llvm-svn: 157685

llvm-svn: 157684

llvm-svn: 157683

Clang docs update: list command-line flags that control the size of generated debug information. Reviewed by chandlerc@ and echristo@.

llvm-svn: 157681

SSE4A should not imply LZCNT and POPCNT. FMA4 should imply SSE4A. Add missing break at the end of btver1 feature list.

llvm-svn: 157680

llvm-svn: 157679

llvm-svn: 157678

llvm-svn: 157677

llvm-svn: 157676

llvm-svn: 157675

llvm-svn: 157674

If integer overflow causes one of the terms to reach zero, that can
force the entire expression to zero.

Fixes PR12929: cast<Ty>() argument of incompatible type

llvm-svn: 157673

No functionality.

llvm-svn: 157672

llvm-svn: 157669

setting breakpoints.  That's dangerous, since while we are setting a breakpoint,
the target might hit the dyld load notification, and start removing modules from
the list.  This change adds a GetMutex accessor to the ModuleList class, and
uses it whenever we are accessing the target's ModuleList (as returned by GetImages().)

<rdar://problem/11552372>

llvm-svn: 157668

llvm-svn: 157667

same value and were initialized with literals.  Clang will warn on code like
this:

enum A {
  FIRST = 1,
  SECOND = 1
};

llvm-svn: 157666

llvm-svn: 157664

llvm-svn: 157663

If-converter models predicated defs as read + write. The read should be marked as 'undef' since it may not already be live. This appeases -verify-machineinstrs.

llvm-svn: 157662

llvm-svn: 157661

Also, do not display the builtin name and macro expansion when the
function is a builtin.

llvm-svn: 157659

llvm-svn: 157656

Besides adding the new insertPass function, this patch uses it to
enhance the existing -print-machineinstrs so that the MachineInstrs
after a specific pass can be printed.

Patch by Bin Zeng!

llvm-svn: 157655

Apparently, a friend can't derive from a private class according to gcc.

llvm-svn: 157654

Check for NULL modules coming into the SearchFilter's ModulePasses & PlatformDarwin::ModuleIsExcludedForNonModuleSpecificSearches functions.

llvm-svn: 157653

checker functions exist.

llvm-svn: 157652

The register unit lists are typically much shorter than the register
overlap lists, and the backing table for register units has better cache
locality because it is smaller.

This makes llc about 0.5% faster. The regsOverlap() function isn't that hot.

llvm-svn: 157651

Register units are already used internally in TableGen to compute
register pressure sets and overlapping registers. This patch makes them
available to the code generators.

The register unit lists are differentially encoded so they can be reused
for many related registers. This keeps the total size of the lists below
200 bytes for most targets. ARM has the largest table at 560 bytes.

Add an MCRegUnitIterator for traversing the register unit lists. It
provides an abstract interface so the representation can be changed in
the future without changing all clients.

llvm-svn: 157650

 - hoist checks out of loops where SCEV is smart enough
 - add additional statistics to measure how much we loose for not supporting interprocedural and pointers loaded from memory

llvm-svn: 157649

Do not show the derived registers like "eax", ... for the vanilla "register read" command.
Also add a test scenario for that.

llvm-svn: 157647

This patch fixes it. // rdar://11488351

llvm-svn: 157646

llvm-svn: 157644

llvm-svn: 157643

llvm-svn: 157642

getter result type is safe but does not match with property 
type resulting in spurious warning followed by crash in
IRGen. // rdar://11515196

llvm-svn: 157641

llvm-svn: 157640

llvm-svn: 157639

This is a clean-room implementation based on public documentation and
I tried to validate it as much as possible against gcc.

llvm-svn: 157638

llvm-svn: 157635