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

llvm-svn: 138319

For the test case added to function-redecl.cpp, we were previously complaining
about a mismatch in the parameter types, since the definition used the
typedef'd type.

llvm-svn: 138318

SSE transition penalty. The pass is enabled through the "x86-use-vzeroupper"
llc command line option. This is only the first step (very naive and
conservative one) to sketch out the idea, but proper DFA is coming next
to allow smarter decisions. Comments and ideas now and in further commits
will be very appreciated.

llvm-svn: 138317

llvm-svn: 138316

Additional code cleanups ; Short option name for --python-script in type summary add moved from -s to -o (this is a preliminary step in moving the short option for --summary-string from -f to -s) ; Accordingly updated the test suite

llvm-svn: 138315

is captured by a given statement expression. // rdar://10001085

llvm-svn: 138314

[analyzer] Rename CFRefCount's evalSummary method to evalCallOrMessage, since it no longer, uh, evaluates call summaries.

llvm-svn: 138313

llvm-svn: 138312

llvm-svn: 138311

llvm-svn: 138310

[analyzer] Migrate the handling of retain-count-related RetEffects and ArgEffects from CFRefCount to RetainReleaseChecker. No intended functionality change.

llvm-svn: 138309

llvm-svn: 138308

Separated FormatNavigator and FormatManager in two different files ; moved FormatCategoryItem enum out of FormatManager.h as a debugger-wide lldb_private enum ; minor style cleanups

llvm-svn: 138307

llvm-svn: 138306

llvm-svn: 138305

llvm-svn: 138304

llvm-svn: 138303

value, we insert a load of the exception object and selector object from memory,
which is where it actually resides. If it's used by a PHI node, we follow that
to where it is being used. Eventually, all landingpad instructions should have
no uses. Any PHI nodes that were associated with those landingpads should be
removed.

llvm-svn: 138302

llvm-svn: 138301

llvm-svn: 138300

llvm-svn: 138299

llvm-svn: 138298

llvm-svn: 138297

[analyzer] MacOSKeychainAPIChecker: Users of KeyChain API often use free() to deallocate the password. Catch this error explicitly and generate the error message at the place where free() is called.

llvm-svn: 138296

llvm-svn: 138295

llvm-svn: 138294

llvm-svn: 138293

llvm-svn: 138292

llvm-svn: 138291

llvm-svn: 138290

Follow up to Jim's r138278. This fixes commuteInstruction so it handles two-address instructions correctly. I'll let Jim add a test case. :-)

llvm-svn: 138289

llvm-svn: 138288

llvm-svn: 138287

llvm-svn: 138286

llvm-svn: 138285

rolled back, and the testcase that the rollback
broke.

The new LLVM has a new ARM disassembler, which
may cause instability.  Keeping the old one would
force us into a contorted position vis-a-vis the
LLVM sources we bring in, so we will address
issues on the new one rather than keeping the old
one around.

llvm-svn: 138284

llvm-svn: 138283