Just use a native "load" instead of translating the builtin later. Clang can

take it!

I wasn't able to get __builtin_ia32_loaddqu to transform into an unaligned
load...I'll have to look into it further.

llvm-svn: 129427
This commit is contained in:
Bill Wendling 2011-04-13 05:58:17 +00:00
parent efc283c076
commit b9c9e34cb3
3 changed files with 2 additions and 4 deletions

View File

@ -2092,8 +2092,6 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
// If palignr is shifting the pair of vectors more than 32 bytes, emit zero.
return llvm::Constant::getNullValue(ConvertType(E->getType()));
}
case X86::BI__builtin_ia32_loadups:
case X86::BI__builtin_ia32_loadupd:
case X86::BI__builtin_ia32_loaddqu: {
const llvm::Type *VecTy = ConvertType(E->getType());
const llvm::Type *IntTy = llvm::IntegerType::get(getLLVMContext(), 128);

View File

@ -466,7 +466,7 @@ _mm_loadr_pd(double const *dp)
static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
_mm_loadu_pd(double const *dp)
{
return __builtin_ia32_loadupd(dp);
return (__m128d){ dp[0], dp[1] };
}
static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))

View File

@ -539,7 +539,7 @@ _mm_load_ps(const float *p)
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
_mm_loadu_ps(const float *p)
{
return __builtin_ia32_loadups(p);
return (__m128){ p[0], p[1], p[2], p[3] }; /* __builtin_ia32_loadups(p); */
}
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))