No need recognize load from a fixed argument slot as re-materializable. LiveIntervalAnalysis already handles it as a special case.

llvm-svn: 47522

llvm-svn: 47521

Recognize loads of arguments as re-materializable first. Therefore if isReallyTriviallyReMaterializable() returns true it doesn't confuse it as a "normal" re-materializable instruction.

llvm-svn: 47520

llvm-svn: 47519

llvm-svn: 47518

llvm-svn: 47517

llvm-svn: 47516

llvm-svn: 47514

llvm-svn: 47513

llvm-svn: 47511

llvm-svn: 47510

llvm-svn: 47509

To support multiple return values, now ret instruction supports multiple operands instead of one aggregate operand.

llvm-svn: 47508

llvm-svn: 47507

stuff into ParamAttrsList.h.  Per feedback from
ParamAttrs changes.

llvm-svn: 47504

llvm-svn: 47503

llvm-svn: 47502

llvm-svn: 47501

llvm-svn: 47500

Really really bad local register allocator bug. On X86, it was never using ESI, EDI, and EBP because of a bug in RALocal::isPhysRegAvailable(). For example, when
it checks if ESI is available, it then looks at registers aliases to ESI. SIL is marked -2 (not allocatable) but isPhysRegAvailable() incorrectly assumes it is in use and returns false for ESI.

llvm-svn: 47499

llvm-svn: 47496

llvm-svn: 47495

llvm-svn: 47494

llvm-svn: 47493

llvm-svn: 47492

llvm-svn: 47488

llvm-svn: 47485

the way through.  It is now used for codegen.

llvm-svn: 47484

llvm-svn: 47483

llvm-svn: 47482

llvm-svn: 47481

llvm-svn: 47480

llvm-svn: 47479

llvm-svn: 47478

llvm-svn: 47476

the definition of the operand also reaches its uses.

llvm-svn: 47475

instead of with mmx registers.  This horribleness is apparently
done by gcc to avoid having to insert emms in places that really 
should have it.  This is the second half of rdar://5741668.

llvm-svn: 47474

llvm-svn: 47473

llvm-svn: 47472

llvm-svn: 47471