and O0 + debug codegen.

llvm-svn: 168680

If the Src and Dst are the same instruction,
no loop-independent dependence is possible,
so we force the PossiblyLoopIndependent flag to false.

The test case results are updated appropriately.

llvm-svn: 168678

This patch migrates the sprintf optimizations from the simplify-libcalls
pass into the instcombine library call simplifier.

llvm-svn: 168677

llvm-svn: 168670

llvm-svn: 168666

llvm-svn: 168664

llvm-svn: 168663

llvm-svn: 168661

llvm-svn: 168660

llvm-svn: 168659

classes.  The associated test case still doesn't pass, but it does have far
fewer issues.
rdar://12719844

llvm-svn: 168657

Revert r168635 "Step towards implementation of pass manager with doInitialization and doFinalization per module detangled from runOn?? calls, still has temporary code not to break ASAN to be removed when that pass conforms to the proposed model".
It appears to have broken at least one buildbot.

llvm-svn: 168654

Added in first optimization using fast-math flags to serve as an example for following optimizations. SimplifyInstruction will now try to optimize an fmul observing its FastMathFlags to see if it can fold multiply by zero when 'nnan' and 'nsz' flags are set.

llvm-svn: 168648

Added in bitcode enum for the serializing of fast-math flags. Added in the reading/writing of fast-math flags from the OptimizationFlags record for BinaryOps.

llvm-svn: 168646

Added in the ability to read LLVM IR text that contains fast-math flags as a sequence of capital letters separated by spaces in any order. Added in the printing of the fast-math flags in a canonical order, and don't print the other flags when 'fast' is specified, as 'fast' implies all the others.

llvm-svn: 168645

llvm-svn: 168644

add a TODO for starting.

llvm-svn: 168643

Add in getter/setter methods for Instructions, allowing them to be the interface to FPMathOperator similarly to now NUS/NSW is handled.

llvm-svn: 168642

llvm-svn: 168638

llvm-svn: 168637

Step towards implementation of pass manager with doInitialization and doFinalization per module detangled from runOn?? calls, still has temporary code not to break ASAN to be removed when that pass conforms to the proposed model

Patch by Pedro Artigas, with feedback from by Chandler Carruth.

llvm-svn: 168635

llvm-svn: 168633

r168627), we no longer need to call the freezeReservedRegs() function a second
time.  Previously, this pass was conservatively adding the FP to the set of
reserved registers, requiring the second update to the reserved registers.
rdar://12719844

llvm-svn: 168631

r168627), we no longer need to call the freezeReservedRegs() function a second
time.  Previously, this pass was conservatively adding the FP to the set of
reserved registers, requiring the second update to the reserved registers.
rdar://12719844

llvm-svn: 168630

InstCombineLoadStoreAlloca.cpp, which had many issues.
(At least two bugs were noted on llvm-commits, and it was overly conservative.)
Instead, use getOrEnforceKnownAlignment.

llvm-svn: 168629

This pass was conservative in that it always reserved the FP to enable dynamic
stack realignment, which allowed the RA to use aligned spills for vector
registers.  This happens even when spills were not necessary.  The RA has 
since been improved to use unaligned spills when necessary.

The new behavior is to realign the stack if the frame pointer was already
reserved for some other reason, but don't reserve the frame pointer just
because a function contains vector virtual registers.

Part of rdar://12719844

llvm-svn: 168627

llvm-svn: 168622

llvm-svn: 168617

Enhancement to InstCombine. Try to catch this opportunity:
  
 ---------------------------------------------------------------
 ((X^C1) >> C2) ^ C3  => (X>>C2) ^ ((C1>>C2)^C3)
  where the subexpression "X ^ C1" has more than one uses, and
  "(X^C1) >> C2" has single use. 
 ---------------------------------------------------------------- 

 Reviewed by Nadav (with minor change per his request).

llvm-svn: 168615

llvm-svn: 168608

When I migrated the toascii simplifier in r168580 Benjamin Kramer noticed
a bug in one of the comments that I migrated.

llvm-svn: 168605

This patch migrates the printf optimizations from the simplify-libcalls
pass into the instcombine library call simplifier.

llvm-svn: 168604

llvm-svn: 168601

llvm-svn: 168600

llvm-svn: 168597

llvm-svn: 168592

llvm-svn: 168590

Simplify some repetitive code with it. No functionality change.

llvm-svn: 168587

llvm-svn: 168586

The bug can be triggered when we require LoopInfo analysis ahead of DominatorTree construction in a Module Pass. The cause is that the LoopInfo analysis itself also invokes DominatorTree construction, therefore, when PassManager schedules LoopInfo, it will add DominatorTree first. Then after that, when the PassManger turns to schedule DominatorTree invoked by the above ModulePass, it finds there is already a DominatorTree, so it delete the redundant one. However, somehow it still try to access that pass pointer after free as code pasted below, which results in segment fault.

llvm-svn: 168581