Fix PR3241: Currently EmitCopyFromReg emits a copy from the physical register to a virtual register unless it requires an expensive cross class copy. That means we are only treating "expensive to copy" register dependency as physical register dependency.
Also future proof the scheduler to handle "normal" physical register dependencies. The code is not exercised yet.

llvm-svn: 62074

llvm-svn: 62073

llvm.memcpy/memset/memmove.  This allows removal of some 
hackish code from basicaa.

llvm-svn: 62071

present.

llvm-svn: 62070

llvm-svn: 62069

llvm-svn: 62068

change.

llvm-svn: 62067

not a random piece of it.  No functionality change.

llvm-svn: 62066

parse them.  tblgen doesn't yet do anything with this info though.

llvm-svn: 62065

llvm-svn: 62064

llvm-svn: 62063

llvm-svn: 62062

llvm-svn: 62061

make ScalarExprEmitter::EmitCompare() emit the expression with the correct type instead of always zext it to an int
this fixes codegen of simple exprs in C++ like 'if (x != 0)'

llvm-svn: 62060

 - Some things are still hardcoded, and macosx-version-min comparison
   isn't implemented, but otherwise this very closely matches gcc.

 - The one exception is that arguments (like -framework or -Wl,) which are
   treated as linker inputs instead of options are not being
   forwarded yet.

llvm-svn: 62059

compensation for turning off gcc's inliner.  This gets
us closer to the amount of inlining we were getting before.
It is not a win on everything, of course, but seems to
gain overall.

llvm-svn: 62058

llvm-svn: 62057

llvm-svn: 62056

opcode on each delegation.
Instead the information is cached on construction and the cached flag used thereafter.
Introduced two predicates: isCall and isInvoke.

llvm-svn: 62055

 - Also, fix bug in MultipleValuesOption which was accepting joined
   arguments.

 - Add ArgList::getArgs, provides iterator over all arg instances for a
   given option.
 
 - Option definition is very much in need of cleaning...

llvm-svn: 62054

archs (as a driver driver).

llvm-svn: 62053

llvm-svn: 62052

llvm-svn: 62049

llvm-svn: 62048

not thrilled about 64-bit % in general, so rewrite to use * instead.

llvm-svn: 62047

We should treat vectors as atomic types, not like arrays.

llvm-svn: 62046

canonicalization transform based on duncan's comments:

1) improve the comment about %.
2) within our index loop make sure the offset stays 
   within the *type size*, instead of within the *abi size*.
   This allows us to reason explicitly about landing in tail
   padding and means that issues like non-zero offsets into
   [0 x foo] types don't occur anymore.

llvm-svn: 62045

respectively.  Inspired by a patch by Dan Villiom Podlaski Christiansen.

llvm-svn: 62044

llvm-svn: 62043

llvm-svn: 62042

llvm-svn: 62041

this patch "adds support for specifying multiple dependancy targets using ‘-MT’. Using this patch, I was able to use ‘ccc’ as C compiler for the few C files in LLVM. In my brief testing, the output remains identical to that of GCC."

Patch by Dan Villiom Podlaski Christiansen!

llvm-svn: 62040

llvm-svn: 62039

llvm-svn: 62038

This completes the ObjCContainerDecl AST cleanup (for now).

llvm-svn: 62037

llvm-svn: 62036

llvm-svn: 62035

llvm-svn: 62034

llvm-svn: 62033

No performance regression in my basic test.
Also fixed a type error in ActOnFinishSwitchStmt's arguments (body is a stmt).

llvm-svn: 62032