Summary:
Regardless how the loop body weight is distributed, we should preserve
total loop body weight. i.e. we should have same weight reaching the body of the loop
or its duplicates in peeled and unpeeled case.

Reviewers: mkuper, davidxl, anemet

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D28179

llvm-svn: 290833

Attempts to fix Windows build breakage caused by r290818.

llvm-svn: 290832

In the previous fix I used a PMF type as a semi-safe bool type in C++03.
However immediately after committing I realized clang offered explicit
conversion operators as an extension. This patch removes the old fix and
enables _LIBCPP_EXPLICIT using __has_extension instead.

This change also affects the following other classes, which have
'_LIBCPP_EXPLICIT operator bool()'.

* shared_ptr
* unique_ptr
* error_condition
* basic_ios
* function (already C++11 only)
* istream::sentry
* experimental::string_view.

In all of the above cases I believe it is safe to enable the extension, except
in the experimental::string_view case. There seem to be some Clang bugs
affecting the experimental::string_view conversion to std::basic_string. To
work around that I manually disabled _LIBCPP_EXPLICIT in that case.

llvm-svn: 290831

llvm-svn: 290830

llvm-svn: 290829

llvm-svn: 290828

llvm-svn: 290827

The checks were improved with:
https://reviews.llvm.org/rL290194

llvm-svn: 290826

It placates some bots which complain because they compile the
assertion out and think the variable is unused.

llvm-svn: 290825

As pointed out by Howard, this is actually 134774 days (* 24 * 3600),
and therefore seconds, not 100ns units.  Adjust the units to reflect
reality.

llvm-svn: 290824

Visual C++ 14 and newer split msvcrt into msvcrt and ucrt with flavours
of the ucrt for different environments.  This changed the access to the
ctype table by introducing the `__pctype_func` and `__pwctype_func`
accessors.  Use this rather than directly accessing `_ctype` which
allows us to be safer in threaded situations by going through the libc
locking.

llvm-svn: 290823

Apparently my suggestion of using ternary doesn't really work
as clang complains about incompatible types on LHS and RHS. Some
GCC versions happen to accept the code but clang behaviour is
correct here.

llvm-svn: 290822

TestNoreturnUnwind fails on arm/aarch32 linux targets.

Bug is already described for x86_64 android targets in llvm.org/pr31192.

llvm-svn: 290821

llvm-svn: 290820

Split the PDB tests into DWARF test and actual PDB tests, the latter
requiring DIA SDK. Use the new LLVMConfig.cmake LLVM_ENABLE_DIA_SDK
symbol to enable the PDB tests rather than relying on
llvm/Config/config.h private include file that is not available when
building standalone.

Differential Revision: https://reviews.llvm.org/D26249

llvm-svn: 290819

Add an explicit LLVM_ENABLE_DIA_SDK option to control building support
for DIA SDK-based debugging. Control its value to match whether DIA SDK
support was found and expose it in LLVMConfig (alike LLVM_ENABLE_ZLIB).

Its value is needed for LLDB to determine whether to run tests requiring
DIA support. Currently it is obtained from llvm/Config/config.h;
however, this file is not available for standalone builds. Following
this change, LLDB will be modified to use the value from LLVMConfig.

Differential Revision: https://reviews.llvm.org/D26255

llvm-svn: 290818

GNU as rejects input where .cfi_sections is used after .cfi_startproc,
if the new section differs from the old. Adjust our output to always
emit .cfi_sections before the first .cfi_startproc to minimize necessary
code.

Differential Revision: https://reviews.llvm.org/D28011

llvm-svn: 290817

NewGVN: Add UnknownExpression and create them for things we can't symbolize. Kill fragile machinery for handling null expressions.

Summary:
This avoids the very fragile code for null expressions. We could also use a denseset that tracks which things have null expressions instead, but that seems pretty fragile and premature optimization.

This resolves a number of infinite loop cases, test reductions coming.

Reviewers: davide

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D28193

llvm-svn: 290816

Summary: Previously, we tried to fix up the equivalences during symbolic evaluation.  This does not work. Now, we change the equivalences during congruence finding, where it belongs.  We also initialize the equivalence table to give a maximal answer.

Reviewers: davide

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D28192

llvm-svn: 290815

The pass has been on by default for a long time without problems.

llvm-svn: 290814

[OpenMP] Add test cases for the proc_bind and schedule clauses with 'teams distribute parallel for' pragma.

https://reviews.llvm.org/D28205

llvm-svn: 290813

(This changed was approved in https://reviews.llvm.org/D28118, but Simon asked to submit it separately).

llvm-svn: 290812

Revert "DR1391: Check for implicit conversion sequences for non-dependent function template parameters between deduction and substitution. The idea is to accept as many cases as possible, on the basis that substitution failure outside the immediate context is much more common during substitution than during implicit conversion sequence formation."

This reverts commit r290808, as it broken all ARM and AArch64 test-suite
test: MultiSource/UnitTests/C++11/frame_layout

Also, please, next time, try to write a commit message in according to
our guidelines:

http://llvm.org/docs/DeveloperPolicy.html#commit-messages

llvm-svn: 290811

X86 target does not provide any target specific cost calculation for interleave patterns.It uses the common target-independent calculation, which gives very high numbers. As a result, the scalar version is chosen in many cases. The situation on AVX-512 is even worse, since we have 3-src shuffles that significantly reduce the cost.

In this patch I calculate the cost on AVX-512. It will allow to compare interleave pattern with gather/scatter and choose a better solution (PR31426).

* Shiffle-broadcast cost will be changed in Simon's upcoming patch.

Differential Revision: https://reviews.llvm.org/D28118

llvm-svn: 290810

This reapplies rL289013 (reverted in rL289014) with the fixes identified
in D21731. Should hopefully pass the buildbots this time.

llvm-svn: 290809

template parameters between deduction and substitution. The idea is to accept
as many cases as possible, on the basis that substitution failure outside
the immediate context is much more common during substitution than during
implicit conversion sequence formation.

This does not implement the partial ordering portion of DR1391, which so
far appears to be misguided.

llvm-svn: 290808

llvm-svn: 290807

Drawing some inspiration from code from Bill O'Neal as pointed out by
Howard, rework the code to avoid an overflow in the duration.  Adjust
the style to match libc++ style as well.

Create a local typedef for the FILETIME duration (100-ns units).  Use
this to define the difference between the NT and the UNIX epochs (which
previously overflowed due to the representation limits due to the
bouncing to ns).  Return the FILETIME duration biased by the NT-to-UNIX
epoch conversion.

Use of the custom duration makes it easier to read and reason about the
code.

llvm-svn: 290806

Correct style to match libc++ style as pointed out by David Majnemer on
IRC.  NFC.

llvm-svn: 290805

Provide a definition for a steady monotonic clock by wrapping
QueryPerformanceCounter.

llvm-svn: 290804

system_clock::now is not entirely straight forward on Windows, which
does not have a clock_gettime function.

GetSystemTimeAsFileTime gives us the value relative to the NT epoch (Jan
1 1601) rather than the Unix epoch (Jan 1 1970).  However, this function
has a low resolution (~10ms).  Newer versions of Windows provide
GetSystemTimePreciseAsFileTime which gives us a much more accurate time
(<1us).  Unfortunately, the latter is only available on Windows 8+ when
targeting desktop apps.

llvm-svn: 290803

This allows us to build with cl (or rather clang-cl) by using the
correct spelling for `-include` (`/FI` for cl).  clang-cl and cl default
to C++11/C++14 as they support it rather than permitting an explicit
language standard.

llvm-svn: 290802

Hard code the defaults for Windows for the time being.  The checks
really are always going to return the same value.  Technically, the
pthread linkage is possible, however, it seems better to use the Win32
threading along with the external threading support that we have added.

llvm-svn: 290801

This is necessary to support cross-compiling a Windows libc++ from
Linux.  The CMAKE_SYSTEM_HOST_NAME tells you what, in autotools
parlance, is known as the build as opposed to WIN32 which maps to, in
autotools parlance, host.

llvm-svn: 290800

The MS ABI RTTI has a reserved field which is used as a cache for the
demangled name.  It must be zero-initialized, which is used as a hint by
the runtime to say that the cache has not been populated.  Since this
field is populated at runtime, the RTTI structures must be placed in the
.data section rather than .rdata.  NFC

llvm-svn: 290799

The test was added at r290797, and a patch to enable the transform is proposed in D28204.

llvm-svn: 290798

This will change with the current draft of:
https://reviews.llvm.org/D28204

llvm-svn: 290797

Summary:
`PromotedFloats` needs to be checked in 
`DAGTypeLegalizer::PerformExpensiveChecks`. This patch fixes a few type
legalization failures with expansive checks for ARM fp16 tests.

Reviewers: baldrick, bogner, arsenm

Subscribers: arsenm, aemerson, llvm-commits

Differential Revision: https://reviews.llvm.org/D28187

llvm-svn: 290796

llvm-svn: 290795

I'm not sure if this was intentional, but today
isGuaranteedToTransferExecutionToSuccessor returns true for readonly and
argmemonly calls that may throw.  This commit changes the function to
not implicitly infer nounwind this way.

Even if we eventually specify readonly calls as not throwing,
isGuaranteedToTransferExecutionToSuccessor is not the best place to
infer that.  We should instead teach FunctionAttrs or some other such
pass to tag readonly functions / calls as nounwind instead.

llvm-svn: 290794