llvm-svn: 162683

llvm-svn: 162682

This allows us to better reason about status objects, like Clang's own
llvm::Optional (when its contents are trivially destructible), which are
often intended to be passed around by value.

We still don't inline constructors for temporaries in the general case.

<rdar://problem/11986434>

llvm-svn: 162681

llvm-svn: 162680

llvm-svn: 162679

llvm-svn: 162678

llvm-svn: 162677

llvm-svn: 162676

llvm-svn: 162675

llvm-svn: 162674

[Sanitizer] Use low-level allocator in flag parsing to avoid calling malloc() before ASan/TSan initialization is done

llvm-svn: 162673

llvm-svn: 162672

[Sanitizer] Rename ProcessMaps to MemoryMappingLayout and fix Windows build by providing stub implementation

llvm-svn: 162671

llvm-svn: 162670

The patch reviewed by Akira Hatanaka.

llvm-svn: 162669

The patch reviewed by Akira Hatanaka.

llvm-svn: 162668

llvm-svn: 162667

llvm-svn: 162666

llvm-svn: 162665

llvm/test/CodeGen/X86/fma_patterns.ll: Add -mtriple=x86_64. It was incompatible on i686 and Windows x64.

llvm-svn: 162664

llvm-svn: 162663

The dependency printing was accidentally removed in during a previous
restructuring.

llvm-svn: 162662

llvm-svn: 162661

llvm-svn: 162660

llvm-svn: 162659

Don't allow vextractf128 to be folded with unaligned stores. We don't fold unaligned loads so shouldn't fold unaligned stores as it can cause an alignment fault to occur.

llvm-svn: 162658

This section (introduced in DWARF-3) is used to define instruction address
ranges for functions that are not contiguous and can't be described
by low_pc/high_pc attributes (this is the usual case for inlined subroutines).
The patch is the first step to support fetching complete inlining info from DWARF.

Reviewed by Benjamin Kramer.

llvm-svn: 162657

Fold some patterns into instruction definitons so tablegen can infer flags removing the need for an explicit 'neverHasSideEffects = 1'

llvm-svn: 162656

FMA3 tests on bdver2 target for changes made in rev 162012. Also made
corresponding changes to existing tests for darwin triple to ensure that
same pattern is tested for bdver2 target.

llvm-svn: 162655

Add HasAVX1Only predicate and use it for patterns that have an AVX1 instruction and an AVX2 instruction rather than relying on AddedComplexity.

llvm-svn: 162654

llvm-svn: 162653

--as-needed.
Patch by Felix Geyer. Fixes pr13262.

llvm-svn: 162652

llvm-svn: 162651

llvm-svn: 162650

llvm-svn: 162649

provided char type other than char or wchar_t. It throw exception during
construction, so there is no chance to imbue own ctype.

This fixes http://llvm.org/bugs/show_bug.cgi?id=13698

llvm-svn: 162648

llvm-svn: 162647

It does not consider user-defined conversions that convert an rvalue
into an lvalue and works incorrectly for types with such a conversion
operator.
For example, 

    struct foo
    {
        operator int&();
    };

 returns false_type. 
Attached a patch that fixes this problem. 
http://llvm.org/bugs/show_bug.cgi?id=13601

llvm-svn: 162644

array new with a non-trivial constructor. Pointed out in PR13380.

llvm-svn: 162643

llvm-svn: 162642