This is based on the discussions on: [LLVMdev] [RFC] Raising LLVM minimum required MSVC version to 2013 for trunk

llvm-svn: 229185

With this commit, llvm-dsymutil learns how to choose which DIEs
it will link in the final output and which ones it won't. This
is based on the 'valid relocation' information that has been
built in the previous commits.

The test only tests that we choose the right 'root DIEs'. The
selection algorithm (and especially the part that walk the
dependencies of a root DIE) lacks a bit test coverage. This
will be much easier to cover when we output actual Dwarf and
thus can use llvm-dwarfdump to verify the structure of the
emitted DIE trees. I'll add more tests then.

llvm-svn: 229183

llvm-svn: 229182

To be used in subsequent commits (separated to keep only core logic
in the follow-ups).

llvm-svn: 229181

llvm-svn: 229180

To be used in dsymutil (or any other client that wants to take
advantage of the fact that DIEs are stored in a vector).

llvm-svn: 229179

These 'valid relocations' in the debug_info section will be how
dsymutil identifies the DIEs it needs to keep in the linked debug
information.

llvm-svn: 229178

It turns out the debug map will be interogated both by name and
by object file address. Add the latter capability.

llvm-svn: 229177

llvm-svn: 229176

Two minor tweaks I noticed when reading through the code:
- No need to recompute begin() on every iteration.  We're not modifying the instructions in this loop.
- We can ignore PHINodes and Dbg intrinsics.  The current code does this anyways, but it will spend slightly more time doing so and will count towards the limit of instructions in the block.  It seems really silly to give up due the presence of PHIs...

Differential Revision: http://reviews.llvm.org/D7624

llvm-svn: 229175

llvm-svn: 229172

llvm-svn: 229171

llvm-svn: 229170

do so and fix up the calls.

llvm-svn: 229169

enough we can use it to initialize frame lowering.

llvm-svn: 229168

Clean up a subtarget function that has it passed in while we're at it.

llvm-svn: 229164

llvm-svn: 229163

Discovered by Halide users who had C++ code like this:
  Triple.setArch(Triple::x86);
  Triple.setOS(Triple::Windows);
  Triple.setObjectFormat(Triple::ELF);
  Triple.setEnvironment(Triple::MSVC);

This would produce the stringified triple of x86-windows-msvc, instead
of the x86-windows-msvc-elf string needed to run MCJIT.

With this change, they retain the -elf suffix.

llvm-svn: 229160

This takes the preposterous number of patterns in this section
that were last added to in r219033 down to just plain obnoxious.

With a little more work, we might get this down to just comical.

I've added more test cases to the existing file that checks these
patterns, but it seems that some of these patterns simply don't
exist with today's shuffle lowering.

llvm-svn: 229158

llc would hang trying to write output to a full pipe that FileCheck
wasn't reading. FileCheck wasn't reading from stdin because it needs a
file as a positional argument.

llvm-svn: 229157

because I didn't have binutils set up properly to build the gold plugin.

Fixes PR22581 which was filed because this broke the build for folks
relying on the plugin. Very sorry! =]

I've gotten the plugin stuff building now as well so it shouldn't keep
happening.

llvm-svn: 229156

llvm-svn: 229155

llvm-svn: 229154

these are expected to fix any 64->32 bit real truncation issues.

llvm-svn: 229153

llvm-svn: 229152

llvm-svn: 229151

llvm-svn: 229150

Original commit message:
SmallVector: Resolve a long-standing fixme by using the existing unitialized_copy dispatch.

This makes append() use memcpy for trivially copyable types.

llvm-svn: 229149

llvm-svn: 229144

compensate for slow machines, and the growing number of tests.

Prodded by:	Hans Wennborg

llvm-svn: 229143

This makes append() use memcpy for trivially copyable types.

llvm-svn: 229142

Reviewers: dsanders

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D7618

llvm-svn: 229140

Reviewers: dsanders

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D7615

llvm-svn: 229138

llvm-svn: 229136

This requires considering the size of the operand when
checking immediate legality.

llvm-svn: 229135

This prevents these from failing in a later commit.

llvm-svn: 229134

llvm-svn: 229130

This correctly prints the function pointers, and also prints
function signatures for symbols as opposed to just types.  So
actual functions in your program will now be printed with full
name and signature, as opposed to just name as before.

llvm-svn: 229129

This patch adds functionality in MIPS delay slot filler such as if delay slot
filler have to put NOP instruction into the delay slot of microMIPS JR
instruction, then instead of emitting NOP this instruction is replaced by
compact jump instruction JRC.

Differential Revision: http://reviews.llvm.org/D7522

llvm-svn: 229128

This patch fixes a problem I accidentally introduced in an instruction combine
on select instructions added at r227197. That revision taught the instruction
combiner how to fold a cttz/ctlz followed by a icmp plus select into a single
cttz/ctlz with flag 'is_zero_undef' cleared.

However, the new rule added at r227197 would have produced wrong results in the
case where a cttz/ctlz with flag 'is_zero_undef' cleared was follwed by a
zero-extend or truncate. In that case, the folded instruction would have
been inserted in a wrong location thus leaving the CFG in an inconsistent
state.

This patch fixes the problem and add two reproducible test cases to
existing test 'InstCombine/select-cmp-cttz-ctlz.ll'.

llvm-svn: 229124