Headers: be a bit more careful about inline asm

[compnerd]

Summary

Conditionally include x86intrin.h if we are building for x86 or x86_64.
Conditionalise definition of inline assembly routines which use x86 or x86_64
inline assembly. This is needed as clang can target Windows on ARM where these
definitions may be included into user code.

[rnk]

IMO intrin.h should avoid including x86intrin.h on non-x86 platforms, rather than making ia32intrin.h do nothing on non-x86. Does that work for you?

[compnerd]

I think that is perfectly reasonable. I did do that, but then backtracked away from it. Would you be against these changes to ia32intrin.h in addition to the inclusion guard?

[rnk]

lgtm

[abdulras]

SVN r211716.

lib/Headers/Intrin.h

 #define __INTRIN_H
 
 /* First include the standard intrinsics. */
+#if defined(__i386__) || defined(__x86_64__)
 #include <x86intrin.h>
+#endif
 
 /* For the definition of jmp_buf. */
 #include <setjmp.h>

   *a = *a | (1 << b);
   return x;
 }
+
+#if defined(__i386__) || defined(__x86_64__)
 static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
 _interlockedbittestandset(long volatile *__BitBase, long __BitPos) {
   unsigned char __Res;

            : "Ir"(__BitPos));
   return __Res;
 }
+#endif
+
 #ifdef __x86_64__
 static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
 _BitScanForward64(unsigned long *_Index, unsigned __int64 _Mask) {

 /*----------------------------------------------------------------------------*\
 |* Barriers
 \*----------------------------------------------------------------------------*/
+
+#if defined(__i386__) || defined(__x86_64__)
 static __inline__ void __attribute__((__always_inline__, __nodebug__))
 __attribute__((deprecated("use other intrinsics or C++11 atomics instead")))
 _ReadWriteBarrier(void) {

 _WriteBarrier(void) {
   __asm__ volatile ("" : : : "memory");
 }
+#endif
+
 #ifdef __x86_64__
 static __inline__ void __attribute__((__always_inline__, __nodebug__))
 __faststorefence(void) {
   __asm__ volatile("lock orq $0, (%%rsp)" : : : "memory");
 }
 #endif
+
 /*----------------------------------------------------------------------------*\
 |* readfs, readgs
 |* (Pointers in address space #256 and #257 are relative to the GS and FS

 /*----------------------------------------------------------------------------*\
 |* movs, stos
 \*----------------------------------------------------------------------------*/
+
+#if defined(__i386__) || defined(__x86_64__)
 static __inline__ void __attribute__((__always_inline__, __nodebug__))
 __movsb(unsigned char *__dst, unsigned char const *__src, size_t __n) {
   __asm__("rep movsb" : : "D"(__dst), "S"(__src), "c"(__n)

   __asm__("rep stosh" : : "D"(__dst), "a"(__x), "c"(__n)
                         : "%edi", "%ecx");
 }
+#endif
+
 #ifdef __x86_64__
 static __inline__ void __attribute__((__always_inline__, __nodebug__))
 __movsq(unsigned long long *__dst, unsigned long long const *__src, size_t __n) {

 _ReturnAddress(void) {
   return __builtin_return_address(0);
 }
+
+#if defined(__i386__) || defined(__x86_64__)
 static __inline__ void __attribute__((__always_inline__, __nodebug__))
 __cpuid(int __info[4], int __level) {
   __asm__ ("cpuid" : "=a"(__info[0]), "=b" (__info[1]), "=c"(__info[2]), "=d"(__info[3])

   __asm__ ("cpuid" : "=a"(__info[0]), "=b" (__info[1]), "=c"(__info[2]), "=d"(__info[3])
                    : "a"(__level), "c"(__ecx));
 }
+
 static __inline__ unsigned __int64 __cdecl __attribute__((__always_inline__, __nodebug__))
 _xgetbv(unsigned int __xcr_no) {
   unsigned int __eax, __edx;
   __asm__ ("xgetbv" : "=a" (__eax), "=d" (__edx) : "c" (__xcr_no));
   return ((unsigned __int64)__edx << 32) | __eax;
 }
+
 static __inline__ void __attribute__((__always_inline__, __nodebug__))
 __halt(void) {
   __asm__ volatile ("hlt");
 }
+#endif
 
 /*----------------------------------------------------------------------------*\
 |* Privileged intrinsics
 \*----------------------------------------------------------------------------*/
+#if defined(__i386__) || defined(__x86_64__)
 static __inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
 __readmsr(unsigned long __register) {
   // Loads the contents of a 64-bit model specific register (MSR) specified in

 __writecr3(unsigned int __cr3_val) {
   __asm__ ("mov %0, %%cr3" : : "q"(__cr3_val) : "memory");
 }
+#endif
 
 #ifdef __cplusplus
 }