fix vector fmin/fmax non-fast/fast intrinsics NaN handling
This bugs shows up in release mode tests of `stdsimd`: https://github.com/rust-lang-nursery/stdsimd/pull/391 . The intrinsics are thoroughly tested there for roundoff errors, NaN, and overflow behavior.
The problem was that the non-fast intrinsics where specifying `NoNaNs == true`, which meant that they don't support NaNs. This is incorrect, the non-fast intrinsics should handle NaNs properly.
Also, the "fast" intrinsics where specifying `NoNaNs == false` which meant that they support NaNs and then fast-math, which probably disables this support. This was not intended either.
I've added a comment specifying what the boolean flags do.
whitelist every target feature for rustdoc
When https://github.com/rust-lang-nursery/stdsimd/pull/367 was attempted to be upstreamed, it failed to document on non-x86 targets because it made every intrinsic visible, even the ones on foreign arches. This change makes it so that whenever rustdoc asks for the target feature whitelist, it gets a list of every feature known to every arch in `rustc_trans/llvm_util.rs`.
Before pushing, i temporarily updated the `stdsimd` submodule to include the `doc(cfg)` change, generated documentation for `aarch64-unknown-linux-gnu`, and it completed without a problem. The generated `core::arch` docs contained complete submodules for all main arches.
rustc: Enable embedding LLVM bitcode for iOS
This commit updates rustc to embed bitcode in each object file generated by
default when compiling for iOS. This was determined in #35968 as a step
towards better compatibility with the iOS toolchain, so let's give it a spin and
see how it turns out!
Note that this also updates the `cc` dependency which should propagate this
change of embedding bitcode for C dependencies as well.
rustc_trans: fix small aggregate returns for big-endian mips64 FFI
Current model of threating small aggregate returns as smallest encompassing integer works only for little-endian mips64.
The patch forces small aggregate return values to be viewed as one or two i64 chunks leaving to the casting implementation
to handle endianes differences.
rustc: Start a custom cabi module for wasm32
It actually was already using the `cabi_asmjs` module but that was by accident,
so route the new `wasm32-unknown-unknown` target to a new `cabi_wasm32` module.
The first entries in this module are to use `signext` and `zeroext` for types
that are under 32 bytes in size
Closesrust-lang-nursery/rust-wasm#88
Stabilise feature(never_type). Introduce feature(exhaustive_patterns)
This stabilizes `!`, removing the feature gate as well as the old defaulting-to-`()` behavior. The pattern exhaustiveness checks which were covered by `feature(never_type)` have been moved behind a new `feature(exhaustive_patterns)` gate.
Current model of threating small aggregate returns as smallest encompassing integer works only for little-endian mips64.
The patch forces small aggregate return values to be viewed as one or two i64 chunks leaving to the casting implementation
to handle endianes differences.
bump mipsel isa leval and enable fpxx
This PR:
* Bumps the default ISA level of the mipsel targets to `mips32r2`. The big endian mips targets are already built with `mips32r2`. This is the usual baseline for the MIPS ISA these days used by other projects, although it does drop support for the 4K processor (which was the only processor released with mips32 r1). Debian no longer supports pre-R2 processors. Using R2 also improves code generation in FPXX in certain circumstances.
* Enables the FPXX floating point ABI[1] on 32-bit hard-float targets by default. This ABI adds some extra restrictions to the existing ABI which allows code to run on the two main floating point modes found on MIPS (FR0 and FR1) and remains compatible with the FR32 ABI currently in use. All code within an executable (including all shared libraries) must be compiled with FPXX/FP64 to be able to use MSA on 32-bit MIPS.
* Enables the "nooddspreg" feature with FPXX. This feature is usually enabled whenever FPXX is. It also helps workaround some issues on Loongson processors. I'm hoping this will fix some test failures mentioned in #39013.
* Adds the `fp64` feature to the MIPS whitelist. This feature must be enabled to use MSA on 32-bit MIPS, otherwise LLVM will complain.
[1] See https://dmz-portal.mips.com/wiki/MIPS_O32_ABI_-_FR0_and_FR1_Interlinking
rustc: Don't invoke `lld` with an `@`-file
Looks like LLD doesn't support this yet, so always try to use the OS before we
fall back to using `@`
cc https://github.com/rust-lang/rust/issues/48948
This commit updates rustc to embed bitcode in each object file generated by
default when compiling for iOS. This was determined in #35968 as a step
towards better compatibility with the iOS toolchain, so let's give it a spin and
see how it turns out!
Note that this also updates the `cc` dependency which should propagate this
change of embedding bitcode for C dependencies as well.
- `ParamEnv::empty()` -- does not reveal all, good for typeck
- `ParamEnv::reveal_all()` -- does, good for trans
- `param_env.with_reveal_all()` -- converts an existing parameter environment
It actually was already using the `cabi_asmjs` module but that was by accident,
so route the new `wasm32-unknown-unknown` target to a new `cabi_wasm32` module.
The first entries in this module are to use `signext` and `zeroext` for types
that are under 32 bytes in size
Closesrust-lang-nursery/rust-wasm#88
Add relro-level tests
The `relro-level` debugging flag was added in #43170 which was merged in July 2017. This PR moves this flag to be a proper codegen flag.
