Preston Gurd 1edadea42f For the current Atom processor, the fastest way to handle a call ... indirect through a memory address is to load the memory address into a register and then call indirect through the register. This patch implements this improvement by modifying SelectionDAG to force a function address which is a memory reference to be loaded into a virtual register. Patch by Sriram Murali. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178171 91177308-0d34-0410-b5e6-96231b3b80d8		2013-03-27 19:14:02 +00:00
..
Analysis	BasicAA: Only query twice if the result of the more general query was MayAlias	2013-03-26 18:07:53 +00:00
Archive
AsmParser
Bitcode
CodeGen	Fix target-customized spilling in the register scavenger	2013-03-27 13:00:56 +00:00
DebugInfo	Fix missing std::. Not sure how this compiles for anyone else.	2013-03-21 00:57:21 +00:00
ExecutionEngine
IR	Fix comment	2013-03-27 17:50:12 +00:00
IRReader	Split out the IRReader header and the utility functions it provides into	2013-03-26 02:25:37 +00:00
Linker	Specutively revert r178130.	2013-03-27 17:54:41 +00:00
MC	Dead code.	2013-03-19 22:13:05 +00:00
Object
Option
Support	Don't try to generate crash diagnostics if we had an I/O failure. It's very	2013-03-27 18:30:00 +00:00
TableGen	Allow TableGen DAG arguments to be just a name.	2013-03-24 19:36:51 +00:00
Target	For the current Atom processor, the fastest way to handle a call	2013-03-27 19:14:02 +00:00
Transforms	Use the full path when outputting the `.gcda' file.	2013-03-26 22:47:50 +00:00
CMakeLists.txt	Split out the IRReader header and the utility functions it provides into	2013-03-26 02:25:37 +00:00
LLVMBuild.txt	Split out the IRReader header and the utility functions it provides into	2013-03-26 02:25:37 +00:00
Makefile	Split out the IRReader header and the utility functions it provides into	2013-03-26 02:25:37 +00:00