From 9f0928782b051edb6add78e9310472abface1e5c Mon Sep 17 00:00:00 2001
From: bjorn3 <17426603+bjorn3@users.noreply.github.com>
Date: Mon, 5 Sep 2022 15:38:33 +0000
Subject: [PATCH] Use simd_bitmask intrinsic in a couple of places

---
 crates/core_arch/src/x86/avx2.rs | 4 +---
 crates/core_arch/src/x86/sse2.rs | 4 +---
 2 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/crates/core_arch/src/x86/avx2.rs b/crates/core_arch/src/x86/avx2.rs
index 24f9c03017..16add3dbb1 100644
--- a/crates/core_arch/src/x86/avx2.rs
+++ b/crates/core_arch/src/x86/avx2.rs
@@ -2001,7 +2001,7 @@ pub unsafe fn _mm256_min_epu8(a: __m256i, b: __m256i) -> __m256i {
 #[cfg_attr(test, assert_instr(vpmovmskb))]
 #[stable(feature = "simd_x86", since = "1.27.0")]
 pub unsafe fn _mm256_movemask_epi8(a: __m256i) -> i32 {
-    pmovmskb(a.as_i8x32())
+    simd_bitmask::<_, u32>(a.as_i8x32()) as i32
 }
 
 /// Computes the sum of absolute differences (SADs) of quadruplets of unsigned
@@ -3642,8 +3642,6 @@ extern "C" {
     fn pminud(a: u32x8, b: u32x8) -> u32x8;
     #[link_name = "llvm.x86.avx2.pminu.b"]
     fn pminub(a: u8x32, b: u8x32) -> u8x32;
-    #[link_name = "llvm.x86.avx2.pmovmskb"]
-    fn pmovmskb(a: i8x32) -> i32;
     #[link_name = "llvm.x86.avx2.mpsadbw"]
     fn mpsadbw(a: u8x32, b: u8x32, imm8: i32) -> u16x16;
     #[link_name = "llvm.x86.avx2.pmulhu.w"]
diff --git a/crates/core_arch/src/x86/sse2.rs b/crates/core_arch/src/x86/sse2.rs
index d82b8641f6..3e79b35395 100644
--- a/crates/core_arch/src/x86/sse2.rs
+++ b/crates/core_arch/src/x86/sse2.rs
@@ -1378,7 +1378,7 @@ pub unsafe fn _mm_insert_epi16<const IMM8: i32>(a: __m128i, i: i32) -> __m128i {
 #[cfg_attr(test, assert_instr(pmovmskb))]
 #[stable(feature = "simd_x86", since = "1.27.0")]
 pub unsafe fn _mm_movemask_epi8(a: __m128i) -> i32 {
-    pmovmskb(a.as_i8x16())
+    simd_bitmask::<_, u16>(a.as_i8x16()) as u32 as i32
 }
 
 /// Shuffles 32-bit integers in `a` using the control in `IMM8`.
@@ -2856,8 +2856,6 @@ extern "C" {
     fn packssdw(a: i32x4, b: i32x4) -> i16x8;
     #[link_name = "llvm.x86.sse2.packuswb.128"]
     fn packuswb(a: i16x8, b: i16x8) -> u8x16;
-    #[link_name = "llvm.x86.sse2.pmovmskb.128"]
-    fn pmovmskb(a: i8x16) -> i32;
     #[link_name = "llvm.x86.sse2.max.sd"]
     fn maxsd(a: __m128d, b: __m128d) -> __m128d;
     #[link_name = "llvm.x86.sse2.max.pd"]