Vendor import of llvm trunk r304149:

https://llvm.org/svn/llvm-project/llvm/trunk@304149

CMakeLists.txt

@@ -44,6 +44,13 @@ if (NOT PACKAGE_VERSION)
     "${LLVM_VERSION_MAJOR}.${LLVM_VERSION_MINOR}.${LLVM_VERSION_PATCH}${LLVM_VERSION_SUFFIX}")
 endif()
 
+if ((CMAKE_GENERATOR MATCHES "Visual Studio") AND (CMAKE_GENERATOR_TOOLSET STREQUAL ""))
+  message(WARNING "Visual Studio generators use the x86 host compiler by "
+                  "default, even for 64-bit targets. This can result in linker "
+                  "instability and out of memory errors. To use the 64-bit "
+                  "host compiler, pass -Thost=x64 on the CMake command line.")
+endif()
+
 project(LLVM
   ${cmake_3_0_PROJ_VERSION}
   ${cmake_3_0_LANGUAGES}

docs/Benchmarking.rst

@@ -0,0 +1,87 @@
+==================================
+Benchmarking tips
+==================================
+
+
+Introduction
+============
+
+For benchmarking a patch we want to reduce all possible sources of
+noise as much as possible. How to do that is very OS dependent.
+
+Note that low noise is required, but not sufficient. It does not
+exclude measurement bias. See
+https://www.cis.upenn.edu/~cis501/papers/producing-wrong-data.pdf for
+example.
+
+General
+================================
+
+* Use a high resolution timer, e.g. perf under linux.
+
+* Run the benchmark multiple times to be able to recognize noise.
+
+* Disable as many processes or services as possible on the target system.
+
+* Disable frequency scaling, turbo boost and address space
+  randomization (see OS specific section).
+
+* Static link if the OS supports it. That avoids any variation that
+  might be introduced by loading dynamic libraries. This can be done
+  by passing ``-DLLVM_BUILD_STATIC=ON`` to cmake.
+
+* Try to avoid storage. On some systems you can use tmpfs. Putting the
+  program, inputs and outputs on tmpfs avoids touching a real storage
+  system, which can have a pretty big variability.
+
+  To mount it (on linux and freebsd at least)::
+
+    mount -t tmpfs -o size=<XX>g none dir_to_mount
+
+Linux
+=====
+
+* Disable address space randomization::
+
+    echo 0 > /proc/sys/kernel/randomize_va_space
+
+* Set scaling_governor to performance::
+
+   for i in /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
+   do
+     echo performance > /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
+   done
+
+* Use https://github.com/lpechacek/cpuset to reserve cpus for just the
+  program you are benchmarking. If using perf, leave at least 2 cores
+  so that perf runs in one and your program in another::
+
+    cset shield -c N1,N2 -k on
+
+  This will move all threads out of N1 and N2. The ``-k on`` means
+  that even kernel threads are moved out.
+
+* Disable the SMT pair of the cpus you will use for the benchmark. The
+  pair of cpu N can be found in
+  ``/sys/devices/system/cpu/cpuN/topology/thread_siblings_list`` and
+  disabled with::
+
+    echo 0 > /sys/devices/system/cpu/cpuX/online
+
+
+* Run the program with::
+
+    cset shield --exec -- perf stat -r 10 <cmd>
+
+  This will run the command after ``--`` in the isolated cpus. The
+  particular perf command runs the ``<cmd>`` 10 times and reports
+  statistics.
+
+With these in place you can expect perf variations of less than 0.1%.
+
+Linux Intel
+-----------
+
+* Disable turbo mode::
+
+    echo 1 > /sys/devices/system/cpu/intel_pstate/no_turbo

docs/GettingStartedVS.rst

@@ -100,6 +100,10 @@ Here's the short story for getting up and running quickly with LLVM:
    * CMake generates project files for all build types. To select a specific
      build type, use the Configuration manager from the VS IDE or the 
      ``/property:Configuration`` command line option when using MSBuild.
+   * By default, the Visual Studio project files generated by CMake use the
+     32-bit toolset. If you are developing on a 64-bit version of Windows and
+     want to use the 64-bit toolset, pass the ``-Thost=x64`` flag when
+     generating the Visual Studio solution. This requires CMake 3.8.0 or later.
 
 6. Start Visual Studio
 

docs/LangRef.rst

@@ -4415,12 +4415,6 @@ The current supported vocabulary is limited:
   address space identifier.
 - ``DW_OP_stack_value`` marks a constant value.
 
-DIExpression nodes that contain a ``DW_OP_stack_value`` operator are standalone
-location descriptions that describe constant values. This form is used to
-describe global constants that have been optimized away. All other expressions
-are modifiers to another location: A debug intrinsic ties a location and a
-DIExpression together.
-
 DWARF specifies three kinds of simple location descriptions: Register, memory,
 and implicit location descriptions. Register and memory location descriptions
 describe the *location* of a source variable (in the sense that a debugger might
@@ -12722,7 +12716,7 @@ Syntax:
       declare <type> 
       @llvm.experimental.constrained.fadd(<type> <op1>, <type> <op2>,
                                           metadata <rounding mode>,
-                                          metadata  <exception behavior>)
+                                          metadata <exception behavior>)
 
 Overview:
 """""""""
@@ -12759,7 +12753,7 @@ Syntax:
       declare <type> 
       @llvm.experimental.constrained.fsub(<type> <op1>, <type> <op2>,
                                           metadata <rounding mode>,
-                                          metadata  <exception behavior>)
+                                          metadata <exception behavior>)
 
 Overview:
 """""""""
@@ -12796,7 +12790,7 @@ Syntax:
       declare <type> 
       @llvm.experimental.constrained.fmul(<type> <op1>, <type> <op2>,
                                           metadata <rounding mode>,
-                                          metadata  <exception behavior>)
+                                          metadata <exception behavior>)
 
 Overview:
 """""""""
@@ -12833,7 +12827,7 @@ Syntax:
       declare <type> 
       @llvm.experimental.constrained.fdiv(<type> <op1>, <type> <op2>,
                                           metadata <rounding mode>,
-                                          metadata  <exception behavior>)
+                                          metadata <exception behavior>)
 
 Overview:
 """""""""
@@ -12870,7 +12864,7 @@ Syntax:
       declare <type> 
       @llvm.experimental.constrained.frem(<type> <op1>, <type> <op2>,
                                           metadata <rounding mode>,
-                                          metadata  <exception behavior>)
+                                          metadata <exception behavior>)
 
 Overview:
 """""""""
@@ -12899,6 +12893,461 @@ value operands and has the same type as the operands.  The remainder has the
 same sign as the dividend. 
 
 
+Constrained libm-equivalent Intrinsics
+--------------------------------------
+
+In addition to the basic floating point operations for which constrained
+intrinsics are described above, there are constrained versions of various
+operations which provide equivalent behavior to a corresponding libm function.
+These intrinsics allow the precise behavior of these operations with respect to
+rounding mode and exception behavior to be controlled.
+
+As with the basic constrained floating point intrinsics, the rounding mode
+and exception behavior arguments only control the behavior of the optimizer.
+They do not change the runtime floating point environment.
+
+
+'``llvm.experimental.constrained.sqrt``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.sqrt(<type> <op1>,
+                                          metadata <rounding mode>,
+                                          metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.sqrt``' intrinsic returns the square root
+of the specified value, returning the same value as the libm '``sqrt``'
+functions would, but without setting ``errno``.
+
+Arguments:
+""""""""""
+
+The first argument and the return type are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the nonnegative square root of the specified value.
+If the value is less than negative zero, a floating point exception occurs
+and the the return value is architecture specific.
+
+
+'``llvm.experimental.constrained.pow``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.pow(<type> <op1>, <type> <op2>,
+                                         metadata <rounding mode>,
+                                         metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.pow``' intrinsic returns the first operand
+raised to the (positive or negative) power specified by the second operand.
+
+Arguments:
+""""""""""
+
+The first two arguments and the return value are floating point numbers of the
+same type.  The second argument specifies the power to which the first argument
+should be raised.
+
+The third and fourth arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the first value raised to the second power,
+returning the same values as the libm ``pow`` functions would, and
+handles error conditions in the same way.
+
+
+'``llvm.experimental.constrained.powi``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.powi(<type> <op1>, i32 <op2>,
+                                          metadata <rounding mode>,
+                                          metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.powi``' intrinsic returns the first operand
+raised to the (positive or negative) power specified by the second operand. The
+order of evaluation of multiplications is not defined. When a vector of floating
+point type is used, the second argument remains a scalar integer value.
+
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.  The second argument is a 32-bit signed integer specifying the power to
+which the first argument should be raised.
+
+The third and fourth arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the first value raised to the second power with an
+unspecified sequence of rounding operations.
+
+
+'``llvm.experimental.constrained.sin``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.sin(<type> <op1>,
+                                         metadata <rounding mode>,
+                                         metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.sin``' intrinsic returns the sine of the
+first operand.
+
+Arguments:
+""""""""""
+
+The first argument and the return type are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the sine of the specified operand, returning the
+same values as the libm ``sin`` functions would, and handles error
+conditions in the same way.
+
+
+'``llvm.experimental.constrained.cos``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.cos(<type> <op1>,
+                                         metadata <rounding mode>,
+                                         metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.cos``' intrinsic returns the cosine of the
+first operand.
+
+Arguments:
+""""""""""
+
+The first argument and the return type are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the cosine of the specified operand, returning the
+same values as the libm ``cos`` functions would, and handles error
+conditions in the same way.
+
+
+'``llvm.experimental.constrained.exp``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.exp(<type> <op1>,
+                                         metadata <rounding mode>,
+                                         metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.exp``' intrinsic computes the base-e
+exponential of the specified value.
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``exp`` functions
+would, and handles error conditions in the same way.
+
+
+'``llvm.experimental.constrained.exp2``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.exp2(<type> <op1>,
+                                          metadata <rounding mode>,
+                                          metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.exp2``' intrinsic computes the base-2
+exponential of the specified value.
+
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``exp2`` functions
+would, and handles error conditions in the same way.
+
+
+'``llvm.experimental.constrained.log``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.log(<type> <op1>,
+                                         metadata <rounding mode>,
+                                         metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.log``' intrinsic computes the base-e
+logarithm of the specified value.
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``log`` functions
+would, and handles error conditions in the same way.
+
+
+'``llvm.experimental.constrained.log10``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.log10(<type> <op1>,
+                                           metadata <rounding mode>,
+                                           metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.log10``' intrinsic computes the base-10
+logarithm of the specified value.
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``log10`` functions
+would, and handles error conditions in the same way.
+
+
+'``llvm.experimental.constrained.log2``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.log2(<type> <op1>,
+                                          metadata <rounding mode>,
+                                          metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.log2``' intrinsic computes the base-2
+logarithm of the specified value.
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``log2`` functions
+would, and handles error conditions in the same way.
+
+
+'``llvm.experimental.constrained.rint``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.rint(<type> <op1>,
+                                          metadata <rounding mode>,
+                                          metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.rint``' intrinsic returns the first
+operand rounded to the nearest integer. It may raise an inexact floating point
+exception if the operand is not an integer.
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``rint`` functions
+would, and handles error conditions in the same way.  The rounding mode is
+described, not determined, by the rounding mode argument.  The actual rounding
+mode is determined by the runtime floating point environment.  The rounding
+mode argument is only intended as information to the compiler.
+
+
+'``llvm.experimental.constrained.nearbyint``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <type> 
+      @llvm.experimental.constrained.nearbyint(<type> <op1>,
+                                               metadata <rounding mode>,
+                                               metadata <exception behavior>)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.constrained.nearbyint``' intrinsic returns the first
+operand rounded to the nearest integer. It will not raise an inexact floating
+point exception if the operand is not an integer.
+
+
+Arguments:
+""""""""""
+
+The first argument and the return value are floating point numbers of the same
+type.
+
+The second and third arguments specify the rounding mode and exception
+behavior as described above.
+
+Semantics:
+""""""""""
+
+This function returns the same values as the libm ``nearbyint`` functions
+would, and handles error conditions in the same way.  The rounding mode is
+described, not determined, by the rounding mode argument.  The actual rounding
+mode is determined by the runtime floating point environment.  The rounding
+mode argument is only intended as information to the compiler.
+
+
 General Intrinsics
 ------------------
 

docs/Vectorizers.rst

@@ -99,7 +99,9 @@ Optimization remarks are enabled using:
 indicates if vectorization was specified.
 
 ``-Rpass-analysis=loop-vectorize`` identifies the statements that caused
-vectorization to fail.
+vectorization to fail. If in addition ``-fsave-optimization-record`` is
+provided, multiple causes of vectorization failure may be listed (this behavior
+might change in the future).
 
 Consider the following loop:
 

docs/index.rst

@@ -90,6 +90,7 @@ representation.
    CodeOfConduct
    CompileCudaWithLLVM
    ReportingGuide
+   Benchmarking
 
 :doc:`GettingStarted`
    Discusses how to get up and running quickly with the LLVM infrastructure.

examples/Kaleidoscope/BuildingAJIT/Chapter1/KaleidoscopeJIT.h

@@ -1,4 +1,4 @@
-//===----- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope ----*- C++ -*-===//
+//===- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -17,7 +17,7 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/ExecutionEngine/JITSymbol.h"
-#include "llvm/ExecutionEngine/RuntimeDyld.h"
+#include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
 #include "llvm/ExecutionEngine/SectionMemoryManager.h"
 #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
 #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h"
@@ -44,7 +44,7 @@ private:
   IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
 
 public:
-  typedef decltype(CompileLayer)::ModuleSetHandleT ModuleHandle;
+  using ModuleHandle = decltype(CompileLayer)::ModuleSetHandleT;
 
   KaleidoscopeJIT()
       : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),

examples/Kaleidoscope/BuildingAJIT/Chapter2/KaleidoscopeJIT.h

@@ -1,4 +1,4 @@
-//===----- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope ----*- C++ -*-===//
+//===- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -17,7 +17,7 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/ExecutionEngine/JITSymbol.h"
-#include "llvm/ExecutionEngine/RuntimeDyld.h"
+#include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
 #include "llvm/ExecutionEngine/SectionMemoryManager.h"
 #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
 #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h"
@@ -47,13 +47,13 @@ private:
   RTDyldObjectLinkingLayer<> ObjectLayer;
   IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
 
-  typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
+  using OptimizeFunction =
-    OptimizeFunction;
+      std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>;
 
   IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
 
 public:
-  typedef decltype(OptimizeLayer)::ModuleSetHandleT ModuleHandle;
+  using ModuleHandle = decltype(OptimizeLayer)::ModuleSetHandleT;
 
   KaleidoscopeJIT()
       : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),

examples/Kaleidoscope/BuildingAJIT/Chapter3/KaleidoscopeJIT.h

@@ -1,4 +1,4 @@
-//===----- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope ----*- C++ -*-===//
+//===- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -17,6 +17,7 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/ExecutionEngine/JITSymbol.h"
+#include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
 #include "llvm/ExecutionEngine/RuntimeDyld.h"
 #include "llvm/ExecutionEngine/SectionMemoryManager.h"
 #include "llvm/ExecutionEngine/Orc/CompileOnDemandLayer.h"
@@ -49,8 +50,8 @@ private:
   RTDyldObjectLinkingLayer<> ObjectLayer;
   IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
 
-  typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
+  using OptimizeFunction =
-    OptimizeFunction;
+      std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>;
 
   IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
 
@@ -58,7 +59,7 @@ private:
   CompileOnDemandLayer<decltype(OptimizeLayer)> CODLayer;
 
 public:
-  typedef decltype(CODLayer)::ModuleSetHandleT ModuleHandle;
+  using ModuleHandle = decltype(CODLayer)::ModuleSetHandleT;
 
   KaleidoscopeJIT()
       : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),

examples/Kaleidoscope/BuildingAJIT/Chapter4/KaleidoscopeJIT.h

@@ -1,4 +1,4 @@
-//===----- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope ----*- C++ -*-===//
+//===- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -17,10 +17,10 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/ExecutionEngine/JITSymbol.h"
-#include "llvm/ExecutionEngine/RuntimeDyld.h"
+#include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
 #include "llvm/ExecutionEngine/SectionMemoryManager.h"
-#include "llvm/ExecutionEngine/Orc/CompileOnDemandLayer.h"
 #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
+#include "llvm/ExecutionEngine/Orc/IndirectionUtils.h"
 #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h"
 #include "llvm/ExecutionEngine/Orc/IRTransformLayer.h"
 #include "llvm/ExecutionEngine/Orc/LambdaResolver.h"
@@ -76,8 +76,8 @@ private:
   RTDyldObjectLinkingLayer<> ObjectLayer;
   IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
 
-  typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
+  using OptimizeFunction =
-    OptimizeFunction;
+      std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>;
 
   IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
 
@@ -85,7 +85,7 @@ private:
   std::unique_ptr<IndirectStubsManager> IndirectStubsMgr;
 
 public:
-  typedef decltype(OptimizeLayer)::ModuleSetHandleT ModuleHandle;
+  using ModuleHandle = decltype(OptimizeLayer)::ModuleSetHandleT;
 
   KaleidoscopeJIT()
       : TM(EngineBuilder().selectTarget()),
@@ -106,7 +106,6 @@ public:
   TargetMachine &getTargetMachine() { return *TM; }
 
   ModuleHandle addModule(std::unique_ptr<Module> M) {
-
     // Build our symbol resolver:
     // Lambda 1: Look back into the JIT itself to find symbols that are part of
     //           the same "logical dylib".

examples/Kaleidoscope/BuildingAJIT/Chapter5/KaleidoscopeJIT.h

@@ -1,4 +1,4 @@
-//===----- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope ----*- C++ -*-===//
+//===- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -20,9 +20,8 @@
 #include "llvm/ADT/Triple.h"
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/ExecutionEngine/JITSymbol.h"
-#include "llvm/ExecutionEngine/RuntimeDyld.h"
-#include "llvm/ExecutionEngine/Orc/CompileOnDemandLayer.h"
 #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
+#include "llvm/ExecutionEngine/Orc/IndirectionUtils.h"
 #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h"
 #include "llvm/ExecutionEngine/Orc/IRTransformLayer.h"
 #include "llvm/ExecutionEngine/Orc/LambdaResolver.h"
@@ -73,7 +72,7 @@ namespace llvm {
 namespace orc {
 
 // Typedef the remote-client API.
-typedef remote::OrcRemoteTargetClient<FDRPCChannel> MyRemote;
+using MyRemote = remote::OrcRemoteTargetClient<FDRPCChannel>;
 
 class KaleidoscopeJIT {
 private:
@@ -82,8 +81,8 @@ private:
   RTDyldObjectLinkingLayer<> ObjectLayer;
   IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
 
-  typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
+  using OptimizeFunction =
-    OptimizeFunction;
+      std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>;
 
   IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
 
@@ -92,7 +91,7 @@ private:
   MyRemote &Remote;
 
 public:
-  typedef decltype(OptimizeLayer)::ModuleSetHandleT ModuleHandle;
+  using ModuleHandle = decltype(OptimizeLayer)::ModuleSetHandleT;
 
   KaleidoscopeJIT(MyRemote &Remote)
       : TM(EngineBuilder().selectTarget(Triple(Remote.getTargetTriple()), "",
@@ -124,7 +123,6 @@ public:
   TargetMachine &getTargetMachine() { return *TM; }
 
   ModuleHandle addModule(std::unique_ptr<Module> M) {
-
     // Build our symbol resolver:
     // Lambda 1: Look back into the JIT itself to find symbols that are part of
     //           the same "logical dylib".

examples/Kaleidoscope/BuildingAJIT/Chapter5/Server/server.cpp

@@ -1,17 +1,19 @@
-#include "llvm/Support/CommandLine.h"
+#include "../RemoteJITUtils.h"
-#include "llvm/Support/DynamicLibrary.h"
+#include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
-#include "llvm/Support/TargetSelect.h"
 #include "llvm/ExecutionEngine/Orc/OrcRemoteTargetServer.h"
 #include "llvm/ExecutionEngine/Orc/OrcABISupport.h"
-
+#include "llvm/Support/CommandLine.h"
-#include "../RemoteJITUtils.h"
+#include "llvm/Support/DynamicLibrary.h"
-
+#include "llvm/Support/Error.h"
+#include "llvm/Support/raw_ostream.h"
+#include "llvm/Support/TargetSelect.h"
+#include <cstdint>
+#include <cstdio>
 #include <cstring>
-#include <unistd.h>
+#include <string>
 #include <netinet/in.h>
 #include <sys/socket.h>
 
-
 using namespace llvm;
 using namespace llvm::orc;
 
@@ -22,7 +24,7 @@ cl::opt<uint32_t> Port("port",
 
 ExitOnError ExitOnErr;
 
-typedef int (*MainFun)(int, const char*[]);
+using MainFun = int (*)(int, const char*[]);
 
 template <typename NativePtrT>
 NativePtrT MakeNative(uint64_t P) {
@@ -36,7 +38,6 @@ void printExprResult(double Val) {
 
 // --- LAZY COMPILE TEST ---
 int main(int argc, char* argv[]) {
-
   if (argc == 0)
     ExitOnErr.setBanner("jit_server: ");
   else
@@ -59,14 +60,14 @@ int main(int argc, char* argv[]) {
   int sockfd = socket(PF_INET, SOCK_STREAM, 0);
   sockaddr_in servAddr, clientAddr;
   socklen_t clientAddrLen = sizeof(clientAddr);
-  bzero(&servAddr, sizeof(servAddr));
+  memset(&servAddr, 0, sizeof(servAddr));
   servAddr.sin_family = PF_INET;
   servAddr.sin_family = INADDR_ANY;
   servAddr.sin_port = htons(Port);
 
   {
     // avoid "Address already in use" error.
-    int yes=1;
+    int yes = 1;
     if (setsockopt(sockfd,SOL_SOCKET,SO_REUSEADDR,&yes,sizeof(int)) == -1) {
       errs() << "Error calling setsockopt.\n";
       return 1;
@@ -98,7 +99,8 @@ int main(int argc, char* argv[]) {
     };
 
   FDRPCChannel TCPChannel(newsockfd, newsockfd);
-  typedef remote::OrcRemoteTargetServer<FDRPCChannel, OrcX86_64_SysV> MyServerT;
+
+  using MyServerT = remote::OrcRemoteTargetServer<FDRPCChannel, OrcX86_64_SysV>;
 
   MyServerT Server(TCPChannel, SymbolLookup, RegisterEHFrames, DeregisterEHFrames);
 

examples/Kaleidoscope/include/KaleidoscopeJIT.h

@@ -1,4 +1,4 @@
-//===----- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope ----*- C++ -*-===//
+//===- KaleidoscopeJIT.h - A simple JIT for Kaleidoscope --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -19,7 +19,6 @@
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/ExecutionEngine/JITSymbol.h"
 #include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
-#include "llvm/ExecutionEngine/RuntimeDyld.h"
 #include "llvm/ExecutionEngine/SectionMemoryManager.h"
 #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
 #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h"
@@ -40,9 +39,9 @@ namespace orc {
 
 class KaleidoscopeJIT {
 public:
-  typedef RTDyldObjectLinkingLayer<> ObjLayerT;
+  using ObjLayerT = RTDyldObjectLinkingLayer<>;
-  typedef IRCompileLayer<ObjLayerT> CompileLayerT;
+  using CompileLayerT = IRCompileLayer<ObjLayerT>;
-  typedef CompileLayerT::ModuleSetHandleT ModuleHandleT;
+  using ModuleHandleT = CompileLayerT::ModuleSetHandleT;
 
   KaleidoscopeJIT()
       : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),

include/llvm/ADT/Triple.h

@@ -59,6 +59,7 @@ public:
     mips64,         // MIPS64: mips64
     mips64el,       // MIPS64EL: mips64el
     msp430,         // MSP430: msp430
+    nios2,          // NIOSII: nios2
     ppc,            // PPC: powerpc
     ppc64,          // PPC64: powerpc64, ppu
     ppc64le,        // PPC64LE: powerpc64le

include/llvm/Analysis/InstructionSimplify.h

@@ -70,174 +70,173 @@ struct SimplifyQuery {
     Copy.CxtI = I;
     return Copy;
   }
-  };
+};
 
-  // NOTE: the explicit multiple argument versions of these functions are
+// NOTE: the explicit multiple argument versions of these functions are
-  // deprecated.
+// deprecated.
-  // Please use the SimplifyQuery versions in new code.
+// Please use the SimplifyQuery versions in new code.
 
-  /// Given operands for an Add, fold the result or return null.
+/// Given operands for an Add, fold the result or return null.
-  Value *SimplifyAddInst(Value *LHS, Value *RHS, bool isNSW, bool isNUW,
+Value *SimplifyAddInst(Value *LHS, Value *RHS, bool isNSW, bool isNUW,
                        const SimplifyQuery &Q);
 
-  /// Given operands for a Sub, fold the result or return null.
+/// Given operands for a Sub, fold the result or return null.
-  Value *SimplifySubInst(Value *LHS, Value *RHS, bool isNSW, bool isNUW,
+Value *SimplifySubInst(Value *LHS, Value *RHS, bool isNSW, bool isNUW,
-                         const SimplifyQuery &Q);
+                       const SimplifyQuery &Q);
 
-  /// Given operands for an FAdd, fold the result or return null.
+/// Given operands for an FAdd, fold the result or return null.
-  Value *SimplifyFAddInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+Value *SimplifyFAddInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+                        const SimplifyQuery &Q);
+
+/// Given operands for an FSub, fold the result or return null.
+Value *SimplifyFSubInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+                        const SimplifyQuery &Q);
+
+/// Given operands for an FMul, fold the result or return null.
+Value *SimplifyFMulInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+                        const SimplifyQuery &Q);
+
+/// Given operands for a Mul, fold the result or return null.
+Value *SimplifyMulInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for an SDiv, fold the result or return null.
+Value *SimplifySDivInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for a UDiv, fold the result or return null.
+Value *SimplifyUDivInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for an FDiv, fold the result or return null.
+Value *SimplifyFDivInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+                        const SimplifyQuery &Q);
+
+/// Given operands for an SRem, fold the result or return null.
+Value *SimplifySRemInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for a URem, fold the result or return null.
+Value *SimplifyURemInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for an FRem, fold the result or return null.
+Value *SimplifyFRemInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+                        const SimplifyQuery &Q);
+
+/// Given operands for a Shl, fold the result or return null.
+Value *SimplifyShlInst(Value *Op0, Value *Op1, bool isNSW, bool isNUW,
+                       const SimplifyQuery &Q);
+
+/// Given operands for a LShr, fold the result or return null.
+Value *SimplifyLShrInst(Value *Op0, Value *Op1, bool isExact,
+                        const SimplifyQuery &Q);
+
+/// Given operands for a AShr, fold the result or return nulll.
+Value *SimplifyAShrInst(Value *Op0, Value *Op1, bool isExact,
+                        const SimplifyQuery &Q);
+
+/// Given operands for an And, fold the result or return null.
+Value *SimplifyAndInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for an Or, fold the result or return null.
+Value *SimplifyOrInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for an Xor, fold the result or return null.
+Value *SimplifyXorInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+
+/// Given operands for an ICmpInst, fold the result or return null.
+Value *SimplifyICmpInst(unsigned Predicate, Value *LHS, Value *RHS,
+                        const SimplifyQuery &Q);
+
+/// Given operands for an FCmpInst, fold the result or return null.
+Value *SimplifyFCmpInst(unsigned Predicate, Value *LHS, Value *RHS,
+                        FastMathFlags FMF, const SimplifyQuery &Q);
+
+/// Given operands for a SelectInst, fold the result or return null.
+Value *SimplifySelectInst(Value *Cond, Value *TrueVal, Value *FalseVal,
                           const SimplifyQuery &Q);
 
-  /// Given operands for an FSub, fold the result or return null.
+/// Given operands for a GetElementPtrInst, fold the result or return null.
-  Value *SimplifyFSubInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+Value *SimplifyGEPInst(Type *SrcTy, ArrayRef<Value *> Ops,
-                          const SimplifyQuery &Q);
+                       const SimplifyQuery &Q);
 
-  /// Given operands for an FMul, fold the result or return null.
+/// Given operands for an InsertValueInst, fold the result or return null.
-  Value *SimplifyFMulInst(Value *LHS, Value *RHS, FastMathFlags FMF,
+Value *SimplifyInsertValueInst(Value *Agg, Value *Val, ArrayRef<unsigned> Idxs,
-                          const SimplifyQuery &Q);
+                               const SimplifyQuery &Q);
 
-  /// Given operands for a Mul, fold the result or return null.
+/// Given operands for an ExtractValueInst, fold the result or return null.
-  Value *SimplifyMulInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+Value *SimplifyExtractValueInst(Value *Agg, ArrayRef<unsigned> Idxs,
+                                const SimplifyQuery &Q);
 
-  /// Given operands for an SDiv, fold the result or return null.
+/// Given operands for an ExtractElementInst, fold the result or return null.
-  Value *SimplifySDivInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
+Value *SimplifyExtractElementInst(Value *Vec, Value *Idx,
-
-  /// Given operands for a UDiv, fold the result or return null.
-  Value *SimplifyUDivInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
-
-  /// Given operands for an FDiv, fold the result or return null.
-  Value *SimplifyFDivInst(Value *LHS, Value *RHS, FastMathFlags FMF,
-                          const SimplifyQuery &Q);
-
-  /// Given operands for an SRem, fold the result or return null.
-  Value *SimplifySRemInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
-
-  /// Given operands for a URem, fold the result or return null.
-  Value *SimplifyURemInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
-
-  /// Given operands for an FRem, fold the result or return null.
-  Value *SimplifyFRemInst(Value *LHS, Value *RHS, FastMathFlags FMF,
-                          const SimplifyQuery &Q);
-
-  /// Given operands for a Shl, fold the result or return null.
-  Value *SimplifyShlInst(Value *Op0, Value *Op1, bool isNSW, bool isNUW,
-                         const SimplifyQuery &Q);
-
-  /// Given operands for a LShr, fold the result or return null.
-  Value *SimplifyLShrInst(Value *Op0, Value *Op1, bool isExact,
-                          const SimplifyQuery &Q);
-
-  /// Given operands for a AShr, fold the result or return nulll.
-  Value *SimplifyAShrInst(Value *Op0, Value *Op1, bool isExact,
-                          const SimplifyQuery &Q);
-
-  /// Given operands for an And, fold the result or return null.
-  Value *SimplifyAndInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
-
-  /// Given operands for an Or, fold the result or return null.
-  Value *SimplifyOrInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
-
-  /// Given operands for an Xor, fold the result or return null.
-  Value *SimplifyXorInst(Value *LHS, Value *RHS, const SimplifyQuery &Q);
-
-  /// Given operands for an ICmpInst, fold the result or return null.
-  Value *SimplifyICmpInst(unsigned Predicate, Value *LHS, Value *RHS,
-                          const SimplifyQuery &Q);
-
-  /// Given operands for an FCmpInst, fold the result or return null.
-  Value *SimplifyFCmpInst(unsigned Predicate, Value *LHS, Value *RHS,
-                          FastMathFlags FMF, const SimplifyQuery &Q);
-
-  /// Given operands for a SelectInst, fold the result or return null.
-  Value *SimplifySelectInst(Value *Cond, Value *TrueVal, Value *FalseVal,
-                            const SimplifyQuery &Q);
-
-  /// Given operands for a GetElementPtrInst, fold the result or return null. 
-  Value *SimplifyGEPInst(Type *SrcTy, ArrayRef<Value *> Ops,
-                         const SimplifyQuery &Q);
-
-  /// Given operands for an InsertValueInst, fold the result or return null.
-  Value *SimplifyInsertValueInst(Value *Agg, Value *Val,
-                                 ArrayRef<unsigned> Idxs,
-                                 const SimplifyQuery &Q);
-
-  /// Given operands for an ExtractValueInst, fold the result or return null.
-  Value *SimplifyExtractValueInst(Value *Agg, ArrayRef<unsigned> Idxs,
                                   const SimplifyQuery &Q);
 
-  /// Given operands for an ExtractElementInst, fold the result or return null.
+/// Given operands for a CastInst, fold the result or return null.
-  Value *SimplifyExtractElementInst(Value *Vec, Value *Idx,
+Value *SimplifyCastInst(unsigned CastOpc, Value *Op, Type *Ty,
-                                    const SimplifyQuery &Q);
+                        const SimplifyQuery &Q);
 
-  /// Given operands for a CastInst, fold the result or return null.
+/// Given operands for a ShuffleVectorInst, fold the result or return null.
-  Value *SimplifyCastInst(unsigned CastOpc, Value *Op, Type *Ty,
+Value *SimplifyShuffleVectorInst(Value *Op0, Value *Op1, Constant *Mask,
-                          const SimplifyQuery &Q);
+                                 Type *RetTy, const SimplifyQuery &Q);
 
-  /// Given operands for a ShuffleVectorInst, fold the result or return null.
+//=== Helper functions for higher up the class hierarchy.
-  Value *SimplifyShuffleVectorInst(Value *Op0, Value *Op1, Constant *Mask,
-                                   Type *RetTy, const SimplifyQuery &Q);
 
-  //=== Helper functions for higher up the class hierarchy.
+/// Given operands for a CmpInst, fold the result or return null.
-
+Value *SimplifyCmpInst(unsigned Predicate, Value *LHS, Value *RHS,
-
-  /// Given operands for a CmpInst, fold the result or return null.
-  Value *SimplifyCmpInst(unsigned Predicate, Value *LHS, Value *RHS,
-                         const SimplifyQuery &Q);
-
-  /// Given operands for a BinaryOperator, fold the result or return null.
-  Value *SimplifyBinOp(unsigned Opcode, Value *LHS, Value *RHS,
                        const SimplifyQuery &Q);
 
-  /// Given operands for an FP BinaryOperator, fold the result or return null.
+/// Given operands for a BinaryOperator, fold the result or return null.
-  /// In contrast to SimplifyBinOp, try to use FastMathFlag when folding the
+Value *SimplifyBinOp(unsigned Opcode, Value *LHS, Value *RHS,
-  /// result. In case we don't need FastMathFlags, simply fall to SimplifyBinOp.
+                     const SimplifyQuery &Q);
-  Value *SimplifyFPBinOp(unsigned Opcode, Value *LHS, Value *RHS,
-                         FastMathFlags FMF, const SimplifyQuery &Q);
 
-  /// Given a function and iterators over arguments, fold the result or return
+/// Given operands for an FP BinaryOperator, fold the result or return null.
-  /// null.
+/// In contrast to SimplifyBinOp, try to use FastMathFlag when folding the
-  Value *SimplifyCall(Value *V, User::op_iterator ArgBegin,
+/// result. In case we don't need FastMathFlags, simply fall to SimplifyBinOp.
-                      User::op_iterator ArgEnd, const SimplifyQuery &Q);
+Value *SimplifyFPBinOp(unsigned Opcode, Value *LHS, Value *RHS,
+                       FastMathFlags FMF, const SimplifyQuery &Q);
 
-  /// Given a function and set of arguments, fold the result or return null.
+/// Given a function and iterators over arguments, fold the result or return
-  Value *SimplifyCall(Value *V, ArrayRef<Value *> Args, const SimplifyQuery &Q);
+/// null.
+Value *SimplifyCall(Value *V, User::op_iterator ArgBegin,
+                    User::op_iterator ArgEnd, const SimplifyQuery &Q);
 
-  /// See if we can compute a simplified version of this instruction. If not,
+/// Given a function and set of arguments, fold the result or return null.
-  /// return null.
+Value *SimplifyCall(Value *V, ArrayRef<Value *> Args, const SimplifyQuery &Q);
-  Value *SimplifyInstruction(Instruction *I, const SimplifyQuery &Q,
-                             OptimizationRemarkEmitter *ORE = nullptr);
 
-  /// Replace all uses of 'I' with 'SimpleV' and simplify the uses recursively.
+/// See if we can compute a simplified version of this instruction. If not,
-  ///
+/// return null.
-  /// This first performs a normal RAUW of I with SimpleV. It then recursively
+Value *SimplifyInstruction(Instruction *I, const SimplifyQuery &Q,
-  /// attempts to simplify those users updated by the operation. The 'I'
+                           OptimizationRemarkEmitter *ORE = nullptr);
-  /// instruction must not be equal to the simplified value 'SimpleV'.
-  ///
-  /// The function returns true if any simplifications were performed.
-  bool replaceAndRecursivelySimplify(Instruction *I, Value *SimpleV,
-                                     const TargetLibraryInfo *TLI = nullptr,
-                                     const DominatorTree *DT = nullptr,
-                                     AssumptionCache *AC = nullptr);
 
-  /// Recursively attempt to simplify an instruction.
+/// Replace all uses of 'I' with 'SimpleV' and simplify the uses recursively.
-  ///
+///
-  /// This routine uses SimplifyInstruction to simplify 'I', and if successful
+/// This first performs a normal RAUW of I with SimpleV. It then recursively
-  /// replaces uses of 'I' with the simplified value. It then recurses on each
+/// attempts to simplify those users updated by the operation. The 'I'
-  /// of the users impacted. It returns true if any simplifications were
+/// instruction must not be equal to the simplified value 'SimpleV'.
-  /// performed.
+///
-  bool recursivelySimplifyInstruction(Instruction *I,
+/// The function returns true if any simplifications were performed.
-                                      const TargetLibraryInfo *TLI = nullptr,
+bool replaceAndRecursivelySimplify(Instruction *I, Value *SimpleV,
-                                      const DominatorTree *DT = nullptr,
+                                   const TargetLibraryInfo *TLI = nullptr,
-                                      AssumptionCache *AC = nullptr);
+                                   const DominatorTree *DT = nullptr,
-  // These helper functions return a SimplifyQuery structure that contains as
+                                   AssumptionCache *AC = nullptr);
-  // many of the optional analysis we use as are currently valid.  This is the
+
-  // strongly preferred way of constructing SimplifyQuery in passes.
+/// Recursively attempt to simplify an instruction.
-  const SimplifyQuery getBestSimplifyQuery(Pass &, Function &);
+///
-  template <class T, class... TArgs>
+/// This routine uses SimplifyInstruction to simplify 'I', and if successful
-  const SimplifyQuery getBestSimplifyQuery(AnalysisManager<T, TArgs...> &,
+/// replaces uses of 'I' with the simplified value. It then recurses on each
-                                           Function &);
+/// of the users impacted. It returns true if any simplifications were
-  const SimplifyQuery getBestSimplifyQuery(LoopStandardAnalysisResults &,
+/// performed.
-                                           const DataLayout &);
+bool recursivelySimplifyInstruction(Instruction *I,
+                                    const TargetLibraryInfo *TLI = nullptr,
+                                    const DominatorTree *DT = nullptr,
+                                    AssumptionCache *AC = nullptr);
+
+// These helper functions return a SimplifyQuery structure that contains as
+// many of the optional analysis we use as are currently valid.  This is the
+// strongly preferred way of constructing SimplifyQuery in passes.
+const SimplifyQuery getBestSimplifyQuery(Pass &, Function &);
+template <class T, class... TArgs>
+const SimplifyQuery getBestSimplifyQuery(AnalysisManager<T, TArgs...> &,
+                                         Function &);
+const SimplifyQuery getBestSimplifyQuery(LoopStandardAnalysisResults &,
+                                         const DataLayout &);
 } // end namespace llvm
 
 #endif

include/llvm/Analysis/LoopPass.h

@@ -126,9 +126,8 @@ public:
   }
 
 public:
-  // Add a new loop into the loop queue as a child of the given parent, or at
+  // Add a new loop into the loop queue.
-  // the top level if \c ParentLoop is null.
+  void addLoop(Loop &L);
-  Loop &addLoop(Loop *ParentLoop);
 
   //===--------------------------------------------------------------------===//
   /// SimpleAnalysis - Provides simple interface to update analysis info

include/llvm/Analysis/ScalarEvolution.h

@@ -1533,6 +1533,12 @@ public:
   /// specified loop.
   bool isLoopInvariant(const SCEV *S, const Loop *L);
 
+  /// Determine if the SCEV can be evaluated at loop's entry. It is true if it
+  /// doesn't depend on a SCEVUnknown of an instruction which is dominated by
+  /// the header of loop L.
+  bool isAvailableAtLoopEntry(const SCEV *S, const Loop *L, DominatorTree &DT,
+                              LoopInfo &LI);
+
   /// Return true if the given SCEV changes value in a known way in the
   /// specified loop.  This property being true implies that the value is
   /// variant in the loop AND that we can emit an expression to compute the

include/llvm/Analysis/TargetTransformInfo.h

@@ -396,6 +396,9 @@ public:
   bool isLegalMaskedScatter(Type *DataType) const;
   bool isLegalMaskedGather(Type *DataType) const;
 
+  /// Return true if target doesn't mind addresses in vectors.
+  bool prefersVectorizedAddressing() const;
+
   /// \brief Return the cost of the scaling factor used in the addressing
   /// mode represented by AM for this target, for a load/store
   /// of the specified type.
@@ -807,6 +810,7 @@ public:
   virtual bool isLegalMaskedLoad(Type *DataType) = 0;
   virtual bool isLegalMaskedScatter(Type *DataType) = 0;
   virtual bool isLegalMaskedGather(Type *DataType) = 0;
+  virtual bool prefersVectorizedAddressing() = 0;
   virtual int getScalingFactorCost(Type *Ty, GlobalValue *BaseGV,
                                    int64_t BaseOffset, bool HasBaseReg,
                                    int64_t Scale, unsigned AddrSpace) = 0;
@@ -1000,6 +1004,9 @@ public:
   bool isLegalMaskedGather(Type *DataType) override {
     return Impl.isLegalMaskedGather(DataType);
   }
+  bool prefersVectorizedAddressing() override {
+    return Impl.prefersVectorizedAddressing();
+  }
   int getScalingFactorCost(Type *Ty, GlobalValue *BaseGV, int64_t BaseOffset,
                            bool HasBaseReg, int64_t Scale,
                            unsigned AddrSpace) override {

include/llvm/Analysis/TargetTransformInfoImpl.h

@@ -237,6 +237,8 @@ public:
 
   bool isLegalMaskedGather(Type *DataType) { return false; }
 
+  bool prefersVectorizedAddressing() { return true; }
+
   int getScalingFactorCost(Type *Ty, GlobalValue *BaseGV, int64_t BaseOffset,
                            bool HasBaseReg, int64_t Scale, unsigned AddrSpace) {
     // Guess that all legal addressing mode are free.

include/llvm/Analysis/ValueTracking.h

@@ -60,7 +60,8 @@ template <typename T> class ArrayRef;
   KnownBits computeKnownBits(const Value *V, const DataLayout &DL,
                              unsigned Depth = 0, AssumptionCache *AC = nullptr,
                              const Instruction *CxtI = nullptr,
-                             const DominatorTree *DT = nullptr);
+                             const DominatorTree *DT = nullptr,
+                             OptimizationRemarkEmitter *ORE = nullptr);
   /// Compute known bits from the range metadata.
   /// \p KnownZero the set of bits that are known to be zero
   /// \p KnownOne the set of bits that are known to be one

include/llvm/CodeGen/AsmPrinter.h

@@ -34,6 +34,7 @@
 namespace llvm {
 
 class AsmPrinterHandler;
+class BasicBlock;
 class BlockAddress;
 class Constant;
 class ConstantArray;
@@ -43,6 +44,7 @@ class DIEAbbrev;
 class DwarfDebug;
 class GCMetadataPrinter;
 class GlobalIndirectSymbol;
+class GlobalObject;
 class GlobalValue;
 class GlobalVariable;
 class GCStrategy;
@@ -65,6 +67,8 @@ class MCSubtargetInfo;
 class MCSymbol;
 class MCTargetOptions;
 class MDNode;
+class Module;
+class raw_ostream;
 class TargetLoweringObjectFile;
 class TargetMachine;
 
@@ -109,7 +113,7 @@ public:
 
   /// Map global GOT equivalent MCSymbols to GlobalVariables and keep track of
   /// its number of uses by other globals.
-  typedef std::pair<const GlobalVariable *, unsigned> GOTEquivUsePair;
+  using GOTEquivUsePair = std::pair<const GlobalVariable *, unsigned>;
   MapVector<const MCSymbol *, GOTEquivUsePair> GlobalGOTEquivs;
 
   /// Enable print [latency:throughput] in output

include/llvm/CodeGen/AtomicExpandUtils.h

@@ -1,4 +1,4 @@
-//===-- AtomicExpandUtils.h - Utilities for expanding atomic instructions -===//
+//===- AtomicExpandUtils.h - Utilities for expanding atomic instructions --===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -7,19 +7,24 @@
 //
 //===----------------------------------------------------------------------===//
 
+#ifndef LLVM_CODEGEN_ATOMICEXPANDUTILS_H
+#define LLVM_CODEGEN_ATOMICEXPANDUTILS_H
+
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/IR/IRBuilder.h"
+#include "llvm/Support/AtomicOrdering.h"
 
 namespace llvm {
-class Value;
-class AtomicRMWInst;
 
+class AtomicRMWInst;
+class Value;
 
 /// Parameters (see the expansion example below):
 /// (the builder, %addr, %loaded, %new_val, ordering,
 ///  /* OUT */ %success, /* OUT */ %new_loaded)
-typedef function_ref<void(IRBuilder<> &, Value *, Value *, Value *,
+using CreateCmpXchgInstFun =
-                          AtomicOrdering, Value *&, Value *&)> CreateCmpXchgInstFun;
+    function_ref<void(IRBuilder<> &, Value *, Value *, Value *, AtomicOrdering,
+                      Value *&, Value *&)>;
 
 /// \brief Expand an atomic RMW instruction into a loop utilizing
 /// cmpxchg. You'll want to make sure your target machine likes cmpxchg
@@ -42,7 +47,8 @@ typedef function_ref<void(IRBuilder<> &, Value *, Value *, Value *,
 /// loop:
 ///     %loaded = phi iN [ %init_loaded, %entry ], [ %new_loaded, %loop ]
 ///     %new = some_op iN %loaded, %incr
-/// ; This is what -atomic-expand will produce using this function on i686 targets:
+/// ; This is what -atomic-expand will produce using this function on i686
+/// targets:
 ///     %pair = cmpxchg iN* %addr, iN %loaded, iN %new_val
 ///     %new_loaded = extractvalue { iN, i1 } %pair, 0
 ///     %success = extractvalue { iN, i1 } %pair, 1
@@ -52,6 +58,8 @@ typedef function_ref<void(IRBuilder<> &, Value *, Value *, Value *,
 ///     [...]
 ///
 /// Returns true if the containing function was modified.
-bool
+bool expandAtomicRMWToCmpXchg(AtomicRMWInst *AI, CreateCmpXchgInstFun Factory);
-expandAtomicRMWToCmpXchg(AtomicRMWInst *AI, CreateCmpXchgInstFun Factory);
+
-}
+} // end namespace llvm
+
+#endif // LLVM_CODEGEN_ATOMICEXPANDUTILS_H

include/llvm/CodeGen/DIE.h

@@ -1,4 +1,4 @@
-//===--- lib/CodeGen/DIE.h - DWARF Info Entries -----------------*- C++ -*-===//
+//===- lib/CodeGen/DIE.h - DWARF Info Entries -------------------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -31,6 +31,7 @@
 #include <iterator>
 #include <new>
 #include <type_traits>
+#include <utility>
 #include <vector>
 
 namespace llvm {
@@ -53,11 +54,11 @@ class DIEAbbrevData {
   dwarf::Form Form;
 
   /// Dwarf attribute value for DW_FORM_implicit_const
-  int64_t Value;
+  int64_t Value = 0;
 
 public:
   DIEAbbrevData(dwarf::Attribute A, dwarf::Form F)
-      : Attribute(A), Form(F), Value(0) {}
+      : Attribute(A), Form(F) {}
   DIEAbbrevData(dwarf::Attribute A, int64_t V)
       : Attribute(A), Form(dwarf::DW_FORM_implicit_const), Value(V) {}
 
@@ -136,13 +137,14 @@ class DIEAbbrevSet {
   /// storage container.
   BumpPtrAllocator &Alloc;
   /// \brief FoldingSet that uniques the abbreviations.
-  llvm::FoldingSet<DIEAbbrev> AbbreviationsSet;
+  FoldingSet<DIEAbbrev> AbbreviationsSet;
   /// A list of all the unique abbreviations in use.
   std::vector<DIEAbbrev *> Abbreviations;
 
 public:
   DIEAbbrevSet(BumpPtrAllocator &A) : Alloc(A) {}
   ~DIEAbbrevSet();
+
   /// Generate the abbreviation declaration for a DIE and return a pointer to
   /// the generated abbreviation.
   ///
@@ -289,13 +291,11 @@ public:
 /// A pointer to another debug information entry.  An instance of this class can
 /// also be used as a proxy for a debug information entry not yet defined
 /// (ie. types.)
-class DIE;
 class DIEEntry {
   DIE *Entry;
 
-  DIEEntry() = delete;
-
 public:
+  DIEEntry() = delete;
   explicit DIEEntry(DIE &E) : Entry(&E) {}
 
   DIE &getEntry() const { return *Entry; }
@@ -348,10 +348,10 @@ private:
   ///
   /// All values that aren't standard layout (or are larger than 8 bytes)
   /// should be stored by reference instead of by value.
-  typedef AlignedCharArrayUnion<DIEInteger, DIEString, DIEExpr, DIELabel,
+  using ValTy = AlignedCharArrayUnion<DIEInteger, DIEString, DIEExpr, DIELabel,
-                                DIEDelta *, DIEEntry, DIEBlock *, DIELoc *,
+                                      DIEDelta *, DIEEntry, DIEBlock *,
-                                DIELocList>
+                                      DIELoc *, DIELocList>;
-      ValTy;
+
   static_assert(sizeof(ValTy) <= sizeof(uint64_t) ||
                     sizeof(ValTy) <= sizeof(void *),
                 "Expected all large types to be stored via pointer");
@@ -486,10 +486,12 @@ struct IntrusiveBackListNode {
 };
 
 struct IntrusiveBackListBase {
-  typedef IntrusiveBackListNode Node;
+  using Node = IntrusiveBackListNode;
+
   Node *Last = nullptr;
 
   bool empty() const { return !Last; }
+
   void push_back(Node &N) {
     assert(N.Next.getPointer() == &N && "Expected unlinked node");
     assert(N.Next.getInt() == true && "Expected unlinked node");
@@ -505,6 +507,7 @@ struct IntrusiveBackListBase {
 template <class T> class IntrusiveBackList : IntrusiveBackListBase {
 public:
   using IntrusiveBackListBase::empty;
+
   void push_back(T &N) { IntrusiveBackListBase::push_back(N); }
   T &back() { return *static_cast<T *>(Last); }
   const T &back() const { return *static_cast<T *>(Last); }
@@ -513,6 +516,7 @@ public:
   class iterator
       : public iterator_facade_base<iterator, std::forward_iterator_tag, T> {
     friend class const_iterator;
+
     Node *N = nullptr;
 
   public:
@@ -585,10 +589,12 @@ public:
 class DIEValueList {
   struct Node : IntrusiveBackListNode {
     DIEValue V;
+
     explicit Node(DIEValue V) : V(V) {}
   };
 
-  typedef IntrusiveBackList<Node> ListTy;
+  using ListTy = IntrusiveBackList<Node>;
+
   ListTy List;
 
 public:
@@ -597,9 +603,10 @@ public:
       : public iterator_adaptor_base<value_iterator, ListTy::iterator,
                                      std::forward_iterator_tag, DIEValue> {
     friend class const_value_iterator;
-    typedef iterator_adaptor_base<value_iterator, ListTy::iterator,
+
-                                  std::forward_iterator_tag,
+    using iterator_adaptor =
-                                  DIEValue> iterator_adaptor;
+        iterator_adaptor_base<value_iterator, ListTy::iterator,
+                              std::forward_iterator_tag, DIEValue>;
 
   public:
     value_iterator() = default;
@@ -612,9 +619,9 @@ public:
   class const_value_iterator : public iterator_adaptor_base<
                                    const_value_iterator, ListTy::const_iterator,
                                    std::forward_iterator_tag, const DIEValue> {
-    typedef iterator_adaptor_base<const_value_iterator, ListTy::const_iterator,
+    using iterator_adaptor =
-                                  std::forward_iterator_tag,
+        iterator_adaptor_base<const_value_iterator, ListTy::const_iterator,
-                                  const DIEValue> iterator_adaptor;
+                              std::forward_iterator_tag, const DIEValue>;
 
   public:
     const_value_iterator() = default;
@@ -627,8 +634,8 @@ public:
     const DIEValue &operator*() const { return wrapped()->V; }
   };
 
-  typedef iterator_range<value_iterator> value_range;
+  using value_range = iterator_range<value_iterator>;
-  typedef iterator_range<const_value_iterator> const_value_range;
+  using const_value_range = iterator_range<const_value_iterator>;
 
   value_iterator addValue(BumpPtrAllocator &Alloc, const DIEValue &V) {
     List.push_back(*new (Alloc) Node(V));
@@ -657,15 +664,15 @@ class DIE : IntrusiveBackListNode, public DIEValueList {
   friend class DIEUnit;
 
   /// Dwarf unit relative offset.
-  unsigned Offset;
+  unsigned Offset = 0;
   /// Size of instance + children.
-  unsigned Size;
+  unsigned Size = 0;
   unsigned AbbrevNumber = ~0u;
   /// Dwarf tag code.
   dwarf::Tag Tag = (dwarf::Tag)0;
   /// Set to true to force a DIE to emit an abbreviation that says it has
   /// children even when it doesn't. This is used for unit testing purposes.
-  bool ForceChildren;
+  bool ForceChildren = false;
   /// Children DIEs.
   IntrusiveBackList<DIE> Children;
 
@@ -673,20 +680,19 @@ class DIE : IntrusiveBackListNode, public DIEValueList {
   /// DIEUnit which contains this DIE as its unit DIE.
   PointerUnion<DIE *, DIEUnit *> Owner;
 
-  DIE() = delete;
+  explicit DIE(dwarf::Tag Tag) : Tag(Tag) {}
-  explicit DIE(dwarf::Tag Tag) : Offset(0), Size(0), Tag(Tag),
-      ForceChildren(false) {}
 
 public:
+  DIE() = delete;
+  DIE(const DIE &RHS) = delete;
+  DIE(DIE &&RHS) = delete;
+  DIE &operator=(const DIE &RHS) = delete;
+  DIE &operator=(const DIE &&RHS) = delete;
+
   static DIE *get(BumpPtrAllocator &Alloc, dwarf::Tag Tag) {
     return new (Alloc) DIE(Tag);
   }
 
-  DIE(const DIE &RHS) = delete;
-  DIE(DIE &&RHS) = delete;
-  void operator=(const DIE &RHS) = delete;
-  void operator=(const DIE &&RHS) = delete;
-
   // Accessors.
   unsigned getAbbrevNumber() const { return AbbrevNumber; }
   dwarf::Tag getTag() const { return Tag; }
@@ -696,10 +702,10 @@ public:
   bool hasChildren() const { return ForceChildren || !Children.empty(); }
   void setForceChildren(bool B) { ForceChildren = B; }
 
-  typedef IntrusiveBackList<DIE>::iterator child_iterator;
+  using child_iterator = IntrusiveBackList<DIE>::iterator;
-  typedef IntrusiveBackList<DIE>::const_iterator const_child_iterator;
+  using const_child_iterator = IntrusiveBackList<DIE>::const_iterator;
-  typedef iterator_range<child_iterator> child_range;
+  using child_range = iterator_range<child_iterator>;
-  typedef iterator_range<const_child_iterator> const_child_range;
+  using const_child_range = iterator_range<const_child_iterator>;
 
   child_range children() {
     return make_range(Children.begin(), Children.end());
@@ -838,10 +844,10 @@ struct BasicDIEUnit final : DIEUnit {
 /// DIELoc - Represents an expression location.
 //
 class DIELoc : public DIEValueList {
-  mutable unsigned Size; // Size in bytes excluding size header.
+  mutable unsigned Size = 0; // Size in bytes excluding size header.
 
 public:
-  DIELoc() : Size(0) {}
+  DIELoc() = default;
 
   /// ComputeSize - Calculate the size of the location expression.
   ///
@@ -872,10 +878,10 @@ public:
 /// DIEBlock - Represents a block of values.
 //
 class DIEBlock : public DIEValueList {
-  mutable unsigned Size; // Size in bytes excluding size header.
+  mutable unsigned Size = 0; // Size in bytes excluding size header.
 
 public:
-  DIEBlock() : Size(0) {}
+  DIEBlock() = default;
 
   /// ComputeSize - Calculate the size of the location expression.
   ///

include/llvm/CodeGen/FaultMaps.h

@@ -56,7 +56,7 @@ private:
           HandlerOffsetExpr(HandlerOffset) {}
   };
 
-  typedef std::vector<FaultInfo> FunctionFaultInfos;
+  using FunctionFaultInfos = std::vector<FaultInfo>;
 
   // We'd like to keep a stable iteration order for FunctionInfos to help
   // FileCheck based testing.
@@ -78,20 +78,17 @@ private:
 /// generated by the version of LLVM that includes it.  No guarantees are made
 /// with respect to forward or backward compatibility.
 class FaultMapParser {
-  typedef uint8_t FaultMapVersionType;
+  using FaultMapVersionType = uint8_t;
-  static const size_t FaultMapVersionOffset = 0;
+  using Reserved0Type = uint8_t;
+  using Reserved1Type = uint16_t;
+  using NumFunctionsType = uint32_t;
 
-  typedef uint8_t Reserved0Type;
+  static const size_t FaultMapVersionOffset = 0;
   static const size_t Reserved0Offset =
       FaultMapVersionOffset + sizeof(FaultMapVersionType);
-
-  typedef uint16_t Reserved1Type;
   static const size_t Reserved1Offset = Reserved0Offset + sizeof(Reserved0Type);
-
-  typedef uint32_t NumFunctionsType;
   static const size_t NumFunctionsOffset =
       Reserved1Offset + sizeof(Reserved1Type);
-
   static const size_t FunctionInfosOffset =
       NumFunctionsOffset + sizeof(NumFunctionsType);
 
@@ -105,14 +102,13 @@ class FaultMapParser {
 
 public:
   class FunctionFaultInfoAccessor {
-    typedef uint32_t FaultKindType;
+    using FaultKindType = uint32_t;
-    static const size_t FaultKindOffset = 0;
+    using FaultingPCOffsetType = uint32_t;
+    using HandlerPCOffsetType = uint32_t;
 
-    typedef uint32_t FaultingPCOffsetType;
+    static const size_t FaultKindOffset = 0;
     static const size_t FaultingPCOffsetOffset =
         FaultKindOffset + sizeof(FaultKindType);
-
-    typedef uint32_t HandlerPCOffsetType;
     static const size_t HandlerPCOffsetOffset =
         FaultingPCOffsetOffset + sizeof(FaultingPCOffsetType);
 
@@ -140,20 +136,17 @@ public:
   };
 
   class FunctionInfoAccessor {
-    typedef uint64_t FunctionAddrType;
+    using FunctionAddrType = uint64_t;
-    static const size_t FunctionAddrOffset = 0;
+    using NumFaultingPCsType = uint32_t;
+    using ReservedType = uint32_t;
 
-    typedef uint32_t NumFaultingPCsType;
+    static const size_t FunctionAddrOffset = 0;
     static const size_t NumFaultingPCsOffset =
         FunctionAddrOffset + sizeof(FunctionAddrType);
-
-    typedef uint32_t ReservedType;
     static const size_t ReservedOffset =
         NumFaultingPCsOffset + sizeof(NumFaultingPCsType);
-
     static const size_t FunctionFaultInfosOffset =
         ReservedOffset + sizeof(ReservedType);
-
     static const size_t FunctionInfoHeaderSize = FunctionFaultInfosOffset;
 
     const uint8_t *P = nullptr;

include/llvm/CodeGen/GlobalISel/Localizer.h

@@ -0,0 +1,78 @@
+//== llvm/CodeGen/GlobalISel/Localizer.h - Localizer -------------*- C++ -*-==//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+/// \file This file describes the interface of the Localizer pass.
+/// This pass moves/duplicates constant-like instructions close to their uses.
+/// Its primarily goal is to workaround the deficiencies of the fast register
+/// allocator.
+/// With GlobalISel constants are all materialized in the entry block of
+/// a function. However, the fast allocator cannot rematerialize constants and
+/// has a lot more live-ranges to deal with and will most likely end up
+/// spilling a lot.
+/// By pushing the constants close to their use, we only create small
+/// live-ranges.
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CODEGEN_GLOBALISEL_LOCALIZER_H
+#define LLVM_CODEGEN_GLOBALISEL_LOCALIZER_H
+
+#include "llvm/CodeGen/GlobalISel/MachineIRBuilder.h"
+#include "llvm/CodeGen/MachineFunctionPass.h"
+
+namespace llvm {
+// Forward declarations.
+class MachineRegisterInfo;
+
+/// This pass implements the localization mechanism described at the
+/// top of this file. One specificity of the implementation is that
+/// it will materialize one and only one instance of a constant per
+/// basic block, thus enabling reuse of that constant within that block.
+/// Moreover, it only materializes constants in blocks where they
+/// are used. PHI uses are considered happening at the end of the
+/// related predecessor.
+class Localizer : public MachineFunctionPass {
+public:
+  static char ID;
+
+private:
+  /// MRI contains all the register class/bank information that this
+  /// pass uses and updates.
+  MachineRegisterInfo *MRI;
+
+  /// Check whether or not \p MI needs to be moved close to its uses.
+  static bool shouldLocalize(const MachineInstr &MI);
+
+  /// Check if \p MOUse is used in the same basic block as \p Def.
+  /// If the use is in the same block, we say it is local.
+  /// When the use is not local, \p InsertMBB will contain the basic
+  /// block when to insert \p Def to have a local use.
+  static bool isLocalUse(MachineOperand &MOUse, const MachineInstr &Def,
+                         MachineBasicBlock *&InsertMBB);
+
+  /// Initialize the field members using \p MF.
+  void init(MachineFunction &MF);
+
+public:
+  Localizer();
+
+  StringRef getPassName() const override { return "Localizer"; }
+
+  MachineFunctionProperties getRequiredProperties() const override {
+    return MachineFunctionProperties()
+        .set(MachineFunctionProperties::Property::IsSSA)
+        .set(MachineFunctionProperties::Property::Legalized)
+        .set(MachineFunctionProperties::Property::RegBankSelected);
+  }
+
+  bool runOnMachineFunction(MachineFunction &MF) override;
+};
+
+} // End namespace llvm.
+
+#endif

include/llvm/CodeGen/ISDOpcodes.h

@@ -264,6 +264,14 @@ namespace ISD {
     /// optimized.
     STRICT_FADD, STRICT_FSUB, STRICT_FMUL, STRICT_FDIV, STRICT_FREM,
 
+    /// Constrained versions of libm-equivalent floating point intrinsics.
+    /// These will be lowered to the equivalent non-constrained pseudo-op
+    /// (or expanded to the equivalent library call) before final selection.
+    /// They are used to limit optimizations while the DAG is being optimized.
+    STRICT_FSQRT, STRICT_FPOW, STRICT_FPOWI, STRICT_FSIN, STRICT_FCOS,
+    STRICT_FEXP, STRICT_FEXP2, STRICT_FLOG, STRICT_FLOG10, STRICT_FLOG2,
+    STRICT_FRINT, STRICT_FNEARBYINT,
+
     /// FMA - Perform a * b + c with no intermediate rounding step.
     FMA,
 

include/llvm/CodeGen/LiveInterval.h

@@ -1,4 +1,4 @@
-//===-- llvm/CodeGen/LiveInterval.h - Interval representation ---*- C++ -*-===//
+//===- llvm/CodeGen/LiveInterval.h - Interval representation ----*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -21,22 +21,30 @@
 #ifndef LLVM_CODEGEN_LIVEINTERVAL_H
 #define LLVM_CODEGEN_LIVEINTERVAL_H
 
+#include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/IntEqClasses.h"
+#include "llvm/ADT/iterator_range.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/STLExtras.h"
 #include "llvm/CodeGen/SlotIndexes.h"
+#include "llvm/MC/LaneBitmask.h"
 #include "llvm/Support/Allocator.h"
-#include "llvm/Target/TargetRegisterInfo.h"
+#include "llvm/Support/MathExtras.h"
+#include <algorithm>
 #include <cassert>
-#include <climits>
+#include <cstddef>
+#include <functional>
+#include <memory>
 #include <set>
+#include <tuple>
+#include <utility>
 
 namespace llvm {
+
   class CoalescerPair;
   class LiveIntervals;
-  class MachineInstr;
   class MachineRegisterInfo;
-  class TargetRegisterInfo;
   class raw_ostream;
-  template <typename T, unsigned Small> class SmallPtrSet;
 
   /// VNInfo - Value Number Information.
   /// This class holds information about a machine level values, including
@@ -44,7 +52,7 @@ namespace llvm {
   ///
   class VNInfo {
   public:
-    typedef BumpPtrAllocator Allocator;
+    using Allocator = BumpPtrAllocator;
 
     /// The ID number of this value.
     unsigned id;
@@ -53,14 +61,10 @@ namespace llvm {
     SlotIndex def;
 
     /// VNInfo constructor.
-    VNInfo(unsigned i, SlotIndex d)
+    VNInfo(unsigned i, SlotIndex d) : id(i), def(d) {}
-      : id(i), def(d)
-    { }
 
     /// VNInfo constructor, copies values from orig, except for the value number.
-    VNInfo(unsigned i, const VNInfo &orig)
+    VNInfo(unsigned i, const VNInfo &orig) : id(i), def(orig.def) {}
-      : id(i), def(orig.def)
-    { }
 
     /// Copy from the parameter into this VNInfo.
     void copyFrom(VNInfo &src) {
@@ -152,16 +156,16 @@ namespace llvm {
   /// segment with a new value number is used.
   class LiveRange {
   public:
-
     /// This represents a simple continuous liveness interval for a value.
     /// The start point is inclusive, the end point exclusive. These intervals
     /// are rendered as [start,end).
     struct Segment {
       SlotIndex start;  // Start point of the interval (inclusive)
       SlotIndex end;    // End point of the interval (exclusive)
-      VNInfo *valno;    // identifier for the value contained in this segment.
+      VNInfo *valno = nullptr; // identifier for the value contained in this
+                               // segment.
 
-      Segment() : valno(nullptr) {}
+      Segment() = default;
 
       Segment(SlotIndex S, SlotIndex E, VNInfo *V)
         : start(S), end(E), valno(V) {
@@ -189,8 +193,8 @@ namespace llvm {
       void dump() const;
     };
 
-    typedef SmallVector<Segment, 2> Segments;
+    using Segments = SmallVector<Segment, 2>;
-    typedef SmallVector<VNInfo *, 2> VNInfoList;
+    using VNInfoList = SmallVector<VNInfo *, 2>;
 
     Segments segments;   // the liveness segments
     VNInfoList valnos;   // value#'s
@@ -198,22 +202,24 @@ namespace llvm {
     // The segment set is used temporarily to accelerate initial computation
     // of live ranges of physical registers in computeRegUnitRange.
     // After that the set is flushed to the segment vector and deleted.
-    typedef std::set<Segment> SegmentSet;
+    using SegmentSet = std::set<Segment>;
     std::unique_ptr<SegmentSet> segmentSet;
 
-    typedef Segments::iterator iterator;
+    using iterator = Segments::iterator;
+    using const_iterator = Segments::const_iterator;
+
     iterator begin() { return segments.begin(); }
     iterator end()   { return segments.end(); }
 
-    typedef Segments::const_iterator const_iterator;
     const_iterator begin() const { return segments.begin(); }
     const_iterator end() const  { return segments.end(); }
 
-    typedef VNInfoList::iterator vni_iterator;
+    using vni_iterator = VNInfoList::iterator;
+    using const_vni_iterator = VNInfoList::const_iterator;
+
     vni_iterator vni_begin() { return valnos.begin(); }
     vni_iterator vni_end()   { return valnos.end(); }
 
-    typedef VNInfoList::const_iterator const_vni_iterator;
     const_vni_iterator vni_begin() const { return valnos.begin(); }
     const_vni_iterator vni_end() const   { return valnos.end(); }
 
@@ -631,40 +637,37 @@ namespace llvm {
   /// or stack slot.
   class LiveInterval : public LiveRange {
   public:
-    typedef LiveRange super;
+    using super = LiveRange;
 
     /// A live range for subregisters. The LaneMask specifies which parts of the
     /// super register are covered by the interval.
     /// (@sa TargetRegisterInfo::getSubRegIndexLaneMask()).
     class SubRange : public LiveRange {
     public:
-      SubRange *Next;
+      SubRange *Next = nullptr;
       LaneBitmask LaneMask;
 
       /// Constructs a new SubRange object.
-      SubRange(LaneBitmask LaneMask)
+      SubRange(LaneBitmask LaneMask) : LaneMask(LaneMask) {}
-        : Next(nullptr), LaneMask(LaneMask) {
-      }
 
       /// Constructs a new SubRange object by copying liveness from @p Other.
       SubRange(LaneBitmask LaneMask, const LiveRange &Other,
                BumpPtrAllocator &Allocator)
-        : LiveRange(Other, Allocator), Next(nullptr), LaneMask(LaneMask) {
+        : LiveRange(Other, Allocator), LaneMask(LaneMask) {}
-      }
 
       void print(raw_ostream &OS) const;
       void dump() const;
     };
 
   private:
-    SubRange *SubRanges; ///< Single linked list of subregister live ranges.
+    SubRange *SubRanges = nullptr; ///< Single linked list of subregister live
+                                   /// ranges.
 
   public:
     const unsigned reg;  // the register or stack slot of this interval.
     float weight;        // weight of this interval
 
-    LiveInterval(unsigned Reg, float Weight)
+    LiveInterval(unsigned Reg, float Weight) : reg(Reg), weight(Weight) {}
-      : SubRanges(nullptr), reg(Reg), weight(Weight) {}
 
     ~LiveInterval() {
       clearSubRanges();
@@ -673,8 +676,10 @@ namespace llvm {
     template<typename T>
     class SingleLinkedListIterator {
       T *P;
+
     public:
       SingleLinkedListIterator<T>(T *P) : P(P) {}
+
       SingleLinkedListIterator<T> &operator++() {
         P = P->Next;
         return *this;
@@ -698,7 +703,9 @@ namespace llvm {
       }
     };
 
-    typedef SingleLinkedListIterator<SubRange> subrange_iterator;
+    using subrange_iterator = SingleLinkedListIterator<SubRange>;
+    using const_subrange_iterator = SingleLinkedListIterator<const SubRange>;
+
     subrange_iterator subrange_begin() {
       return subrange_iterator(SubRanges);
     }
@@ -706,7 +713,6 @@ namespace llvm {
       return subrange_iterator(nullptr);
     }
 
-    typedef SingleLinkedListIterator<const SubRange> const_subrange_iterator;
     const_subrange_iterator subrange_begin() const {
       return const_subrange_iterator(SubRanges);
     }
@@ -759,12 +765,12 @@ namespace llvm {
 
     /// isSpillable - Can this interval be spilled?
     bool isSpillable() const {
-      return weight != llvm::huge_valf;
+      return weight != huge_valf;
     }
 
     /// markNotSpillable - Mark interval as not spillable
     void markNotSpillable() {
-      weight = llvm::huge_valf;
+      weight = huge_valf;
     }
 
     /// For a given lane mask @p LaneMask, compute indexes at which the
@@ -931,5 +937,7 @@ namespace llvm {
     void Distribute(LiveInterval &LI, LiveInterval *LIV[],
                     MachineRegisterInfo &MRI);
   };
-}
+
-#endif
+} // end namespace llvm
+
+#endif // LLVM_CODEGEN_LIVEINTERVAL_H

include/llvm/CodeGen/LiveIntervalAnalysis.h

@@ -1,4 +1,4 @@
-//===-- LiveIntervalAnalysis.h - Live Interval Analysis ---------*- C++ -*-===//
+//===- LiveIntervalAnalysis.h - Live Interval Analysis ----------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -20,6 +20,7 @@
 #ifndef LLVM_CODEGEN_LIVEINTERVALANALYSIS_H
 #define LLVM_CODEGEN_LIVEINTERVALANALYSIS_H
 
+#include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/IndexedMap.h"
 #include "llvm/ADT/SmallVector.h"
 #include "llvm/Analysis/AliasAnalysis.h"
@@ -27,27 +28,29 @@
 #include "llvm/CodeGen/MachineBasicBlock.h"
 #include "llvm/CodeGen/MachineFunctionPass.h"
 #include "llvm/CodeGen/SlotIndexes.h"
-#include "llvm/Support/Allocator.h"
+#include "llvm/MC/LaneBitmask.h"
 #include "llvm/Support/CommandLine.h"
+#include "llvm/Support/Compiler.h"
+#include "llvm/Support/ErrorHandling.h"
 #include "llvm/Target/TargetRegisterInfo.h"
-#include <cmath>
+#include <cassert>
+#include <cstdint>
+#include <utility>
 
 namespace llvm {
 
 extern cl::opt<bool> UseSegmentSetForPhysRegs;
 
-  class BitVector;
+class BitVector;
-  class BlockFrequency;
+class LiveRangeCalc;
-  class LiveRangeCalc;
+class MachineBlockFrequencyInfo;
-  class LiveVariables;
+class MachineDominatorTree;
-  class MachineDominatorTree;
+class MachineFunction;
-  class MachineLoopInfo;
+class MachineInstr;
-  class TargetRegisterInfo;
+class MachineRegisterInfo;
-  class MachineRegisterInfo;
+class raw_ostream;
-  class TargetInstrInfo;
+class TargetInstrInfo;
-  class TargetRegisterClass;
+class VirtRegMap;
-  class VirtRegMap;
-  class MachineBlockFrequencyInfo;
 
   class LiveIntervals : public MachineFunctionPass {
     MachineFunction* MF;
@@ -56,8 +59,8 @@ extern cl::opt<bool> UseSegmentSetForPhysRegs;
     const TargetInstrInfo* TII;
     AliasAnalysis *AA;
     SlotIndexes* Indexes;
-    MachineDominatorTree *DomTree;
+    MachineDominatorTree *DomTree = nullptr;
-    LiveRangeCalc *LRCalc;
+    LiveRangeCalc *LRCalc = nullptr;
 
     /// Special pool allocator for VNInfo's (LiveInterval val#).
     VNInfo::Allocator VNInfoAllocator;
@@ -95,6 +98,7 @@ extern cl::opt<bool> UseSegmentSetForPhysRegs;
 
   public:
     static char ID;
+
     LiveIntervals();
     ~LiveIntervals() override;
 
@@ -466,6 +470,7 @@ extern cl::opt<bool> UseSegmentSetForPhysRegs;
 
     class HMEditor;
   };
-} // End llvm namespace
 
-#endif
+} // end namespace llvm
+
+#endif // LLVM_CODEGEN_LIVEINTERVALANALYSIS_H

include/llvm/CodeGen/LiveIntervalUnion.h

@@ -26,12 +26,14 @@
 
 namespace llvm {
 
+class raw_ostream;
 class TargetRegisterInfo;
 
 #ifndef NDEBUG
 // forward declaration
 template <unsigned Element> class SparseBitVector;
-typedef SparseBitVector<128> LiveVirtRegBitSet;
+
+using LiveVirtRegBitSet = SparseBitVector<128>;
 #endif
 
 /// Union of live intervals that are strong candidates for coalescing into a
@@ -42,19 +44,19 @@ class LiveIntervalUnion {
   // A set of live virtual register segments that supports fast insertion,
   // intersection, and removal.
   // Mapping SlotIndex intervals to virtual register numbers.
-  typedef IntervalMap<SlotIndex, LiveInterval*> LiveSegments;
+  using LiveSegments = IntervalMap<SlotIndex, LiveInterval*>;
 
 public:
   // SegmentIter can advance to the next segment ordered by starting position
   // which may belong to a different live virtual register. We also must be able
   // to reach the current segment's containing virtual register.
-  typedef LiveSegments::iterator SegmentIter;
+  using SegmentIter = LiveSegments::iterator;
 
   /// Const version of SegmentIter.
-  typedef LiveSegments::const_iterator ConstSegmentIter;
+  using ConstSegmentIter = LiveSegments::const_iterator;
 
   // LiveIntervalUnions share an external allocator.
-  typedef LiveSegments::Allocator Allocator;
+  using Allocator = LiveSegments::Allocator;
 
 private:
   unsigned Tag = 0;       // unique tag for current contents.
@@ -76,7 +78,7 @@ public:
   SlotIndex startIndex() const { return Segments.start(); }
 
   // Provide public access to the underlying map to allow overlap iteration.
-  typedef LiveSegments Map;
+  using Map = LiveSegments;
   const Map &getMap() const { return Segments; }
 
   /// getTag - Return an opaque tag representing the current state of the union.

include/llvm/CodeGen/LivePhysRegs.h

@@ -7,23 +7,24 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file implements the LivePhysRegs utility for tracking liveness of
+/// \file
-// physical registers. This can be used for ad-hoc liveness tracking after
+/// This file implements the LivePhysRegs utility for tracking liveness of
-// register allocation. You can start with the live-ins/live-outs at the
+/// physical registers. This can be used for ad-hoc liveness tracking after
-// beginning/end of a block and update the information while walking the
+/// register allocation. You can start with the live-ins/live-outs at the
-// instructions inside the block. This implementation tracks the liveness on a
+/// beginning/end of a block and update the information while walking the
-// sub-register granularity.
+/// instructions inside the block. This implementation tracks the liveness on a
-//
+/// sub-register granularity.
-// We assume that the high bits of a physical super-register are not preserved
+///
-// unless the instruction has an implicit-use operand reading the super-
+/// We assume that the high bits of a physical super-register are not preserved
-// register.
+/// unless the instruction has an implicit-use operand reading the super-
-//
+/// register.
-// X86 Example:
+///
-// %YMM0<def> = ...
+/// X86 Example:
-// %XMM0<def> = ... (Kills %XMM0, all %XMM0s sub-registers, and %YMM0)
+/// %YMM0<def> = ...
-//
+/// %XMM0<def> = ... (Kills %XMM0, all %XMM0s sub-registers, and %YMM0)
-// %YMM0<def> = ...
+///
-// %XMM0<def> = ..., %YMM0<imp-use> (%YMM0 and all its sub-registers are alive)
+/// %YMM0<def> = ...
+/// %XMM0<def> = ..., %YMM0<imp-use> (%YMM0 and all its sub-registers are alive)
 //===----------------------------------------------------------------------===//
 
 #ifndef LLVM_CODEGEN_LIVEPHYSREGS_H
@@ -39,40 +40,42 @@
 namespace llvm {
 
 class MachineInstr;
+class MachineOperand;
+class MachineRegisterInfo;
+class raw_ostream;
 
-/// \brief A set of live physical registers with functions to track liveness
+/// \brief A set of physical registers with utility functions to track liveness
 /// when walking backward/forward through a basic block.
 class LivePhysRegs {
   const TargetRegisterInfo *TRI = nullptr;
   SparseSet<unsigned> LiveRegs;
 
+public:
+  /// Constructs an unitialized set. init() needs to be called to initialize it.
+  LivePhysRegs() = default;
+
+  /// Constructs and initializes an empty set.
+  LivePhysRegs(const TargetRegisterInfo &TRI) : TRI(&TRI) {
+    LiveRegs.setUniverse(TRI.getNumRegs());
+  }
+
   LivePhysRegs(const LivePhysRegs&) = delete;
   LivePhysRegs &operator=(const LivePhysRegs&) = delete;
 
-public:
+  /// (re-)initializes and clears the set.
-  /// \brief Constructs a new empty LivePhysRegs set.
-  LivePhysRegs() = default;
-
-  /// \brief Constructs and initialize an empty LivePhysRegs set.
-  LivePhysRegs(const TargetRegisterInfo *TRI) : TRI(TRI) {
-    assert(TRI && "Invalid TargetRegisterInfo pointer.");
-    LiveRegs.setUniverse(TRI->getNumRegs());
-  }
-
-  /// \brief Clear and initialize the LivePhysRegs set.
   void init(const TargetRegisterInfo &TRI) {
     this->TRI = &TRI;
     LiveRegs.clear();
     LiveRegs.setUniverse(TRI.getNumRegs());
   }
 
-  /// \brief Clears the LivePhysRegs set.
+  /// Clears the set.
   void clear() { LiveRegs.clear(); }
 
-  /// \brief Returns true if the set is empty.
+  /// Returns true if the set is empty.
   bool empty() const { return LiveRegs.empty(); }
 
-  /// \brief Adds a physical register and all its sub-registers to the set.
+  /// Adds a physical register and all its sub-registers to the set.
   void addReg(unsigned Reg) {
     assert(TRI && "LivePhysRegs is not initialized.");
     assert(Reg <= TRI->getNumRegs() && "Expected a physical register.");
@@ -90,12 +93,13 @@ public:
       LiveRegs.erase(*R);
   }
 
-  /// \brief Removes physical registers clobbered by the regmask operand @p MO.
+  /// Removes physical registers clobbered by the regmask operand \p MO.
   void removeRegsInMask(const MachineOperand &MO,
-        SmallVectorImpl<std::pair<unsigned, const MachineOperand*>> *Clobbers);
+        SmallVectorImpl<std::pair<unsigned, const MachineOperand*>> *Clobbers =
+        nullptr);
 
-  /// \brief Returns true if register @p Reg is contained in the set. This also
+  /// \brief Returns true if register \p Reg is contained in the set. This also
-  /// works if only the super register of @p Reg has been defined, because
+  /// works if only the super register of \p Reg has been defined, because
   /// addReg() always adds all sub-registers to the set as well.
   /// Note: Returns false if just some sub registers are live, use available()
   /// when searching a free register.
@@ -104,48 +108,48 @@ public:
   /// Returns true if register \p Reg and no aliasing register is in the set.
   bool available(const MachineRegisterInfo &MRI, unsigned Reg) const;
 
-  /// \brief Simulates liveness when stepping backwards over an
+  /// Simulates liveness when stepping backwards over an instruction(bundle).
-  /// instruction(bundle): Remove Defs, add uses. This is the recommended way of
+  /// Remove Defs, add uses. This is the recommended way of calculating
-  /// calculating liveness.
+  /// liveness.
   void stepBackward(const MachineInstr &MI);
 
-  /// \brief Simulates liveness when stepping forward over an
+  /// Simulates liveness when stepping forward over an instruction(bundle).
-  /// instruction(bundle): Remove killed-uses, add defs. This is the not
+  /// Remove killed-uses, add defs. This is the not recommended way, because it
-  /// recommended way, because it depends on accurate kill flags. If possible
+  /// depends on accurate kill flags. If possible use stepBackward() instead of
-  /// use stepBackward() instead of this function.
+  /// this function. The clobbers set will be the list of registers either
-  /// The clobbers set will be the list of registers either defined or clobbered
+  /// defined or clobbered by a regmask.  The operand will identify whether this
-  /// by a regmask.  The operand will identify whether this is a regmask or
+  /// is a regmask or register operand.
-  /// register operand.
   void stepForward(const MachineInstr &MI,
         SmallVectorImpl<std::pair<unsigned, const MachineOperand*>> &Clobbers);
 
-  /// Adds all live-in registers of basic block @p MBB.
+  /// Adds all live-in registers of basic block \p MBB.
   /// Live in registers are the registers in the blocks live-in list and the
   /// pristine registers.
   void addLiveIns(const MachineBasicBlock &MBB);
 
-  /// Adds all live-out registers of basic block @p MBB.
+  /// Adds all live-out registers of basic block \p MBB.
   /// Live out registers are the union of the live-in registers of the successor
   /// blocks and pristine registers. Live out registers of the end block are the
   /// callee saved registers.
   void addLiveOuts(const MachineBasicBlock &MBB);
 
-  /// Like addLiveOuts() but does not add pristine registers/callee saved
+  /// Adds all live-out registers of basic block \p MBB but skips pristine
   /// registers.
   void addLiveOutsNoPristines(const MachineBasicBlock &MBB);
 
-  typedef SparseSet<unsigned>::const_iterator const_iterator;
+  using const_iterator = SparseSet<unsigned>::const_iterator;
+
   const_iterator begin() const { return LiveRegs.begin(); }
   const_iterator end() const { return LiveRegs.end(); }
 
-  /// \brief Prints the currently live registers to @p OS.
+  /// Prints the currently live registers to \p OS.
   void print(raw_ostream &OS) const;
 
-  /// \brief Dumps the currently live registers to the debug output.
+  /// Dumps the currently live registers to the debug output.
   void dump() const;
 
 private:
-  /// Adds live-in registers from basic block @p MBB, taking associated
+  /// \brief Adds live-in registers from basic block \p MBB, taking associated
   /// lane masks into consideration.
   void addBlockLiveIns(const MachineBasicBlock &MBB);
 };
@@ -155,11 +159,11 @@ inline raw_ostream &operator<<(raw_ostream &OS, const LivePhysRegs& LR) {
   return OS;
 }
 
-/// Compute the live-in list for \p MBB assuming all of its successors live-in
+/// \brief Computes the live-in list for \p MBB assuming all of its successors
-/// lists are up-to-date. Uses the given LivePhysReg instance \p LiveRegs; This
+/// live-in lists are up-to-date. Uses the given LivePhysReg instance \p
-/// is just here to avoid repeated heap allocations when calling this multiple
+/// LiveRegs; This is just here to avoid repeated heap allocations when calling
-/// times in a pass.
+/// this multiple times in a pass.
-void computeLiveIns(LivePhysRegs &LiveRegs, const TargetRegisterInfo &TRI,
+void computeLiveIns(LivePhysRegs &LiveRegs, const MachineRegisterInfo &MRI,
                     MachineBasicBlock &MBB);
 
 } // end namespace llvm

include/llvm/CodeGen/LiveRangeEdit.h

@@ -1,4 +1,4 @@
-//===---- LiveRangeEdit.h - Basic tools for split and spill -----*- C++ -*-===//
+//===- LiveRangeEdit.h - Basic tools for split and spill --------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -19,19 +19,28 @@
 #define LLVM_CODEGEN_LIVERANGEEDIT_H
 
 #include "llvm/ADT/ArrayRef.h"
+#include "llvm/ADT/None.h"
 #include "llvm/ADT/SetVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
+#include "llvm/ADT/SmallVector.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/CodeGen/LiveInterval.h"
+#include "llvm/CodeGen/MachineBasicBlock.h"
+#include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
-#include "llvm/Target/TargetMachine.h"
+#include "llvm/CodeGen/SlotIndexes.h"
 #include "llvm/Target/TargetSubtargetInfo.h"
+#include <cassert>
 
 namespace llvm {
 
 class LiveIntervals;
 class MachineBlockFrequencyInfo;
+class MachineInstr;
 class MachineLoopInfo;
+class MachineOperand;
+class TargetInstrInfo;
+class TargetRegisterInfo;
 class VirtRegMap;
 
 class LiveRangeEdit : private MachineRegisterInfo::Delegate {
@@ -39,7 +48,10 @@ public:
   /// Callback methods for LiveRangeEdit owners.
   class Delegate {
     virtual void anchor();
+
   public:
+    virtual ~Delegate() = default;
+
     /// Called immediately before erasing a dead machine instruction.
     virtual void LRE_WillEraseInstruction(MachineInstr *MI) {}
 
@@ -53,8 +65,6 @@ public:
     /// Called after cloning a virtual register.
     /// This is used for new registers representing connected components of Old.
     virtual void LRE_DidCloneVirtReg(unsigned New, unsigned Old) {}
-
-    virtual ~Delegate() {}
   };
 
 private:
@@ -70,7 +80,7 @@ private:
   const unsigned FirstNew;
 
   /// ScannedRemattable - true when remattable values have been identified.
-  bool ScannedRemattable;
+  bool ScannedRemattable = false;
 
   /// DeadRemats - The saved instructions which have already been dead after
   /// rematerialization but not deleted yet -- to be done in postOptimization.
@@ -78,11 +88,11 @@ private:
 
   /// Remattable - Values defined by remattable instructions as identified by
   /// tii.isTriviallyReMaterializable().
-  SmallPtrSet<const VNInfo*,4> Remattable;
+  SmallPtrSet<const VNInfo *, 4> Remattable;
 
   /// Rematted - Values that were actually rematted, and so need to have their
   /// live range trimmed or entirely removed.
-  SmallPtrSet<const VNInfo*,4> Rematted;
+  SmallPtrSet<const VNInfo *, 4> Rematted;
 
   /// scanRemattable - Identify the Parent values that may rematerialize.
   void scanRemattable(AliasAnalysis *aa);
@@ -94,11 +104,11 @@ private:
 
   /// foldAsLoad - If LI has a single use and a single def that can be folded as
   /// a load, eliminate the register by folding the def into the use.
-  bool foldAsLoad(LiveInterval *LI, SmallVectorImpl<MachineInstr*> &Dead);
+  bool foldAsLoad(LiveInterval *LI, SmallVectorImpl<MachineInstr *> &Dead);
+
+  using ToShrinkSet = SetVector<LiveInterval *, SmallVector<LiveInterval *, 8>,
+                                SmallPtrSet<LiveInterval *, 8>>;
 
-  typedef SetVector<LiveInterval*,
-                    SmallVector<LiveInterval*, 8>,
-                    SmallPtrSet<LiveInterval*, 8> > ToShrinkSet;
   /// Helper for eliminateDeadDefs.
   void eliminateDeadDef(MachineInstr *MI, ToShrinkSet &ToShrink,
                         AliasAnalysis *AA);
@@ -129,26 +139,26 @@ public:
                 SmallPtrSet<MachineInstr *, 32> *deadRemats = nullptr)
       : Parent(parent), NewRegs(newRegs), MRI(MF.getRegInfo()), LIS(lis),
         VRM(vrm), TII(*MF.getSubtarget().getInstrInfo()), TheDelegate(delegate),
-        FirstNew(newRegs.size()), ScannedRemattable(false),
+        FirstNew(newRegs.size()), DeadRemats(deadRemats) {
-        DeadRemats(deadRemats) {
     MRI.setDelegate(this);
   }
 
   ~LiveRangeEdit() override { MRI.resetDelegate(this); }
 
   LiveInterval &getParent() const {
-   assert(Parent && "No parent LiveInterval");
+    assert(Parent && "No parent LiveInterval");
-   return *Parent;
+    return *Parent;
   }
+
   unsigned getReg() const { return getParent().reg; }
 
   /// Iterator for accessing the new registers added by this edit.
-  typedef SmallVectorImpl<unsigned>::const_iterator iterator;
+  using iterator = SmallVectorImpl<unsigned>::const_iterator;
-  iterator begin() const { return NewRegs.begin()+FirstNew; }
+  iterator begin() const { return NewRegs.begin() + FirstNew; }
   iterator end() const { return NewRegs.end(); }
-  unsigned size() const { return NewRegs.size()-FirstNew; }
+  unsigned size() const { return NewRegs.size() - FirstNew; }
   bool empty() const { return size() == 0; }
-  unsigned get(unsigned idx) const { return NewRegs[idx+FirstNew]; }
+  unsigned get(unsigned idx) const { return NewRegs[idx + FirstNew]; }
 
   /// pop_back - It allows LiveRangeEdit users to drop new registers.
   /// The context is when an original def instruction of a register is
@@ -176,26 +186,25 @@ public:
     return createEmptyIntervalFrom(getReg());
   }
 
-  unsigned create() {
+  unsigned create() { return createFrom(getReg()); }
-    return createFrom(getReg());
-  }
 
   /// anyRematerializable - Return true if any parent values may be
   /// rematerializable.
   /// This function must be called before any rematerialization is attempted.
-  bool anyRematerializable(AliasAnalysis*);
+  bool anyRematerializable(AliasAnalysis *);
 
   /// checkRematerializable - Manually add VNI to the list of rematerializable
   /// values if DefMI may be rematerializable.
   bool checkRematerializable(VNInfo *VNI, const MachineInstr *DefMI,
-                             AliasAnalysis*);
+                             AliasAnalysis *);
 
   /// Remat - Information needed to rematerialize at a specific location.
   struct Remat {
-    VNInfo *ParentVNI;      // parent_'s value at the remat location.
+    VNInfo *ParentVNI;              // parent_'s value at the remat location.
-    MachineInstr *OrigMI;   // Instruction defining OrigVNI. It contains the
+    MachineInstr *OrigMI = nullptr; // Instruction defining OrigVNI. It contains
-                            // real expr for remat.
+                                    // the real expr for remat.
-    explicit Remat(VNInfo *ParentVNI) : ParentVNI(ParentVNI), OrigMI(nullptr) {}
+
+    explicit Remat(VNInfo *ParentVNI) : ParentVNI(ParentVNI) {}
   };
 
   /// canRematerializeAt - Determine if ParentVNI can be rematerialized at
@@ -209,10 +218,8 @@ public:
   /// liveness is not updated.
   /// Return the SlotIndex of the new instruction.
   SlotIndex rematerializeAt(MachineBasicBlock &MBB,
-                            MachineBasicBlock::iterator MI,
+                            MachineBasicBlock::iterator MI, unsigned DestReg,
-                            unsigned DestReg,
+                            const Remat &RM, const TargetRegisterInfo &,
-                            const Remat &RM,
-                            const TargetRegisterInfo&,
                             bool Late = false);
 
   /// markRematerialized - explicitly mark a value as rematerialized after doing
@@ -248,11 +255,10 @@ public:
 
   /// calculateRegClassAndHint - Recompute register class and hint for each new
   /// register.
-  void calculateRegClassAndHint(MachineFunction&,
+  void calculateRegClassAndHint(MachineFunction &, const MachineLoopInfo &,
-                                const MachineLoopInfo&,
+                                const MachineBlockFrequencyInfo &);
-                                const MachineBlockFrequencyInfo&);
 };
 
-}
+} // end namespace llvm
 
-#endif
+#endif // LLVM_CODEGEN_LIVERANGEEDIT_H

include/llvm/CodeGen/LiveStackAnalysis.h

@@ -1,4 +1,4 @@
-//===-- LiveStackAnalysis.h - Live Stack Slot Analysis ----------*- C++ -*-===//
+//===- LiveStackAnalysis.h - Live Stack Slot Analysis -----------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -18,13 +18,16 @@
 
 #include "llvm/CodeGen/LiveInterval.h"
 #include "llvm/CodeGen/MachineFunctionPass.h"
-#include "llvm/Support/Allocator.h"
+#include "llvm/Pass.h"
-#include "llvm/Target/TargetRegisterInfo.h"
+#include <cassert>
 #include <map>
 #include <unordered_map>
 
 namespace llvm {
 
+class TargetRegisterClass;
+class TargetRegisterInfo;
+
 class LiveStacks : public MachineFunctionPass {
   const TargetRegisterInfo *TRI;
 
@@ -33,8 +36,7 @@ class LiveStacks : public MachineFunctionPass {
   VNInfo::Allocator VNInfoAllocator;
 
   /// S2IMap - Stack slot indices to live interval mapping.
-  ///
+  using SS2IntervalMap = std::unordered_map<int, LiveInterval>;
-  typedef std::unordered_map<int, LiveInterval> SS2IntervalMap;
   SS2IntervalMap S2IMap;
 
   /// S2RCMap - Stack slot indices to register class mapping.
@@ -42,12 +44,14 @@ class LiveStacks : public MachineFunctionPass {
 
 public:
   static char ID; // Pass identification, replacement for typeid
+
   LiveStacks() : MachineFunctionPass(ID) {
     initializeLiveStacksPass(*PassRegistry::getPassRegistry());
   }
 
-  typedef SS2IntervalMap::iterator iterator;
+  using iterator = SS2IntervalMap::iterator;
-  typedef SS2IntervalMap::const_iterator const_iterator;
+  using const_iterator = SS2IntervalMap::const_iterator;
+
   const_iterator begin() const { return S2IMap.begin(); }
   const_iterator end() const { return S2IMap.end(); }
   iterator begin() { return S2IMap.begin(); }
@@ -93,6 +97,7 @@ public:
   /// print - Implement the dump method.
   void print(raw_ostream &O, const Module * = nullptr) const override;
 };
-}
 
-#endif /* LLVM_CODEGEN_LIVESTACK_ANALYSIS_H */
+} // end namespace llvm
+
+#endif // LLVM_CODEGEN_LIVESTACK_ANALYSIS_H

include/llvm/CodeGen/MachineBasicBlock.h

@@ -1,4 +1,4 @@
-//===-- llvm/CodeGen/MachineBasicBlock.h ------------------------*- C++ -*-===//
+//===- llvm/CodeGen/MachineBasicBlock.h -------------------------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -15,41 +15,50 @@
 #define LLVM_CODEGEN_MACHINEBASICBLOCK_H
 
 #include "llvm/ADT/GraphTraits.h"
+#include "llvm/ADT/ilist.h"
+#include "llvm/ADT/ilist_node.h"
 #include "llvm/ADT/iterator_range.h"
+#include "llvm/ADT/simple_ilist.h"
 #include "llvm/CodeGen/MachineInstrBundleIterator.h"
 #include "llvm/CodeGen/MachineInstr.h"
+#include "llvm/IR/DebugLoc.h"
 #include "llvm/Support/BranchProbability.h"
 #include "llvm/MC/LaneBitmask.h"
 #include "llvm/MC/MCRegisterInfo.h"
-#include "llvm/Support/DataTypes.h"
+#include <cassert>
+#include <cstdint>
 #include <functional>
+#include <iterator>
+#include <string>
+#include <vector>
 
 namespace llvm {
 
-class Pass;
 class BasicBlock;
 class MachineFunction;
 class MCSymbol;
-class MIPrinter;
+class ModuleSlotTracker;
+class Pass;
 class SlotIndexes;
 class StringRef;
 class raw_ostream;
-class MachineBranchProbabilityInfo;
+class TargetRegisterClass;
+class TargetRegisterInfo;
 
 template <> struct ilist_traits<MachineInstr> {
 private:
   friend class MachineBasicBlock; // Set by the owning MachineBasicBlock.
+
   MachineBasicBlock *Parent;
 
-  typedef simple_ilist<MachineInstr, ilist_sentinel_tracking<true>>::iterator
+  using instr_iterator =
-      instr_iterator;
+      simple_ilist<MachineInstr, ilist_sentinel_tracking<true>>::iterator;
 
 public:
   void addNodeToList(MachineInstr *N);
   void removeNodeFromList(MachineInstr *N);
   void transferNodesFromList(ilist_traits &OldList, instr_iterator First,
                              instr_iterator Last);
-
   void deleteNode(MachineInstr *MI);
 };
 
@@ -69,7 +78,8 @@ public:
   };
 
 private:
-  typedef ilist<MachineInstr, ilist_sentinel_tracking<true>> Instructions;
+  using Instructions = ilist<MachineInstr, ilist_sentinel_tracking<true>>;
+
   Instructions Insts;
   const BasicBlock *BB;
   int Number;
@@ -83,12 +93,12 @@ private:
   /// same order as Successors, or it is empty if we don't use it (disable
   /// optimization).
   std::vector<BranchProbability> Probs;
-  typedef std::vector<BranchProbability>::iterator probability_iterator;
+  using probability_iterator = std::vector<BranchProbability>::iterator;
-  typedef std::vector<BranchProbability>::const_iterator
+  using const_probability_iterator =
-      const_probability_iterator;
+      std::vector<BranchProbability>::const_iterator;
 
   /// Keep track of the physical registers that are livein of the basicblock.
-  typedef std::vector<RegisterMaskPair> LiveInVector;
+  using LiveInVector = std::vector<RegisterMaskPair>;
   LiveInVector LiveIns;
 
   /// Alignment of the basic block. Zero if the basic block does not need to be
@@ -113,7 +123,7 @@ private:
   mutable MCSymbol *CachedMCSymbol = nullptr;
 
   // Intrusive list support
-  MachineBasicBlock() {}
+  MachineBasicBlock() = default;
 
   explicit MachineBasicBlock(MachineFunction &MF, const BasicBlock *BB);
 
@@ -145,16 +155,16 @@ public:
   const MachineFunction *getParent() const { return xParent; }
   MachineFunction *getParent() { return xParent; }
 
-  typedef Instructions::iterator                                 instr_iterator;
+  using instr_iterator = Instructions::iterator;
-  typedef Instructions::const_iterator                     const_instr_iterator;
+  using const_instr_iterator = Instructions::const_iterator;
-  typedef Instructions::reverse_iterator reverse_instr_iterator;
+  using reverse_instr_iterator = Instructions::reverse_iterator;
-  typedef Instructions::const_reverse_iterator const_reverse_instr_iterator;
+  using const_reverse_instr_iterator = Instructions::const_reverse_iterator;
 
-  typedef MachineInstrBundleIterator<MachineInstr> iterator;
+  using iterator = MachineInstrBundleIterator<MachineInstr>;
-  typedef MachineInstrBundleIterator<const MachineInstr> const_iterator;
+  using const_iterator = MachineInstrBundleIterator<const MachineInstr>;
-  typedef MachineInstrBundleIterator<MachineInstr, true> reverse_iterator;
+  using reverse_iterator = MachineInstrBundleIterator<MachineInstr, true>;
-  typedef MachineInstrBundleIterator<const MachineInstr, true>
+  using const_reverse_iterator =
-      const_reverse_iterator;
+      MachineInstrBundleIterator<const MachineInstr, true>;
 
   unsigned size() const { return (unsigned)Insts.size(); }
   bool empty() const { return Insts.empty(); }
@@ -178,8 +188,8 @@ public:
   reverse_instr_iterator       instr_rend  ()       { return Insts.rend();   }
   const_reverse_instr_iterator instr_rend  () const { return Insts.rend();   }
 
-  typedef iterator_range<instr_iterator> instr_range;
+  using instr_range = iterator_range<instr_iterator>;
-  typedef iterator_range<const_instr_iterator> const_instr_range;
+  using const_instr_range = iterator_range<const_instr_iterator>;
   instr_range instrs() { return instr_range(instr_begin(), instr_end()); }
   const_instr_range instrs() const {
     return const_instr_range(instr_begin(), instr_end());
@@ -213,18 +223,18 @@ public:
   }
 
   // Machine-CFG iterators
-  typedef std::vector<MachineBasicBlock *>::iterator       pred_iterator;
+  using pred_iterator = std::vector<MachineBasicBlock *>::iterator;
-  typedef std::vector<MachineBasicBlock *>::const_iterator const_pred_iterator;
+  using const_pred_iterator = std::vector<MachineBasicBlock *>::const_iterator;
-  typedef std::vector<MachineBasicBlock *>::iterator       succ_iterator;
+  using succ_iterator = std::vector<MachineBasicBlock *>::iterator;
-  typedef std::vector<MachineBasicBlock *>::const_iterator const_succ_iterator;
+  using const_succ_iterator = std::vector<MachineBasicBlock *>::const_iterator;
-  typedef std::vector<MachineBasicBlock *>::reverse_iterator
+  using pred_reverse_iterator =
-                                                         pred_reverse_iterator;
+      std::vector<MachineBasicBlock *>::reverse_iterator;
-  typedef std::vector<MachineBasicBlock *>::const_reverse_iterator
+  using const_pred_reverse_iterator =
-                                                   const_pred_reverse_iterator;
+      std::vector<MachineBasicBlock *>::const_reverse_iterator;
-  typedef std::vector<MachineBasicBlock *>::reverse_iterator
+  using succ_reverse_iterator =
-                                                         succ_reverse_iterator;
+      std::vector<MachineBasicBlock *>::reverse_iterator;
-  typedef std::vector<MachineBasicBlock *>::const_reverse_iterator
+  using const_succ_reverse_iterator =
-                                                   const_succ_reverse_iterator;
+      std::vector<MachineBasicBlock *>::const_reverse_iterator;
   pred_iterator        pred_begin()       { return Predecessors.begin(); }
   const_pred_iterator  pred_begin() const { return Predecessors.begin(); }
   pred_iterator        pred_end()         { return Predecessors.end();   }
@@ -307,7 +317,7 @@ public:
 
   // Iteration support for live in sets.  These sets are kept in sorted
   // order by their register number.
-  typedef LiveInVector::const_iterator livein_iterator;
+  using livein_iterator = LiveInVector::const_iterator;
 #ifndef NDEBUG
   /// Unlike livein_begin, this method does not check that the liveness
   /// information is accurate. Still for debug purposes it may be useful
@@ -455,7 +465,6 @@ public:
   /// other block.
   bool isLayoutSuccessor(const MachineBasicBlock *MBB) const;
 
-
   /// Return the fallthrough block if the block can implicitly
   /// transfer control to the block after it by falling off the end of
   /// it.  This should return null if it can reach the block after
@@ -695,7 +704,7 @@ public:
   LivenessQueryResult computeRegisterLiveness(const TargetRegisterInfo *TRI,
                                               unsigned Reg,
                                               const_iterator Before,
-                                              unsigned Neighborhood=10) const;
+                                              unsigned Neighborhood = 10) const;
 
   // Debugging methods.
   void dump() const;
@@ -714,7 +723,6 @@ public:
   /// Return the MCSymbol for this basic block.
   MCSymbol *getSymbol() const;
 
-
 private:
   /// Return probability iterator corresponding to the I successor iterator.
   probability_iterator getProbabilityIterator(succ_iterator I);
@@ -764,8 +772,8 @@ struct MBB2NumberFunctor :
 //
 
 template <> struct GraphTraits<MachineBasicBlock *> {
-  typedef MachineBasicBlock *NodeRef;
+  using NodeRef = MachineBasicBlock *;
-  typedef MachineBasicBlock::succ_iterator ChildIteratorType;
+  using ChildIteratorType = MachineBasicBlock::succ_iterator;
 
   static NodeRef getEntryNode(MachineBasicBlock *BB) { return BB; }
   static ChildIteratorType child_begin(NodeRef N) { return N->succ_begin(); }
@@ -773,8 +781,8 @@ template <> struct GraphTraits<MachineBasicBlock *> {
 };
 
 template <> struct GraphTraits<const MachineBasicBlock *> {
-  typedef const MachineBasicBlock *NodeRef;
+  using NodeRef = const MachineBasicBlock *;
-  typedef MachineBasicBlock::const_succ_iterator ChildIteratorType;
+  using ChildIteratorType = MachineBasicBlock::const_succ_iterator;
 
   static NodeRef getEntryNode(const MachineBasicBlock *BB) { return BB; }
   static ChildIteratorType child_begin(NodeRef N) { return N->succ_begin(); }
@@ -787,28 +795,30 @@ template <> struct GraphTraits<const MachineBasicBlock *> {
 // to be when traversing the predecessor edges of a MBB
 // instead of the successor edges.
 //
-template <> struct GraphTraits<Inverse<MachineBasicBlock*> > {
+template <> struct GraphTraits<Inverse<MachineBasicBlock*>> {
-  typedef MachineBasicBlock *NodeRef;
+  using NodeRef = MachineBasicBlock *;
-  typedef MachineBasicBlock::pred_iterator ChildIteratorType;
+  using ChildIteratorType = MachineBasicBlock::pred_iterator;
+
   static NodeRef getEntryNode(Inverse<MachineBasicBlock *> G) {
     return G.Graph;
   }
+
   static ChildIteratorType child_begin(NodeRef N) { return N->pred_begin(); }
   static ChildIteratorType child_end(NodeRef N) { return N->pred_end(); }
 };
 
-template <> struct GraphTraits<Inverse<const MachineBasicBlock*> > {
+template <> struct GraphTraits<Inverse<const MachineBasicBlock*>> {
-  typedef const MachineBasicBlock *NodeRef;
+  using NodeRef = const MachineBasicBlock *;
-  typedef MachineBasicBlock::const_pred_iterator ChildIteratorType;
+  using ChildIteratorType = MachineBasicBlock::const_pred_iterator;
+
   static NodeRef getEntryNode(Inverse<const MachineBasicBlock *> G) {
     return G.Graph;
   }
+
   static ChildIteratorType child_begin(NodeRef N) { return N->pred_begin(); }
   static ChildIteratorType child_end(NodeRef N) { return N->pred_end(); }
 };
 
-
-
 /// MachineInstrSpan provides an interface to get an iteration range
 /// containing the instruction it was initialized with, along with all
 /// those instructions inserted prior to or following that instruction
@@ -816,6 +826,7 @@ template <> struct GraphTraits<Inverse<const MachineBasicBlock*> > {
 class MachineInstrSpan {
   MachineBasicBlock &MBB;
   MachineBasicBlock::iterator I, B, E;
+
 public:
   MachineInstrSpan(MachineBasicBlock::iterator I)
     : MBB(*I->getParent()),
@@ -854,6 +865,6 @@ inline IterT skipDebugInstructionsBackward(IterT It, IterT Begin) {
   return It;
 }
 
-} // End llvm namespace
+} // end namespace llvm
 
-#endif
+#endif // LLVM_CODEGEN_MACHINEBASICBLOCK_H

include/llvm/CodeGen/MachineBlockFrequencyInfo.h

@@ -1,4 +1,4 @@
-//===- MachineBlockFrequencyInfo.h - MBB Frequency Analysis -*- C++ -*-----===//
+//===- MachineBlockFrequencyInfo.h - MBB Frequency Analysis -----*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -17,26 +17,28 @@
 #include "llvm/ADT/Optional.h"
 #include "llvm/CodeGen/MachineFunctionPass.h"
 #include "llvm/Support/BlockFrequency.h"
-#include <climits>
+#include <cstdint>
+#include <memory>
 
 namespace llvm {
 
+template <class BlockT> class BlockFrequencyInfoImpl;
 class MachineBasicBlock;
 class MachineBranchProbabilityInfo;
+class MachineFunction;
 class MachineLoopInfo;
-template <class BlockT> class BlockFrequencyInfoImpl;
+class raw_ostream;
 
 /// MachineBlockFrequencyInfo pass uses BlockFrequencyInfoImpl implementation
 /// to estimate machine basic block frequencies.
 class MachineBlockFrequencyInfo : public MachineFunctionPass {
-  typedef BlockFrequencyInfoImpl<MachineBasicBlock> ImplType;
+  using ImplType = BlockFrequencyInfoImpl<MachineBasicBlock>;
   std::unique_ptr<ImplType> MBFI;
 
 public:
   static char ID;
 
   MachineBlockFrequencyInfo();
-
   ~MachineBlockFrequencyInfo() override;
 
   void getAnalysisUsage(AnalysisUsage &AU) const override;
@@ -74,9 +76,8 @@ public:
                               const MachineBasicBlock *MBB) const;
 
   uint64_t getEntryFreq() const;
-
 };
 
-}
+} // end namespace llvm
 
-#endif
+#endif // LLVM_CODEGEN_MACHINEBLOCKFREQUENCYINFO_H

include/llvm/CodeGen/MachineDominanceFrontier.h

@@ -11,23 +11,28 @@
 #define LLVM_CODEGEN_MACHINEDOMINANCEFRONTIER_H
 
 #include "llvm/Analysis/DominanceFrontier.h"
+#include "llvm/Analysis/DominanceFrontierImpl.h"
 #include "llvm/CodeGen/MachineBasicBlock.h"
 #include "llvm/CodeGen/MachineFunctionPass.h"
-
+#include "llvm/Support/GenericDomTree.h"
+#include <vector>
 
 namespace llvm {
 
 class MachineDominanceFrontier : public MachineFunctionPass {
   ForwardDominanceFrontierBase<MachineBasicBlock> Base;
-public:
-  typedef DominatorTreeBase<MachineBasicBlock> DomTreeT;
-  typedef DomTreeNodeBase<MachineBasicBlock> DomTreeNodeT;
-  typedef DominanceFrontierBase<MachineBasicBlock>::DomSetType DomSetType;
-  typedef DominanceFrontierBase<MachineBasicBlock>::iterator iterator;
-  typedef DominanceFrontierBase<MachineBasicBlock>::const_iterator const_iterator;
 
-  void operator=(const MachineDominanceFrontier &) = delete;
+public:
+  using DomTreeT = DominatorTreeBase<MachineBasicBlock>;
+  using DomTreeNodeT = DomTreeNodeBase<MachineBasicBlock>;
+  using DomSetType = DominanceFrontierBase<MachineBasicBlock>::DomSetType;
+  using iterator = DominanceFrontierBase<MachineBasicBlock>::iterator;
+  using const_iterator =
+      DominanceFrontierBase<MachineBasicBlock>::const_iterator;
+
   MachineDominanceFrontier(const MachineDominanceFrontier &) = delete;
+  MachineDominanceFrontier &
+  operator=(const MachineDominanceFrontier &) = delete;
 
   static char ID;
 
@@ -104,6 +109,6 @@ public:
   void getAnalysisUsage(AnalysisUsage &AU) const override;
 };
 
-}
+} // end namespace llvm
 
-#endif
+#endif // LLVM_CODEGEN_MACHINEDOMINANCEFRONTIER_H

include/llvm/CodeGen/MachineDominators.h

@@ -1,4 +1,4 @@
-//=- llvm/CodeGen/MachineDominators.h - Machine Dom Calculation --*- C++ -*-==//
+//==- llvm/CodeGen/MachineDominators.h - Machine Dom Calculation -*- C++ -*-==//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -16,12 +16,15 @@
 #define LLVM_CODEGEN_MACHINEDOMINATORS_H
 
 #include "llvm/ADT/SmallSet.h"
+#include "llvm/ADT/SmallVector.h"
 #include "llvm/CodeGen/MachineBasicBlock.h"
-#include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/MachineFunctionPass.h"
+#include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/Support/GenericDomTree.h"
 #include "llvm/Support/GenericDomTreeConstruction.h"
+#include <cassert>
 #include <memory>
+#include <vector>
 
 namespace llvm {
 
@@ -33,7 +36,7 @@ inline void DominatorTreeBase<MachineBasicBlock>::addRoot(MachineBasicBlock* MBB
 extern template class DomTreeNodeBase<MachineBasicBlock>;
 extern template class DominatorTreeBase<MachineBasicBlock>;
 
-typedef DomTreeNodeBase<MachineBasicBlock> MachineDomTreeNode;
+using MachineDomTreeNode = DomTreeNodeBase<MachineBasicBlock>;
 
 //===-------------------------------------
 /// DominatorTree Class - Concrete subclass of DominatorTreeBase that is used to
@@ -52,6 +55,7 @@ class MachineDominatorTree : public MachineFunctionPass {
   /// The splitting of a critical edge is local and thus, it is possible
   /// to apply several of those changes at the same time.
   mutable SmallVector<CriticalEdge, 32> CriticalEdgesToSplit;
+
   /// \brief Remember all the basic blocks that are inserted during
   /// edge splitting.
   /// Invariant: NewBBs == all the basic blocks contained in the NewBB
@@ -259,8 +263,8 @@ public:
 
 template <class Node, class ChildIterator>
 struct MachineDomTreeGraphTraitsBase {
-  typedef Node *NodeRef;
+  using NodeRef = Node *;
-  typedef ChildIterator ChildIteratorType;
+  using ChildIteratorType = ChildIterator;
 
   static NodeRef getEntryNode(NodeRef N) { return N; }
   static ChildIteratorType child_begin(NodeRef N) { return N->begin(); }
@@ -287,6 +291,6 @@ template <> struct GraphTraits<MachineDominatorTree*>
   }
 };
 
-}
+} // end namespace llvm
 
-#endif
+#endif // LLVM_CODEGEN_MACHINEDOMINATORS_H

include/llvm/CodeGen/MachineInstr.h

@@ -826,20 +826,12 @@ public:
       getOperand(0).getSubReg() == getOperand(1).getSubReg();
   }
 
-  /// Return true if this is a transient instruction that is
+  /// Return true if this instruction doesn't produce any output in the form of
-  /// either very likely to be eliminated during register allocation (such as
+  /// executable instructions.
-  /// copy-like instructions), or if this instruction doesn't have an
+  bool isMetaInstruction() const {
-  /// execution-time cost.
+    switch (getOpcode()) {
-  bool isTransient() const {
+    default:
-    switch(getOpcode()) {
+      return false;
-    default: return false;
-    // Copy-like instructions are usually eliminated during register allocation.
-    case TargetOpcode::PHI:
-    case TargetOpcode::COPY:
-    case TargetOpcode::INSERT_SUBREG:
-    case TargetOpcode::SUBREG_TO_REG:
-    case TargetOpcode::REG_SEQUENCE:
-    // Pseudo-instructions that don't produce any real output.
     case TargetOpcode::IMPLICIT_DEF:
     case TargetOpcode::KILL:
     case TargetOpcode::CFI_INSTRUCTION:
@@ -850,6 +842,23 @@ public:
     }
   }
 
+  /// Return true if this is a transient instruction that is either very likely
+  /// to be eliminated during register allocation (such as copy-like
+  /// instructions), or if this instruction doesn't have an execution-time cost.
+  bool isTransient() const {
+    switch (getOpcode()) {
+    default:
+      return isMetaInstruction();
+    // Copy-like instructions are usually eliminated during register allocation.
+    case TargetOpcode::PHI:
+    case TargetOpcode::COPY:
+    case TargetOpcode::INSERT_SUBREG:
+    case TargetOpcode::SUBREG_TO_REG:
+    case TargetOpcode::REG_SEQUENCE:
+      return true;
+    }
+  }
+
   /// Return the number of instructions inside the MI bundle, excluding the
   /// bundle header.
   ///

include/llvm/CodeGen/MachineRegisterInfo.h

@@ -642,6 +642,11 @@ public:
   ///
   void setRegBank(unsigned Reg, const RegisterBank &RegBank);
 
+  void setRegClassOrRegBank(unsigned Reg,
+                            const RegClassOrRegBank &RCOrRB){
+    VRegInfo[Reg].first = RCOrRB;
+  }
+
   /// constrainRegClass - Constrain the register class of the specified virtual
   /// register to be a common subclass of RC and the current register class,
   /// but only if the new class has at least MinNumRegs registers.  Return the

include/llvm/CodeGen/MachineValueType.h

@@ -26,7 +26,7 @@ namespace llvm {
   /// Machine Value Type. Every type that is supported natively by some
   /// processor targeted by LLVM occurs here. This means that any legal value
   /// type can be represented by an MVT.
-class MVT {
+  class MVT {
   public:
     enum SimpleValueType : uint8_t {
       // Simple value types that aren't explicitly part of this enumeration

include/llvm/CodeGen/ScheduleDAG.h

@@ -52,14 +52,14 @@ class TargetRegisterInfo;
     /// These are the different kinds of scheduling dependencies.
     enum Kind {
       Data,        ///< Regular data dependence (aka true-dependence).
-      Anti,        ///< A register anti-dependedence (aka WAR).
+      Anti,        ///< A register anti-dependence (aka WAR).
       Output,      ///< A register output-dependence (aka WAW).
       Order        ///< Any other ordering dependency.
     };
 
     // Strong dependencies must be respected by the scheduler. Artificial
     // dependencies may be removed only if they are redundant with another
-    // strong depedence.
+    // strong dependence.
     //
     // Weak dependencies may be violated by the scheduling strategy, but only if
     // the strategy can prove it is correct to do so.
@@ -342,7 +342,7 @@ class TargetRegisterInfo;
     /// BoundaryNodes can have DAG edges, including Data edges, but they do not
     /// correspond to schedulable entities (e.g. instructions) and do not have a
     /// valid ID. Consequently, always check for boundary nodes before accessing
-    /// an assoicative data structure keyed on node ID.
+    /// an associative data structure keyed on node ID.
     bool isBoundaryNode() const { return NodeNum == BoundaryID; }
 
     /// Assigns the representative SDNode for this SUnit. This may be used

include/llvm/CodeGen/ScheduleDAGInstrs.h

@@ -18,6 +18,7 @@
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SparseMultiSet.h"
 #include "llvm/ADT/SparseSet.h"
+#include "llvm/CodeGen/LivePhysRegs.h"
 #include "llvm/CodeGen/ScheduleDAG.h"
 #include "llvm/CodeGen/TargetSchedule.h"
 #include "llvm/Support/Compiler.h"
@@ -224,7 +225,7 @@ namespace llvm {
     MachineInstr *FirstDbgValue;
 
     /// Set of live physical registers for updating kill flags.
-    BitVector LiveRegs;
+    LivePhysRegs LiveRegs;
 
   public:
     explicit ScheduleDAGInstrs(MachineFunction &mf,
@@ -311,7 +312,7 @@ namespace llvm {
     std::string getDAGName() const override;
 
     /// Fixes register kill flags that scheduling has made invalid.
-    void fixupKills(MachineBasicBlock *MBB);
+    void fixupKills(MachineBasicBlock &MBB);
 
   protected:
     void initSUnits();

include/llvm/CodeGen/SelectionDAG.h

@@ -1070,6 +1070,11 @@ public:
   SDNode *MorphNodeTo(SDNode *N, unsigned Opc, SDVTList VTs,
                       ArrayRef<SDValue> Ops);
 
+  /// Mutate the specified strict FP node to its non-strict equivalent,
+  /// unlinking the node from its chain and dropping the metadata arguments.
+  /// The node must be a strict FP node.
+  SDNode *mutateStrictFPToFP(SDNode *Node);
+
   /// These are used for target selectors to create a new node
   /// with specified return type(s), MachineInstr opcode, and operands.
   ///

include/llvm/CodeGen/SelectionDAGNodes.h

@@ -612,6 +612,32 @@ public:
            SDNodeBits.IsMemIntrinsic;
   }
 
+  /// Test if this node is a strict floating point pseudo-op.
+  bool isStrictFPOpcode() {
+    switch (NodeType) {
+      default: 
+        return false;
+      case ISD::STRICT_FADD:
+      case ISD::STRICT_FSUB:
+      case ISD::STRICT_FMUL:
+      case ISD::STRICT_FDIV:
+      case ISD::STRICT_FREM:
+      case ISD::STRICT_FSQRT:
+      case ISD::STRICT_FPOW:
+      case ISD::STRICT_FPOWI:
+      case ISD::STRICT_FSIN:
+      case ISD::STRICT_FCOS:
+      case ISD::STRICT_FEXP:
+      case ISD::STRICT_FEXP2:
+      case ISD::STRICT_FLOG:
+      case ISD::STRICT_FLOG10:
+      case ISD::STRICT_FLOG2:
+      case ISD::STRICT_FRINT:
+      case ISD::STRICT_FNEARBYINT:
+        return true;
+    }
+  }
+
   /// Test if this node has a post-isel opcode, directly
   /// corresponding to a MachineInstr opcode.
   bool isMachineOpcode() const { return NodeType < 0; }

include/llvm/DebugInfo/CodeView/CVRecord.h

@@ -14,6 +14,7 @@
 #include "llvm/ADT/Optional.h"
 #include "llvm/DebugInfo/CodeView/CodeViewError.h"
 #include "llvm/DebugInfo/CodeView/RecordSerialization.h"
+#include "llvm/DebugInfo/CodeView/TypeIndex.h"
 #include "llvm/Support/BinaryStreamReader.h"
 #include "llvm/Support/BinaryStreamRef.h"
 #include "llvm/Support/Endian.h"
@@ -50,6 +51,13 @@ public:
   Optional<uint32_t> Hash;
 };
 
+template <typename Kind> struct RemappedRecord {
+  explicit RemappedRecord(const CVRecord<Kind> &R) : OriginalRecord(R) {}
+
+  CVRecord<Kind> OriginalRecord;
+  SmallVector<std::pair<uint32_t, TypeIndex>, 8> Mappings;
+};
+
 } // end namespace codeview
 
 template <typename Kind>

include/llvm/DebugInfo/CodeView/CVTypeVisitor.h

@@ -46,6 +46,7 @@ Error visitMemberRecordStream(ArrayRef<uint8_t> FieldList,
                               TypeVisitorCallbacks &Callbacks);
 
 Error visitTypeStream(const CVTypeArray &Types, TypeVisitorCallbacks &Callbacks,
+                      VisitorDataSource Source = VDS_BytesPresent,
                       TypeServerHandler *TS = nullptr);
 Error visitTypeStream(CVTypeRange Types, TypeVisitorCallbacks &Callbacks,
                       TypeServerHandler *TS = nullptr);

include/llvm/DebugInfo/CodeView/TypeDeserializer.h

@@ -40,6 +40,17 @@ class TypeDeserializer : public TypeVisitorCallbacks {
 public:
   TypeDeserializer() = default;
 
+  template <typename T> static Error deserializeAs(CVType &CVT, T &Record) {
+    MappingInfo I(CVT.content());
+    if (auto EC = I.Mapping.visitTypeBegin(CVT))
+      return EC;
+    if (auto EC = I.Mapping.visitKnownRecord(CVT, Record))
+      return EC;
+    if (auto EC = I.Mapping.visitTypeEnd(CVT))
+      return EC;
+    return Error::success();
+  }
+
   Error visitTypeBegin(CVType &Record) override {
     assert(!Mapping && "Already in a type mapping!");
     Mapping = llvm::make_unique<MappingInfo>(Record.content());

include/llvm/DebugInfo/CodeView/TypeIndexDiscovery.h

@@ -0,0 +1,33 @@
+//===- TypeIndexDiscovery.h -------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_DEBUGINFO_CODEVIEW_TYPEINDEXDISCOVERY_H
+#define LLVM_DEBUGINFO_CODEVIEW_TYPEINDEXDISCOVERY_H
+
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/DebugInfo/CodeView/TypeRecord.h"
+#include "llvm/Support/Error.h"
+
+namespace llvm {
+namespace codeview {
+enum class TiRefKind { TypeRef, IndexRef };
+struct TiReference {
+  TiRefKind Kind;
+  uint32_t Offset;
+  uint32_t Count;
+};
+
+void discoverTypeIndices(ArrayRef<uint8_t> RecordData,
+                         SmallVectorImpl<TiReference> &Refs);
+void discoverTypeIndices(const CVType &Type,
+                         SmallVectorImpl<TiReference> &Refs);
+}
+}
+
+#endif

include/llvm/DebugInfo/CodeView/TypeRecord.h

@@ -35,6 +35,7 @@ using support::ulittle16_t;
 using support::ulittle32_t;
 
 typedef CVRecord<TypeLeafKind> CVType;
+typedef RemappedRecord<TypeLeafKind> RemappedType;
 
 struct CVMemberRecord {
   TypeLeafKind Kind;
@@ -278,15 +279,9 @@ public:
         Attrs(calcAttrs(PK, PM, PO, Size)) {}
 
   PointerRecord(TypeIndex ReferentType, PointerKind PK, PointerMode PM,
-                PointerOptions PO, uint8_t Size,
+                PointerOptions PO, uint8_t Size, const MemberPointerInfo &MPI)
-                const MemberPointerInfo &Member)
       : TypeRecord(TypeRecordKind::Pointer), ReferentType(ReferentType),
-        Attrs(calcAttrs(PK, PM, PO, Size)), MemberInfo(Member) {}
+        Attrs(calcAttrs(PK, PM, PO, Size)), MemberInfo(MPI) {}
-
-  PointerRecord(TypeIndex ReferentType, uint32_t Attrs,
-                const MemberPointerInfo &Member)
-      : TypeRecord(TypeRecordKind::Pointer), ReferentType(ReferentType),
-        Attrs(Attrs), MemberInfo(Member) {}
 
   TypeIndex getReferentType() const { return ReferentType; }
 

include/llvm/DebugInfo/CodeView/TypeSerializer.h

@@ -17,7 +17,6 @@
 
 #include "llvm/ADT/Optional.h"
 #include "llvm/ADT/SmallVector.h"
-#include "llvm/ADT/StringMap.h"
 #include "llvm/ADT/StringRef.h"
 #include "llvm/Support/Allocator.h"
 #include "llvm/Support/Error.h"
@@ -26,6 +25,8 @@ namespace llvm {
 
 namespace codeview {
 
+class TypeHasher;
+
 class TypeSerializer : public TypeVisitorCallbacks {
   struct SubRecord {
     SubRecord(TypeLeafKind K, uint32_t S) : Kind(K), Size(S) {}
@@ -45,14 +46,13 @@ class TypeSerializer : public TypeVisitorCallbacks {
     }
   };
 
-  typedef SmallVector<MutableArrayRef<uint8_t>, 2> RecordList;
+  typedef SmallVector<MutableArrayRef<uint8_t>, 2> MutableRecordList;
 
   static constexpr uint8_t ContinuationLength = 8;
   BumpPtrAllocator &RecordStorage;
   RecordSegment CurrentSegment;
-  RecordList FieldListSegments;
+  MutableRecordList FieldListSegments;
 
-  TypeIndex LastTypeIndex;
   Optional<TypeLeafKind> TypeKind;
   Optional<TypeLeafKind> MemberKind;
   std::vector<uint8_t> RecordBuffer;
@@ -60,28 +60,35 @@ class TypeSerializer : public TypeVisitorCallbacks {
   BinaryStreamWriter Writer;
   TypeRecordMapping Mapping;
 
-  RecordList SeenRecords;
+  /// Private type record hashing implementation details are handled here.
-  StringMap<TypeIndex> HashedRecords;
+  std::unique_ptr<TypeHasher> Hasher;
+
+  /// Contains a list of all records indexed by TypeIndex.toArrayIndex().
+  SmallVector<ArrayRef<uint8_t>, 2> SeenRecords;
+
+  /// Temporary storage that we use to copy a record's data while re-writing
+  /// its type indices.
+  SmallVector<uint8_t, 256> RemapStorage;
+
+  TypeIndex nextTypeIndex() const;
 
   bool isInFieldList() const;
-  TypeIndex calcNextTypeIndex() const;
-  TypeIndex incrementTypeIndex();
   MutableArrayRef<uint8_t> getCurrentSubRecordData();
   MutableArrayRef<uint8_t> getCurrentRecordData();
   Error writeRecordPrefix(TypeLeafKind Kind);
-  TypeIndex insertRecordBytesPrivate(MutableArrayRef<uint8_t> Record);
-  TypeIndex insertRecordBytesWithCopy(CVType &Record,
-                                      MutableArrayRef<uint8_t> Data);
 
   Expected<MutableArrayRef<uint8_t>>
   addPadding(MutableArrayRef<uint8_t> Record);
 
 public:
-  explicit TypeSerializer(BumpPtrAllocator &Storage);
+  explicit TypeSerializer(BumpPtrAllocator &Storage, bool Hash = true);
+  ~TypeSerializer();
 
-  ArrayRef<MutableArrayRef<uint8_t>> records() const;
+  void reset();
-  TypeIndex getLastTypeIndex() const;
+
-  TypeIndex insertRecordBytes(MutableArrayRef<uint8_t> Record);
+  ArrayRef<ArrayRef<uint8_t>> records() const;
+  TypeIndex insertRecordBytes(ArrayRef<uint8_t> &Record);
+  TypeIndex insertRecord(const RemappedType &Record);
   Expected<TypeIndex> visitTypeEndGetIndex(CVType &Record);
 
   Error visitTypeBegin(CVType &Record) override;

include/llvm/DebugInfo/CodeView/TypeStreamMerger.h

@@ -22,12 +22,75 @@ class TypeIndex;
 class TypeServerHandler;
 class TypeTableBuilder;
 
-/// Merges one type stream into another. Returns true on success.
+/// \brief Merge one set of type records into another.  This method assumes
-Error mergeTypeStreams(TypeTableBuilder &DestIdStream,
+/// that all records are type records, and there are no Id records present.
-                       TypeTableBuilder &DestTypeStream,
+///
+/// \param Dest The table to store the re-written type records into.
+///
+/// \param SourceToDest A vector, indexed by the TypeIndex in the source
+/// type stream, that contains the index of the corresponding type record
+/// in the destination stream.
+///
+/// \param Handler (optional) If non-null, an interface that gets invoked
+/// to handle type server records.
+///
+/// \param Types The collection of types to merge in.
+///
+/// \returns Error::success() if the operation succeeded, otherwise an
+/// appropriate error code.
+Error mergeTypeRecords(TypeTableBuilder &Dest,
                        SmallVectorImpl<TypeIndex> &SourceToDest,
                        TypeServerHandler *Handler, const CVTypeArray &Types);
 
+/// \brief Merge one set of id records into another.  This method assumes
+/// that all records are id records, and there are no Type records present.
+/// However, since Id records can refer back to Type records, this method
+/// assumes that the referenced type records have also been merged into
+/// another type stream (for example using the above method), and accepts
+/// the mapping from source to dest for that stream so that it can re-write
+/// the type record mappings accordingly.
+///
+/// \param Dest The table to store the re-written id records into.
+///
+/// \param Types The mapping to use for the type records that these id
+/// records refer to.
+///
+/// \param SourceToDest A vector, indexed by the TypeIndex in the source
+/// id stream, that contains the index of the corresponding id record
+/// in the destination stream.
+///
+/// \param Ids The collection of id records to merge in.
+///
+/// \returns Error::success() if the operation succeeded, otherwise an
+/// appropriate error code.
+Error mergeIdRecords(TypeTableBuilder &Dest, ArrayRef<TypeIndex> Types,
+                     SmallVectorImpl<TypeIndex> &SourceToDest,
+  const CVTypeArray &Ids);
+
+/// \brief Merge a unified set of type and id records, splitting them into
+/// separate output streams.
+///
+/// \param DestIds The table to store the re-written id records into.
+///
+/// \param DestTypes the table to store the re-written type records into.
+///
+/// \param SourceToDest A vector, indexed by the TypeIndex in the source
+/// id stream, that contains the index of the corresponding id record
+/// in the destination stream.
+///
+/// \param Handler (optional) If non-null, an interface that gets invoked
+/// to handle type server records.
+///
+/// \param IdsAndTypes The collection of id records to merge in.
+///
+/// \returns Error::success() if the operation succeeded, otherwise an
+/// appropriate error code.
+Error mergeTypeAndIdRecords(TypeTableBuilder &DestIds,
+                            TypeTableBuilder &DestTypes,
+                            SmallVectorImpl<TypeIndex> &SourceToDest,
+                            TypeServerHandler *Handler,
+  const CVTypeArray &IdsAndTypes);
+
 } // end namespace codeview
 } // end namespace llvm
 

include/llvm/DebugInfo/CodeView/TypeTableBuilder.h

@@ -64,10 +64,14 @@ public:
     return *ExpectedIndex;
   }
 
-  TypeIndex writeSerializedRecord(MutableArrayRef<uint8_t> Record) {
+  TypeIndex writeSerializedRecord(ArrayRef<uint8_t> Record) {
     return Serializer.insertRecordBytes(Record);
   }
 
+  TypeIndex writeSerializedRecord(const RemappedType &Record) {
+    return Serializer.insertRecord(Record);
+  }
+
   template <typename TFunc> void ForEachRecord(TFunc Func) {
     uint32_t Index = TypeIndex::FirstNonSimpleIndex;
 
@@ -77,23 +81,24 @@ public:
     }
   }
 
-  ArrayRef<MutableArrayRef<uint8_t>> records() const {
+  ArrayRef<ArrayRef<uint8_t>> records() const { return Serializer.records(); }
-    return Serializer.records();
-  }
 };
 
 class FieldListRecordBuilder {
   TypeTableBuilder &TypeTable;
+  BumpPtrAllocator Allocator;
   TypeSerializer TempSerializer;
   CVType Type;
 
 public:
   explicit FieldListRecordBuilder(TypeTableBuilder &TypeTable)
-      : TypeTable(TypeTable), TempSerializer(TypeTable.getAllocator()) {
+      : TypeTable(TypeTable), TempSerializer(Allocator, false) {
     Type.Type = TypeLeafKind::LF_FIELDLIST;
   }
 
   void begin() {
+    TempSerializer.reset();
+
     if (auto EC = TempSerializer.visitTypeBegin(Type))
       consumeError(std::move(EC));
   }
@@ -109,23 +114,19 @@ public:
       consumeError(std::move(EC));
   }
 
-  TypeIndex end() {
+  TypeIndex end(bool Write) {
+    TypeIndex Index;
     if (auto EC = TempSerializer.visitTypeEnd(Type)) {
       consumeError(std::move(EC));
       return TypeIndex();
     }
 
-    TypeIndex Index;
+    if (Write) {
-    for (auto Record : TempSerializer.records()) {
+      for (auto Record : TempSerializer.records())
-      Index = TypeTable.writeSerializedRecord(Record);
+        Index = TypeTable.writeSerializedRecord(Record);
     }
-    return Index;
-  }
 
-  /// Stop building the record.
+    return Index;
-  void reset() {
-    if (auto EC = TempSerializer.visitTypeEnd(Type))
-      consumeError(std::move(EC));
   }
 };
 

include/llvm/DebugInfo/CodeView/TypeTableCollection.h

@@ -18,7 +18,7 @@ namespace codeview {
 
 class TypeTableCollection : public TypeCollection {
 public:
-  explicit TypeTableCollection(ArrayRef<MutableArrayRef<uint8_t>> Records);
+  explicit TypeTableCollection(ArrayRef<ArrayRef<uint8_t>> Records);
 
   Optional<TypeIndex> getFirst() override;
   Optional<TypeIndex> getNext(TypeIndex Prev) override;
@@ -33,7 +33,7 @@ private:
   bool hasCapacityFor(TypeIndex Index) const;
   void ensureTypeExists(TypeIndex Index);
 
-  ArrayRef<MutableArrayRef<uint8_t>> Records;
+  ArrayRef<ArrayRef<uint8_t>> Records;
   TypeDatabase Database;
 };
 }

include/llvm/DebugInfo/DWARF/DWARFContext.h

@@ -46,7 +46,8 @@ class raw_ostream;
 /// Reads a value from data extractor and applies a relocation to the result if
 /// one exists for the given offset.
 uint64_t getRelocatedValue(const DataExtractor &Data, uint32_t Size,
-                           uint32_t *Off, const RelocAddrMap *Relocs);
+                           uint32_t *Off, const RelocAddrMap *Relocs,
+                           uint64_t *SecNdx = nullptr);
 
 /// DWARFContext
 /// This data structure is the top level entity that deals with dwarf debug
@@ -71,6 +72,14 @@ class DWARFContext : public DIContext {
   std::unique_ptr<DWARFDebugAbbrev> AbbrevDWO;
   std::unique_ptr<DWARFDebugLocDWO> LocDWO;
 
+  struct DWOFile {
+    object::OwningBinary<object::ObjectFile> File;
+    std::unique_ptr<DWARFContext> Context;
+  };
+  StringMap<std::weak_ptr<DWOFile>> DWOFiles;
+  std::weak_ptr<DWOFile> DWP;
+  bool CheckedForDWP = false;
+
   /// Read compile units from the debug_info section (if necessary)
   /// and store them in CUs.
   void parseCompileUnits();
@@ -165,6 +174,8 @@ public:
     return DWOCUs[index].get();
   }
 
+  DWARFCompileUnit *getDWOCompileUnitForHash(uint64_t Hash);
+
   /// Get a DIE given an exact offset.
   DWARFDie getDIEForOffset(uint32_t Offset);
 
@@ -206,6 +217,7 @@ public:
   DIInliningInfo getInliningInfoForAddress(uint64_t Address,
       DILineInfoSpecifier Specifier = DILineInfoSpecifier()) override;
 
+  virtual StringRef getFileName() const = 0;
   virtual bool isLittleEndian() const = 0;
   virtual uint8_t getAddressSize() const = 0;
   virtual const DWARFSection &getInfoSection() = 0;
@@ -248,6 +260,8 @@ public:
     return version == 2 || version == 3 || version == 4 || version == 5;
   }
 
+  std::shared_ptr<DWARFContext> getDWOContext(StringRef AbsolutePath);
+
 private:
   /// Return the compile unit that includes an offset (relative to .debug_info).
   DWARFCompileUnit *getCompileUnitForOffset(uint32_t Offset);
@@ -263,6 +277,7 @@ private:
 class DWARFContextInMemory : public DWARFContext {
   virtual void anchor();
 
+  StringRef FileName;
   bool IsLittleEndian;
   uint8_t AddressSize;
   DWARFSection InfoSection;
@@ -316,6 +331,7 @@ public:
                        uint8_t AddrSize,
                        bool isLittleEndian = sys::IsLittleEndianHost);
 
+  StringRef getFileName() const override { return FileName; }
   bool isLittleEndian() const override { return IsLittleEndian; }
   uint8_t getAddressSize() const override { return AddressSize; }
   const DWARFSection &getInfoSection() override { return InfoSection; }

include/llvm/DebugInfo/DWARF/DWARFDebugRangeList.h

@@ -25,6 +25,7 @@ class raw_ostream;
 struct DWARFAddressRange {
   uint64_t LowPC;
   uint64_t HighPC;
+  uint64_t SectionIndex;
 };
 
 /// DWARFAddressRangesVector - represents a set of absolute address ranges.
@@ -44,6 +45,8 @@ public:
     /// address past the end of the address range. The ending address must
     /// be greater than or equal to the beginning address.
     uint64_t EndAddress;
+    /// A section index this range belongs to.
+    uint64_t SectionIndex;
 
     /// The end of any given range list is marked by an end of list entry,
     /// which consists of a 0 for the beginning address offset

include/llvm/DebugInfo/DWARF/DWARFDie.h

@@ -195,7 +195,8 @@ public:
 
   /// Retrieves DW_AT_low_pc and DW_AT_high_pc from CU.
   /// Returns true if both attributes are present.
-  bool getLowAndHighPC(uint64_t &LowPC, uint64_t &HighPC) const;
+  bool getLowAndHighPC(uint64_t &LowPC, uint64_t &HighPC,
+                       uint64_t &SectionIndex) const;
 
   /// Get the address ranges for this DIE.
   ///

include/llvm/DebugInfo/DWARF/DWARFFormValue.h

@@ -47,6 +47,7 @@ private:
       const char *cstr;
     };
     const uint8_t *data = nullptr;
+    uint64_t SectionIndex;      /// Section index for reference forms.
   };
 
   dwarf::Form Form;             /// Form for this value.
@@ -58,6 +59,7 @@ public:
 
   dwarf::Form getForm() const { return Form; }
   uint64_t getRawUValue() const { return Value.uval; }
+  uint64_t getSectionIndex() const { return Value.SectionIndex; }
   void setForm(dwarf::Form F) { Form = F; }
   void setUValue(uint64_t V) { Value.uval = V; }
   void setSValue(int64_t V) { Value.sval = V; }

include/llvm/DebugInfo/DWARF/DWARFRelocMap.h

@@ -16,7 +16,10 @@
 
 namespace llvm {
 
+/// RelocAddrEntry contains relocated value and section index.
+/// Section index is -1LL if relocation points to absolute symbol.
 struct RelocAddrEntry {
+  uint64_t SectionIndex;
   uint64_t Value;
 };
 

include/llvm/DebugInfo/DWARF/DWARFUnit.h

@@ -143,17 +143,7 @@ class DWARFUnit {
   typedef iterator_range<std::vector<DWARFDebugInfoEntry>::iterator>
       die_iterator_range;
 
-  class DWOHolder {
+  std::shared_ptr<DWARFUnit> DWO;
-    object::OwningBinary<object::ObjectFile> DWOFile;
-    std::unique_ptr<DWARFContext> DWOContext;
-    DWARFUnit *DWOU = nullptr;
-
-  public:
-    DWOHolder(StringRef DWOPath, uint64_t DWOId);
-
-    DWARFUnit *getUnit() const { return DWOU; }
-  };
-  std::unique_ptr<DWOHolder> DWO;
 
   const DWARFUnitIndex::Entry *IndexEntry;
 

include/llvm/DebugInfo/MSF/MappedBlockStream.h

@@ -43,8 +43,8 @@ class MappedBlockStream : public BinaryStream {
   friend class WritableMappedBlockStream;
 public:
   static std::unique_ptr<MappedBlockStream>
-  createStream(uint32_t BlockSize, uint32_t NumBlocks,
+  createStream(uint32_t BlockSize, const MSFStreamLayout &Layout,
-               const MSFStreamLayout &Layout, BinaryStreamRef MsfData);
+               BinaryStreamRef MsfData);
 
   static std::unique_ptr<MappedBlockStream>
   createIndexedStream(const MSFLayout &Layout, BinaryStreamRef MsfData,
@@ -74,12 +74,11 @@ public:
   void invalidateCache();
 
   uint32_t getBlockSize() const { return BlockSize; }
-  uint32_t getNumBlocks() const { return NumBlocks; }
+  uint32_t getNumBlocks() const { return StreamLayout.Blocks.size(); }
   uint32_t getStreamLength() const { return StreamLayout.Length; }
 
 protected:
-  MappedBlockStream(uint32_t BlockSize, uint32_t NumBlocks,
+  MappedBlockStream(uint32_t BlockSize, const MSFStreamLayout &StreamLayout,
-                    const MSFStreamLayout &StreamLayout,
                     BinaryStreamRef MsfData);
 
 private:
@@ -91,7 +90,6 @@ private:
                            ArrayRef<uint8_t> &Buffer);
 
   const uint32_t BlockSize;
-  const uint32_t NumBlocks;
   const MSFStreamLayout StreamLayout;
   BinaryStreamRef MsfData;
 
@@ -103,8 +101,8 @@ private:
 class WritableMappedBlockStream : public WritableBinaryStream {
 public:
   static std::unique_ptr<WritableMappedBlockStream>
-  createStream(uint32_t BlockSize, uint32_t NumBlocks,
+  createStream(uint32_t BlockSize, const MSFStreamLayout &Layout,
-               const MSFStreamLayout &Layout, WritableBinaryStreamRef MsfData);
+               WritableBinaryStreamRef MsfData);
 
   static std::unique_ptr<WritableMappedBlockStream>
   createIndexedStream(const MSFLayout &Layout, WritableBinaryStreamRef MsfData,
@@ -139,7 +137,7 @@ public:
   uint32_t getStreamLength() const { return ReadInterface.getStreamLength(); }
 
 protected:
-  WritableMappedBlockStream(uint32_t BlockSize, uint32_t NumBlocks,
+  WritableMappedBlockStream(uint32_t BlockSize,
                             const MSFStreamLayout &StreamLayout,
                             WritableBinaryStreamRef MsfData);
 

include/llvm/DebugInfo/PDB/Native/DbiStreamBuilder.h

@@ -82,6 +82,7 @@ private:
 
   Error finalize();
   uint32_t calculateModiSubstreamSize() const;
+  uint32_t calculateNamesOffset() const;
   uint32_t calculateSectionContribsStreamSize() const;
   uint32_t calculateSectionMapStreamSize() const;
   uint32_t calculateFileInfoSubstreamSize() const;

include/llvm/DebugInfo/PDB/Native/PDBTypeServerHandler.h

@@ -11,8 +11,7 @@
 #define LLVM_DEBUGINFO_PDB_PDBTYPESERVERHANDLER_H
 
 #include "llvm/ADT/SmallString.h"
-#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/StringSet.h"
-#include "llvm/ADT/StringMap.h"
 #include "llvm/DebugInfo/CodeView/TypeRecord.h"
 #include "llvm/DebugInfo/CodeView/TypeServerHandler.h"
 #include "llvm/DebugInfo/PDB/Native/NativeSession.h"
@@ -39,7 +38,7 @@ private:
 
   bool RevisitAlways;
   std::unique_ptr<NativeSession> Session;
-  SmallVector<SmallString<64>, 4> SearchPaths;
+  StringSet<> SearchPaths;
 };
 }
 }

include/llvm/DebugInfo/PDB/Native/TpiStream.h

@@ -21,6 +21,9 @@
 #include "llvm/Support/Error.h"
 
 namespace llvm {
+namespace codeview {
+class LazyRandomTypeCollection;
+}
 namespace msf {
 class MappedBlockStream;
 }
@@ -53,12 +56,16 @@ public:
   codeview::CVTypeRange types(bool *HadError) const;
   const codeview::CVTypeArray &typeArray() const { return TypeRecords; }
 
+  codeview::LazyRandomTypeCollection &typeCollection() { return *Types; }
+
   Error commit();
 
 private:
   const PDBFile &Pdb;
   std::unique_ptr<msf::MappedBlockStream> Stream;
 
+  std::unique_ptr<codeview::LazyRandomTypeCollection> Types;
+
   codeview::CVTypeArray TypeRecords;
 
   std::unique_ptr<BinaryStream> HashStream;

include/llvm/IR/Attributes.h

@@ -322,7 +322,7 @@ template <> struct DenseMapInfo<AttributeSet> {
 /// the AttributeList object. The function attributes are at index
 /// `AttributeList::FunctionIndex', the return value is at index
 /// `AttributeList::ReturnIndex', and the attributes for the parameters start at
-/// index `1'.
+/// index `AttributeList::FirstArgIndex'.
 class AttributeList {
 public:
   enum AttrIndex : unsigned {
@@ -347,8 +347,8 @@ public:
   /// \brief Create an AttributeList with the specified parameters in it.
   static AttributeList get(LLVMContext &C,
                            ArrayRef<std::pair<unsigned, Attribute>> Attrs);
-  static AttributeList
+  static AttributeList get(LLVMContext &C,
-  get(LLVMContext &C, ArrayRef<std::pair<unsigned, AttributeSet>> Attrs);
+                           ArrayRef<std::pair<unsigned, AttributeSet>> Attrs);
 
   /// \brief Create an AttributeList from attribute sets for a function, its
   /// return value, and all of its arguments.
@@ -356,13 +356,11 @@ public:
                            AttributeSet RetAttrs,
                            ArrayRef<AttributeSet> ArgAttrs);
 
-  static AttributeList
-  getImpl(LLVMContext &C,
-          ArrayRef<std::pair<unsigned, AttributeSet>> Attrs);
-
 private:
   explicit AttributeList(AttributeListImpl *LI) : pImpl(LI) {}
 
+  static AttributeList getImpl(LLVMContext &C, ArrayRef<AttributeSet> AttrSets);
+
 public:
   AttributeList() = default;
 
@@ -521,39 +519,31 @@ public:
   /// \brief Return the attributes at the index as a string.
   std::string getAsString(unsigned Index, bool InAttrGrp = false) const;
 
-  using iterator = ArrayRef<Attribute>::iterator;
+  //===--------------------------------------------------------------------===//
+  // AttributeList Introspection
+  //===--------------------------------------------------------------------===//
 
-  iterator begin(unsigned Slot) const;
+  typedef const AttributeSet *iterator;
-  iterator end(unsigned Slot) const;
+  iterator begin() const;
+  iterator end() const;
+
+  unsigned getNumAttrSets() const;
+
+  /// Use these to iterate over the valid attribute indices.
+  unsigned index_begin() const { return AttributeList::FunctionIndex; }
+  unsigned index_end() const { return getNumAttrSets() - 1; }
 
   /// operator==/!= - Provide equality predicates.
   bool operator==(const AttributeList &RHS) const { return pImpl == RHS.pImpl; }
   bool operator!=(const AttributeList &RHS) const { return pImpl != RHS.pImpl; }
 
-  //===--------------------------------------------------------------------===//
-  // AttributeList Introspection
-  //===--------------------------------------------------------------------===//
-
   /// \brief Return a raw pointer that uniquely identifies this attribute list.
   void *getRawPointer() const {
     return pImpl;
   }
 
   /// \brief Return true if there are no attributes.
-  bool isEmpty() const {
+  bool isEmpty() const { return pImpl == nullptr; }
-    return getNumSlots() == 0;
-  }
-
-  /// \brief Return the number of slots used in this attribute list.  This is
-  /// the number of arguments that have an attribute set on them (including the
-  /// function itself).
-  unsigned getNumSlots() const;
-
-  /// \brief Return the index for the given slot.
-  unsigned getSlotIndex(unsigned Slot) const;
-
-  /// \brief Return the attributes at the given slot.
-  AttributeSet getSlotAttributes(unsigned Slot) const;
 
   void dump() const;
 };

include/llvm/IR/BasicBlock.h

@@ -33,6 +33,7 @@ class Function;
 class LandingPadInst;
 class LLVMContext;
 class Module;
+class PHINode;
 class TerminatorInst;
 class ValueSymbolTable;
 
@@ -261,6 +262,50 @@ public:
   inline const Instruction       &back() const { return InstList.back();  }
   inline       Instruction       &back()       { return InstList.back();  }
 
+  /// Iterator to walk just the phi nodes in the basic block.
+  template <typename PHINodeT = PHINode, typename BBIteratorT = iterator>
+  class phi_iterator_impl
+      : public iterator_facade_base<phi_iterator_impl<PHINodeT, BBIteratorT>,
+                                    std::forward_iterator_tag, PHINodeT> {
+    friend BasicBlock;
+
+    PHINodeT *PN;
+
+    phi_iterator_impl(PHINodeT *PN) : PN(PN) {}
+
+  public:
+    // Allow default construction to build variables, but this doesn't build
+    // a useful iterator.
+    phi_iterator_impl() = default;
+
+    // Allow conversion between instantiations where valid.
+    template <typename PHINodeU, typename BBIteratorU>
+    phi_iterator_impl(const phi_iterator_impl<PHINodeU, BBIteratorU> &Arg)
+        : PN(Arg.PN) {}
+
+    bool operator==(const phi_iterator_impl &Arg) const { return PN == Arg.PN; }
+
+    PHINodeT &operator*() const { return *PN; }
+
+    using phi_iterator_impl::iterator_facade_base::operator++;
+    phi_iterator_impl &operator++() {
+      assert(PN && "Cannot increment the end iterator!");
+      PN = dyn_cast<PHINodeT>(std::next(BBIteratorT(PN)));
+      return *this;
+    }
+  };
+  typedef phi_iterator_impl<> phi_iterator;
+  typedef phi_iterator_impl<const PHINode, BasicBlock::const_iterator>
+      const_phi_iterator;
+
+  /// Returns a range that iterates over the phis in the basic block.
+  ///
+  /// Note that this cannot be used with basic blocks that have no terminator.
+  iterator_range<const_phi_iterator> phis() const {
+    return const_cast<BasicBlock *>(this)->phis();
+  }
+  iterator_range<phi_iterator> phis();
+
   /// \brief Return the underlying instruction list container.
   ///
   /// Currently you need to access the underlying instruction list container

include/llvm/IR/IntrinsicInst.h

@@ -171,6 +171,7 @@ namespace llvm {
       ebStrict
     };
 
+    bool isUnaryOp() const;
     RoundingMode getRoundingMode() const;
     ExceptionBehavior getExceptionBehavior() const;
 
@@ -182,6 +183,18 @@ namespace llvm {
       case Intrinsic::experimental_constrained_fmul:
       case Intrinsic::experimental_constrained_fdiv:
       case Intrinsic::experimental_constrained_frem:
+      case Intrinsic::experimental_constrained_sqrt:
+      case Intrinsic::experimental_constrained_pow:
+      case Intrinsic::experimental_constrained_powi:
+      case Intrinsic::experimental_constrained_sin:
+      case Intrinsic::experimental_constrained_cos:
+      case Intrinsic::experimental_constrained_exp:
+      case Intrinsic::experimental_constrained_exp2:
+      case Intrinsic::experimental_constrained_log:
+      case Intrinsic::experimental_constrained_log10:
+      case Intrinsic::experimental_constrained_log2:
+      case Intrinsic::experimental_constrained_rint:
+      case Intrinsic::experimental_constrained_nearbyint:
         return true;
       default: return false;
       }

include/llvm/IR/Intrinsics.td

@@ -489,8 +489,64 @@ let IntrProperties = [IntrInaccessibleMemOnly] in {
                                                       LLVMMatchType<0>,
                                                       llvm_metadata_ty,
                                                       llvm_metadata_ty ]>;
+
+  // These intrinsics are sensitive to the rounding mode so we need constrained
+  // versions of each of them.  When strict rounding and exception control are
+  // not required the non-constrained versions of these intrinsics should be
+  // used.
+  def int_experimental_constrained_sqrt : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_powi : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_i32_ty,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_sin  : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_cos  : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_pow  : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_log  : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_log10: Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_log2 : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_exp  : Intrinsic<[ llvm_anyfloat_ty ], 
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_exp2 : Intrinsic<[ llvm_anyfloat_ty ],
+                                                    [ LLVMMatchType<0>,
+                                                      llvm_metadata_ty,
+                                                      llvm_metadata_ty ]>;
+  def int_experimental_constrained_rint  : Intrinsic<[ llvm_anyfloat_ty ],
+                                                     [ LLVMMatchType<0>,
+                                                       llvm_metadata_ty,
+                                                       llvm_metadata_ty ]>;
+  def int_experimental_constrained_nearbyint : Intrinsic<[ llvm_anyfloat_ty ],
+                                                         [ LLVMMatchType<0>,
+                                                           llvm_metadata_ty,
+                                                           llvm_metadata_ty ]>;
 }
-// FIXME: Add intrinsic for fcmp, fptrunc, fpext, fptoui and fptosi.
+// FIXME: Add intrinsics for fcmp, fptrunc, fpext, fptoui and fptosi.
+// FIXME: Add intrinsics for fabs, copysign, floor, ceil, trunc and round?
 
 
 //===------------------------- Expect Intrinsics --------------------------===//

include/llvm/IR/IntrinsicsAMDGPU.td

@@ -566,6 +566,16 @@ def int_amdgcn_s_getreg :
   [IntrReadMem, IntrSpeculatable]
 >;
 
+// int_amdgcn_s_getpc is provided to allow a specific style of position
+// independent code to determine the high part of its address when it is
+// known (through convention) that the code and any data of interest does
+// not cross a 4Gb address boundary. Use for any other purpose may not
+// produce the desired results as optimizations may cause code movement,
+// especially as we explicitly use IntrNoMem to allow optimizations.
+def int_amdgcn_s_getpc :
+  GCCBuiltin<"__builtin_amdgcn_s_getpc">,
+  Intrinsic<[llvm_i64_ty], [], [IntrNoMem, IntrSpeculatable]>;
+
 // __builtin_amdgcn_interp_mov <param>, <attr_chan>, <attr>, <m0>
 // param values: 0 = P10, 1 = P20, 2 = P0
 def int_amdgcn_interp_mov :

include/llvm/IR/Metadata.h

@@ -1223,6 +1223,7 @@ public:
 
   // FIXME: Fix callers and remove condition on N.
   unsigned size() const { return N ? N->getNumOperands() : 0u; }
+  bool empty() const { return N ? N->getNumOperands() == 0 : true; }
   T *operator[](unsigned I) const { return cast_or_null<T>(N->getOperand(I)); }
 
   // FIXME: Fix callers and remove condition on N.

include/llvm/IR/Module.h

@@ -139,9 +139,12 @@ public:
     /// during the append operation.
     AppendUnique = 6,
 
+    /// Takes the max of the two values, which are required to be integers.
+    Max = 7,
+
     // Markers:
     ModFlagBehaviorFirstVal = Error,
-    ModFlagBehaviorLastVal = AppendUnique
+    ModFlagBehaviorLastVal = Max
   };
 
   /// Checks if Metadata represents a valid ModFlagBehavior, and stores the

include/llvm/InitializePasses.h

@@ -144,6 +144,7 @@ void initializeGCMachineCodeAnalysisPass(PassRegistry&);
 void initializeGCModuleInfoPass(PassRegistry&);
 void initializeGCOVProfilerLegacyPassPass(PassRegistry&);
 void initializeGVNHoistLegacyPassPass(PassRegistry&);
+void initializeGVNSinkLegacyPassPass(PassRegistry&);
 void initializeGVNLegacyPassPass(PassRegistry&);
 void initializeGlobalDCELegacyPassPass(PassRegistry&);
 void initializeGlobalMergePass(PassRegistry&);
@@ -193,6 +194,7 @@ void initializeLiveVariablesPass(PassRegistry&);
 void initializeLoadCombinePass(PassRegistry&);
 void initializeLoadStoreVectorizerPass(PassRegistry&);
 void initializeLoaderPassPass(PassRegistry&);
+void initializeLocalizerPass(PassRegistry&);
 void initializeLocalStackSlotPassPass(PassRegistry&);
 void initializeLoopAccessLegacyAnalysisPass(PassRegistry&);
 void initializeLoopDataPrefetchLegacyPassPass(PassRegistry&);

include/llvm/LTO/Config.h

@@ -39,7 +39,7 @@ struct Config {
   std::string CPU;
   TargetOptions Options;
   std::vector<std::string> MAttrs;
-  Reloc::Model RelocModel = Reloc::PIC_;
+  Optional<Reloc::Model> RelocModel = Reloc::PIC_;
   CodeModel::Model CodeModel = CodeModel::Default;
   CodeGenOpt::Level CGOptLevel = CodeGenOpt::Default;
   TargetMachine::CodeGenFileType CGFileType = TargetMachine::CGFT_ObjectFile;

include/llvm/Object/Binary.h

@@ -95,9 +95,7 @@ public:
     return TypeID > ID_StartObjects && TypeID < ID_EndObjects;
   }
 
-  bool isSymbolic() const {
+  bool isSymbolic() const { return isIR() || isObject() || isCOFFImportFile(); }
-    return isIR() || isObject();
-  }
 
   bool isArchive() const {
     return TypeID == ID_Archive;

include/llvm/Object/COFF.h

@@ -782,6 +782,7 @@ protected:
   std::error_code getSectionName(DataRefImpl Sec,
                                  StringRef &Res) const override;
   uint64_t getSectionAddress(DataRefImpl Sec) const override;
+  uint64_t getSectionIndex(DataRefImpl Sec) const override;
   uint64_t getSectionSize(DataRefImpl Sec) const override;
   std::error_code getSectionContents(DataRefImpl Sec,
                                      StringRef &Res) const override;

include/llvm/Object/ELFObjectFile.h

@@ -235,6 +235,7 @@ protected:
   std::error_code getSectionName(DataRefImpl Sec,
                                  StringRef &Res) const override;
   uint64_t getSectionAddress(DataRefImpl Sec) const override;
+  uint64_t getSectionIndex(DataRefImpl Sec) const override;
   uint64_t getSectionSize(DataRefImpl Sec) const override;
   std::error_code getSectionContents(DataRefImpl Sec,
                                      StringRef &Res) const override;
@@ -645,6 +646,17 @@ uint64_t ELFObjectFile<ELFT>::getSectionAddress(DataRefImpl Sec) const {
   return getSection(Sec)->sh_addr;
 }
 
+template <class ELFT>
+uint64_t ELFObjectFile<ELFT>::getSectionIndex(DataRefImpl Sec) const {
+  auto SectionsOrErr = EF.sections();
+  handleAllErrors(std::move(SectionsOrErr.takeError()),
+                  [](const ErrorInfoBase &) {
+                    llvm_unreachable("unable to get section index");
+                  });
+  const Elf_Shdr *First = SectionsOrErr->begin();
+  return getSection(Sec) - First;
+}
+
 template <class ELFT>
 uint64_t ELFObjectFile<ELFT>::getSectionSize(DataRefImpl Sec) const {
   return getSection(Sec)->sh_size;

include/llvm/Object/MachO.h

@@ -290,6 +290,7 @@ public:
   std::error_code getSectionName(DataRefImpl Sec,
                                  StringRef &Res) const override;
   uint64_t getSectionAddress(DataRefImpl Sec) const override;
+  uint64_t getSectionIndex(DataRefImpl Sec) const override;
   uint64_t getSectionSize(DataRefImpl Sec) const override;
   std::error_code getSectionContents(DataRefImpl Sec,
                                      StringRef &Res) const override;

include/llvm/Object/ObjectFile.h

@@ -95,6 +95,7 @@ public:
 
   std::error_code getName(StringRef &Result) const;
   uint64_t getAddress() const;
+  uint64_t getIndex() const;
   uint64_t getSize() const;
   std::error_code getContents(StringRef &Result) const;
 
@@ -222,6 +223,7 @@ protected:
   virtual std::error_code getSectionName(DataRefImpl Sec,
                                          StringRef &Res) const = 0;
   virtual uint64_t getSectionAddress(DataRefImpl Sec) const = 0;
+  virtual uint64_t getSectionIndex(DataRefImpl Sec) const = 0;
   virtual uint64_t getSectionSize(DataRefImpl Sec) const = 0;
   virtual std::error_code getSectionContents(DataRefImpl Sec,
                                              StringRef &Res) const = 0;
@@ -393,6 +395,10 @@ inline uint64_t SectionRef::getAddress() const {
   return OwningObject->getSectionAddress(SectionPimpl);
 }
 
+inline uint64_t SectionRef::getIndex() const {
+  return OwningObject->getSectionIndex(SectionPimpl);
+}
+
 inline uint64_t SectionRef::getSize() const {
   return OwningObject->getSectionSize(SectionPimpl);
 }

include/llvm/Object/RelocVisitor.h

@@ -40,13 +40,13 @@ public:
   // TODO: Should handle multiple applied relocations via either passing in the
   // previously computed value or just count paired relocations as a single
   // visit.
-  uint64_t visit(uint32_t RelocType, RelocationRef R, uint64_t Value = 0) {
+  uint64_t visit(uint32_t Rel, RelocationRef R, uint64_t Value = 0) {
     if (isa<ELFObjectFileBase>(ObjToVisit))
-      return visitELF(RelocType, R, Value);
+      return visitELF(Rel, R, Value);
     if (isa<COFFObjectFile>(ObjToVisit))
-      return visitCOFF(RelocType, R, Value);
+      return visitCOFF(Rel, R, Value);
     if (isa<MachOObjectFile>(ObjToVisit))
-      return visitMachO(RelocType, R, Value);
+      return visitMachO(Rel, R, Value);
 
     HasError = true;
     return 0;
@@ -58,214 +58,60 @@ private:
   const ObjectFile &ObjToVisit;
   bool HasError = false;
 
-  uint64_t visitELF(uint32_t RelocType, RelocationRef R, uint64_t Value) {
+  uint64_t visitELF(uint32_t Rel, RelocationRef R, uint64_t Value) {
     if (ObjToVisit.getBytesInAddress() == 8) { // 64-bit object file
       switch (ObjToVisit.getArch()) {
       case Triple::x86_64:
-        switch (RelocType) {
+        return visitX86_64(Rel, R, Value);
-        case ELF::R_X86_64_NONE:
-          return visitELF_X86_64_NONE(R);
-        case ELF::R_X86_64_64:
-          return visitELF_X86_64_64(R, Value);
-        case ELF::R_X86_64_PC32:
-          return visitELF_X86_64_PC32(R, Value);
-        case ELF::R_X86_64_32:
-          return visitELF_X86_64_32(R, Value);
-        case ELF::R_X86_64_32S:
-          return visitELF_X86_64_32S(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::aarch64:
       case Triple::aarch64_be:
-        switch (RelocType) {
+        return visitAarch64(Rel, R, Value);
-        case ELF::R_AARCH64_ABS32:
-          return visitELF_AARCH64_ABS32(R, Value);
-        case ELF::R_AARCH64_ABS64:
-          return visitELF_AARCH64_ABS64(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::bpfel:
       case Triple::bpfeb:
-        switch (RelocType) {
+        return visitBpf(Rel, R, Value);
-        case ELF::R_BPF_64_64:
-          return visitELF_BPF_64_64(R, Value);
-        case ELF::R_BPF_64_32:
-          return visitELF_BPF_64_32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::mips64el:
       case Triple::mips64:
-        switch (RelocType) {
+        return visitMips64(Rel, R, Value);
-        case ELF::R_MIPS_32:
-          return visitELF_MIPS64_32(R, Value);
-        case ELF::R_MIPS_64:
-          return visitELF_MIPS64_64(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::ppc64le:
       case Triple::ppc64:
-        switch (RelocType) {
+        return visitPPC64(Rel, R, Value);
-        case ELF::R_PPC64_ADDR32:
-          return visitELF_PPC64_ADDR32(R, Value);
-        case ELF::R_PPC64_ADDR64:
-          return visitELF_PPC64_ADDR64(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::systemz:
-        switch (RelocType) {
+        return visitSystemz(Rel, R, Value);
-        case ELF::R_390_32:
-          return visitELF_390_32(R, Value);
-        case ELF::R_390_64:
-          return visitELF_390_64(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::sparcv9:
-        switch (RelocType) {
+        return visitSparc64(Rel, R, Value);
-        case ELF::R_SPARC_32:
-        case ELF::R_SPARC_UA32:
-          return visitELF_SPARCV9_32(R, Value);
-        case ELF::R_SPARC_64:
-        case ELF::R_SPARC_UA64:
-          return visitELF_SPARCV9_64(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       case Triple::amdgcn:
-        switch (RelocType) {
+        return visitAmdgpu(Rel, R, Value);
-        case ELF::R_AMDGPU_ABS32:
-          return visitELF_AMDGPU_ABS32(R, Value);
-        case ELF::R_AMDGPU_ABS64:
-          return visitELF_AMDGPU_ABS64(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
       default:
         HasError = true;
         return 0;
       }
-    } else if (ObjToVisit.getBytesInAddress() == 4) { // 32-bit object file
-      switch (ObjToVisit.getArch()) {
-      case Triple::x86:
-        switch (RelocType) {
-        case ELF::R_386_NONE:
-          return visitELF_386_NONE(R);
-        case ELF::R_386_32:
-          return visitELF_386_32(R, Value);
-        case ELF::R_386_PC32:
-          return visitELF_386_PC32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
-      case Triple::ppc:
-        switch (RelocType) {
-        case ELF::R_PPC_ADDR32:
-          return visitELF_PPC_ADDR32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
-      case Triple::arm:
-      case Triple::armeb:
-        switch (RelocType) {
-        default:
-          HasError = true;
-          return 0;
-        case ELF::R_ARM_ABS32:
-          return visitELF_ARM_ABS32(R, Value);
-        }
-      case Triple::lanai:
-        switch (RelocType) {
-        case ELF::R_LANAI_32:
-          return visitELF_Lanai_32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
-      case Triple::mipsel:
-      case Triple::mips:
-        switch (RelocType) {
-        case ELF::R_MIPS_32:
-          return visitELF_MIPS_32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
-      case Triple::sparc:
-        switch (RelocType) {
-        case ELF::R_SPARC_32:
-        case ELF::R_SPARC_UA32:
-          return visitELF_SPARC_32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
-      case Triple::hexagon:
-        switch (RelocType) {
-        case ELF::R_HEX_32:
-          return visitELF_HEX_32(R, Value);
-        default:
-          HasError = true;
-          return 0;
-        }
-      default:
-        HasError = true;
-        return 0;
-      }
-    } else {
-      report_fatal_error("Invalid word size in object file");
     }
-  }
 
-  uint64_t visitCOFF(uint32_t RelocType, RelocationRef R, uint64_t Value) {
+    // 32-bit object file
+    assert(ObjToVisit.getBytesInAddress() == 4 &&
+           "Invalid word size in object file");
+
     switch (ObjToVisit.getArch()) {
     case Triple::x86:
-      switch (RelocType) {
+      return visitX86(Rel, R, Value);
-      case COFF::IMAGE_REL_I386_SECREL:
+    case Triple::ppc:
-        return visitCOFF_I386_SECREL(R, Value);
+      return visitPPC32(Rel, R, Value);
-      case COFF::IMAGE_REL_I386_DIR32:
+    case Triple::arm:
-        return visitCOFF_I386_DIR32(R, Value);
+    case Triple::armeb:
-      }
+      return visitARM(Rel, R, Value);
-      break;
+    case Triple::lanai:
-    case Triple::x86_64:
+      return visitLanai(Rel, R, Value);
-      switch (RelocType) {
+    case Triple::mipsel:
-      case COFF::IMAGE_REL_AMD64_SECREL:
+    case Triple::mips:
-        return visitCOFF_AMD64_SECREL(R, Value);
+      return visitMips32(Rel, R, Value);
-      case COFF::IMAGE_REL_AMD64_ADDR64:
+    case Triple::sparc:
-        return visitCOFF_AMD64_ADDR64(R, Value);
+      return visitSparc32(Rel, R, Value);
-      }
+    case Triple::hexagon:
-      break;
+      return visitHexagon(Rel, R, Value);
+    default:
+      HasError = true;
+      return 0;
     }
-    HasError = true;
-    return 0;
-  }
-
-  uint64_t visitMachO(uint32_t RelocType, RelocationRef R, uint64_t Value) {
-    switch (ObjToVisit.getArch()) {
-    default: break;
-    case Triple::x86_64:
-      switch (RelocType) {
-        default: break;
-        case MachO::X86_64_RELOC_UNSIGNED:
-          return visitMACHO_X86_64_UNSIGNED(R, Value);
-      }
-    }
-    HasError = true;
-    return 0;
   }
 
   int64_t getELFAddend(RelocationRef R) {
@@ -275,176 +121,193 @@ private:
     return *AddendOrErr;
   }
 
-  /// Operations
+  uint64_t visitX86_64(uint32_t Rel, RelocationRef R, uint64_t Value) {
-
+    switch (Rel) {
-  /// 386-ELF
+    case ELF::R_X86_64_NONE:
-  uint64_t visitELF_386_NONE(RelocationRef R) {
+      return 0;
+    case ELF::R_X86_64_64:
+      return Value + getELFAddend(R);
+    case ELF::R_X86_64_PC32:
+      return Value + getELFAddend(R) - R.getOffset();
+    case ELF::R_X86_64_32:
+    case ELF::R_X86_64_32S:
+      return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    }
+    HasError = true;
     return 0;
   }
 
-  // Ideally the Addend here will be the addend in the data for
+  uint64_t visitAarch64(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  // the relocation. It's not actually the case for Rel relocations.
+    switch (Rel) {
-  uint64_t visitELF_386_32(RelocationRef R, uint64_t Value) {
+    case ELF::R_AARCH64_ABS32: {
-    return Value;
+      int64_t Res = Value + getELFAddend(R);
-  }
+      if (Res < INT32_MIN || Res > UINT32_MAX)
-
+        HasError = true;
-  uint64_t visitELF_386_PC32(RelocationRef R, uint64_t Value) {
+      return static_cast<uint32_t>(Res);
-    return Value - R.getOffset();
+    }
-  }
+    case ELF::R_AARCH64_ABS64:
-
+      return Value + getELFAddend(R);
-  /// X86-64 ELF
+    }
-  uint64_t visitELF_X86_64_NONE(RelocationRef R) {
+    HasError = true;
     return 0;
   }
 
-  uint64_t visitELF_X86_64_64(RelocationRef R, uint64_t Value) {
+  uint64_t visitBpf(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
+    switch (Rel) {
+    case ELF::R_BPF_64_32:
+      return Value & 0xFFFFFFFF;
+    case ELF::R_BPF_64_64:
+      return Value;
+    }
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_X86_64_PC32(RelocationRef R, uint64_t Value) {
+  uint64_t visitMips64(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R) - R.getOffset();
+    switch (Rel) {
+    case ELF::R_MIPS_32:
+      return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    case ELF::R_MIPS_64:
+      return Value + getELFAddend(R);
+    }
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_X86_64_32(RelocationRef R, uint64_t Value) {
+  uint64_t visitPPC64(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    switch (Rel) {
+    case ELF::R_PPC64_ADDR32:
+      return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    case ELF::R_PPC64_ADDR64:
+      return Value + getELFAddend(R);
+    }
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_X86_64_32S(RelocationRef R, uint64_t Value) {
+  uint64_t visitSystemz(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    switch (Rel) {
+    case ELF::R_390_32: {
+      int64_t Res = Value + getELFAddend(R);
+      if (Res < INT32_MIN || Res > UINT32_MAX)
+        HasError = true;
+      return static_cast<uint32_t>(Res);
+    }
+    case ELF::R_390_64:
+      return Value + getELFAddend(R);
+    }
+    HasError = true;
+    return 0;
   }
 
-  /// BPF ELF
+  uint64_t visitSparc64(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_BPF_64_32(RelocationRef R, uint64_t Value) {
+    switch (Rel) {
-    return Value & 0xFFFFFFFF;
+    case ELF::R_SPARC_32:
+    case ELF::R_SPARC_64:
+    case ELF::R_SPARC_UA32:
+    case ELF::R_SPARC_UA64:
+      return Value + getELFAddend(R);
+    }
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_BPF_64_64(RelocationRef R, uint64_t Value) {
+  uint64_t visitAmdgpu(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return Value;
+    switch (Rel) {
+    case ELF::R_AMDGPU_ABS32:
+    case ELF::R_AMDGPU_ABS64:
+      return Value + getELFAddend(R);
+    }
+    HasError = true;
+    return 0;
   }
 
-  /// PPC64 ELF
+  uint64_t visitX86(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_PPC64_ADDR32(RelocationRef R, uint64_t Value) {
+    switch (Rel) {
-    return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    case ELF::R_386_NONE:
+      return 0;
+    case ELF::R_386_32:
+      return Value;
+    case ELF::R_386_PC32:
+      return Value - R.getOffset();
+    }
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_PPC64_ADDR64(RelocationRef R, uint64_t Value) {
+  uint64_t visitPPC32(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
+    if (Rel == ELF::R_PPC_ADDR32)
+      return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    HasError = true;
+    return 0;
   }
 
-  /// PPC32 ELF
+  uint64_t visitARM(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_PPC_ADDR32(RelocationRef R, uint64_t Value) {
+    if (Rel == ELF::R_ARM_ABS32) {
-    return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+      if ((int64_t)Value < INT32_MIN || (int64_t)Value > UINT32_MAX)
+        HasError = true;
+      return static_cast<uint32_t>(Value);
+    }
+    HasError = true;
+    return 0;
   }
 
-  /// Lanai ELF
+  uint64_t visitLanai(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_Lanai_32(RelocationRef R, uint64_t Value) {
+    if (Rel == ELF::R_LANAI_32)
-    return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+      return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+    HasError = true;
+    return 0;
   }
 
-  /// MIPS ELF
+  uint64_t visitMips32(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_MIPS_32(RelocationRef R, uint64_t Value) {
+    if (Rel == ELF::R_MIPS_32)
-    return Value & 0xFFFFFFFF;
+      return Value & 0xFFFFFFFF;
+    HasError = true;
+    return 0;
   }
 
-  /// MIPS64 ELF
+  uint64_t visitSparc32(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_MIPS64_32(RelocationRef R, uint64_t Value) {
+    if (Rel == ELF::R_SPARC_32 || Rel == ELF::R_SPARC_UA32)
-    return (Value + getELFAddend(R)) & 0xFFFFFFFF;
+      return Value + getELFAddend(R);
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_MIPS64_64(RelocationRef R, uint64_t Value) {
+  uint64_t visitHexagon(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
+    if (Rel == ELF::R_HEX_32)
+      return Value + getELFAddend(R);
+    HasError = true;
+    return 0;
   }
 
-  // AArch64 ELF
+  uint64_t visitCOFF(uint32_t Rel, RelocationRef R, uint64_t Value) {
-  uint64_t visitELF_AARCH64_ABS32(RelocationRef R, uint64_t Value) {
+    switch (ObjToVisit.getArch()) {
-    int64_t Addend = getELFAddend(R);
+    case Triple::x86:
-    int64_t Res =  Value + Addend;
+      switch (Rel) {
-
+      case COFF::IMAGE_REL_I386_SECREL:
-    // Overflow check allows for both signed and unsigned interpretation.
+      case COFF::IMAGE_REL_I386_DIR32:
-    if (Res < INT32_MIN || Res > UINT32_MAX)
+        return static_cast<uint32_t>(Value);
-      HasError = true;
+      }
-
+      break;
-    return static_cast<uint32_t>(Res);
+    case Triple::x86_64:
+      switch (Rel) {
+      case COFF::IMAGE_REL_AMD64_SECREL:
+        return static_cast<uint32_t>(Value);
+      case COFF::IMAGE_REL_AMD64_ADDR64:
+        return Value;
+      }
+      break;
+    }
+    HasError = true;
+    return 0;
   }
 
-  uint64_t visitELF_AARCH64_ABS64(RelocationRef R, uint64_t Value) {
+  uint64_t visitMachO(uint32_t Rel, RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
+    if (ObjToVisit.getArch() == Triple::x86_64 &&
-  }
+        Rel == MachO::X86_64_RELOC_UNSIGNED)
-
+      return Value;
-  // SystemZ ELF
+    HasError = true;
-  uint64_t visitELF_390_32(RelocationRef R, uint64_t Value) {
+    return 0;
-    int64_t Addend = getELFAddend(R);
-    int64_t Res = Value + Addend;
-
-    // Overflow check allows for both signed and unsigned interpretation.
-    if (Res < INT32_MIN || Res > UINT32_MAX)
-      HasError = true;
-
-    return static_cast<uint32_t>(Res);
-  }
-
-  uint64_t visitELF_390_64(RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
-  }
-
-  uint64_t visitELF_SPARC_32(RelocationRef R, uint32_t Value) {
-    return Value + getELFAddend(R);
-  }
-
-  uint64_t visitELF_SPARCV9_32(RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
-  }
-
-  uint64_t visitELF_SPARCV9_64(RelocationRef R, uint64_t Value) {
-    return Value + getELFAddend(R);
-  }
-
-  uint64_t visitELF_ARM_ABS32(RelocationRef R, uint64_t Value) {
-    int64_t Res = Value;
-
-    // Overflow check allows for both signed and unsigned interpretation.
-    if (Res < INT32_MIN || Res > UINT32_MAX)
-      HasError = true;
-
-    return static_cast<uint32_t>(Res);
-  }
-
-  uint64_t visitELF_HEX_32(RelocationRef R, uint64_t Value) {
-    int64_t Addend = getELFAddend(R);
-    return Value + Addend;
-  }
-
-  uint64_t visitELF_AMDGPU_ABS32(RelocationRef R, uint64_t Value) {
-    int64_t Addend = getELFAddend(R);
-    return Value + Addend;
-  }
-
-  uint64_t visitELF_AMDGPU_ABS64(RelocationRef R, uint64_t Value) {
-    int64_t Addend = getELFAddend(R);
-    return Value + Addend;
-  }
-
-  /// I386 COFF
-  uint64_t visitCOFF_I386_SECREL(RelocationRef R, uint64_t Value) {
-    return static_cast<uint32_t>(Value);
-  }
-
-  uint64_t visitCOFF_I386_DIR32(RelocationRef R, uint64_t Value) {
-    return static_cast<uint32_t>(Value);
-  }
-
-  /// AMD64 COFF
-  uint64_t visitCOFF_AMD64_SECREL(RelocationRef R, uint64_t Value) {
-    return static_cast<uint32_t>(Value);
-  }
-
-  uint64_t visitCOFF_AMD64_ADDR64(RelocationRef R, uint64_t Value) {
-    return Value;
-  }
-
-  // X86_64 MachO
-  uint64_t visitMACHO_X86_64_UNSIGNED(RelocationRef R, uint64_t Value) {
-    return Value;
   }
 };
 

include/llvm/Object/Wasm.h

@@ -119,6 +119,7 @@ public:
   std::error_code getSectionName(DataRefImpl Sec,
                                  StringRef &Res) const override;
   uint64_t getSectionAddress(DataRefImpl Sec) const override;
+  uint64_t getSectionIndex(DataRefImpl Sec) const override;
   uint64_t getSectionSize(DataRefImpl Sec) const override;
   std::error_code getSectionContents(DataRefImpl Sec,
                                      StringRef &Res) const override;

include/llvm/Option/OptTable.h

@@ -113,6 +113,14 @@ public:
     return getInfo(id).MetaVar;
   }
 
+  /// Find flags from OptTable which starts with Cur.
+  ///
+  /// \param [in] Cur - String prefix that all returned flags need
+  //  to start with.
+  ///
+  /// \return The vector of flags which start with Cur.
+  std::vector<std::string> findByPrefix(StringRef Cur) const;
+
   /// \brief Parse a single argument; returning the new argument and
   /// updating Index.
   ///

include/llvm/ProfileData/InstrProf.h

@@ -212,12 +212,12 @@ StringRef getFuncNameWithoutPrefix(StringRef PGOFuncName,
 ///  third field is the uncompressed strings; otherwise it is the
 /// compressed string. When the string compression is off, the
 /// second field will have value zero.
-Error collectPGOFuncNameStrings(const std::vector<std::string> &NameStrs,
+Error collectPGOFuncNameStrings(ArrayRef<std::string> NameStrs,
                                 bool doCompression, std::string &Result);
 
 /// Produce \c Result string with the same format described above. The input
 /// is vector of PGO function name variables that are referenced.
-Error collectPGOFuncNameStrings(const std::vector<GlobalVariable *> &NameVars,
+Error collectPGOFuncNameStrings(ArrayRef<GlobalVariable *> NameVars,
                                 std::string &Result, bool doCompression = true);
 
 /// \c NameStrings is a string composed of one of more sub-strings encoded in
@@ -967,7 +967,7 @@ struct Header {
 } // end namespace RawInstrProf
 
 // Parse MemOP Size range option.
-void getMemOPSizeRangeFromOption(std::string Str, int64_t &RangeStart,
+void getMemOPSizeRangeFromOption(StringRef Str, int64_t &RangeStart,
                                  int64_t &RangeLast);
 
 } // end namespace llvm

include/llvm/TableGen/Record.h

@@ -671,7 +671,7 @@ public:
 /// [AL, AH, CL] - Represent a list of defs
 ///
 class ListInit final : public TypedInit, public FoldingSetNode,
-                       public TrailingObjects<BitsInit, Init *> {
+                       public TrailingObjects<ListInit, Init *> {
   unsigned NumValues;
 
 public:
@@ -1137,17 +1137,19 @@ public:
 /// to have at least one value then a (possibly empty) list of arguments.  Each
 /// argument can have a name associated with it.
 ///
-class DagInit : public TypedInit, public FoldingSetNode {
+class DagInit final : public TypedInit, public FoldingSetNode,
+                      public TrailingObjects<DagInit, Init *, StringInit *> {
   Init *Val;
   StringInit *ValName;
-  SmallVector<Init*, 4> Args;
+  unsigned NumArgs;
-  SmallVector<StringInit*, 4> ArgNames;
+  unsigned NumArgNames;
 
-  DagInit(Init *V, StringInit *VN, ArrayRef<Init *> ArgRange,
+  DagInit(Init *V, StringInit *VN, unsigned NumArgs, unsigned NumArgNames)
-          ArrayRef<StringInit *> NameRange)
       : TypedInit(IK_DagInit, DagRecTy::get()), Val(V), ValName(VN),
-          Args(ArgRange.begin(), ArgRange.end()),
+        NumArgs(NumArgs), NumArgNames(NumArgNames) {}
-          ArgNames(NameRange.begin(), NameRange.end()) {}
+
+  friend TrailingObjects;
+  size_t numTrailingObjects(OverloadToken<Init *>) const { return NumArgs; }
 
 public:
   DagInit(const DagInit &Other) = delete;
@@ -1173,20 +1175,24 @@ public:
     return ValName ? ValName->getValue() : StringRef();
   }
 
-  unsigned getNumArgs() const { return Args.size(); }
+  unsigned getNumArgs() const { return NumArgs; }
   Init *getArg(unsigned Num) const {
-    assert(Num < Args.size() && "Arg number out of range!");
+    assert(Num < NumArgs && "Arg number out of range!");
-    return Args[Num];
+    return getTrailingObjects<Init *>()[Num];
   }
   StringInit *getArgName(unsigned Num) const {
-    assert(Num < ArgNames.size() && "Arg number out of range!");
+    assert(Num < NumArgNames && "Arg number out of range!");
-    return ArgNames[Num];
+    return getTrailingObjects<StringInit *>()[Num];
   }
   StringRef getArgNameStr(unsigned Num) const {
     StringInit *Init = getArgName(Num);
     return Init ? Init->getValue() : StringRef();
   }
 
+  ArrayRef<StringInit *> getArgNames() const {
+    return makeArrayRef(getTrailingObjects<StringInit *>(), NumArgNames);
+  }
+
   Init *resolveReferences(Record &R, const RecordVal *RV) const override;
 
   std::string getAsString() const override;
@@ -1194,20 +1200,20 @@ public:
   typedef SmallVectorImpl<Init*>::const_iterator       const_arg_iterator;
   typedef SmallVectorImpl<StringInit*>::const_iterator const_name_iterator;
 
-  inline const_arg_iterator  arg_begin() const { return Args.begin(); }
+  inline const_arg_iterator  arg_begin() const { return getTrailingObjects<Init *>(); }
-  inline const_arg_iterator  arg_end  () const { return Args.end();   }
+  inline const_arg_iterator  arg_end  () const { return arg_begin() + NumArgs;   }
   inline iterator_range<const_arg_iterator> args() const {
     return llvm::make_range(arg_begin(), arg_end());
   }
 
-  inline size_t              arg_size () const { return Args.size();  }
+  inline size_t              arg_size () const { return NumArgs;  }
-  inline bool                arg_empty() const { return Args.empty(); }
+  inline bool                arg_empty() const { return NumArgs == 0; }
 
-  inline const_name_iterator name_begin() const { return ArgNames.begin(); }
+  inline const_name_iterator name_begin() const { return getTrailingObjects<StringInit *>(); }
-  inline const_name_iterator name_end  () const { return ArgNames.end();   }
+  inline const_name_iterator name_end  () const { return name_begin() + NumArgNames;   }
 
-  inline size_t              name_size () const { return ArgNames.size();  }
+  inline size_t              name_size () const { return NumArgNames;  }
-  inline bool                name_empty() const { return ArgNames.empty(); }
+  inline bool                name_empty() const { return NumArgNames == 0; }
 
   Init *getBit(unsigned Bit) const override {
     llvm_unreachable("Illegal bit reference off dag");

include/llvm/Target/TargetLowering.h

@@ -405,7 +405,9 @@ public:
   }
 
   /// Returns if it's reasonable to merge stores to MemVT size.
-  virtual bool canMergeStoresTo(EVT MemVT) const { return true; }
+  virtual bool canMergeStoresTo(unsigned AddressSpace, EVT MemVT) const {
+    return true;
+  }
 
   /// \brief Return true if it is cheap to speculate a call to intrinsic cttz.
   virtual bool isCheapToSpeculateCttz() const {
@@ -736,7 +738,7 @@ public:
     if (VT.isExtended()) return Expand;
     // If a target-specific SDNode requires legalization, require the target
     // to provide custom legalization for it.
-    if (Op > array_lengthof(OpActions[0])) return Custom;
+    if (Op >= array_lengthof(OpActions[0])) return Custom;
     return OpActions[(unsigned)VT.getSimpleVT().SimpleTy][Op];
   }
 

include/llvm/Transforms/Scalar.h

@@ -354,6 +354,13 @@ FunctionPass *createEarlyCSEPass(bool UseMemorySSA = false);
 //
 FunctionPass *createGVNHoistPass();
 
+//===----------------------------------------------------------------------===//
+//
+// GVNSink - This pass uses an "inverted" value numbering to decide the
+// similarity of expressions and sinks similar expressions into successors.
+//
+FunctionPass *createGVNSinkPass();
+
 //===----------------------------------------------------------------------===//
 //
 // MergedLoadStoreMotion - This pass merges loads and stores in diamonds. Loads

include/llvm/Transforms/Scalar/GVN.h

@@ -68,6 +68,24 @@ public:
   class ValueTable {
     DenseMap<Value *, uint32_t> valueNumbering;
     DenseMap<Expression, uint32_t> expressionNumbering;
+
+    // Expressions is the vector of Expression. ExprIdx is the mapping from
+    // value number to the index of Expression in Expressions. We use it
+    // instead of a DenseMap because filling such mapping is faster than
+    // filling a DenseMap and the compile time is a little better.
+    uint32_t nextExprNumber;
+    std::vector<Expression> Expressions;
+    std::vector<uint32_t> ExprIdx;
+    // Value number to PHINode mapping. Used for phi-translate in scalarpre.
+    DenseMap<uint32_t, PHINode *> NumberingPhi;
+    // Cache for phi-translate in scalarpre.
+    typedef DenseMap<std::pair<uint32_t, const BasicBlock *>, uint32_t>
+        PhiTranslateMap;
+    PhiTranslateMap PhiTranslateTable;
+    // Map the block to reversed postorder traversal number. It is used to
+    // find back edge easily.
+    DenseMap<const BasicBlock *, uint32_t> BlockRPONumber;
+
     AliasAnalysis *AA;
     MemoryDependenceResults *MD;
     DominatorTree *DT;
@@ -79,6 +97,10 @@ public:
                              Value *LHS, Value *RHS);
     Expression createExtractvalueExpr(ExtractValueInst *EI);
     uint32_t lookupOrAddCall(CallInst *C);
+    uint32_t phiTranslateImpl(const BasicBlock *BB, const BasicBlock *PhiBlock,
+                              uint32_t Num, GVN &Gvn);
+    std::pair<uint32_t, bool> assignExpNewValueNum(Expression &exp);
+    bool areAllValsInBB(uint32_t num, const BasicBlock *BB, GVN &Gvn);
 
   public:
     ValueTable();
@@ -87,9 +109,12 @@ public:
     ~ValueTable();
 
     uint32_t lookupOrAdd(Value *V);
-    uint32_t lookup(Value *V) const;
+    uint32_t lookup(Value *V, bool Verify = true) const;
     uint32_t lookupOrAddCmp(unsigned Opcode, CmpInst::Predicate Pred,
                             Value *LHS, Value *RHS);
+    uint32_t phiTranslate(const BasicBlock *BB, const BasicBlock *PhiBlock,
+                          uint32_t Num, GVN &Gvn);
+    void assignBlockRPONumber(Function &F);
     bool exists(Value *V) const;
     void add(Value *V, uint32_t num);
     void clear();
@@ -238,7 +263,12 @@ struct GVNHoistPass : PassInfoMixin<GVNHoistPass> {
   /// \brief Run the pass over the function.
   PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
 };
-
+/// \brief Uses an "inverted" value numbering to decide the similarity of
+/// expressions and sinks similar expressions into successors.
+struct GVNSinkPass : PassInfoMixin<GVNSinkPass> {
+  /// \brief Run the pass over the function.
+  PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
+};
 }
 
 #endif

include/llvm/Transforms/Utils/Local.h

@@ -356,6 +356,10 @@ void combineMetadata(Instruction *K, const Instruction *J, ArrayRef<unsigned> Kn
 /// Unknown metadata is removed.
 void combineMetadataForCSE(Instruction *K, const Instruction *J);
 
+// Replace each use of 'From' with 'To', if that use does not belong to basic
+// block where 'From' is defined. Returns the number of replacements made.
+unsigned replaceNonLocalUsesWith(Instruction *From, Value *To);
+
 /// Replace each use of 'From' with 'To' if that use is dominated by
 /// the given edge.  Returns the number of replacements made.
 unsigned replaceDominatedUsesWith(Value *From, Value *To, DominatorTree &DT,
@@ -406,6 +410,14 @@ bool recognizeBSwapOrBitReverseIdiom(
 void maybeMarkSanitizerLibraryCallNoBuiltin(CallInst *CI,
                                             const TargetLibraryInfo *TLI);
 
+//===----------------------------------------------------------------------===//
+//  Transform predicates
+//
+
+/// Given an instruction, is it legal to set operand OpIdx to a non-constant
+/// value?
+bool canReplaceOperandWithVariable(const Instruction *I, unsigned OpIdx);
+
 } // End llvm namespace
 
 #endif

lib/Analysis/ConstantFolding.cpp

@@ -687,11 +687,8 @@ Constant *SymbolicallyEvaluateBinop(unsigned Opc, Constant *Op0, Constant *Op1,
   // bits.
 
   if (Opc == Instruction::And) {
-    unsigned BitWidth = DL.getTypeSizeInBits(Op0->getType()->getScalarType());
+    KnownBits Known0 = computeKnownBits(Op0, DL);
-    KnownBits Known0(BitWidth);
+    KnownBits Known1 = computeKnownBits(Op1, DL);
-    KnownBits Known1(BitWidth);
-    computeKnownBits(Op0, Known0, DL);
-    computeKnownBits(Op1, Known1, DL);
     if ((Known1.One | Known0.Zero).isAllOnesValue()) {
       // All the bits of Op0 that the 'and' could be masking are already zero.
       return Op0;

lib/Analysis/InstructionSimplify.cpp

@@ -688,9 +688,7 @@ static Value *SimplifySubInst(Value *Op0, Value *Op1, bool isNSW, bool isNUW,
     if (isNUW)
       return Op0;
 
-    unsigned BitWidth = Op1->getType()->getScalarSizeInBits();
+    KnownBits Known = computeKnownBits(Op1, Q.DL, 0, Q.AC, Q.CxtI, Q.DT);
-    KnownBits Known(BitWidth);
-    computeKnownBits(Op1, Known, Q.DL, 0, Q.AC, Q.CxtI, Q.DT);
     if (Known.Zero.isMaxSignedValue()) {
       // Op1 is either 0 or the minimum signed value. If the sub is NSW, then
       // Op1 must be 0 because negating the minimum signed value is undefined.
@@ -1309,15 +1307,13 @@ static Value *SimplifyShift(Instruction::BinaryOps Opcode, Value *Op0,
 
   // If any bits in the shift amount make that value greater than or equal to
   // the number of bits in the type, the shift is undefined.
-  unsigned BitWidth = Op1->getType()->getScalarSizeInBits();
+  KnownBits Known = computeKnownBits(Op1, Q.DL, 0, Q.AC, Q.CxtI, Q.DT);
-  KnownBits Known(BitWidth);
+  if (Known.One.getLimitedValue() >= Known.getBitWidth())
-  computeKnownBits(Op1, Known, Q.DL, 0, Q.AC, Q.CxtI, Q.DT);
-  if (Known.One.getLimitedValue() >= BitWidth)
     return UndefValue::get(Op0->getType());
 
   // If all valid bits in the shift amount are known zero, the first operand is
   // unchanged.
-  unsigned NumValidShiftBits = Log2_32_Ceil(BitWidth);
+  unsigned NumValidShiftBits = Log2_32_Ceil(Known.getBitWidth());
   if (Known.countMinTrailingZeros() >= NumValidShiftBits)
     return Op0;
 
@@ -1343,9 +1339,7 @@ static Value *SimplifyRightShift(Instruction::BinaryOps Opcode, Value *Op0,
 
   // The low bit cannot be shifted out of an exact shift if it is set.
   if (isExact) {
-    unsigned BitWidth = Op0->getType()->getScalarSizeInBits();
+    KnownBits Op0Known = computeKnownBits(Op0, Q.DL, /*Depth=*/0, Q.AC, Q.CxtI, Q.DT);
-    KnownBits Op0Known(BitWidth);
-    computeKnownBits(Op0, Op0Known, Q.DL, /*Depth=*/0, Q.AC, Q.CxtI, Q.DT);
     if (Op0Known.One[0])
       return Op0;
   }
@@ -1428,6 +1422,8 @@ Value *llvm::SimplifyAShrInst(Value *Op0, Value *Op1, bool isExact,
   return ::SimplifyAShrInst(Op0, Op1, isExact, Q, RecursionLimit);
 }
 
+/// Commuted variants are assumed to be handled by calling this function again
+/// with the parameters swapped.
 static Value *simplifyUnsignedRangeCheck(ICmpInst *ZeroICmp,
                                          ICmpInst *UnsignedICmp, bool IsAnd) {
   Value *X, *Y;
@@ -1560,20 +1556,8 @@ static Value *simplifyAndOrOfICmpsWithConstants(ICmpInst *Cmp0, ICmpInst *Cmp1,
   return nullptr;
 }
 
-/// Commuted variants are assumed to be handled by calling this function again
+static Value *simplifyAndOfICmpsWithAdd(ICmpInst *Op0, ICmpInst *Op1) {
-/// with the parameters swapped.
-static Value *simplifyAndOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
-  if (Value *X = simplifyUnsignedRangeCheck(Op0, Op1, /*IsAnd=*/true))
-    return X;
-
-  if (Value *X = simplifyAndOfICmpsWithSameOperands(Op0, Op1))
-    return X;
-
-  if (Value *X = simplifyAndOrOfICmpsWithConstants(Op0, Op1, true))
-    return X;
-
   // (icmp (add V, C0), C1) & (icmp V, C0)
-  Type *ITy = Op0->getType();
   ICmpInst::Predicate Pred0, Pred1;
   const APInt *C0, *C1;
   Value *V;
@@ -1587,6 +1571,7 @@ static Value *simplifyAndOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
   if (AddInst->getOperand(1) != Op1->getOperand(1))
     return nullptr;
 
+  Type *ITy = Op0->getType();
   bool isNSW = AddInst->hasNoSignedWrap();
   bool isNUW = AddInst->hasNoUnsignedWrap();
 
@@ -1617,18 +1602,29 @@ static Value *simplifyAndOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
   return nullptr;
 }
 
-/// Commuted variants are assumed to be handled by calling this function again
+static Value *simplifyAndOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
-/// with the parameters swapped.
+  if (Value *X = simplifyUnsignedRangeCheck(Op0, Op1, /*IsAnd=*/true))
-static Value *simplifyOrOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
+    return X;
-  if (Value *X = simplifyUnsignedRangeCheck(Op0, Op1, /*IsAnd=*/false))
+  if (Value *X = simplifyUnsignedRangeCheck(Op1, Op0, /*IsAnd=*/true))
     return X;
 
-  if (Value *X = simplifyOrOfICmpsWithSameOperands(Op0, Op1))
+  if (Value *X = simplifyAndOfICmpsWithSameOperands(Op0, Op1))
+    return X;
+  if (Value *X = simplifyAndOfICmpsWithSameOperands(Op1, Op0))
     return X;
 
-  if (Value *X = simplifyAndOrOfICmpsWithConstants(Op0, Op1, false))
+  if (Value *X = simplifyAndOrOfICmpsWithConstants(Op0, Op1, true))
     return X;
 
+  if (Value *X = simplifyAndOfICmpsWithAdd(Op0, Op1))
+    return X;
+  if (Value *X = simplifyAndOfICmpsWithAdd(Op1, Op0))
+    return X;
+
+  return nullptr;
+}
+
+static Value *simplifyOrOfICmpsWithAdd(ICmpInst *Op0, ICmpInst *Op1) {
   // (icmp (add V, C0), C1) | (icmp V, C0)
   ICmpInst::Predicate Pred0, Pred1;
   const APInt *C0, *C1;
@@ -1674,19 +1670,24 @@ static Value *simplifyOrOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
   return nullptr;
 }
 
-static Value *simplifyPossiblyCastedAndOrOfICmps(ICmpInst *Cmp0, ICmpInst *Cmp1,
+static Value *simplifyOrOfICmps(ICmpInst *Op0, ICmpInst *Op1) {
-                                                 bool IsAnd, CastInst *Cast) {
+  if (Value *X = simplifyUnsignedRangeCheck(Op0, Op1, /*IsAnd=*/false))
-  Value *V =
+    return X;
-      IsAnd ? simplifyAndOfICmps(Cmp0, Cmp1) : simplifyOrOfICmps(Cmp0, Cmp1);
+  if (Value *X = simplifyUnsignedRangeCheck(Op1, Op0, /*IsAnd=*/false))
-  if (!V)
+    return X;
-    return nullptr;
-  if (!Cast)
-    return V;
 
-  // If we looked through casts, we can only handle a constant simplification
+  if (Value *X = simplifyOrOfICmpsWithSameOperands(Op0, Op1))
-  // because we are not allowed to create a cast instruction here.
+    return X;
-  if (auto *C = dyn_cast<Constant>(V))
+  if (Value *X = simplifyOrOfICmpsWithSameOperands(Op1, Op0))
-    return ConstantExpr::getCast(Cast->getOpcode(), C, Cast->getType());
+    return X;
+
+  if (Value *X = simplifyAndOrOfICmpsWithConstants(Op0, Op1, false))
+    return X;
+
+  if (Value *X = simplifyOrOfICmpsWithAdd(Op0, Op1))
+    return X;
+  if (Value *X = simplifyOrOfICmpsWithAdd(Op1, Op0))
+    return X;
 
   return nullptr;
 }
@@ -1706,11 +1707,18 @@ static Value *simplifyAndOrOfICmps(Value *Op0, Value *Op1, bool IsAnd) {
   if (!Cmp0 || !Cmp1)
     return nullptr;
 
-  if (Value *V = simplifyPossiblyCastedAndOrOfICmps(Cmp0, Cmp1, IsAnd, Cast0))
+  Value *V =
-    return V;
+      IsAnd ? simplifyAndOfICmps(Cmp0, Cmp1) : simplifyOrOfICmps(Cmp0, Cmp1);
-  if (Value *V = simplifyPossiblyCastedAndOrOfICmps(Cmp1, Cmp0, IsAnd, Cast0))
+  if (!V)
+    return nullptr;
+  if (!Cast0)
     return V;
 
+  // If we looked through casts, we can only handle a constant simplification
+  // because we are not allowed to create a cast instruction here.
+  if (auto *C = dyn_cast<Constant>(V))
+    return ConstantExpr::getCast(Cast0->getOpcode(), C, Cast0->getType());
+
   return nullptr;
 }
 
@@ -1927,37 +1935,27 @@ static Value *SimplifyOrInst(Value *Op0, Value *Op1, const SimplifyQuery &Q,
                                          MaxRecurse))
       return V;
 
-  // (A & C)|(B & D)
+  // (A & C1)|(B & C2)
-  Value *C = nullptr, *D = nullptr;
+  const APInt *C1, *C2;
-  if (match(Op0, m_And(m_Value(A), m_Value(C))) &&
+  if (match(Op0, m_And(m_Value(A), m_APInt(C1))) &&
-      match(Op1, m_And(m_Value(B), m_Value(D)))) {
+      match(Op1, m_And(m_Value(B), m_APInt(C2)))) {
-    ConstantInt *C1 = dyn_cast<ConstantInt>(C);
+    if (*C1 == ~*C2) {
-    ConstantInt *C2 = dyn_cast<ConstantInt>(D);
-    if (C1 && C2 && (C1->getValue() == ~C2->getValue())) {
       // (A & C1)|(B & C2)
       // If we have: ((V + N) & C1) | (V & C2)
       // .. and C2 = ~C1 and C2 is 0+1+ and (N & C2) == 0
       // replace with V+N.
-      Value *V1, *V2;
+      Value *N;
-      if ((C2->getValue() & (C2->getValue() + 1)) == 0 && // C2 == 0+1+
+      if (C2->isMask() && // C2 == 0+1+
-          match(A, m_Add(m_Value(V1), m_Value(V2)))) {
+          match(A, m_c_Add(m_Specific(B), m_Value(N)))) {
         // Add commutes, try both ways.
-        if (V1 == B &&
+        if (MaskedValueIsZero(N, *C2, Q.DL, 0, Q.AC, Q.CxtI, Q.DT))
-            MaskedValueIsZero(V2, C2->getValue(), Q.DL, 0, Q.AC, Q.CxtI, Q.DT))
-          return A;
-        if (V2 == B &&
-            MaskedValueIsZero(V1, C2->getValue(), Q.DL, 0, Q.AC, Q.CxtI, Q.DT))
           return A;
       }
       // Or commutes, try both ways.
-      if ((C1->getValue() & (C1->getValue() + 1)) == 0 &&
+      if (C1->isMask() &&
-          match(B, m_Add(m_Value(V1), m_Value(V2)))) {
+          match(B, m_c_Add(m_Specific(A), m_Value(N)))) {
         // Add commutes, try both ways.
-        if (V1 == A &&
+        if (MaskedValueIsZero(N, *C1, Q.DL, 0, Q.AC, Q.CxtI, Q.DT))
-            MaskedValueIsZero(V2, C1->getValue(), Q.DL, 0, Q.AC, Q.CxtI, Q.DT))
-          return B;
-        if (V2 == A &&
-            MaskedValueIsZero(V1, C1->getValue(), Q.DL, 0, Q.AC, Q.CxtI, Q.DT))
           return B;
       }
     }
@@ -3372,9 +3370,7 @@ static Value *SimplifyICmpInst(unsigned Predicate, Value *LHS, Value *RHS,
   if (ICmpInst::isEquality(Pred)) {
     const APInt *RHSVal;
     if (match(RHS, m_APInt(RHSVal))) {
-      unsigned BitWidth = RHSVal->getBitWidth();
+      KnownBits LHSKnown = computeKnownBits(LHS, Q.DL, /*Depth=*/0, Q.AC, Q.CxtI, Q.DT);
-      KnownBits LHSKnown(BitWidth);
-      computeKnownBits(LHS, LHSKnown, Q.DL, /*Depth=*/0, Q.AC, Q.CxtI, Q.DT);
       if (LHSKnown.Zero.intersects(*RHSVal) ||
           !LHSKnown.One.isSubsetOf(*RHSVal))
         return Pred == ICmpInst::ICMP_EQ ? ConstantInt::getFalse(ITy)
@@ -3539,6 +3535,10 @@ static const Value *SimplifyWithOpReplaced(Value *V, Value *Op, Value *RepOp,
   if (V == Op)
     return RepOp;
 
+  // We cannot replace a constant, and shouldn't even try.
+  if (isa<Constant>(Op))
+    return nullptr;
+
   auto *I = dyn_cast<Instruction>(V);
   if (!I)
     return nullptr;
@@ -4444,19 +4444,21 @@ static Value *SimplifyIntrinsic(Function *F, IterTy ArgBegin, IterTy ArgEnd,
     case Intrinsic::uadd_with_overflow:
     case Intrinsic::sadd_with_overflow: {
       // X + undef -> undef
-      if (isa<UndefValue>(RHS))
+      if (isa<UndefValue>(LHS) || isa<UndefValue>(RHS))
         return UndefValue::get(ReturnType);
 
       return nullptr;
     }
     case Intrinsic::umul_with_overflow:
     case Intrinsic::smul_with_overflow: {
+      // 0 * X -> { 0, false }
       // X * 0 -> { 0, false }
-      if (match(RHS, m_Zero()))
+      if (match(LHS, m_Zero()) || match(RHS, m_Zero()))
         return Constant::getNullValue(ReturnType);
 
+      // undef * X -> { 0, false }
       // X * undef -> { 0, false }
-      if (match(RHS, m_Undef()))
+      if (match(LHS, m_Undef()) || match(RHS, m_Undef()))
         return Constant::getNullValue(ReturnType);
 
       return nullptr;
@@ -4680,9 +4682,7 @@ Value *llvm::SimplifyInstruction(Instruction *I, const SimplifyQuery &SQ,
   // In general, it is possible for computeKnownBits to determine all bits in a
   // value even when the operands are not all constants.
   if (!Result && I->getType()->isIntOrIntVectorTy()) {
-    unsigned BitWidth = I->getType()->getScalarSizeInBits();
+    KnownBits Known = computeKnownBits(I, Q.DL, /*Depth*/ 0, Q.AC, I, Q.DT, ORE);
-    KnownBits Known(BitWidth);
-    computeKnownBits(I, Known, Q.DL, /*Depth*/ 0, Q.AC, I, Q.DT, ORE);
     if (Known.isConstant())
       Result = ConstantInt::get(I->getType(), Known.getConstant());
   }

lib/Analysis/Lint.cpp

@@ -534,9 +534,7 @@ static bool isZero(Value *V, const DataLayout &DL, DominatorTree *DT,
 
   VectorType *VecTy = dyn_cast<VectorType>(V->getType());
   if (!VecTy) {
-    unsigned BitWidth = V->getType()->getIntegerBitWidth();
+    KnownBits Known = computeKnownBits(V, DL, 0, AC, dyn_cast<Instruction>(V), DT);
-    KnownBits Known(BitWidth);
-    computeKnownBits(V, Known, DL, 0, AC, dyn_cast<Instruction>(V), DT);
     return Known.isZero();
   }
 
@@ -550,14 +548,12 @@ static bool isZero(Value *V, const DataLayout &DL, DominatorTree *DT,
 
   // For a vector, KnownZero will only be true if all values are zero, so check
   // this per component
-  unsigned BitWidth = VecTy->getElementType()->getIntegerBitWidth();
   for (unsigned I = 0, N = VecTy->getNumElements(); I != N; ++I) {
     Constant *Elem = C->getAggregateElement(I);
     if (isa<UndefValue>(Elem))
       return true;
 
-    KnownBits Known(BitWidth);
+    KnownBits Known = computeKnownBits(Elem, DL);
-    computeKnownBits(Elem, Known, DL);
     if (Known.isZero())
       return true;
   }

lib/Analysis/LoopPass.cpp

@@ -73,30 +73,23 @@ LPPassManager::LPPassManager()
   CurrentLoop = nullptr;
 }
 
-// Inset loop into loop nest (LoopInfo) and loop queue (LQ).
+// Insert loop into loop nest (LoopInfo) and loop queue (LQ).
-Loop &LPPassManager::addLoop(Loop *ParentLoop) {
+void LPPassManager::addLoop(Loop &L) {
-  // Create a new loop. LI will take ownership.
+  if (!L.getParentLoop()) {
-  Loop *L = new Loop();
-
-  // Insert into the loop nest and the loop queue.
-  if (!ParentLoop) {
     // This is the top level loop.
-    LI->addTopLevelLoop(L);
+    LQ.push_front(&L);
-    LQ.push_front(L);
+    return;
-    return *L;
   }
 
-  ParentLoop->addChildLoop(L);
   // Insert L into the loop queue after the parent loop.
   for (auto I = LQ.begin(), E = LQ.end(); I != E; ++I) {
-    if (*I == L->getParentLoop()) {
+    if (*I == L.getParentLoop()) {
       // deque does not support insert after.
       ++I;
-      LQ.insert(I, 1, L);
+      LQ.insert(I, 1, &L);
-      break;
+      return;
     }
   }
-  return *L;
 }
 
 /// cloneBasicBlockSimpleAnalysis - Invoke cloneBasicBlockAnalysis hook for

lib/Analysis/ScalarEvolution.cpp

@@ -2178,6 +2178,63 @@ StrengthenNoWrapFlags(ScalarEvolution *SE, SCEVTypes Type,
   return Flags;
 }
 
+bool ScalarEvolution::isAvailableAtLoopEntry(const SCEV *S, const Loop *L,
+                                             DominatorTree &DT, LoopInfo &LI) {
+  if (!isLoopInvariant(S, L))
+    return false;
+  // If a value depends on a SCEVUnknown which is defined after the loop, we
+  // conservatively assume that we cannot calculate it at the loop's entry.
+  struct FindDominatedSCEVUnknown {
+    bool Found = false;
+    const Loop *L;
+    DominatorTree &DT;
+    LoopInfo &LI;
+
+    FindDominatedSCEVUnknown(const Loop *L, DominatorTree &DT, LoopInfo &LI)
+        : L(L), DT(DT), LI(LI) {}
+
+    bool checkSCEVUnknown(const SCEVUnknown *SU) {
+      if (auto *I = dyn_cast<Instruction>(SU->getValue())) {
+        if (DT.dominates(L->getHeader(), I->getParent()))
+          Found = true;
+        else
+          assert(DT.dominates(I->getParent(), L->getHeader()) &&
+                 "No dominance relationship between SCEV and loop?");
+      }
+      return false;
+    }
+
+    bool follow(const SCEV *S) {
+      switch (static_cast<SCEVTypes>(S->getSCEVType())) {
+      case scConstant:
+        return false;
+      case scAddRecExpr:
+      case scTruncate:
+      case scZeroExtend:
+      case scSignExtend:
+      case scAddExpr:
+      case scMulExpr:
+      case scUMaxExpr:
+      case scSMaxExpr:
+      case scUDivExpr:
+        return true;
+      case scUnknown:
+        return checkSCEVUnknown(cast<SCEVUnknown>(S));
+      case scCouldNotCompute:
+        llvm_unreachable("Attempt to use a SCEVCouldNotCompute object!");
+      }
+      return false;
+    }
+
+    bool isDone() { return Found; }
+  };
+
+  FindDominatedSCEVUnknown FSU(L, DT, LI);
+  SCEVTraversal<FindDominatedSCEVUnknown> ST(FSU);
+  ST.visitAll(S);
+  return !FSU.Found;
+}
+
 /// Get a canonical add expression, or something simpler if possible.
 const SCEV *ScalarEvolution::getAddExpr(SmallVectorImpl<const SCEV *> &Ops,
                                         SCEV::NoWrapFlags Flags,
@@ -2459,7 +2516,7 @@ const SCEV *ScalarEvolution::getAddExpr(SmallVectorImpl<const SCEV *> &Ops,
     const SCEVAddRecExpr *AddRec = cast<SCEVAddRecExpr>(Ops[Idx]);
     const Loop *AddRecLoop = AddRec->getLoop();
     for (unsigned i = 0, e = Ops.size(); i != e; ++i)
-      if (isLoopInvariant(Ops[i], AddRecLoop)) {
+      if (isAvailableAtLoopEntry(Ops[i], AddRecLoop, DT, LI)) {
         LIOps.push_back(Ops[i]);
         Ops.erase(Ops.begin()+i);
         --i; --e;
@@ -2734,7 +2791,7 @@ const SCEV *ScalarEvolution::getMulExpr(SmallVectorImpl<const SCEV *> &Ops,
     const SCEVAddRecExpr *AddRec = cast<SCEVAddRecExpr>(Ops[Idx]);
     const Loop *AddRecLoop = AddRec->getLoop();
     for (unsigned i = 0, e = Ops.size(); i != e; ++i)
-      if (isLoopInvariant(Ops[i], AddRecLoop)) {
+      if (isAvailableAtLoopEntry(Ops[i], AddRecLoop, DT, LI)) {
         LIOps.push_back(Ops[i]);
         Ops.erase(Ops.begin()+i);
         --i; --e;
@@ -4648,10 +4705,7 @@ uint32_t ScalarEvolution::GetMinTrailingZerosImpl(const SCEV *S) {
 
   if (const SCEVUnknown *U = dyn_cast<SCEVUnknown>(S)) {
     // For a SCEVUnknown, ask ValueTracking.
-    unsigned BitWidth = getTypeSizeInBits(U->getType());
+    KnownBits Known = computeKnownBits(U->getValue(), getDataLayout(), 0, &AC, nullptr, &DT);
-    KnownBits Known(BitWidth);
-    computeKnownBits(U->getValue(), Known, getDataLayout(), 0, &AC,
-                     nullptr, &DT);
     return Known.countMinTrailingZeros();
   }
 
@@ -4831,8 +4885,7 @@ ScalarEvolution::getRange(const SCEV *S,
     const DataLayout &DL = getDataLayout();
     if (SignHint == ScalarEvolution::HINT_RANGE_UNSIGNED) {
       // For a SCEVUnknown, ask ValueTracking.
-      KnownBits Known(BitWidth);
+      KnownBits Known = computeKnownBits(U->getValue(), DL, 0, &AC, nullptr, &DT);
-      computeKnownBits(U->getValue(), Known, DL, 0, &AC, nullptr, &DT);
       if (Known.One != ~Known.Zero + 1)
         ConservativeResult =
             ConservativeResult.intersectWith(ConstantRange(Known.One,
@@ -9537,8 +9590,11 @@ struct SCEVCollectAddRecMultiplies {
       bool HasAddRec = false;
       SmallVector<const SCEV *, 0> Operands;
       for (auto Op : Mul->operands()) {
-        if (isa<SCEVUnknown>(Op)) {
+        const SCEVUnknown *Unknown = dyn_cast<SCEVUnknown>(Op);
+        if (Unknown && !isa<CallInst>(Unknown->getValue())) {
           Operands.push_back(Op);
+        } else if (Unknown) {
+          HasAddRec = true;
         } else {
           bool ContainsAddRec;
           SCEVHasAddRec ContiansAddRec(ContainsAddRec);

lib/Analysis/ScalarEvolutionExpander.cpp

@@ -1305,12 +1305,17 @@ Value *SCEVExpander::expandAddRecExprLiterally(const SCEVAddRecExpr *S) {
   // Expand the core addrec. If we need post-loop scaling, force it to
   // expand to an integer type to avoid the need for additional casting.
   Type *ExpandTy = PostLoopScale ? IntTy : STy;
+  // We can't use a pointer type for the addrec if the pointer type is
+  // non-integral.
+  Type *AddRecPHIExpandTy =
+      DL.isNonIntegralPointerType(STy) ? Normalized->getType() : ExpandTy;
+
   // In some cases, we decide to reuse an existing phi node but need to truncate
   // it and/or invert the step.
   Type *TruncTy = nullptr;
   bool InvertStep = false;
-  PHINode *PN = getAddRecExprPHILiterally(Normalized, L, ExpandTy, IntTy,
+  PHINode *PN = getAddRecExprPHILiterally(Normalized, L, AddRecPHIExpandTy,
-                                          TruncTy, InvertStep);
+                                          IntTy, TruncTy, InvertStep);
 
   // Accommodate post-inc mode, if necessary.
   Value *Result;
@@ -1383,8 +1388,15 @@ Value *SCEVExpander::expandAddRecExprLiterally(const SCEVAddRecExpr *S) {
   // Re-apply any non-loop-dominating offset.
   if (PostLoopOffset) {
     if (PointerType *PTy = dyn_cast<PointerType>(ExpandTy)) {
-      const SCEV *const OffsetArray[1] = { PostLoopOffset };
+      if (Result->getType()->isIntegerTy()) {
-      Result = expandAddToGEP(OffsetArray, OffsetArray+1, PTy, IntTy, Result);
+        Value *Base = expandCodeFor(PostLoopOffset, ExpandTy);
+        const SCEV *const OffsetArray[1] = {SE.getUnknown(Result)};
+        Result = expandAddToGEP(OffsetArray, OffsetArray + 1, PTy, IntTy, Base);
+      } else {
+        const SCEV *const OffsetArray[1] = {PostLoopOffset};
+        Result =
+            expandAddToGEP(OffsetArray, OffsetArray + 1, PTy, IntTy, Result);
+      }
     } else {
       Result = InsertNoopCastOfTo(Result, IntTy);
       Result = Builder.CreateAdd(Result,

lib/Analysis/TargetTransformInfo.cpp

@@ -149,6 +149,10 @@ bool TargetTransformInfo::isLegalMaskedScatter(Type *DataType) const {
   return TTIImpl->isLegalMaskedGather(DataType);
 }
 
+bool TargetTransformInfo::prefersVectorizedAddressing() const {
+  return TTIImpl->prefersVectorizedAddressing();
+}
+
 int TargetTransformInfo::getScalingFactorCost(Type *Ty, GlobalValue *BaseGV,
                                               int64_t BaseOffset,
                                               bool HasBaseReg,

lib/Analysis/ValueTracking.cpp

@@ -149,8 +149,10 @@ static KnownBits computeKnownBits(const Value *V, unsigned Depth,
 KnownBits llvm::computeKnownBits(const Value *V, const DataLayout &DL,
                                  unsigned Depth, AssumptionCache *AC,
                                  const Instruction *CxtI,
-                                 const DominatorTree *DT) {
+                                 const DominatorTree *DT,
-  return ::computeKnownBits(V, Depth, Query(DL, AC, safeCxtI(V, CxtI), DT));
+                                 OptimizationRemarkEmitter *ORE) {
+  return ::computeKnownBits(V, Depth,
+                            Query(DL, AC, safeCxtI(V, CxtI), DT, ORE));
 }
 
 bool llvm::haveNoCommonBitsSet(const Value *LHS, const Value *RHS,

lib/Bitcode/Writer/BitcodeWriter.cpp

@@ -660,10 +660,12 @@ void ModuleBitcodeWriter::writeAttributeTable() {
 
   SmallVector<uint64_t, 64> Record;
   for (unsigned i = 0, e = Attrs.size(); i != e; ++i) {
-    const AttributeList &A = Attrs[i];
+    AttributeList AL = Attrs[i];
-    for (unsigned i = 0, e = A.getNumSlots(); i != e; ++i)
+    for (unsigned i = AL.index_begin(), e = AL.index_end(); i != e; ++i) {
-      Record.push_back(
+      AttributeSet AS = AL.getAttributes(i);
-          VE.getAttributeGroupID({A.getSlotIndex(i), A.getSlotAttributes(i)}));
+      if (AS.hasAttributes())
+        Record.push_back(VE.getAttributeGroupID({i, AS}));
+    }
 
     Stream.EmitRecord(bitc::PARAMATTR_CODE_ENTRY, Record);
     Record.clear();
@@ -3413,30 +3415,8 @@ void IndexBitcodeWriter::writeCombinedGlobalValueSummary() {
 
   // Create value IDs for undefined references.
   forEachSummary([&](GVInfo I) {
-    if (auto *VS = dyn_cast<GlobalVarSummary>(I.second)) {
+    for (auto &RI : I.second->refs())
-      for (auto &RI : VS->refs())
-        assignValueId(RI.getGUID());
-      return;
-    }
-
-    auto *FS = dyn_cast<FunctionSummary>(I.second);
-    if (!FS)
-      return;
-    for (auto &RI : FS->refs())
       assignValueId(RI.getGUID());
-
-    for (auto &EI : FS->calls()) {
-      GlobalValue::GUID GUID = EI.first.getGUID();
-      if (!hasValueId(GUID)) {
-        // For SamplePGO, the indirect call targets for local functions will
-        // have its original name annotated in profile. We try to find the
-        // corresponding PGOFuncName as the GUID.
-        GUID = Index.getGUIDFromOriginalID(GUID);
-        if (GUID == 0 || !hasValueId(GUID))
-          continue;
-      }
-      assignValueId(GUID);
-    }
   });
 
   for (const auto &GVI : valueIds()) {

lib/Bitcode/Writer/ValueEnumerator.cpp

@@ -902,8 +902,11 @@ void ValueEnumerator::EnumerateAttributes(AttributeList PAL) {
   }
 
   // Do lookups for all attribute groups.
-  for (unsigned i = 0, e = PAL.getNumSlots(); i != e; ++i) {
+  for (unsigned i = PAL.index_begin(), e = PAL.index_end(); i != e; ++i) {
-    IndexAndAttrSet Pair = {PAL.getSlotIndex(i), PAL.getSlotAttributes(i)};
+    AttributeSet AS = PAL.getAttributes(i);
+    if (!AS.hasAttributes())
+      continue;
+    IndexAndAttrSet Pair = {i, AS};
     unsigned &Entry = AttributeGroupMap[Pair];
     if (Entry == 0) {
       AttributeGroups.push_back(Pair);

lib/CodeGen/AsmPrinter/AsmPrinter.cpp

@@ -628,12 +628,15 @@ void AsmPrinter::EmitDebugThreadLocal(const MCExpr *Value,
 /// EmitFunctionHeader - This method emits the header for the current
 /// function.
 void AsmPrinter::EmitFunctionHeader() {
+  const Function *F = MF->getFunction();
+
+  if (isVerbose())
+    OutStreamer->GetCommentOS() << "-- Begin function " << F->getName() << '\n';
+
   // Print out constants referenced by the function
   EmitConstantPool();
 
   // Print the 'header' of function.
-  const Function *F = MF->getFunction();
-
   OutStreamer->SwitchSection(getObjFileLowering().SectionForGlobal(F, TM));
   EmitVisibility(CurrentFnSym, F->getVisibility());
 
@@ -1107,6 +1110,9 @@ void AsmPrinter::EmitFunctionBody() {
     HI.Handler->endFunction(MF);
   }
 
+  if (isVerbose())
+    OutStreamer->GetCommentOS() << "-- End function\n";
+
   OutStreamer->AddBlankLine();
 }
 

lib/CodeGen/AsmPrinter/CodeViewDebug.cpp

@@ -1025,11 +1025,11 @@ void CodeViewDebug::beginFunctionImpl(const MachineFunction *MF) {
   bool EmptyPrologue = true;
   for (const auto &MBB : *MF) {
     for (const auto &MI : MBB) {
-      if (!MI.isDebugValue() && !MI.getFlag(MachineInstr::FrameSetup) &&
+      if (!MI.isMetaInstruction() && !MI.getFlag(MachineInstr::FrameSetup) &&
           MI.getDebugLoc()) {
         PrologEndLoc = MI.getDebugLoc();
         break;
-      } else if (!MI.isDebugValue()) {
+      } else if (!MI.isMetaInstruction()) {
         EmptyPrologue = false;
       }
     }
@@ -1562,7 +1562,7 @@ TypeIndex CodeViewDebug::lowerTypeEnum(const DICompositeType *Ty) {
         EnumeratorCount++;
       }
     }
-    FTI = FLRB.end();
+    FTI = FLRB.end(true);
   }
 
   std::string FullName = getFullyQualifiedName(Ty);
@@ -1869,7 +1869,7 @@ CodeViewDebug::lowerRecordFieldList(const DICompositeType *Ty) {
     MemberCount++;
   }
 
-  TypeIndex FieldTI = FLBR.end();
+  TypeIndex FieldTI = FLBR.end(true);
   return std::make_tuple(FieldTI, Info.VShapeTI, MemberCount,
                          !Info.NestedClasses.empty());
 }

lib/CodeGen/AsmPrinter/DIEHash.cpp

@@ -116,65 +116,17 @@ void DIEHash::addParentContext(const DIE &Parent) {
 
 // Collect all of the attributes for a particular DIE in single structure.
 void DIEHash::collectAttributes(const DIE &Die, DIEAttrs &Attrs) {
-#define COLLECT_ATTR(NAME)                                                     \
-  case dwarf::NAME:                                                            \
-    Attrs.NAME = V;                                                            \
-    break
 
   for (const auto &V : Die.values()) {
     DEBUG(dbgs() << "Attribute: "
                  << dwarf::AttributeString(V.getAttribute())
                  << " added.\n");
     switch (V.getAttribute()) {
-      COLLECT_ATTR(DW_AT_name);
+#define HANDLE_DIE_HASH_ATTR(NAME)                                             \
-      COLLECT_ATTR(DW_AT_accessibility);
+  case dwarf::NAME:                                                            \
-      COLLECT_ATTR(DW_AT_address_class);
+    Attrs.NAME = V;                                                            \
-      COLLECT_ATTR(DW_AT_allocated);
+    break;
-      COLLECT_ATTR(DW_AT_artificial);
+#include "DIEHashAttributes.def"
-      COLLECT_ATTR(DW_AT_associated);
-      COLLECT_ATTR(DW_AT_binary_scale);
-      COLLECT_ATTR(DW_AT_bit_offset);
-      COLLECT_ATTR(DW_AT_bit_size);
-      COLLECT_ATTR(DW_AT_bit_stride);
-      COLLECT_ATTR(DW_AT_byte_size);
-      COLLECT_ATTR(DW_AT_byte_stride);
-      COLLECT_ATTR(DW_AT_const_expr);
-      COLLECT_ATTR(DW_AT_const_value);
-      COLLECT_ATTR(DW_AT_containing_type);
-      COLLECT_ATTR(DW_AT_count);
-      COLLECT_ATTR(DW_AT_data_bit_offset);
-      COLLECT_ATTR(DW_AT_data_location);
-      COLLECT_ATTR(DW_AT_data_member_location);
-      COLLECT_ATTR(DW_AT_decimal_scale);
-      COLLECT_ATTR(DW_AT_decimal_sign);
-      COLLECT_ATTR(DW_AT_default_value);
-      COLLECT_ATTR(DW_AT_digit_count);
-      COLLECT_ATTR(DW_AT_discr);
-      COLLECT_ATTR(DW_AT_discr_list);
-      COLLECT_ATTR(DW_AT_discr_value);
-      COLLECT_ATTR(DW_AT_encoding);
-      COLLECT_ATTR(DW_AT_enum_class);
-      COLLECT_ATTR(DW_AT_endianity);
-      COLLECT_ATTR(DW_AT_explicit);
-      COLLECT_ATTR(DW_AT_is_optional);
-      COLLECT_ATTR(DW_AT_location);
-      COLLECT_ATTR(DW_AT_lower_bound);
-      COLLECT_ATTR(DW_AT_mutable);
-      COLLECT_ATTR(DW_AT_ordering);
-      COLLECT_ATTR(DW_AT_picture_string);
-      COLLECT_ATTR(DW_AT_prototyped);
-      COLLECT_ATTR(DW_AT_small);
-      COLLECT_ATTR(DW_AT_segment);
-      COLLECT_ATTR(DW_AT_string_length);
-      COLLECT_ATTR(DW_AT_threads_scaled);
-      COLLECT_ATTR(DW_AT_upper_bound);
-      COLLECT_ATTR(DW_AT_use_location);
-      COLLECT_ATTR(DW_AT_use_UTF8);
-      COLLECT_ATTR(DW_AT_variable_parameter);
-      COLLECT_ATTR(DW_AT_virtuality);
-      COLLECT_ATTR(DW_AT_visibility);
-      COLLECT_ATTR(DW_AT_vtable_elem_location);
-      COLLECT_ATTR(DW_AT_type);
     default:
       break;
     }
@@ -366,62 +318,12 @@ void DIEHash::hashAttribute(const DIEValue &Value, dwarf::Tag Tag) {
 // Go through the attributes from \param Attrs in the order specified in 7.27.4
 // and hash them.
 void DIEHash::hashAttributes(const DIEAttrs &Attrs, dwarf::Tag Tag) {
-#define ADD_ATTR(ATTR)                                                         \
+#define HANDLE_DIE_HASH_ATTR(NAME)                                             \
   {                                                                            \
-    if (ATTR)                                                                  \
+    if (Attrs.NAME)                                                           \
-      hashAttribute(ATTR, Tag);                                                \
+      hashAttribute(Attrs.NAME, Tag);                                         \
   }
-
+#include "DIEHashAttributes.def"
-  ADD_ATTR(Attrs.DW_AT_name);
-  ADD_ATTR(Attrs.DW_AT_accessibility);
-  ADD_ATTR(Attrs.DW_AT_address_class);
-  ADD_ATTR(Attrs.DW_AT_allocated);
-  ADD_ATTR(Attrs.DW_AT_artificial);
-  ADD_ATTR(Attrs.DW_AT_associated);
-  ADD_ATTR(Attrs.DW_AT_binary_scale);
-  ADD_ATTR(Attrs.DW_AT_bit_offset);
-  ADD_ATTR(Attrs.DW_AT_bit_size);
-  ADD_ATTR(Attrs.DW_AT_bit_stride);
-  ADD_ATTR(Attrs.DW_AT_byte_size);
-  ADD_ATTR(Attrs.DW_AT_byte_stride);
-  ADD_ATTR(Attrs.DW_AT_const_expr);
-  ADD_ATTR(Attrs.DW_AT_const_value);
-  ADD_ATTR(Attrs.DW_AT_containing_type);
-  ADD_ATTR(Attrs.DW_AT_count);
-  ADD_ATTR(Attrs.DW_AT_data_bit_offset);
-  ADD_ATTR(Attrs.DW_AT_data_location);
-  ADD_ATTR(Attrs.DW_AT_data_member_location);
-  ADD_ATTR(Attrs.DW_AT_decimal_scale);
-  ADD_ATTR(Attrs.DW_AT_decimal_sign);
-  ADD_ATTR(Attrs.DW_AT_default_value);
-  ADD_ATTR(Attrs.DW_AT_digit_count);
-  ADD_ATTR(Attrs.DW_AT_discr);
-  ADD_ATTR(Attrs.DW_AT_discr_list);
-  ADD_ATTR(Attrs.DW_AT_discr_value);
-  ADD_ATTR(Attrs.DW_AT_encoding);
-  ADD_ATTR(Attrs.DW_AT_enum_class);
-  ADD_ATTR(Attrs.DW_AT_endianity);
-  ADD_ATTR(Attrs.DW_AT_explicit);
-  ADD_ATTR(Attrs.DW_AT_is_optional);
-  ADD_ATTR(Attrs.DW_AT_location);
-  ADD_ATTR(Attrs.DW_AT_lower_bound);
-  ADD_ATTR(Attrs.DW_AT_mutable);
-  ADD_ATTR(Attrs.DW_AT_ordering);
-  ADD_ATTR(Attrs.DW_AT_picture_string);
-  ADD_ATTR(Attrs.DW_AT_prototyped);
-  ADD_ATTR(Attrs.DW_AT_small);
-  ADD_ATTR(Attrs.DW_AT_segment);
-  ADD_ATTR(Attrs.DW_AT_string_length);
-  ADD_ATTR(Attrs.DW_AT_threads_scaled);
-  ADD_ATTR(Attrs.DW_AT_upper_bound);
-  ADD_ATTR(Attrs.DW_AT_use_location);
-  ADD_ATTR(Attrs.DW_AT_use_UTF8);
-  ADD_ATTR(Attrs.DW_AT_variable_parameter);
-  ADD_ATTR(Attrs.DW_AT_virtuality);
-  ADD_ATTR(Attrs.DW_AT_visibility);
-  ADD_ATTR(Attrs.DW_AT_vtable_elem_location);
-  ADD_ATTR(Attrs.DW_AT_type);
-
   // FIXME: Add the extended attributes.
 }
 
@@ -478,10 +380,12 @@ void DIEHash::computeHash(const DIE &Die) {
 /// DWARF4 standard. It is an md5 hash of the flattened description of the DIE
 /// with the inclusion of the full CU and all top level CU entities.
 // TODO: Initialize the type chain at 0 instead of 1 for CU signatures.
-uint64_t DIEHash::computeCUSignature(const DIE &Die) {
+uint64_t DIEHash::computeCUSignature(StringRef DWOName, const DIE &Die) {
   Numbering.clear();
   Numbering[&Die] = 1;
 
+  if (!DWOName.empty())
+    Hash.update(DWOName);
   // Hash the DIE.
   computeHash(Die);
 

lib/CodeGen/AsmPrinter/DIEHash.h

@@ -28,64 +28,15 @@ class CompileUnit;
 class DIEHash {
   // Collection of all attributes used in hashing a particular DIE.
   struct DIEAttrs {
-    DIEValue DW_AT_name;
+#define HANDLE_DIE_HASH_ATTR(NAME) DIEValue NAME;
-    DIEValue DW_AT_accessibility;
+#include "DIEHashAttributes.def"
-    DIEValue DW_AT_address_class;
-    DIEValue DW_AT_allocated;
-    DIEValue DW_AT_artificial;
-    DIEValue DW_AT_associated;
-    DIEValue DW_AT_binary_scale;
-    DIEValue DW_AT_bit_offset;
-    DIEValue DW_AT_bit_size;
-    DIEValue DW_AT_bit_stride;
-    DIEValue DW_AT_byte_size;
-    DIEValue DW_AT_byte_stride;
-    DIEValue DW_AT_const_expr;
-    DIEValue DW_AT_const_value;
-    DIEValue DW_AT_containing_type;
-    DIEValue DW_AT_count;
-    DIEValue DW_AT_data_bit_offset;
-    DIEValue DW_AT_data_location;
-    DIEValue DW_AT_data_member_location;
-    DIEValue DW_AT_decimal_scale;
-    DIEValue DW_AT_decimal_sign;
-    DIEValue DW_AT_default_value;
-    DIEValue DW_AT_digit_count;
-    DIEValue DW_AT_discr;
-    DIEValue DW_AT_discr_list;
-    DIEValue DW_AT_discr_value;
-    DIEValue DW_AT_encoding;
-    DIEValue DW_AT_enum_class;
-    DIEValue DW_AT_endianity;
-    DIEValue DW_AT_explicit;
-    DIEValue DW_AT_is_optional;
-    DIEValue DW_AT_location;
-    DIEValue DW_AT_lower_bound;
-    DIEValue DW_AT_mutable;
-    DIEValue DW_AT_ordering;
-    DIEValue DW_AT_picture_string;
-    DIEValue DW_AT_prototyped;
-    DIEValue DW_AT_small;
-    DIEValue DW_AT_segment;
-    DIEValue DW_AT_string_length;
-    DIEValue DW_AT_threads_scaled;
-    DIEValue DW_AT_upper_bound;
-    DIEValue DW_AT_use_location;
-    DIEValue DW_AT_use_UTF8;
-    DIEValue DW_AT_variable_parameter;
-    DIEValue DW_AT_virtuality;
-    DIEValue DW_AT_visibility;
-    DIEValue DW_AT_vtable_elem_location;
-    DIEValue DW_AT_type;
-
-    // Insert any additional ones here...
   };
 
 public:
   DIEHash(AsmPrinter *A = nullptr) : AP(A) {}
 
   /// \brief Computes the CU signature.
-  uint64_t computeCUSignature(const DIE &Die);
+  uint64_t computeCUSignature(StringRef DWOName, const DIE &Die);
 
   /// \brief Computes the type signature.
   uint64_t computeTypeSignature(const DIE &Die);

lib/CodeGen/AsmPrinter/DIEHashAttributes.def

@@ -0,0 +1,55 @@
+#ifndef HANDLE_DIE_HASH_ATTR
+#error "Missing macro definition of HANDLE_DIE_HASH_ATTR"
+#endif
+
+HANDLE_DIE_HASH_ATTR(DW_AT_name)
+HANDLE_DIE_HASH_ATTR(DW_AT_accessibility)
+HANDLE_DIE_HASH_ATTR(DW_AT_address_class)
+HANDLE_DIE_HASH_ATTR(DW_AT_allocated)
+HANDLE_DIE_HASH_ATTR(DW_AT_artificial)
+HANDLE_DIE_HASH_ATTR(DW_AT_associated)
+HANDLE_DIE_HASH_ATTR(DW_AT_binary_scale)
+HANDLE_DIE_HASH_ATTR(DW_AT_bit_offset)
+HANDLE_DIE_HASH_ATTR(DW_AT_bit_size)
+HANDLE_DIE_HASH_ATTR(DW_AT_bit_stride)
+HANDLE_DIE_HASH_ATTR(DW_AT_byte_size)
+HANDLE_DIE_HASH_ATTR(DW_AT_byte_stride)
+HANDLE_DIE_HASH_ATTR(DW_AT_const_expr)
+HANDLE_DIE_HASH_ATTR(DW_AT_const_value)
+HANDLE_DIE_HASH_ATTR(DW_AT_containing_type)
+HANDLE_DIE_HASH_ATTR(DW_AT_count)
+HANDLE_DIE_HASH_ATTR(DW_AT_data_bit_offset)
+HANDLE_DIE_HASH_ATTR(DW_AT_data_location)
+HANDLE_DIE_HASH_ATTR(DW_AT_data_member_location)
+HANDLE_DIE_HASH_ATTR(DW_AT_decimal_scale)
+HANDLE_DIE_HASH_ATTR(DW_AT_decimal_sign)
+HANDLE_DIE_HASH_ATTR(DW_AT_default_value)
+HANDLE_DIE_HASH_ATTR(DW_AT_digit_count)
+HANDLE_DIE_HASH_ATTR(DW_AT_discr)
+HANDLE_DIE_HASH_ATTR(DW_AT_discr_list)
+HANDLE_DIE_HASH_ATTR(DW_AT_discr_value)
+HANDLE_DIE_HASH_ATTR(DW_AT_encoding)
+HANDLE_DIE_HASH_ATTR(DW_AT_enum_class)
+HANDLE_DIE_HASH_ATTR(DW_AT_endianity)
+HANDLE_DIE_HASH_ATTR(DW_AT_explicit)
+HANDLE_DIE_HASH_ATTR(DW_AT_is_optional)
+HANDLE_DIE_HASH_ATTR(DW_AT_location)
+HANDLE_DIE_HASH_ATTR(DW_AT_lower_bound)
+HANDLE_DIE_HASH_ATTR(DW_AT_mutable)
+HANDLE_DIE_HASH_ATTR(DW_AT_ordering)
+HANDLE_DIE_HASH_ATTR(DW_AT_picture_string)
+HANDLE_DIE_HASH_ATTR(DW_AT_prototyped)
+HANDLE_DIE_HASH_ATTR(DW_AT_small)
+HANDLE_DIE_HASH_ATTR(DW_AT_segment)
+HANDLE_DIE_HASH_ATTR(DW_AT_string_length)
+HANDLE_DIE_HASH_ATTR(DW_AT_threads_scaled)
+HANDLE_DIE_HASH_ATTR(DW_AT_upper_bound)
+HANDLE_DIE_HASH_ATTR(DW_AT_use_location)
+HANDLE_DIE_HASH_ATTR(DW_AT_use_UTF8)
+HANDLE_DIE_HASH_ATTR(DW_AT_variable_parameter)
+HANDLE_DIE_HASH_ATTR(DW_AT_virtuality)
+HANDLE_DIE_HASH_ATTR(DW_AT_visibility)
+HANDLE_DIE_HASH_ATTR(DW_AT_vtable_elem_location)
+HANDLE_DIE_HASH_ATTR(DW_AT_type)
+
+#undef HANDLE_DIE_HASH_ATTR