Update stdarch submodule

[sayantn]

Updates the stdarch submodule.

Merged PRs

• std_detect: RISC-V platform guide documentation (non-table part) stdarch#1797
• Re-organize intrinsic-test to enable seamless addition of behaviour testing for more architectures stdarch#1758
• Check cfg on features that stage0 compiler support stdarch#1798
• cmpxchg16b: use atomic_compare_exchange from libcore stdarch#1811
• mark gfni, vaes, vpclmulqdq intrinsics as safe stdarch#1810
• Fix uses of invalid LLVM intrinsics stdarch#1807
• RISC-V: Linux: Imply Zicntr from the IMA base behavior stdarch#1806
• Coding style modernization to deal with Clippy warnings stdarch#1812
• Stabilize keylocker intrinsics and runtime detection stdarch#1795
• Stabilize sha512, sm3 and sm4 intrinsics and runtime detection stdarch#1796
• intrinsic-test: Minor fixes (2025-05) stdarch#1813
• RISC-V: Linux 6.15 riscv_hwprobe support stdarch#1816
• Fix Neon f16 stability stdarch#1818
• Fix some more uses of LLVM intrinsics stdarch#1820
• Stabilize AVX512 intrinsics stdarch#1819

r? @Amanieu
@rustbot label T-libs-api

Closes #111137

[rustbot]

⚠️ Warning ⚠️

• Some commits in this PR modify submodules.

[sayantn]

cc @bjorn3 I guess.

[bjorn3]

That is caused by rust-lang/stdarch#1807. You can try replacing the llvm.aarch64.neon.frintn.v string in compiler/rustc_codegen_cranelift/src/intrinsics/llvm_aarch64.rs with llvm.roundeven.v and likely move it to compiler/rustc_codegen_cranelift/src/intrinsics/llvm.rs.

[rustbot]

Some changes occurred in compiler/rustc_codegen_cranelift

cc @bjorn3

[rustbot]

Portable SIMD is developed in its own repository. If possible, consider making this change to rust-lang/portable-simd instead.

cc @calebzulawski, @programmerjake

The Miri subtree was changed

cc @rust-lang/miri

[Amanieu]

@bors r+ rollup=never

[bors]

📌 Commit fbc15c1 has been approved by Amanieu

It is now in the queue for this repository.

[Amanieu]

Beta-nominating specifically for rust-lang/stdarch#1818: we want that to be backported to beta since we accidentally stabilized some f16 types and intrinsics in beta. However it's not strictly necessary if the FCP on #136306 passes before the release deadline since we would be stabilizing them in the next release anyways.

[adamgemmell]

A minor correction in case it changes something - the f16 types were accidentally introduced originally as stable so those aren't a regression and are currently stable on the stable channel. The couple of f16 intrinsics stabilised are the beta regression.