if-conversion.  The RemoveExtraEdges function doesn't work for blocks that
end with unanalyzable branches, so in those cases, the "extra" edges must
be explicitly removed.  The CopyAndPredicateBlock and MergeBlocks methods
can also avoid copying successor edges due to branches that have already
been removed.  The latter case is especially helpful when MergeBlocks is
called for handling "diamond" if-conversions, where otherwise you can end
up with some weird intermediate states in the CFG.  Unfortunately I've
been unable to find cases where this cleanup actually makes a significant
difference in the code.  There is one test where we manage to remove an
empty block at the end of a function.  Radar 6911268.

llvm-svn: 106939

llvm-svn: 106938

Also, fix mangling of throw specs. Turns out MSVC totally ignores throw
specs when mangling names.

llvm-svn: 106937

llvm-svn: 106935

CopyFromReg nodes for aliasing registers (AX and AL). This confuses the fast
register allocator.

Instead of CopyFromReg(AL), use ExtractSubReg(CopyFromReg(AX), sub_8bit).

This fixes PR7312.

llvm-svn: 106934

llvm-svn: 106932

llvm-svn: 106929

llvm-svn: 106927

It complained with:

llvm[5]: Building Clang arm_neon.h.inc with tblgen
cp: cannot create regular file `/build/buildbot-llvm/clang-x86_64-linux-selfhost-rel/llvm.obj.2/Release/lib/clang/2.0/include/arm_neon.h': No such file or directory

llvm-svn: 106922

"N..." instead of "NEON..." for consistency with the other NEON format names.

llvm-svn: 106921

llvm-svn: 106919

llvm-svn: 106918

llvm-svn: 106917

Renumber MiscFrm to 25.

llvm-svn: 106916

clang is now clang 2.8.

llvm-svn: 106914

llvm-svn: 106913

Add several AVX MOV flavors
Support VEX encoding for MRMDestReg

llvm-svn: 106912

llvm-svn: 106911

llvm-svn: 106910

llvm-svn: 106909

introduced in r106343, but only showed up recently (with a particular compiler &
linker combination) because of the particular check, and because we have no
builtin checking for dereferencing the end of an array, which is truly
unfortunate.

llvm-svn: 106908

just at the head, when doing diamond if-conversion. rdar://7797940

llvm-svn: 106907

llvm-svn: 106906

The VNInfo.kills vector was almost unused except for all the code keeping it
updated. The few places using it were easily rewritten to check for interval
ends instead.

The two new methods LiveInterval::killedAt and killedInRange are replacements.

This brings us down to 3 independent data structures tracking kills.

llvm-svn: 106905

invalid source range for CXXNewExpr.

llvm-svn: 106904

cases where we generated an invalid SourceRange for this expression.  Thanks to John McCall
for helping me figure this out.

llvm-svn: 106903

llvm-svn: 106901

emitNEON2RegInstruction's handling of 2-address operands.

llvm-svn: 106900

llvm-svn: 106898

SCEVUnknown values which are loop-variant, as LSR can't do anything
interesting with these values in any case. This fixes very slow compile
times on loops which have large numbers of such values.

llvm-svn: 106897

Support VEX encoding for MRMDestReg

llvm-svn: 106896

llvm-svn: 106895

llvm-svn: 106894

for an "i" constraint should get lowered; PR 6309.  While
this argument was passed around a lot, this is the only
place it was used, so it goes away from a lot of other
places.

llvm-svn: 106893

llvm-svn: 106891

Tests for lldb are written as python scripts which take advantage of the script
bridging provided by LLDB.framework to interact with lldb core.

A specific naming pattern is followed by the .py script to be recognized as
a module which implements a test scenario, namely, Test*.py.

To specify the directories where "Test*.py" python test scripts are located,
you need to pass in a list of directory names.  By default, the current
working directory is searched if nothing is specified on the command line.

An example:

[14:10:20] johnny:/Volumes/data/lldb/svn/trunk/test $ ./dotest.py -v example
test_choice (TestSequenceFunctions.TestSequenceFunctions) ... ok
test_sample (TestSequenceFunctions.TestSequenceFunctions) ... ok
test_shuffle (TestSequenceFunctions.TestSequenceFunctions) ... ok

----------------------------------------------------------------------
Ran 3 tests in 0.000s

OK

llvm-svn: 106890

llvm-svn: 106887

llvm-svn: 106886

llvm-svn: 106885

a winowed list of checkers that actually do something for a given StmtClass.
As the number of checkers grows, this may potentially significantly reduce
the number of checkers called at any one time.  My own measurements show that
for the ~20 registered Checker objects, only ~5 of them respond at any one time
to a give statement.  While this isn't a net performance win right now (there
is a minor slowdown on sqlite.3) this improvement does greatly improve debugging
when stepping through the checkers used to evaluate a given statement.

llvm-svn: 106884