can run binaries in, which can be a bit difficult to find your bearings in, but I've tried to be as thorough as I could in testing them all.

Tested using roughly the equivalent of ./x test library/std -- platform_version on the following configurations:

• macOS 14.7.3 on a Macbook Pro M2
  • aarch64-apple-darwin
  • x86_64-apple-darwin (under Rosetta)
  • aarch64-apple-ios-macabi
  • x86_64-apple-ios-macabi (under Rosetta)
  • aarch64-apple-ios (using Xcode's "Designed for iPad" setting)
  • aarch64-apple-ios-sim (in iOS Simulator, as iPhone with iOS 17.5)
  • aarch64-apple-ios-sim (in iOS Simulator, as iPad with iOS 18.2)
  • aarch64-apple-tvos-sim (in tvOS Simulator)
  • aarch64-apple-watchos-sim (in watchOS Simulator)
  • aarch64-apple-ios-sim (in visionOS simulator, using Xcode's "Designed for iPad" setting)
  • aarch64-apple-visionos-sim (in visionOS Simulator)
• macOS 15.3.1 VM
  • aarch64-apple-darwin
  • aarch64-apple-ios-macabi
• macOS 10.12.6 on an Intel Macbook from 2013
  • x86_64-apple-darwin
  • i686-apple-darwin
  • x86_64-apple-ios (in iOS Simulator)
• iOS 9.3.6 on a 1st generation iPad Mini
  • armv7-apple-ios with an older compiler

Along with manually inspecting the output of version_from_sysctl() and version_from_plist(), and verifying that they actually match what's expected.

I believe the only real omissions here would be:

• aarch64-apple-ios on a newer iPhone that has sysctl available (iOS 11.4 or above).
• aarch64-apple-ios on a Vision Pro using Xcode's "Designed for iPad" setting.

But I don't have the hardware available to test those.

@rustbot label O-apple A-linkage -T-compiler -A-meta -A-run-make

try-job: aarch64-apple
try-job: x86_64-apple*

[rustbot]

r? @Amanieu

rustbot has assigned @Amanieu.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

This PR modifies run-make tests.

cc @jieyouxu

triagebot.toml has been modified, there may have been changes to the review queue.

cc @davidtwco, @wesleywiser

[Amanieu]

r? @tgross35

[tgross35]

Happy to review the implementation, but @Amanieu mind confirming you are okay providing these symbols from std? It seems reasonable to me but as far as I know we don't have anything like this. (Also no other builtins that effectively need std API).

[Amanieu]

It's a bit of a grey area since it's not clear how much it is Rust's responsibility to provide builtin symbols that are only used by C code. Arguably the more "correct" approach is to tell users of C code that we only provide symbols for Rust code and that they need to separately add a dependency on libclang_rt or libgcc, but that is a lot of hassle and thing usually just work without it.

I'm not opposed to adding this in libstd for now, but it might be worth looking into a better story for how to handle builtins that are needed by C code but not Rust code (e.g. emulated TLS, clear_cache, enable_execute_stack, etc).

[tgross35]

Some initial comments, I still need to take a closer look at the version_from* implementations.

[tgross35]

Cc RFC author @ChrisDenton just so you know this module exists

[madsmtm]

Thanks for the review so far. I've answered or fixed your comments in separate commits, to make it easier for you to review (if you're already halfway through a second review). I can squash once everything is ready.
@rustbot ready

[rust-bors]

⌛ Trying commit 5ec7157 with merge 2008cdd…

To cancel the try build, run the command @bors2 try cancel.

[rust-bors]

💔 Test failed

• CI ❌

[madsmtm]

Error was:

foo.c:11:9: error: unrecognized platform name visionos
        visionos 1000.0,
        ^
1 error generated.

Fixed building test with older Clang (modified tests/run-make/apple-c-available-links/*).
@bors2 try

[rust-bors]

⌛ Trying commit 779ac2f with merge ea9a6dd…

To cancel the try build, run the command @bors2 try cancel.

[rust-bors]

💔 Test failed

• CI ❌

[madsmtm]

Error was:

foo.c:7:10: error: 'TARGET_OS_VISION' is not defined, evaluates to 0 [-Werror,-Wundef-prefix=TARGET_OS_]
    #if !TARGET_OS_VISION
         ^
1 error generated.

Implemented differently with an #ifdef instead.
@bors2 try

[rust-bors]

⌛ Trying commit d543096 with merge 7334def…

To cancel the try build, run the command @bors2 try cancel.

[rust-bors]

☀️ Try build successful (CI)
Build commit: 7334def (7334def256dcd96dafe3a7c2db0126901ab7070f, parent: 015c7770ec0ffdba9ff03f1861144a827497f8ca)

[bors]

☔ The latest upstream changes (presumably #141061) made this pull request unmergeable. Please resolve the merge conflicts.

[tgross35]

I think we still need to discuss the guarantees we want to give by providing this, see the thread in #138944 (comment).

@madsmtm if you're blocking on this, maybe want to start a Zulip discussion? I don't really have any strong opinions here.

[tgross35]

I'm not sure what the status of the above is at this point, seems like there is something worth discussing?

@rustbot author

[rustbot]

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

[madsmtm]

I have thought about it, and the only real guarantee I want from std is that Objective-C's @available and similar will continue to work with whatever minimum Xcode version we currently support. I have tried to encapsulate this in a comment in the code, please verify the wording.

I can still open a Zulip thread if you think it's contentious?

FYI since your last review, apart from a rebase, I also fixed the ABI of the external symbols to use i32 instead of u32 (not that it really matters in practice, the ABI is the same for these types on Apple platforms).

@rustbot ready

[madsmtm]

Right here.

[madsmtm]

So to be clear, the following C code would not be explicitly supported (though due to that being the ABI, it would work for as long as Clang 11 is supported):

int __isOSVersionAtLeast(int Major, int Minor, int Subminor);

int foo() {
    return __isOSVersionAtLeast(10, 12, 0);
}

Whereas the following C code would be (as long as Rust supports interop with LLVM/Clang):

int foo() {
    return __builtin_available(macos 10.12, *);
}

[madsmtm]

Similar to how we do not explicitly support:

int __mulsi3(int a, int b);

int foo() {
    return __mulsi3(1, 2);
}

But we do support:

int foo() {
    return 1 * 2;
}