This reverts commit 20a05be15ea5271ab6185b83200fa88263362400. (svn rev 138340)

Conflicts:

	test/Transforms/InstCombine/bitcast.ll

llvm-svn: 138366

[analyzer] Move helper method handleAutoreleaseCounts from CFRefCount to RetainReleaseChecker. No functionality change.

llvm-svn: 138365

llvm-svn: 138364

llvm-svn: 138363

llvm-svn: 138362

llvm-svn: 138361

llvm-svn: 138360

llvm-svn: 138359

[analyzer] Move ReturnStmt retain-count analysis from CFRefCount to RetainReleaseChecker. Tweak CFRefReport to reflect that fact that ReturnStmt checks are pre-statement, not post-statement.  No intended functionality change.

llvm-svn: 138358

llvm-svn: 138356

llvm-svn: 138354

llvm-svn: 138353

llvm-svn: 138352

llvm-svn: 138351

system flags an error when unlocking a lock which was not held, locking
the same lock twice, having a different lockset on each iteration of a
loop, or going out of scope while still holding a lock. In order to
successfully use the lockset, this patch also makes sure that attribute
arguments are attached correctly for later parsing.

This patch was also worked on by DeLesley Hutchins.

Note: This patch has been reviewed by Chandler Carruth and Jeffrey
Yasskin. Feel free to provide post-commit review comments for a
subsequent patch.

llvm-svn: 138350

llvm-svn: 138349

llvm-svn: 138348

Not including tSTRspi.

llvm-svn: 138347

Patch by Sanjoy Das!

llvm-svn: 138346

llvm-svn: 138345

llvm-svn: 138344

output on darwin so is hard coded there.

As a note this will need a little bit of refactoring in the class
hierarchy to separate it out for different verifiers based on input type.

Fixes rdar://8256258.

llvm-svn: 138343

llvm-svn: 138342

Fix decoding of Thumb2 prefetch instructions, which account for all the remaining Thumb2 decoding failures found by randomized testing so far.

llvm-svn: 138341

1. Cleanup the tests in ConstantFolding.cpp
2. Implement isAllOnes for Constant, ConstantFP, ConstantVector

llvm-svn: 138340

llvm-svn: 138339

There is no non-writeback store multiple instruction in Thumb1, so
don't define one. As a result load multiple is the only instantiation of
the multiclass, so refactor that away entirely.

llvm-svn: 138338

Fix two more instances of mis-matched operand names breaking disassembly.  Found by randomized testing.

llvm-svn: 138337

llvm-svn: 138336

32-bit little-endian CPU. Used by PNaCl and Emscripten.

llvm-svn: 138335

stmt expression, recursively walk down all substatements 
of the stmt expression. // rdar://10001085

llvm-svn: 138334

llvm-svn: 138332

Short option for --summary-string in 'type summary add' is now -s. This might be a breaking change for those who have summaries defined.

llvm-svn: 138331

Patch by Micah Villmow!

llvm-svn: 138330

llvm-svn: 138326

This fixes reference counting if the schedule optimizer is used.

llvm-svn: 138325

This matches gcc's behavior.

llvm-svn: 138324

The motivation to do that:

1. Now, llvm would use the stock config.sub. Before that we had an
uncommitted FreeBSD-related patch. Now, it has been upstreamed and
comes back. It means that it would be easier to update these files in
the next time (less magic knowledge)

2. Fix a typo for pseudo-CPUs: 32e[lb] -> [lb]e32, 64e[lb]->[lb]64.
One of these CPUs is used for PNaCl and it was not really convenient
to have a CPU that starts with a digit.

llvm-svn: 138323

Add support for breaking 256-bit v16i16 and v32i8 VSETCC into two 128-bit ones, avoiding sclarization. Add vex form of pcmpeqq and pcmpgtq. Fixes more cases for PR10712.

llvm-svn: 138321

[analyzer] Only allocate retain summaries for interesting functions/messages. This is a minor saving of memory but doesn't seem to cost any performance.

llvm-svn: 138320