All shift operations will be selected as SALU instructions and then
if necessary lowered to VALU instructions in the SIFixSGPRCopies pass.

This allows us to do more operations on the SALU which will improve
performance and is also required for implementing private memory
using indirect addressing, since the private memory pointers must stay
in the scalar registers.

This patch includes some fixes from Matt Arsenault.

llvm-svn: 194625

llvm-svn: 194624

Add a log message to the console that will display the error code when we fail to reply to a mach message.

llvm-svn: 194623

llvm-svn: 194622

Added two new GDB server packets to debugserver: "QSaveRegisterState" and "QRestoreRegiterState".

"QSaveRegisterState" makes the remote GDB server save all register values and it returns a save identifier as an unsigned integer. This packet can be used prior to running expressions to save all registers.

All registers can them we later restored with "QRestoreRegiterState:SAVEID" what SAVEID is the integer identifier that was returned from the call to QSaveRegisterState.

Cleaned up redundant code in lldb_private::Thread, lldb_private::ThreadPlanCallFunction.
Moved the lldb_private::Thread::RegisterCheckpoint into its own header file and it is now in the lldb_private namespace. Trimmed down the RegisterCheckpoint class to omit stuff that wasn't used (the stack ID).

Added a few new virtual methods to lldb_private::RegisterContext that allow subclasses to efficiently save/restore register states and changed the RegisterContextGDBRemote to take advantage of these new calls.

llvm-svn: 194621

llvm-svn: 194620

llvm-svn: 194619

This patch disables aliasing (and rauw) of derived dtors to base dtors at -O0.
This optimization can have a negative impact on the debug quality.

This was a latent bug for some time with local classes, but got noticed when it
was generalized and broke gdb's destrprint.exp.

llvm-svn: 194618

Differential Revision: http://llvm-reviews.chandlerc.com/D1989

llvm-svn: 194617

This test compares the output of llvm-cov against a coverage file
generated by gcov. Currently, llvm-cov does not work on certain
platforms (namely big-endian architectures such as PowerPC, among
others). These platforms are marked as XFAIL for now, but will be fixed
later.

llvm-svn: 194616

We already have builtins that are only available in GNU mode, so this
mirrors that.

Reviewers: rsmith

Differential Revision: http://llvm-reviews.chandlerc.com/D2128

llvm-svn: 194615

Move <dynarray> into include/experimental, and into the std::experimental namespace, since it's not part of C++14, but of an upcoming TS

llvm-svn: 194614

llvm-svn: 194612

llvm-svn: 194611

llvm-svn: 194610

improving comments to make documentation more accurate.

Differential Revision:http://llvm-reviews.chandlerc.com/D2172

llvm-svn: 194609

llvm-svn: 194607

llvm-svn: 194606

llvm-svn: 194605

Reid pointed out we already have LLVM_USE_CRT_{buildtype} to set it.

llvm-svn: 194604

llvm-svn: 194603

llvm-svn: 194602

llvm-svn: 194601

Remove the --do-read option, and always provide a small dump of memory at each match spot
Add a --dump-offset (-o) option, to specify a byte offset from which to start dumping relative to the matching address

The real solution is to actually provide the format options found on "memory read" and use those as the key to actually printing memory upon each find
That, however, requires a little refactoring work, so put this in for now until I get a chance to do the required shuffling around of moving parts

llvm-svn: 194600

A number of non-overloaded intrinsics have been replaced by thier overloaded
counterparts.

llvm-svn: 194599

instructions.  This patch does not include the shift right and accumulate
instructions.  A number of non-overloaded intrinsics have been remove in favor
of their overloaded counterparts.

llvm-svn: 194598

for release builds.

This is a follow-up to r194589. Aaron pointed out that building
libraries with /MT and using them in an application that uses a
different run-time library can be a bad idea.

Move the option to build with /MT behind a CMake option so it can be
turned on selectively, such as when building the toolchain installer.

llvm-svn: 194596

Now that the relevant tests use -mconstructor-aliases and the missing
features have been implemented, we can just drop this.

No functionality change.

llvm-svn: 194595

To support the generation of IT block in Thumbv7 way or
Thumbv8 way

llvm-svn: 194593

By default, the behavior of IT block generation will be determinated
dynamically base on the arch (armv8 vs armv7). This patch adds backend
options: -arm-restrict-it and -arm-no-restrict-it.  The former one
restricts the generation of IT blocks (the same behavior as thumbv8) for
both arches. The later one allows the generation of legacy IT block (the
same behavior as ARMv7 Thumb2) for both arches.

Clang will support -mrestrict-it and -mno-restrict-it, which is
compatible with GCC.

llvm-svn: 194592

llvm-svn: 194591

No functionality change.

llvm-svn: 194590

This should fix the problem of snapshot builds created with MSVC 2012 not
working for users with MSVC 2010, etc.

Differential Revision: http://llvm-reviews.chandlerc.com/D2157

llvm-svn: 194589

DIEHash: Move header include to be first in the implementation file to flush out header inclusion ordering issues

llvm-svn: 194588

llvm-svn: 194587

At the moment this is just the MC support.

llvm-svn: 194585

llvm-svn: 194584

llvm-svn: 194583

llvm-svn: 194582

Otherwise it's impossible to use it. Also don't include C++ headers in
a C header.

llvm-svn: 194581