[ARM][AArch64] Make ACLE __clzl/__clzll return unsigned int instead of unsigned long/uint64_t.

[craig.topper]

Summary

Use unsigned long in place of uint32_t for both clz and cls.

As far as I can tell this matches what ACLE defines and what gcc implements.

Noticed while investigating fixing https://github.com/llvm/llvm-project/issues/63113

[craig.topper]

Update __clzl test case to pass an unsigned long instead of signed long.

[tmatheson]

LGTM, thanks. Seems odd that the ACLE mixes uint32_t and unsigned int.

clang/lib/Headers/arm_acle.h

#if __SIZEOF_LONG__ == 4
   return __ror(__x, __y);
 #else
   return __rorll(__x, __y);
 #endif
 }
 
 
 /* CLZ */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
 __clz(uint32_t __t) {
-  return (uint32_t)__builtin_clz(__t);
+  return (unsigned int)__builtin_clz(__t);
 }
 
-static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
 __clzl(unsigned long __t) {
-  return (unsigned long)__builtin_clzl(__t);
+  return (unsigned int)__builtin_clzl(__t);
 }
 
-static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__))
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
 __clzll(uint64_t __t) {
-  return (uint64_t)__builtin_clzll(__t);
+  return (unsigned int)__builtin_clzll(__t);
 }
 
 /* CLS */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
 __cls(uint32_t __t) {
   return __builtin_arm_cls(__t);
 }
 
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
 __clsl(unsigned long __t) {
 #if __SIZEOF_LONG__ == 4
   return __builtin_arm_cls(__t);
 #else
   return __builtin_arm_cls64(__t);
 #endif
 }
 
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
 __clsll(uint64_t __t) {
   return __builtin_arm_cls64(__t);
 }
 
 /* REV */
 static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
 __rev(uint32_t __t) {
   return __builtin_bswap32(__t);

clang/test/CodeGen/arm_acle.c

uint64_t test_rorll(uint64_t x, uint32_t y) {
   return __rorll(x, y);
 }
 
 // ARM-LABEL: @test_clz(
 // ARM-NEXT:  entry:
 // ARM-NEXT:    [[TMP0:%.*]] = call i32 @llvm.ctlz.i32(i32 [[T:%.*]], i1 false)
 // ARM-NEXT:    ret i32 [[TMP0]]
 //
-uint32_t test_clz(uint32_t t) {
+unsigned test_clz(uint32_t t) {
   return __clz(t);
 }
 
 // AArch32-LABEL: @test_clzl(
 // AArch32-NEXT:  entry:
 // AArch32-NEXT:    [[TMP0:%.*]] = call i32 @llvm.ctlz.i32(i32 [[T:%.*]], i1 false)
 // AArch32-NEXT:    ret i32 [[TMP0]]
 //
 // AArch64-LABEL: @test_clzl(
 // AArch64-NEXT:  entry:
 // AArch64-NEXT:    [[TMP0:%.*]] = call i64 @llvm.ctlz.i64(i64 [[T:%.*]], i1 false)
 // AArch64-NEXT:    [[CAST_I:%.*]] = trunc i64 [[TMP0]] to i32
-// AArch64-NEXT:    [[CONV_I:%.*]] = sext i32 [[CAST_I]] to i64
-// AArch64-NEXT:    ret i64 [[CONV_I]]
+// AArch64-NEXT:    ret i32 [[CAST_I]]
 //
-long test_clzl(long t) {
+unsigned test_clzl(unsigned long t) {
   return __clzl(t);
 }
 
 // ARM-LABEL: @test_clzll(
 // ARM-NEXT:  entry:
 // ARM-NEXT:    [[TMP0:%.*]] = call i64 @llvm.ctlz.i64(i64 [[T:%.*]], i1 false)
 // ARM-NEXT:    [[CAST_I:%.*]] = trunc i64 [[TMP0]] to i32
-// ARM-NEXT:    [[CONV_I:%.*]] = sext i32 [[CAST_I]] to i64
-// ARM-NEXT:    ret i64 [[CONV_I]]
+// ARM-NEXT:    ret i32 [[CAST_I]]
 //
-uint64_t test_clzll(uint64_t t) {
+unsigned test_clzll(uint64_t t) {
   return __clzll(t);
 }
 
 // AArch32-LABEL: @test_cls(
 // AArch32-NEXT:  entry:
 // AArch32-NEXT:    [[CLS_I:%.*]] = call i32 @llvm.arm.cls(i32 [[T:%.*]])
 // AArch32-NEXT:    ret i32 [[CLS_I]]
 //