llvm-svn: 198291

When there are cycles in the value graph we have to be careful interpreting
"Value*" identity as "value" equivalence. We interpret the value of a phi node
as the value of its operands.
When we check for value equivalence now we make sure that the "Value*" dominates
all cycles (phis).

%0 = phi [%noaliasval, %addr2]
%l = load %ptr
%addr1 = gep @a, 0, %l
%addr2 = gep @a, 0, (%l + 1)
store %ptr ...

Before this patch we would return NoAlias for (%0, %addr1) which is wrong
because the value of the load is from different iterations of the loop.

Tested on x86_64 -mavx at O3 and O3 -flto with no performance or compile time
regressions.

PR18068
radar://15653794

llvm-svn: 198290

Suggested by Aaron Ballman.

llvm-svn: 198288

During the years there have been some attempts at figuring out how to
align byval arguments. A look at the commit log suggests that they
were

* Use the ABI alignment.
* When that was not sufficient for x86-64, I added the 's' specification to
  DataLayout.
* When that was not sufficient Evan added the virtual getByValTypeAlignment.
* When even that was not sufficient, we just got the FE to add the alignment
  to the byval.

This patch is just a simple cleanup that removes my first attempt at fixing the
problem. I also added an AArch64 implementation of getByValTypeAlignment to
make sure this patch is a nop. I also left the 's' parsing for backward
compatibility.

I will send a short email to llvmdev about the change for anyone maintaining
an out of tree target.

llvm-svn: 198287

llvm-svn: 198286

Remove modifierType/Base from X86 disassembler tables as they are no longer used. Removes ~11.5K from static tables.

llvm-svn: 198284

llvm-svn: 198281

[SparcV9]: Use SRL instead of SLL to clear top 32-bits in ctpop:i32. SLL does not clear top 32 bit, only SRL does.

llvm-svn: 198280

llvm-svn: 198279

Remove need for MODIFIER_OPCODE in the disassembler tables. AddRegFrms are really more like OrRegFrm so we don't need a difference since we can just mask bits.

llvm-svn: 198278

Printing rounding control.
Enncoding for EVEX_RC (rounding control).

llvm-svn: 198277

Second attempt at Removing special form of AddRegFrm used by FP instructions. These instructions can be handled by MRMXr instead.

llvm-svn: 198276

llvm-svn: 198275

FIXME: Dragonegg may be updated at non-trivial changes.
llvm-svn: 198274

Stopgap measure until we can just use static_assert().

llvm-svn: 198273

llvm-svn: 198272

Remove UnaryTypeTraitExpr and switch all remaining type trait related handling
over to TypeTraitExpr.

The UTT/BTT/TT enum prefix and evaluation code is retained pending further
cleanup.

This is part of the ongoing work to unify type traits following the removal of
BinaryTypeTraitExpr in r197273.

llvm-svn: 198271

 1) Teach ExpectAndConsume() to emit expected and expected-after diagnostics
    using the generic diagnostic descriptions added in r197972, eliminating another
    set of trivial err_expected_* variations while maintaining existing behaviour.

 2) Lift SkipUntil() recovery out of ExpectAndConsume(). The Expect/Consume
    family of functions are primitive parser operations that now have the
    well-defined property of operating on single tokens. Factoring out recovery
    exposes opportunities for more consistent and tailored error recover at the
    call sites instead of just relying on a bottled SkipUntil formula.

llvm-svn: 198270

llvm-svn: 198269

llvm-svn: 198268

Patch by Ilia Filippov!

llvm-svn: 198267

llvm-svn: 198266

Revert r198238 and add FP disassembler tests. It didn't work and I didn't realized we had no FP disassembler test cases.

llvm-svn: 198265

Patch by joey.gouly@arm.com

llvm-svn: 198264

llvm-svn: 198263

llvm-svn: 198262

Currently fails with:
gotsan.cc:7686:40: error: the address of 'int __sanitizer_pthread_attr_getstack(void*, void**, size_t*)' will always evaluate as 'true' [-Werror=address]

llvm-svn: 198261

Previously any error in enum definition body stopped parsing it. With this
change parser tries to recover from errors.
The patch fixes PR10982.

Differential Revision: http://llvm-reviews.chandlerc.com/D2018

llvm-svn: 198259

llvm-svn: 198258

llvm-svn: 198257

lib/ASTMatchers/ASTMatchFinder.cpp:276:28: error: typedef 'traverse_can_only_be_instantiated_with_base_type' locally defined but not used [-Werror=unused-local-typedefs]
                            traverse_can_only_be_instantiated_with_base_type);

llvm-svn: 198256

lib/Support/ThreadLocal.cpp:53:15: error: typedef 'SIZE_TOO_BIG' locally defined but not used [-Werror=unused-local-typedefs]
   typedef int SIZE_TOO_BIG[sizeof(pthread_key_t) <= sizeof(data) ? 1 : -1];

Done the C++11 way, switching on and using LLVM_STATIC_ASSERT() instead of LLVM_ATTRIBUTE_UNUSED.

llvm-svn: 198255

llvm-svn: 198254

A helper function is a C++ function, and so even though one of the two
definitions is weak, it still technically triggers the ODR. Perhaps
these two definitions are ODR equivalent, but I'm not even confident in
that.

Instead, just define the function once, declare it as weak, and use
a wrapper that is clearly file-local. This avoids two definitions. Also
make the function extern "C" so that we can't even mess up the type
signature somehow or otherwise fail to match up the weak declaration
here with the interceptor defined elsewhere.

llvm-svn: 198253

Includes a fix for a missing highlight range caused by a ',' typo in the PP
diagnostics.

llvm-svn: 198252

llvm-svn: 198249

llvm-svn: 198243

llvm-svn: 198241

Remove special form of AddRegFrm used by FP instructions. These instructions can be handled by MRMXr instead.

llvm-svn: 198238

Checking the trailing letter of the mnemonic is insufficient.  Be more thorough
in the scanning of the instruction to ensure that we correctly work with the
predicated mnemonics.

llvm-svn: 198235