llvm-svn: 141504

A GR8_NOREX virtual register is created when extrating a sub_8bit_hi
sub-register:

  %vreg2<def> = COPY %vreg1:sub_8bit_hi; GR8_NOREX:%vreg2 %GR64_ABCD:%vreg1
  TEST8ri_NOREX %vreg2, 1, %EFLAGS<imp-def>; GR8_NOREX:%vreg2

If such a live range is ever split, its register class must not be
inflated to GR8.  The sub-register copy can only target GR8_NOREX.

I dont have a test case for this theoretical bug.

llvm-svn: 141500

In 64-bit mode, sub_8bit_hi sub-registers can only be used by NOREX
instructions. The COPY created from the EXTRACT_SUBREG DAG node cannot
target all GR8 registers, only those in GR8_NOREX.

TO enforce this, we ensure that all instructions using the
EXTRACT_SUBREG are GR8_NOREX constrained.

This fixes PR11088.

llvm-svn: 141499

llvm-svn: 141498

related to a readwrite property, and
Sub's readwrite property. // rdar://9396329

llvm-svn: 141497

llvm-svn: 141496

llvm-svn: 141495

llvm-svn: 141494

llvm-svn: 141492

llvm-svn: 141490

lib/Frontend/CompilerInstance.cpp: Suppress a "Comparision of unsigned and signed" warning on Cygwin gcc-4.3.4.

llvm-svn: 141488

llvm-svn: 141487

llvm-svn: 141486

llvm-svn: 141485

llvm-svn: 141484

llvm-svn: 141483

Apparently getPtrToInt() takes an explicit type parameter to allow you to generate invalid bitcode, not so that it can actually produce a value of this type.  This should fix PR11085.

llvm-svn: 141482

llvm-svn: 141481

Changed RangeMap over to use llvm::SmallVector and updated the RangeArray and the RangeDataArray to have an extra "unsigned N" template parameter. Updated the lldb_private::Block to use a RangeArray with a uint32_t for both the function base offset and block range size, and then a 1 for the small vector size since most lexical blocks in DWARF only have 1 range. Updates the DWARFDebugRanges RangeArray to use an unsigned of 2 since most blocks that have more than one range usually have 2. Also updated a DWARFDebugAranges to default their RangeArray to use a SmallVector with unsigned size of 1 since this will take care of the .o files when doing DWARF in .o files and since there really isn't any good size we can guess with.

llvm-svn: 141480

Free 'TheTargetCodeGenInfo' in the class CodeGenModule.  This looks like an obvious memory leak that was reported from LLDB devs.  The comment indicates the leak is deliberate, but I have no idea why this needs to be so.  Please comment/revert if you know otherwise.

llvm-svn: 141479

Added some debug code that can verify that all RangeMap.h maps are sorted when needed since RangeArray and RangeDataArray instantiations depend on it. I ran the test suite with ASSERT_RANGEMAP_ARE_SORTED defined and I got no assertions, so I disabled it prior to checkin.

llvm-svn: 141478

llvm-svn: 141477

llvm-svn: 141476

llvm-svn: 141475

llvm-svn: 141474

conversion instructions. 

llvm-svn: 141473

llvm-svn: 141472

instruction selector to generate them.

llvm-svn: 141471

llvm-svn: 141470

llvm-svn: 141469

llvm-svn: 141468

llvm-svn: 141467

across unwind edges. This is for the back-end which expects such things.

The code is from the original SjLj EH pass.

llvm-svn: 141463

llvm-svn: 141462

llvm-svn: 141461

Added more functionality to Range template classes in RangeMap.h and converted remaining DWARF areas that were using ranges over to this class. Also converted lldb_private::Block to use it.

llvm-svn: 141460

llvm-svn: 141459

llvm-svn: 141458

llvm-svn: 141457

llvm-svn: 141456