Mark is_ascii_sse2 as #[inline]

library/core/src/slice/ascii.rs

@@ -465,6 +465,7 @@ const fn is_ascii(s: &[u8]) -> bool {
 const SSE2_CHUNK_SIZE: usize = 64;
 
 #[cfg(all(target_arch = "x86_64", target_feature = "sse2"))]
+#[inline]
 fn is_ascii_sse2(bytes: &[u8]) -> bool {
     use crate::arch::x86_64::{__m128i, _mm_loadu_si128, _mm_movemask_epi8, _mm_or_si128};
 

tests/assembly-llvm/slice-is-ascii.rs

@@ -1,13 +1,28 @@
-//@ revisions: LA64
+//@ revisions: X86_64 LA64
 //@ assembly-output: emit-asm
 //@ compile-flags: -C opt-level=3
 //
+//@ [X86_64] only-x86_64
+//@ [X86_64] compile-flags: -C target-cpu=znver4
+//@ [X86_64] compile-flags: -C llvm-args=-x86-asm-syntax=intel
+//
 //@ [LA64] only-loongarch64
 
 #![crate_type = "lib"]
 
+/// Verify `is_ascii` generates efficient code on different architectures:
+///
+/// - x86_64: Must NOT use `kshiftrd`/`kshiftrq` (broken AVX-512 auto-vectorization).
+///   Good version uses explicit SSE2 intrinsics (`pmovmskb`/`vpmovmskb`).
+///
 /// - loongarch64: Should use `vmskltz.b` instruction for the fast-path.
 
+// X86_64-LABEL: test_is_ascii
+// X86_64-NOT: kshiftrd
+// X86_64-NOT: kshiftrq
+// X86_64: {{vpor|por}}
+// X86_64: {{vpmovmskb|pmovmskb}}
+
 // LA64-LABEL: test_is_ascii
 // LA64: vmskltz.b