| Age | Commit message (Collapse) | Author | Lines |
|---|
|
Unfortunately, multiple people are reporting linker warnings related to
`__rust_no_alloc_shim_is_unstable` after this change. The solution isn't
quite clear yet, let's revert to green for now, and try a reland with a
determined solution for `__rust_no_alloc_shim_is_unstable`.
This reverts commit c8b7f32434c0306db5c1b974ee43443746098a92, reversing
changes made to 667247db71ea18c4130dd018d060e7f09d589490.
|
|
Stage0 bootstrap update
This PR [follows the release process](https://forge.rust-lang.org/release/process.html#master-bootstrap-update-tuesday) to update the stage0 compiler.
The only thing of note is https://github.com/rust-lang/rust/commit/58651d1b316e268fac2100c3ae37bb502a36b8ba, which was flagged by clippy as a correctness fix. I think allowing that lint in our case makes sense, but it's worth to have a second pair of eyes on it.
r? `@Mark-Simulacrum`
|
|
cg_llvm: Clean up some inline assembly bindings
This PR combines a few loosely-related cleanups to LLVM bindings related to inline assembly. These include:
- Replacing `LLVMRustInlineAsm` with LLVM-C's `LLVMGetInlineAsm`
- Adjusting FFI declarations to avoid the need for explicit `as_c_char_ptr` conversions
- Flattening control flow in `inline_asm_call`
There should be no functional changes.
|
|
|
|
As with `DIBuilderBox`, the "Box" suffix does a better job of communicating
that this is an owning pointer to some borrowable resource.
This also renames the `raw` method to `as_ref`, which is what it would have
been named originally if the `Deref` problem had been known at the time.
|
|
|
|
This avoids the need for an explicit `as_c_char_ptr` conversion.
|
|
This patch also changes the Rust-side declaration to take `*const c_uchar`
instead of `*const c_char`, to avoid the need for `AsCCharPtr`.
|
|
This LLVM-C binding replaces the existing `LLVMRustInlineAsm` function.
|
|
This module comment describes why it's OK for LLVM bindings to declare a
parameter type of `*const c_uchar` for pointer/length strings, even though the
corresponding parameter on the C/C++ side uses `const char *`.
Adding a searchable term to each such parameter should make it easier for
future maintainers to understand why `*const c_uchar` is being used instead of
`*const c_char`.
|
|
r=scottmcm,traviscross,tgross35
Use intrinsics for `{f16,f32,f64,f128}::{minimum,maximum}` operations
This PR creates intrinsics for `{f16,f32,f64,f64}::{minimum,maximum}` operations.
This wasn't done when those operations were added as the LLVM support was too weak but now that LLVM has libcalls for unsupported platforms we can finally use them.
Cranelift and GCC[^1] support are partial, Cranelift doesn't support `f16` and `f128`, while GCC doesn't support `f16`.
r? `@tgross35`
try-job: aarch64-gnu
try-job: dist-various-1
try-job: dist-various-2
[^1]: https://www.gnu.org/software///gnulib/manual/html_node/Functions-in-_003cmath_002eh_003e.html
|
|
|
|
remove 'unordered' atomic intrinsics
As their doc comment already indicates, these operations do not currently have a place in our memory model. The intrinsics were introduced to support a hack in compiler-builtins, but that hack recently got removed (see https://github.com/rust-lang/compiler-builtins/issues/788).
|
|
|
|
|
|
Fix linking statics on Arm64EC
Arm64EC builds recently started to fail due to the linker not finding a symbol:
```
symbols.o : error LNK2001: unresolved external symbol #_ZN3std9panicking11EMPTY_PANIC17hc8d2b903527827f1E (EC Symbol)
C:\Code\hello-world\target\arm64ec-pc-windows-msvc\debug\deps\hello_world.exe : fatal error LNK1120: 1 unresolved externals
```
It turns out that `EMPTY_PANIC` is a new static variable that was being exported then imported from the standard library, but when exporting LLVM didn't prepend the name with `#` (as only functions are prefixed with this character), whereas Rust was prefixing with `#` when attempting to import it.
The fix is to have Rust not prefix statics with `#` when importing.
Adding tests discovered another issue: we need to correctly mark static exported from dylibs with `DATA`, otherwise MSVC's linker assumes they are functions and complains that there is no exit thunk for them.
CI found another bug: we only apply `DllImport` to non-local statics that aren't foreign items (i.e., in an `extern` block), that is we want to use `DllImport` for statics coming from other Rust crates. However, `__rust_no_alloc_shim_is_unstable` is a static generated by the Rust compiler if required, but downstream crates consider it a foreign item since it is declared in an `extern "Rust"` block, thus they do not apply `DllImport` to it and so fails to link if it is exported by the previous crate as `DATA`. The fix is to apply `DllImport` to foreign items that are marked with the `rustc_std_internal_symbol` attribute (i.e., we assume they aren't actually foreign and will be in some Rust crate).
Fixes #138541
---
try-job: dist-aarch64-msvc
try-job: dist-x86_64-msvc
try-job: x86_64-msvc-1
try-job: x86_64-msvc-2
|
|
|
|
r=workingjubilee
Added support for `apxf` target feature
|
|
|
|
|
|
|
|
|
|
Add the AVX10 target features
Parent #138843
Adds the `avx10_target_feature` feature gate, and `avx10.1` and `avx10.2` target features.
It is confirmed that Intel is dropping AVX10/256 (see [this comment](https://github.com/rust-lang/rust/issues/111137#issuecomment-2795442288)), so this should be safe to implement now.
The LLVM fix for llvm/llvm-project#135394 was merged, and has been backported to LLVM20, and the patch has also been propagated to rustc in #140502
`@rustbot` label O-x86_64 O-x86_32 A-target-feature A-SIMD
|
|
Fix handling of LoongArch target features not supported by LLVM 19
Fixes #140455
|
|
Share the naked asm impl between cg_ssa and cg_clif
This was introduced in https://github.com/rust-lang/rust/pull/128004.
|
|
|
|
This has been disabled due to an LLVM misoptimization with `powi.f16`
[1]. This was fixed upstream and the fix is included in LLVM20, so tests
no longer need to be disabled.
`f16` still remains disabled on MinGW due to the ABI issue.
[1]: https://github.com/llvm/llvm-project/issues/98665
|
|
Implement the internal feature `cfg_target_has_reliable_f16_f128`
Support for `f16` and `f128` is varied across targets, backends, and backend versions. Eventually we would like to reach a point where all backends support these approximately equally, but until then we have to work around some of these nuances of support being observable.
Introduce the `cfg_target_has_reliable_f16_f128` internal feature, which provides the following new configuration gates:
* `cfg(target_has_reliable_f16)`
* `cfg(target_has_reliable_f16_math)`
* `cfg(target_has_reliable_f128)`
* `cfg(target_has_reliable_f128_math)`
`reliable_f16` and `reliable_f128` indicate that basic arithmetic for the type works correctly. The `_math` versions indicate that anything relying on `libm` works correctly, since sometimes this hits a separate class of codegen bugs.
These options match configuration set by the build script at [1]. The logic for LLVM support is duplicated as-is from the same script. There are a few possible updates that will come as a follow up.
The config introduced here is not planned to ever become stable, it is only intended to replace the build scripts for `std` tests and `compiler-builtins` that don't have any way to configure based on the codegen backend.
MCP: https://github.com/rust-lang/compiler-team/issues/866
Closes: https://github.com/rust-lang/compiler-team/issues/866
[1]: https://github.com/rust-lang/rust/blob/555e1d0386f024a8359645c3217f4b3eae9be042/library/std/build.rs#L84-L186
---
The second commit makes use of this config to replace `cfg_{f16,f128}{,_math}` in `library/`. I omitted providing a `cfg(bootstrap)` configuration to keep things simpler since the next beta branch is in two weeks.
try-job: aarch64-gnu
try-job: i686-msvc-1
try-job: test-various
try-job: x86_64-gnu
try-job: x86_64-msvc-ext2
|
|
add autodiff inline
closes: #138920
r? ```@ZuseZ4```
try-job: dist-aarch64-linux
|
|
|
|
async_drop_in_place::{closure}, scoped async drop added.
|
|
Support for `f16` and `f128` is varied across targets, backends, and
backend versions. Eventually we would like to reach a point where all
backends support these approximately equally, but until then we have to
work around some of these nuances of support being observable.
Introduce the `cfg_target_has_reliable_f16_f128` internal feature, which
provides the following new configuration gates:
* `cfg(target_has_reliable_f16)`
* `cfg(target_has_reliable_f16_math)`
* `cfg(target_has_reliable_f128)`
* `cfg(target_has_reliable_f128_math)`
`reliable_f16` and `reliable_f128` indicate that basic arithmetic for
the type works correctly. The `_math` versions indicate that anything
relying on `libm` works correctly, since sometimes this hits a separate
class of codegen bugs.
These options match configuration set by the build script at [1]. The
logic for LLVM support is duplicated as-is from the same script. There
are a few possible updates that will come as a follow up.
The config introduced here is not planned to ever become stable, it is
only intended to replace the build scripts for `std` tests and
`compiler-builtins` that don't have any way to configure based on the
codegen backend.
MCP: https://github.com/rust-lang/compiler-team/issues/866
Closes: https://github.com/rust-lang/compiler-team/issues/866
[1]: https://github.com/rust-lang/rust/blob/555e1d0386f024a8359645c3217f4b3eae9be042/library/std/build.rs#L84-L186
|
|
|
|
Make #![feature(let_chains)] bootstrap conditional in compiler/
Let chains have been stabilized recently in #132833, so we can remove the gating from our uses in the compiler (as the compiler uses edition 2024).
|
|
|
|
|
|
Autodiff flags
Interestingly, it seems that some other projects have conflicts with exactly the same LLVM optimization passes as autodiff.
At least `LLVMRustOptimize` has exactly the flags that we need to disable problematic opt passes.
This PR enables us to compile code where users differentiate two identical functions in the same module. This has been especially common in test cases, but it's not impossible to encounter in the wild.
It also enables two new flags for testing/debugging. I consider writing an MCP to upgrade PrintPasses to be a standalone -Z flag, since it is *not* the same as `-Z print-llvm-passes`, which IMHO gives less useful output. A discussion can be found here: [#t-compiler/llvm > Print llvm passes. @ 💬](https://rust-lang.zulipchat.com/#narrow/channel/187780-t-compiler.2Fllvm/topic/Print.20llvm.20passes.2E/near/511533038)
Finally, it improves `PrintModBefore` and `PrintModAfter`. They used to work reliable, but now we just schedule enzyme as part of an existing ModulePassManager (MPM). Since Enzyme is last in the MPM scheduling, PrintModBefore became very inaccurate. It used to print the input module, which we gave to the Enzyme and was great to create llvm-ir reproducer. However, lately the MPM would run the whole `default<O3>` pipeline, which heavily modifies the llvm module, before we pass it to Enzyme. That made it impossible to use the flag to create llvm-ir reproducers for Enzyme bugs. We now schedule a PrintModule pass just before Enzyme, solving this problem.
Based on the PrintPass output, it also _seems_ like changing `registerEnzymeAndPassPipeline(PB, true);` to `registerEnzymeAndPassPipeline(PB, false);` has no effect. In theory, the bool should tell Enzyme to schedule some helpful passes in the PassBuilder. However, since it doesn't do anything and I'm not 100% sure anymore on whether we really need it, I'll just disable it for now and postpone investigations.
r? ``@oli-obk``
closes #139471
Tracking:
- https://github.com/rust-lang/rust/issues/124509
|
|
mitigate MSVC alignment issue on x86-32
This implements mitigation for https://github.com/rust-lang/rust/issues/112480 by stopping to emit `align` attributes on loads and function arguments when building for a win32 MSVC target. MSVC is known to not properly align `u64` and similar types, and claiming to LLVM that everything is properly aligned increases the chance that this will cause problems.
Of course, the misalignment is still a bug, but we can't fix that bug, only MSVC can.
Also add an errata note to the platform support page warning users about this known problem.
try-job: `i686-msvc*`
|
|
|
|
simd intrinsics with mask: accept unsigned integer masks, and fix some of the errors
It's not clear at all why the mask would have to be signed, it is anyway interpreted bitwise. The backend should just make sure that works no matter the surface-level type; our LLVM backend already does this correctly. The note of "the mask may be widened, which only has the correct behavior for signed integers" explains... nothing? Why can't the code do the widening correctly? If necessary, just cast to the signed type first...
Also while we are at it, fix the errors. For simd_masked_load/store, the errors talked about the "third argument" but they meant the first argument (the mask is the first argument there). They also used the wrong type for `expected_element`.
I have extremely low confidence in the GCC part of this PR.
See [discussion on Zulip](https://rust-lang.zulipchat.com/#narrow/channel/257879-project-portable-simd/topic/On.20the.20sign.20of.20masks)
|
|
|
|
avoid overflow when generating debuginfo for expanding recursive types
Fixes #135093
Fixes #121538
Fixes #107362
Fixes #100618
Fixes #115994
The overflow happens because expanding recursive types keep creating new nested types when recurring into sub fields.
I fixed that by returning an empty stub node when expanding recursion is detected.
|
|
Autodiff batching2
~I will rebase it once my first PR landed.~ done.
This autodiff batch mode is more similar to scalar autodiff, since it still only takes one shadow argument.
However, that argument is supposed to be `width` times larger.
r? `@oli-obk`
Tracking:
- https://github.com/rust-lang/rust/issues/124509
|
|
|
|
This reverts commit 6adc2c1fd6ecde7bf83c8b8fbc71f402ced87054.
|
|
|
|
|
|
|
|
|
|
before AD
|
