Fix class name in comment.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171096 91177308-0d34-0410-b5e6-96231b3b80d8

Merge SSE/AVX PCMPEQ/PCMPGT instruction definitions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171095 91177308-0d34-0410-b5e6-96231b3b80d8

Doc: add fmuladd to the list of vectorizeable functions. Thanks hfinkel.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171094 91177308-0d34-0410-b5e6-96231b3b80d8

Remove 'v' from mnemonic to fix asm matching failures.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171093 91177308-0d34-0410-b5e6-96231b3b80d8

Use an additional multiclass to merge the 128/256-bit SSE/AVX instruction definitions for a bunch of SSE2 integer arithmetic instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171092 91177308-0d34-0410-b5e6-96231b3b80d8

Reformat the docs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171091 91177308-0d34-0410-b5e6-96231b3b80d8

white space

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171090 91177308-0d34-0410-b5e6-96231b3b80d8

Use an additional multiclass to merge the 128/256-bit SSE/AVX instruction definitions for PAND/POR/PXOR/PANDN

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171087 91177308-0d34-0410-b5e6-96231b3b80d8

Merge an AVX/SSE 256-bit and 128-bit multiclass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171086 91177308-0d34-0410-b5e6-96231b3b80d8

Mark VANDNPD/VANDNPDS as not commutable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171085 91177308-0d34-0410-b5e6-96231b3b80d8

llvm/test/CodeGen/X86: FileCheck-ize two tests in r171083.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171084 91177308-0d34-0410-b5e6-96231b3b80d8

llvm/test/CodeGen/X86: Disable avx in two tests corresponding to r171082.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171083 91177308-0d34-0410-b5e6-96231b3b80d8

Remove alignment from a bunch more VEX encoded operations in the folding tables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171082 91177308-0d34-0410-b5e6-96231b3b80d8

Remove alignment from folding table for VMOVUPD as an unaligned instruction it shouldn't require alignment...

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171081 91177308-0d34-0410-b5e6-96231b3b80d8

Remove alignment requirements from (V)EXTRACTPS. This instruction does 32-bit stores which aren't required to be aligned on SSE or AVX.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171080 91177308-0d34-0410-b5e6-96231b3b80d8

BBVectorize: Use VTTI to compute costs for intrinsics vectorization

For the time being this includes only some dummy test cases. Once the
generic implementation of the intrinsics cost function does something other
than assuming scalarization in all cases, or some target specializes the
interface, some real test cases can be added.

Also, for consistency, I changed the type of IID from unsigned to Intrinsic::ID
in a few other places.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171079 91177308-0d34-0410-b5e6-96231b3b80d8

Remove alignment requirement from VCVTSS2SD in folding tables. Reverting r171049. This instruction doesn't require alignment.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171078 91177308-0d34-0410-b5e6-96231b3b80d8

LoopVectorize: Enable vectorization of the fmuladd intrinsic

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171076 91177308-0d34-0410-b5e6-96231b3b80d8

BBVectorize: Enable vectorization of the fmuladd intrinsic

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171075 91177308-0d34-0410-b5e6-96231b3b80d8

Loosen scheduling restrictions on the PPC dcbt intrinsic

As with the prefetch intrinsic to which it maps, simply have dcbt
marked as reading from and writing to its arguments instead of having
unmodeled side effects. While this might cause unwanted code motion
(because aliasing checks don't really capture cache-line sharing),
it is more important that prefetches in unrolled loops don't block
the scheduler from rearranging the unrolled loop body.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171073 91177308-0d34-0410-b5e6-96231b3b80d8

Expand PPC64 atomic load and store

Use of store or load with the atomic specifier on 64-bit types would
cause instruction-selection failures. As with the 32-bit case, these
can use the default expansion in terms of cmp-and-swap.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171072 91177308-0d34-0410-b5e6-96231b3b80d8

[msan] Fix handling of vectors of pointers.

VectorType::getInteger() can not be used with them, because pointer size
depends on the target.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171070 91177308-0d34-0410-b5e6-96231b3b80d8

[msan] Fix handling of select with vector condition.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171069 91177308-0d34-0410-b5e6-96231b3b80d8

Harden test so it's not affected by changes to compare lowering.

This only failed on hosts that don't have SSE41.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171066 91177308-0d34-0410-b5e6-96231b3b80d8

X86: Shave off one shuffle from the pcmpeqq sequence for SSE2 by making use of and commutativity.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171064 91177308-0d34-0410-b5e6-96231b3b80d8

X86: Custom lower <2 x i64> eq and ne when SSE41 is not available.

pcmpeqd, pshufd, pshufd, pand is cheaper than unpack + cmpq, sbbq, cmpq, sbbq + pack.
Small speedup on loop-vectorized viterbi (-march=core2).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171063 91177308-0d34-0410-b5e6-96231b3b80d8

ASan: initialize callbacks from ASan module pass in a separate function for consistency

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171061 91177308-0d34-0410-b5e6-96231b3b80d8

ASan: move stack poisoning logic into FunctionStackPoisoner struct

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171060 91177308-0d34-0410-b5e6-96231b3b80d8

Fix whitespace. No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171051 91177308-0d34-0410-b5e6-96231b3b80d8

VCVTSS2SD requires a strict alignment. Thanks Elena.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171049 91177308-0d34-0410-b5e6-96231b3b80d8

Rename LLVMContext diagnostic handler types and functions.

These are now generally used for all diagnostics from the backend, not just
for inline assembly, so this drops the "InlineAsm" from the names.  No
functional change.  (I've left aliases for the old names but only for long
enough to let me switch over clang to use the new ones.)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171047 91177308-0d34-0410-b5e6-96231b3b80d8

[CMake] AddLLVM.cmake: Tweak the corner case that "check-all" doesn't have any tests.

"check-all" can be executed with 0 status, "check-all does nothing, no tools built."

LLVM_EXTERNAL_CLANG_BUILD=OFF LLVM_BUILD_TOOLS=OFF can reproduce this.
Oscar Fuentes reported this. Thank you.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171046 91177308-0d34-0410-b5e6-96231b3b80d8

Quiet gcc's -Wparenthesis warning. No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171044 91177308-0d34-0410-b5e6-96231b3b80d8

Fix typo "Makre" -> "Make".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171043 91177308-0d34-0410-b5e6-96231b3b80d8

Use a std::string rather than a dynamically allocated char* buffer.

This affords us to use std::string's allocation routines and use the destructor
for the memory management. Switching to that also means that we can use
operator==(const std::string&, const char *) to perform the string comparison
rather than resorting to libc functionality (i.e. strcmp).

Patch by Saleem Abdulrasool!

Differential Revision: http://llvm-reviews.chandlerc.com/D230

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171042 91177308-0d34-0410-b5e6-96231b3b80d8

Add LLVMContext::emitWarning methods and use them. <rdar://problem/12867368>

When the backend is used from clang, it should produce proper diagnostics
instead of just printing messages to errs(). Other clients may also want to
register their own error handlers with the LLVMContext, and the same handler
should work for warnings in the same way as the existing emitError methods.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171041 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a typo introduced in r168577: FlAGS -> FLAGS (note the lowercase ell)

Now we really pass -Wcovered-switch-default if the compiler supports it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171040 91177308-0d34-0410-b5e6-96231b3b80d8

AutoRegen.sh: update reference to documentation

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171037 91177308-0d34-0410-b5e6-96231b3b80d8

llvm/test/CodeGen/X86/fold-vex.ll: Add explicit triple.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171029 91177308-0d34-0410-b5e6-96231b3b80d8

CostModel: We have API for checking the costs of known shuffles. This patch adds
support for the insert-subvector and extract-subvector kinds.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171027 91177308-0d34-0410-b5e6-96231b3b80d8

Added 6 more value types: v32i1, v64i1, v32i16, v32i8, v64i8, v8f64

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171026 91177308-0d34-0410-b5e6-96231b3b80d8

Removed "static" from "__jit_debug_descriptor" because "static" adds C++ mangling prefix to this symbol.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171025 91177308-0d34-0410-b5e6-96231b3b80d8

Some x86 instructions can load/store one of the operands to memory. On SSE, this memory needs to be aligned.
When these instructions are encoded in VEX (on AVX) there is no such requirement. This changes the folding
tables and removes the alignment restrictions from VEX-encoded instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171024 91177308-0d34-0410-b5e6-96231b3b80d8

LoopVectorizer: When checking for vectorizable types, also check
the StoreInst operands.

PR14705.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171023 91177308-0d34-0410-b5e6-96231b3b80d8

Change the codegen Cost Model API for shuffeles. This patch removes the API for broadcast and adds a more general API that accepts an enum of known shuffles.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171022 91177308-0d34-0410-b5e6-96231b3b80d8

Fix typo in comments

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171021 91177308-0d34-0410-b5e6-96231b3b80d8

Update the docs of the cost model.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171016 91177308-0d34-0410-b5e6-96231b3b80d8

llvm/MC/MCMachObjectWriter.h: ComputeSymbolTable(): Prune one description in the comment. [-Wdocumentation]

/// \param StringIndexMap [out] - Map from symbol names to offsets in the string table.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171010 91177308-0d34-0410-b5e6-96231b3b80d8

LoopVectorizer: Fix an endless loop in the code that looks for reductions.
The bug was in the code that detects PHIs in if-then-else block sequence.

PR14701.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171008 91177308-0d34-0410-b5e6-96231b3b80d8

Documentation: fix typos reported in PR13866

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171006 91177308-0d34-0410-b5e6-96231b3b80d8

CostModel: Change the default target-independent implementation for finding
the cost of arithmetic functions. We now assume that the cost of arithmetic
operations that are marked as Legal or Promote is low, but ops that are
marked as custom are higher.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171002 91177308-0d34-0410-b5e6-96231b3b80d8

LoopVectorize: Fix accidentaly inverted condition.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171001 91177308-0d34-0410-b5e6-96231b3b80d8

LoopVectorize: For scalars and void types there is no need to compute vector insert/extract costs.

Fixes an assert during the build of oggenc in the test suite.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171000 91177308-0d34-0410-b5e6-96231b3b80d8

We are not ready to estimate the cost of integer expansions based on the number of parts. This test is too noisy.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170999 91177308-0d34-0410-b5e6-96231b3b80d8

docs: Add link to external LLVM backend tutorial.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170998 91177308-0d34-0410-b5e6-96231b3b80d8

whitespace

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170997 91177308-0d34-0410-b5e6-96231b3b80d8

Rename a function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170996 91177308-0d34-0410-b5e6-96231b3b80d8

Loop Vectorizer: Update the cost model of scatter/gather operations and make
them more expensive.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170995 91177308-0d34-0410-b5e6-96231b3b80d8

Remove trailing whitespace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170991 91177308-0d34-0410-b5e6-96231b3b80d8

Remove trailing whitespace

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170990 91177308-0d34-0410-b5e6-96231b3b80d8

Remove a special case that doesn't seem necessary any longer.

Back when this exception was added, it was skipping a lot more code, but
now it just looks like a premature optimization.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170989 91177308-0d34-0410-b5e6-96231b3b80d8

Use getNumOperands() instead of Operands.size().

The representation of the Operands array is going to change soon so it
can be allocated from a BumpPtrAllocator.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170988 91177308-0d34-0410-b5e6-96231b3b80d8

X86: Turn mul of <4 x i32> into pmuludq when no SSE4.1 is available.

pmuludq is slow, but it turns out that all the unpacking and packing of the
scalarized mul is even slower. 10% speedup on loop-vectorized paq8p.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170985 91177308-0d34-0410-b5e6-96231b3b80d8

X86: Emit vector sext as shuffle + sra if vpmovsx is not available.

Also loosen the SSSE3 dependency a bit, expanded pshufb + psra is still better
than scalarized loads. Fixes PR14590.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170984 91177308-0d34-0410-b5e6-96231b3b80d8

Add a comma to fix the build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170982 91177308-0d34-0410-b5e6-96231b3b80d8

Use a negative value to represent INVALID_SIMPLE_VALUE_TYPE instead of 256. Its much cheaper for the isSimple() checks to look for values less than 0 rather than a value greater than 255. This shaves ~8k off the size of the llc binary on x86-64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170981 91177308-0d34-0410-b5e6-96231b3b80d8

Add vAny and Metadata to the switch in getSizeInBits for consistency since every other enum was listed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170977 91177308-0d34-0410-b5e6-96231b3b80d8

[utils] Tweak utils/clang-parse-diagnostics-file to ignore autoconf diagnostics.
 - Also, don't print headers if we aren't going to print any diagnostics.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170973 91177308-0d34-0410-b5e6-96231b3b80d8

Change 'AttrVal' to 'AttrKind' to better reflect that it's a kind of attribute instead of the value of the attribute.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170972 91177308-0d34-0410-b5e6-96231b3b80d8

Fix some undefined behavior when parsing YAML input: don't try to compare an
uninitialized value against a default value. Found by -fsanitize=enum.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170970 91177308-0d34-0410-b5e6-96231b3b80d8

Don't call back() on an empty SmallVector. Found by -fsanitize=enum!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170968 91177308-0d34-0410-b5e6-96231b3b80d8

In some cases, due to scheduling constraints we copy the EFLAGS.
The only way to read the eflags is using push and pop. If we don't
adjust the stack then we run over the first frame index. This is
not something that we want to do, so we have to make sure that
our machine function does not copy the flags. If it does then
we have to emit the prolog that adjusts the stack.

rdar://12896831

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170961 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor subword-swap, EXT/INS, load-effective-address and read-hardware
instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170956 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor SYNC and multiply/divide instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170955 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor BAL instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170954 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Fix encoding of BAL instruction. Also, fix assembler test case which
was not catching the error.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170953 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor jump, jump register, jump-and-link and nop instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170952 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor load/store left/right and load-link and store-conditional
instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170950 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor load/store instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170948 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Remove unnecessary isPseudo parameter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170947 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor LUI instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170944 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor count leading zero or one instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170942 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor sign-extension-in-register instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170940 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor instructions which copy from and to HI/LO registers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170939 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Refactor logical NOR instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170937 91177308-0d34-0410-b5e6-96231b3b80d8

[mips] Move instruction definitions in MipsInstrInfo.td.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170936 91177308-0d34-0410-b5e6-96231b3b80d8

R600: Coding style - remove empty spaces from the beginning of functions

No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170923 91177308-0d34-0410-b5e6-96231b3b80d8

R600: Fix MAX_UINT definition

Patch by: Vadim Girlin

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170922 91177308-0d34-0410-b5e6-96231b3b80d8

R600: Add SHADOWCUBE to TEX_SHADOW pattern

Patch by: Vadim Girlin

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170921 91177308-0d34-0410-b5e6-96231b3b80d8

Cleanup compiler warnings on discarding type qualifiers in casts. Switch to C++ style casts.

Patch by Saleem Abdulrasool!

Differential Revision: http://llvm-reviews.chandlerc.com/D204

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170917 91177308-0d34-0410-b5e6-96231b3b80d8

Add a missing assertion, the null register has no register units.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170916 91177308-0d34-0410-b5e6-96231b3b80d8

try to unbreak ppc buildbots.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170913 91177308-0d34-0410-b5e6-96231b3b80d8

Teach sort_includes.py to drop duplicated includes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170911 91177308-0d34-0410-b5e6-96231b3b80d8

X86: Match pmin/pmax as a target specific dag combine. This occurs during vectorization.

Part of PR14667.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170908 91177308-0d34-0410-b5e6-96231b3b80d8

Remove duplicate includes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170902 91177308-0d34-0410-b5e6-96231b3b80d8

R600: Expand vec4 INT <-> FP conversions

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170901 91177308-0d34-0410-b5e6-96231b3b80d8

X86: Match the SSE/AVX min/max vector ops using a custom node instead of intrinsics

This is very mechanical, no functionality change. Preparation for PR14667.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170898 91177308-0d34-0410-b5e6-96231b3b80d8

Test that a landingpad gets the name provided when it was created (see commit
170318).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170886 91177308-0d34-0410-b5e6-96231b3b80d8

[msan] Remove unreachable blocks before instrumenting a function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170883 91177308-0d34-0410-b5e6-96231b3b80d8

Add a missing "virtual" keyword.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170842 91177308-0d34-0410-b5e6-96231b3b80d8