and add some protection against creating such.

llvm-svn: 47957

ExplodedNode that would occasionally result in heap corruption.

llvm-svn: 47956

llvm-svn: 47955

llvm-svn: 47954

llvm-svn: 47953

llvm-svn: 47952

llvm-svn: 47951

were being pruned in patterns where a variable was used more than once, e.g.:

  (or (and R32C:$rA, R32C:$rC), (and R32C:$rB, (not R32C:$rC)))

In this example, $rC is used more than once and is actually significant to
instruction selection pattern matching when commuted variants are produced.
This patch scans the pattern's clauses and collects the variables, creating
a set of variables that are used more than once. TreePatternNode::isIsomorphicTo()
also understands that multiply-used variables are significant.

llvm-svn: 47950

llvm-svn: 47948

llvm-svn: 47947

llvm-svn: 47946

llvm-svn: 47945

llvm-svn: 47944

    http://www.gnu.org/software/autoconf/manual/autoconf.html#Generic-Programs

llvm-svn: 47943

llvm-svn: 47941

llvm-svn: 47940

llvm-svn: 47939

except ppc long double.  This allows us to shrink constant pool
entries for x86 long double constants, which in turn allows us to
use flds/fldl instead of fldt.

llvm-svn: 47938

all the way to float, not stopping at double.

llvm-svn: 47937

Updated CFGStmtVisitor to be in accord with rr47913: CallExprs are no longer (automatically) block-level expressions in the CFG.

llvm-svn: 47935

llvm-svn: 47934

llvm-svn: 47933

bug in r47928 (Int64Ty is the correct type for the constant
pool entry here) and removes the asserts, now that the code
is capable of handling i128.

llvm-svn: 47932

For x86, if sse2 is available, it's not a good idea since cvtss2sd is slower than a movsd load and it prevents load folding. On x87, it's important to shrink fp constant since fldt is very expensive.

llvm-svn: 47931

theoretically useful, but not useful in practice.  It adds a bunch of 
complexity, and not much value.  It's best to nuke it.  One big advantage
is that it means the target interfaces will soon lose their SLoc arguments
and target queries can never emit diagnostics anymore (yay).  Removing this
also simplifies some of the core preprocessor which should make it slightly
faster.

Ted, I didn't simplify TripleProcessor, which can now have at most one
triple, and can probably just be removed.  Please poke at it when you have
time.

llvm-svn: 47930

llvm-svn: 47929

llvm-svn: 47928

findRegisterUseOperandIdx, findRegisterDefOperandIndx. Fix some naming inconsistencies.

llvm-svn: 47927

ExplodedGraph. This allows us to build "layered" ExplodedGraphs where one
simulation is layered on another.

llvm-svn: 47926

llvm-svn: 47925

llvm-svn: 47924

GRStmtNodeBuilder and is now responsible for adding its own nodes to the graph.

llvm-svn: 47923

llvm-svn: 47922

This is conceptually much cleaner, as it shows that the state is always a pointer.
It also makes it easier to marshall these values around across the internal APIs
of the path-sensitive engine.

llvm-svn: 47921

of the entire checker.

llvm-svn: 47920

allow us to pass the builder to plug-in transfer functions and allow those
functions to create intermediate nodes.

llvm-svn: 47919

llvm-svn: 47918

class (cosmetic).  First piece of byval implementation;
this doesn't work yet.  No functional change.

llvm-svn: 47917

llvm-svn: 47916

llvm-svn: 47915