Add instruction dump output.  This helps find bugs.

llvm-svn: 41744

Pluggable coalescers inplementation.

llvm-svn: 41743

llvm-svn: 41739

Use pool allocator for all the VNInfo's to improve memory access locality. This reduces coalescing time on siod Mac OS X PPC by 35%. Also remove the back ptr from VNInfo to LiveInterval and other tweaks.

llvm-svn: 41729

may be the same as the first label for the following
invoke.  Remove a micro-optimization which was wrong
in this case.

llvm-svn: 41720

labels are generated bracketing each call (not just
invokes).  This is used to generate entries in
the exception table required by the C++ personality.
However it gets in the way of tail-merging.  This
patch solves the problem by no longer placing labels
around ordinary calls.  Instead we generate entries
in the exception table that cover every instruction
in the function that wasn't covered by an invoke
range (the range given by the labels around the invoke).
As an optimization, such entries are only generated for
parts of the function that contain a call, since for
the moment those are the only instructions that can
throw an exception [1].  As a happy consequence, we
now get a smaller exception table, since the same
region can cover many calls.  While there, I also
implemented folding of invoke ranges - successive
ranges are merged when safe to do so.  Finally, if
a selector contains only a cleanup, there's a special
shorthand for it - place a 0 in the call-site entry.
I implemented this while there.  As a result, the
exception table output (excluding filters) is now
optimal - it cannot be made smaller [2].  The
problem with throw filters is that folding them
optimally is hard, and the benefit of folding them is
minimal.

[1] I tested that having trapping instructions (eg
divide by zero) in such a region doesn't cause trouble.
[2] It could be made smaller with the help of higher
layers, eg by having branch folding reorder basic blocks
ending in invokes with the same landing pad so they
follow each other.  I don't know if this is worth doing.

llvm-svn: 41718

llvm-svn: 41712

llvm-svn: 41676

Emit proper "secrel" directive, where possible. This fixes invalid asm syntax of debug info on mingw32. Also, cleanup 
some stuff.

llvm-svn: 41675

llvm-svn: 41669

Implement some constant folding in SelectionDAG and
DAGCombiner using APFloat.  Remove double versions
of constructor and getValue from ConstantFPSDNode.

llvm-svn: 41664

llvm-svn: 41650

llvm-svn: 41645

llvm-svn: 41641

Use std::map instead of a (potentially very sparse) array to track val# defined by copy from the other live range. Minor compile time win when number of val# is large.

llvm-svn: 41640

Use APFloat interfaces for more references, mostly
of ConstantFPSDNode.

llvm-svn: 41632

llvm-svn: 41598

Add APFloat interfaces to ConstantFP, SelectionDAG.
Fix integer bit in double->APFloat conversion.
Convert LegalizeDAG to use APFloat interface in
ConstantFPSDNode uses.

llvm-svn: 41587

llvm-svn: 41584

Changes related modules so VNInfo's are not copied. This decrease
copy coalescing time by 45% and overall compilation time by 10% on siod.

llvm-svn: 41579

llvm-svn: 41578

scheduling.

llvm-svn: 41556

1. Eliminate the costly live interval "swapping".
2. Change ValueNumberInfo container from SmallVector to std::vector. The former
   performs slowly when the vector size is very large.

llvm-svn: 41536

where both pointers have non-zero offsets.

llvm-svn: 41491

to not alias each other, it can be translated as an llvm.memcpy.

llvm-svn: 41489

gcc exception handling: if an exception unwinds through
an invoke, then execution must branch to the invoke's
unwind target.  We previously tried to enforce this by
appending a cleanup action to every selector, however
this does not always work correctly due to an optimization
in the C++ unwinding runtime: if only cleanups would be
run while unwinding an exception, then the program just
terminates without actually executing the cleanups, as
invoke semantics would require.  I was hoping this
wouldn't be a problem, but in fact it turns out to be the
cause of all the remaining failures in the LLVM testsuite
(these also fail with -enable-correct-eh-support, so turning
on -enable-eh didn't make things worse!).  Instead we need
to append a full-blown catch-all to the end of each
selector.  The correct way of doing this depends on the
personality function, i.e. it is language dependent, so
can only be done by gcc.  Thus this patch which generalizes
the eh.selector intrinsic so that it can handle all possible
kinds of action table entries (before it didn't accomodate
cleanups): now 0 indicates a cleanup, and filters have to be
specified using the number of type infos plus one rather than
the number of type infos.  Related gcc patches will cause
Ada to pass a cleanup (0) to force the selector to always
fire, while C++ will use a C++ catch-all (null).

llvm-svn: 41484

llvm-svn: 41482

llvm-svn: 41409

over uses in DAGCombiner.  Fix interfaces to work
with APFloats.

llvm-svn: 41407

know how to handle them.  This fixes 
test/CodeGen/Generic/asm-large-immediate.ll

llvm-svn: 41388

llvm-svn: 41386

    .set  Lset0,___gxx_personality_v0-.
    .long Lset0

on some targets. Make it so!

llvm-svn: 41385

changing the interface to allow for future changes.

llvm-svn: 41384

This means backing out the preceding change to Constants.cpp, alas.

llvm-svn: 41378

llvm-svn: 41372

llvm-svn: 41316

llvm-svn: 41286

scheduler implementations.

llvm-svn: 41191

 *) The needed information is propagated to the DAG
 *) The X86-64 backend detects it and aborts

llvm-svn: 41179

llvm-svn: 41163