Dale Johannesen 7609017dc3 Because of the EMMS problem, right now we have to support ... user-defined operations that use MMX register types, but the compiler shouldn't generate them on its own. This adds a Synthesizable abstraction to represent this, and changes the vector widening computation so it won't produce MMX types. (The motivation is to remove noise from the ABI compatibility part of the gcc test suite, which has some breakage right now.) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101951 91177308-0d34-0410-b5e6-96231b3b80d8		2010-04-20 22:34:09 +00:00
..
Analysis	PR6880: Don't dereference CallsExternalNode if it's NULL.	2010-04-20 12:16:50 +00:00
Archive	Don't write into MemoryBuffers.	2010-04-19 16:15:31 +00:00
AsmParser	reapply 'reject forward references to functions whose type don't match'	2010-04-20 04:49:11 +00:00
Bitcode	Revert 101465, it broke internal OpenGL testing.	2010-04-16 23:37:20 +00:00
CodeGen	Because of the EMMS problem, right now we have to support	2010-04-20 22:34:09 +00:00
CompilerDriver
ExecutionEngine	Don't rely upon the MCSymbol "isDefined" method to indicate if a label has been	2010-04-18 00:56:05 +00:00
Linker
MC	refactor .if handling code a bit.	2010-04-17 18:14:27 +00:00
Support	Fix typo. add a test case.	2010-04-19 20:31:39 +00:00
System	This doesn't need SmallVector.h anymore.	2010-04-19 17:51:31 +00:00
Target	Because of the EMMS problem, right now we have to support	2010-04-20 22:34:09 +00:00
Transforms	Rename ValueMapTy as ValueToValueMapTy to clearly indicate that this has no replationship with ADT/ValueMap.	2010-04-20 22:24:18 +00:00
VMCore	Remove the palignr intrinsics now that we lower them to vector shuffles,	2010-04-20 00:59:54 +00:00
Makefile