At least with cmake 3.6.1, the default build type setting was having no
effect; the generated CMakeCache.txt still had an empty CMAKE_BUILD_TYPE.
Force the variable to be set to achieve the desired behavior.

Differential Revision: https://reviews.llvm.org/D26200

llvm-svn: 285789

llvm-svn: 285788

Summary:
We define a new trampoline that's a hybrid between the exit and entry
trampolines with the following properties:

- Saves all of the callee-saved registers according to the x86_64
  calling conventions.
- Indicate to the log handler function being called that this is a
  function exit event.

This fixes a bug that is a result of not saving enough of the register
states, and that the log handler is clobbering registers that would be
used by the function being tail-exited into manifesting as runtime
errors.

Reviewers: rSerge, echristo, majnemer

Subscribers: mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D26020

llvm-svn: 285787

Previously __libcpp_is_constructible checked the validity of reference
construction using 'eat<To>(declval<From>())' but this doesn't consider
From's explicit conversion operators. This patch teaches __libcpp_is_constructible
how to handle these cases. To do this we need to check the validity
using 'static_cast<To>(declval<From>())'. Unfortunately static_cast allows
additional base-to-derived and lvalue-to-rvalue conversions, which have to be
checked for and manually rejected.

While implementing these changes I discovered that Clang incorrectly
rejects `static_cast<int&&>(declval<float&>())` even though
`int &&X(declval<float&>())` is well formed. In order to tolerate this bug
the `__eat<T>(...)` needs to be left in-place. Otherwise it could be replaced
entirely with the new static_cast implementation.

Thanks to Walter Brown for providing the test cases.

llvm-svn: 285786

We need to zero extend the byte in order to correctly shift it into a
64-bit value.

llvm-svn: 285785

Previously, we added strings from DynamicSection::finalize().
It was a bit tricky because finalize() is supposed to fix the final
size of the section, but adding new strings would change the size of
.dynstr section. So there was a dependency between finalize functions
of .dynamic and .dynstr.

However, I noticed that we can elimiante the dependency by simply
add strings early; we don't have to do that in finalize() but can do
from DynamicSection's ctor.

This patch defines a new function, DynamicSection::addEntries, to
add .dynamic entries that doesn't depend on other sections.

llvm-svn: 285784

It was noticed this caused performance regressions and deadlocks. PR30768.

Reorder the code to make it clearer what is tested.

PPC now disables the use of std::call_once only with libstdc++ with
the reordering of the code, as was the original intent.

llvm-svn: 285782

We really shouldn't be sending events for SB API's, dunno when we started
doing that.  We don't do it for other things.  But first restore the status quo.

llvm-svn: 285781

This is the reverse_iterator analogue of getIterator().

llvm-svn: 285780

redeclaration in C++1z mode. We need the exception specification in order for
the function's type to be complete.

llvm-svn: 285779

Release notes checks order and consistent Clang-tidy readability-redundant-declaration description.

llvm-svn: 285778

Should unbreak ocaml binding tests.

Also added an llvm-dis test that checks for the same thing.

llvm-svn: 285777

llvm-svn: 285776

llvm-svn: 285775

These interfaces are no longer used.

Differential Revision: https://reviews.llvm.org/D26222

llvm-svn: 285774

As proposed on llvm-dev:
http://lists.llvm.org/pipermail/llvm-dev/2016-October/106595.html

This change also fixes an API oddity where BitstreamCursor::Read() would
return zero for the first read past the end of the bitstream, but would
report_fatal_error for subsequent reads. Now we always report_fatal_error
for all reads past the end. Updated clients to check for the end of the
bitstream before reading from it.

I also needed to add padding to the invalid bitcode tests in
test/Bitcode/. This is because the streaming interface was not checking that
the file size is a multiple of 4.

Differential Revision: https://reviews.llvm.org/D26219

llvm-svn: 285773

llvm-svn: 285772

This is enough to compile and link but doesn't yet do anything particularly 
useful. Once an ASM parser and printer are added in the next two patches, the 
whole thing can be usefully tested.

Differential Revision: https://reviews.llvm.org/D23562

llvm-svn: 285770

For now, only add instruction definitions for basic ALU operations. Our 
initial target is a working MC layer rather than codegen, so appropriate 
SelectionDAG patterns will come later.

Differential Revision: https://reviews.llvm.org/D23561

llvm-svn: 285769

llvm-svn: 285768

llvm-svn: 285767

This is now doable because this code doesn't have to be in the
dynamic scope of Writer::run().

llvm-svn: 285766

This is already done with VGPR immediates and saves 4 bytes.

llvm-svn: 285765

llvm-svn: 285764

llvm-svn: 285763

This is the conservatively correct way because it's easy to
move or replace a scalar immediate. This was incorrect in the case
when the register class wasn't known from the static instruction
definition, but still needed to be an SGPR. The main example of this
is inlineasm has an SGPR constraint.

Also start verifying the register classes of inlineasm operands.

llvm-svn: 285762

llvm-svn: 285761

You can now write make<T>(Args) instead of new (alloc<T>()) T(Args).

llvm-svn: 285760

Change "use of 'self'..." to "Use of 'self'...". The convention is to
start diagnostics with a capital letter.

rdar://problem/28322494

llvm-svn: 285759

Move the initialization of PreferredLoopExit into runOnMachineFunction to be near the other function specific initializations.

llvm-svn: 285758

Summary:
Currently PreferredLoopExit is set only in buildLoopChains, which is
never called if there are no MachineLoops.

MSan is currently broken by this:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/145/steps/check-llvm%20msan/logs/stdio

This is a naive fix to get things green again. iteratee: you may have a better fix.

This change will also mean PreferredLoopExit will not carry over if
buildCFGChains() is called a second time in runOnMachineFunction, this
appears to be the right thing.

Reviewers: bkramer, iteratee, echristo

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D26069

llvm-svn: 285757

These are only used in the spill to VMEM path. Move them to
the one use.

llvm-svn: 285756

llvm-svn: 285755

We no longer use the section names in this test.

llvm-svn: 285754

This saves a field that is not always used. It also avoids failing a
program that doesn't need the section names.

llvm-svn: 285753

Summary:
Unnamed bitfields cannot be initialized.
Bitfields cannot be in-class initialized.

Reviewers: alexfh, hokein, aaron.ballman

Subscribers: Prazek, nemanjai, cfe-commits

Differential Revision: https://reviews.llvm.org/D26119

llvm-svn: 285752

<N> where "foo (<N>)" is the offset of string "foo" in the string table.

llvm-svn: 285751

Patch by bryant.

Differential Revision: https://reviews.llvm.org/D26126

llvm-svn: 285750

llvm-svn: 285749

This patch replaces GAlloc<ELFT>::<sometype>.Allocate() with
alloc<sometype<ELFT>>().

Patch by Rui!

llvm-svn: 285748