https://git.reactos.org/?p=reactos.git;a=commitdiff;h=dde1b98a783b00cf0965a…
commit dde1b98a783b00cf0965a7d50ec818ee53504f46
Author: Amine Khaldi <amine.khaldi(a)reactos.org>
AuthorDate: Mon Oct 30 22:51:00 2017 +0100
[CRT/INTRIN] Use the gcc intrinsic equivalents when compiling with clang-cl and skip
the ones that are already supported/built-in. CORE-11799 (#94)
---
sdk/include/crt/intrin.h | 2 +-
sdk/include/crt/mingw32/intrin_x86.h | 70 +++++++++++++++++++++++++++++++-----
2 files changed, 63 insertions(+), 9 deletions(-)
diff --git a/sdk/include/crt/intrin.h b/sdk/include/crt/intrin.h
index cc25401f8c..8dc1944fe7 100644
--- a/sdk/include/crt/intrin.h
+++ b/sdk/include/crt/intrin.h
@@ -1012,7 +1012,7 @@ long _InterlockedIncrement(_Interlocked_operand_ long volatile *
_Addend);
}
#endif /* __cplusplus */
-#if defined(__GNUC__) && defined(_WIN32) // We can't use __MINGW32__ here
+#if (defined(__GNUC__) || defined(__clang__)) && defined(_WIN32) // We can't
use __MINGW32__ here
# include "mingw32/intrin.h"
#elif defined(_MSC_VER)
# include "msc/intrin.h"
diff --git a/sdk/include/crt/mingw32/intrin_x86.h b/sdk/include/crt/mingw32/intrin_x86.h
index 1b3e8e0991..f474a818d5 100644
--- a/sdk/include/crt/mingw32/intrin_x86.h
+++ b/sdk/include/crt/mingw32/intrin_x86.h
@@ -89,6 +89,7 @@ __INTRIN_INLINE void* __cdecl memcpy(void* dest, const void* source,
size_t num)
/*** Memory barriers ***/
+#ifndef __clang__
__INTRIN_INLINE void _ReadWriteBarrier(void)
{
__asm__ __volatile__("" : : : "memory");
@@ -116,6 +117,7 @@ __INTRIN_INLINE void _mm_sfence(void)
__asm__ __volatile__("sfence");
_WriteBarrier();
}
+#endif /* !__clang__ */
#ifdef __x86_64__
__INTRIN_INLINE void __faststorefence(void)
@@ -321,6 +323,8 @@ __INTRIN_INLINE long long _InterlockedIncrement64(volatile long long *
lpAddend)
#else /* (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) > 40100 */
+#ifndef __clang__
+
__INTRIN_INLINE char _InterlockedCompareExchange8(volatile char * Destination, char
Exchange, char Comperand)
{
char retval = Comperand;
@@ -571,6 +575,8 @@ __INTRIN_INLINE short _InterlockedIncrement16(volatile short *
lpAddend)
return _InterlockedExchangeAdd16(lpAddend, 1) + 1;
}
+#endif /* !__clang__ */
+
#if defined(__x86_64__)
__INTRIN_INLINE long long _InterlockedDecrement64(volatile long long * lpAddend)
{
@@ -594,6 +600,7 @@ __INTRIN_INLINE long long _InterlockedCompareExchange64(volatile long
long * Des
#else
+#ifndef __clang__
__INTRIN_INLINE long long _InterlockedCompareExchange64(volatile long long * Destination,
long long Exchange, long long Comperand)
{
long long retval = Comperand;
@@ -610,6 +617,7 @@ __INTRIN_INLINE long long _InterlockedCompareExchange64(volatile long
long * Des
return retval;
}
+#endif /* !__clang__ */
#endif
@@ -647,12 +655,14 @@ __INTRIN_INLINE unsigned char _interlockedbittestandreset64(volatile
long long *
}
#endif
+#ifndef __clang__
__INTRIN_INLINE unsigned char _interlockedbittestandset(volatile long * a, long b)
{
unsigned char retval;
__asm__("lock; btsl %[b], %[a]; setc %b[retval]" : [retval] "=q"
(retval), [a] "+m" (*a) : [b] "Ir" (b) : "memory");
return retval;
}
+#endif /* !__clang__ */
#if defined(__x86_64__)
__INTRIN_INLINE unsigned char _interlockedbittestandset64(volatile long long * a, long
long b)
@@ -665,6 +675,7 @@ __INTRIN_INLINE unsigned char _interlockedbittestandset64(volatile
long long * a
/*** String operations ***/
+#ifndef __clang__
/* NOTE: we don't set a memory clobber in the __stosX functions because Visual C++
doesn't */
__INTRIN_INLINE void __stosb(unsigned char * Dest, unsigned char Data, size_t Count)
{
@@ -675,6 +686,7 @@ __INTRIN_INLINE void __stosb(unsigned char * Dest, unsigned char Data,
size_t Co
"[Dest]" (Dest), "a" (Data), "[Count]" (Count)
);
}
+#endif
__INTRIN_INLINE void __stosw(unsigned short * Dest, unsigned short Data, size_t Count)
{
@@ -861,6 +873,8 @@ __INTRIN_INLINE void __writefsdword(unsigned long Offset, unsigned
long Data)
__asm__ __volatile__("movl %k[Data], %%fs:%a[Offset]" : : [Offset]
"ir" (Offset), [Data] "ir" (Data) : "memory");
}
+#ifndef __clang__
+
__INTRIN_INLINE unsigned char __readfsbyte(unsigned long Offset)
{
unsigned char value;
@@ -882,6 +896,8 @@ __INTRIN_INLINE unsigned long __readfsdword(unsigned long Offset)
return value;
}
+#endif /* !__clang__ */
+
__INTRIN_INLINE void __incfsbyte(unsigned long Offset)
{
__asm__ __volatile__("incb %%fs:%a[Offset]" : : [Offset] "ir"
(Offset) : "memory");
@@ -927,6 +943,8 @@ __INTRIN_INLINE void __addfsdword(unsigned long Offset, unsigned long
Data)
/*** Bit manipulation ***/
+#ifndef __clang__
+
__INTRIN_INLINE unsigned char _BitScanForward(unsigned long * Index, unsigned long Mask)
{
__asm__("bsfl %[Mask], %[Index]" : [Index] "=r" (*Index) : [Mask]
"mr" (Mask));
@@ -939,6 +957,8 @@ __INTRIN_INLINE unsigned char _BitScanReverse(unsigned long * Index,
unsigned lo
return Mask ? 1 : 0;
}
+#endif /* !__clang__ */
+
/* NOTE: again, the bizarre implementation follows Visual C++ */
__INTRIN_INLINE unsigned char _bittest(const long * a, long b)
{
@@ -1058,6 +1078,8 @@ __INTRIN_INLINE unsigned char _bittestandcomplement64(long long * a,
long long b
#endif
+#ifndef __clang__
+
__INTRIN_INLINE unsigned char __cdecl _rotl8(unsigned char value, unsigned char shift)
{
unsigned char retval;
@@ -1079,6 +1101,8 @@ __INTRIN_INLINE unsigned int __cdecl _rotl(unsigned int value, int
shift)
return retval;
}
+#endif /* !__clang__ */
+
#ifdef __x86_64__
__INTRIN_INLINE unsigned long long _rotl64(unsigned long long value, int shift)
{
@@ -1086,13 +1110,17 @@ __INTRIN_INLINE unsigned long long _rotl64(unsigned long long
value, int shift)
__asm__("rolq %b[shift], %k[retval]" : [retval] "=rm" (retval) :
"[retval]" (value), [shift] "Nc" (shift));
return retval;
}
-#else
+#else /* __x86_64__ */
+#ifndef __clang__
__INTRIN_INLINE unsigned long long __cdecl _rotl64(unsigned long long value, int shift)
{
/* FIXME: this is probably not optimal */
return (value << shift) | (value >> (64 - shift));
}
-#endif
+#endif /* !__clang__ */
+#endif /* __x86_64__ */
+
+#ifndef __clang__
__INTRIN_INLINE unsigned int __cdecl _rotr(unsigned int value, int shift)
{
@@ -1115,6 +1143,8 @@ __INTRIN_INLINE unsigned short __cdecl _rotr16(unsigned short value,
unsigned ch
return retval;
}
+#endif /* !__clang__ */
+
#ifdef __x86_64__
__INTRIN_INLINE unsigned long long _rotr64(unsigned long long value, int shift)
{
@@ -1122,13 +1152,17 @@ __INTRIN_INLINE unsigned long long _rotr64(unsigned long long
value, int shift)
__asm__("rorq %b[shift], %k[retval]" : [retval] "=rm" (retval) :
"[retval]" (value), [shift] "Nc" (shift));
return retval;
}
-#else
+#else /* __x86_64__ */
+#ifndef __clang__
__INTRIN_INLINE unsigned long long __cdecl _rotr64(unsigned long long value, int shift)
{
/* FIXME: this is probably not optimal */
return (value >> shift) | (value << (64 - shift));
}
-#endif
+#endif /* !__clang__ */
+#endif /* __x86_64__ */
+
+#ifndef __clang__
__INTRIN_INLINE unsigned long __cdecl _lrotl(unsigned long value, int shift)
{
@@ -1144,6 +1178,8 @@ __INTRIN_INLINE unsigned long __cdecl _lrotr(unsigned long value,
int shift)
return retval;
}
+#endif /* !__clang__ */
+
/*
NOTE: in __ll_lshift, __ll_rshift and __ull_rshift we use the "A"
constraint (edx:eax) for the Mask argument, because it's the only way GCC
@@ -1242,6 +1278,8 @@ __INTRIN_INLINE unsigned short __lzcnt16(unsigned short value)
return __builtin_clz(value);
}
+#ifndef __clang__
+
__INTRIN_INLINE unsigned int __popcnt(unsigned int value)
{
return __builtin_popcount(value);
@@ -1252,6 +1290,8 @@ __INTRIN_INLINE unsigned short __popcnt16(unsigned short value)
return __builtin_popcount(value);
}
+#endif /* !__clang__ */
+
#ifdef __x86_64__
__INTRIN_INLINE unsigned long long __lzcnt64(unsigned long long value)
{
@@ -1266,6 +1306,8 @@ __INTRIN_INLINE unsigned long long __popcnt64(unsigned long long
value)
/*** 64-bit math ***/
+#ifndef __clang__
+
__INTRIN_INLINE long long __emul(int a, int b)
{
long long retval;
@@ -1280,6 +1322,8 @@ __INTRIN_INLINE unsigned long long __emulu(unsigned int a, unsigned
int b)
return retval;
}
+#endif /* !__clang__ */
+
__INTRIN_INLINE long long __cdecl _abs64(long long value)
{
return (value >= 0) ? value : -value;
@@ -1435,6 +1479,7 @@ __INTRIN_INLINE void __cpuidex(int CPUInfo[4], int InfoType, int
ECXValue)
__asm__ __volatile__("cpuid" : "=a" (CPUInfo[0]), "=b"
(CPUInfo[1]), "=c" (CPUInfo[2]), "=d" (CPUInfo[3]) : "a"
(InfoType), "c" (ECXValue));
}
+#ifndef __clang__
__INTRIN_INLINE unsigned long long __rdtsc(void)
{
#ifdef __x86_64__
@@ -1447,6 +1492,7 @@ __INTRIN_INLINE unsigned long long __rdtsc(void)
return retval;
#endif
}
+#endif /* !__clang__ */
__INTRIN_INLINE void __writeeflags(uintptr_t Value)
{
@@ -1462,14 +1508,12 @@ __INTRIN_INLINE uintptr_t __readeflags(void)
/*** Interrupts ***/
-#ifdef __clang__
-#define __debugbreak() __asm__("int $3")
-#else
+#ifndef __clang__
+
__INTRIN_INLINE void __cdecl __debugbreak(void)
{
__asm__("int $3");
}
-#endif
__INTRIN_INLINE void __ud2(void)
{
@@ -1481,6 +1525,8 @@ __INTRIN_INLINE void __int2c(void)
__asm__("int $0x2c");
}
+#endif /* !__clang__ */
+
__INTRIN_INLINE void __cdecl _disable(void)
{
__asm__("cli" : : : "memory");
@@ -1496,12 +1542,14 @@ __INTRIN_INLINE void __halt(void)
__asm__("hlt" : : : "memory");
}
+#ifndef __clang__
__declspec(noreturn)
__INTRIN_INLINE void __fastfail(unsigned int Code)
{
__asm__("int $0x29" : : "c"(Code) : "memory");
__builtin_unreachable();
}
+#endif
/*** Protected memory management ***/
@@ -1579,10 +1627,12 @@ __INTRIN_INLINE void __writecr4(unsigned int Data)
__asm__("mov %[Data], %%cr4" : : [Data] "r" (Data) :
"memory");
}
+#ifndef __clang__
__INTRIN_INLINE void __writecr8(unsigned int Data)
{
__asm__("mov %[Data], %%cr8" : : [Data] "r" (Data) :
"memory");
}
+#endif
__INTRIN_INLINE unsigned long __readcr0(void)
{
@@ -1612,12 +1662,14 @@ __INTRIN_INLINE unsigned long __readcr4(void)
return value;
}
+#ifndef __clang__
__INTRIN_INLINE unsigned long __readcr8(void)
{
unsigned long value;
__asm__ __volatile__("mov %%cr8, %[value]" : [value] "=r" (value));
return value;
}
+#endif
#endif /* __x86_64__ */
@@ -1822,10 +1874,12 @@ __INTRIN_INLINE void _sgdt(void *Destination)
/*** Misc operations ***/
+#ifndef __clang__
__INTRIN_INLINE void _mm_pause(void)
{
__asm__ __volatile__("pause" : : : "memory");
}
+#endif
__INTRIN_INLINE void __nop(void)
{