Add vmull_p64 on arm

Author: SparrowLii

crates/core_arch/src/aarch64/neon/generated.rs

@@ -4094,19 +4094,6 @@ pub unsafe fn vmull_high_u32(a: uint32x4_t, b: uint32x4_t) -> uint64x2_t {
     vmull_u32(a, b)
 }
 
-/// Polynomial multiply long
-#[inline]
-#[target_feature(enable = "neon,crypto")]
-#[cfg_attr(test, assert_instr(pmull))]
-pub unsafe fn vmull_p64(a: p64, b: p64) -> p128 {
-    #[allow(improper_ctypes)]
-    extern "C" {
-        #[cfg_attr(target_arch = "aarch64", link_name = "llvm.aarch64.neon.pmull64")]
-        fn vmull_p64_(a: p64, b: p64) -> int8x16_t;
-    }
-    transmute(vmull_p64_(a, b))
-}
-
 /// Polynomial multiply long
 #[inline]
 #[target_feature(enable = "neon")]
@@ -11389,15 +11376,6 @@ mod test {
         assert_eq!(r, e);
     }
 
-    #[simd_test(enable = "neon")]
-    unsafe fn test_vmull_p64() {
-        let a: p64 = 15;
-        let b: p64 = 3;
-        let e: p128 = 17;
-        let r: p128 = transmute(vmull_p64(transmute(a), transmute(b)));
-        assert_eq!(r, e);
-    }
-
     #[simd_test(enable = "neon")]
     unsafe fn test_vmull_high_p8() {
         let a: i8x16 = i8x16::new(1, 2, 9, 10, 9, 10, 11, 12, 9, 10, 11, 12, 13, 14, 15, 16);

crates/core_arch/src/arm_shared/neon/generated.rs

@@ -6062,6 +6062,34 @@ pub unsafe fn vmull_p8(a: poly8x8_t, b: poly8x8_t) -> poly16x8_t {
 vmull_p8_(a, b)
 }
 
+/// Polynomial multiply long
+#[inline]
+#[cfg(target_arch = "arm")]
+#[target_feature(enable = "neon,crypto,v8")]
+#[cfg_attr(all(test, target_arch = "arm"), assert_instr(vmull))]
+pub unsafe fn vmull_p64(a: p64, b: p64) -> p128 {
+    #[allow(improper_ctypes)]
+    extern "C" {
+        #[cfg_attr(target_arch = "arm", link_name = "llvm.arm.neon.vmullp.v2i64")]
+        fn vmull_p64_(a: int64x1_t, b: int64x1_t) -> int64x2_t;
+    }
+transmute(vmull_p64_(transmute(a), transmute(b)))
+}
+
+/// Polynomial multiply long
+#[inline]
+#[cfg(target_arch = "aarch64")]
+#[target_feature(enable = "neon,crypto")]
+#[cfg_attr(all(test, target_arch = "aarch64"), assert_instr(pmull))]
+pub unsafe fn vmull_p64(a: p64, b: p64) -> p128 {
+    #[allow(improper_ctypes)]
+    extern "C" {
+        #[cfg_attr(target_arch = "aarch64", link_name = "llvm.aarch64.neon.pmull64")]
+        fn vmull_p64_(a: p64, b: p64) -> int8x16_t;
+    }
+transmute(vmull_p64_(a, b))
+}
+
 /// Vector long multiply with scalar
 #[inline]
 #[target_feature(enable = "neon")]
@@ -17890,6 +17918,15 @@ mod test {
         assert_eq!(r, e);
     }
 
+    #[simd_test(enable = "neon")]
+    unsafe fn test_vmull_p64() {
+        let a: p64 = 15;
+        let b: p64 = 3;
+        let e: p128 = 17;
+        let r: p128 = transmute(vmull_p64(transmute(a), transmute(b)));
+        assert_eq!(r, e);
+    }
+
     #[simd_test(enable = "neon")]
     unsafe fn test_vmullh_n_s16() {
         let a: i16x4 = i16x4::new(1, 2, 3, 4);

crates/stdarch-gen/neon.spec

@@ -2065,8 +2065,11 @@ target = crypto
 
 aarch64 = pmull
 link-aarch64 = pmull64:p64:p64:p64:int8x16_t
+arm = vmull
+link-arm = vmullp.v2i64:int64x1_t:int64x1_t:int64x1_t:int64x2_t
 generate p64:p64:p128
 
+
 /// Polynomial multiply long
 name = vmull_high
 no-q