WebAssembly: basic instructions todo, and basic register info.

[jfb]

Summary

This code is based on AArch64 for modern backend good practice, and NVPTX for
virtual ISA concerns.

[jfb]

I'll do a follow-up PR to address the FIXME on needsStackRealignment.

[sunfish]

Cool. Here's a first round of comments.

lib/Target/WebAssembly/CMakeLists.txt

3

Please keep these rules alphabetized.

lib/Target/WebAssembly/Makefile

16

Common practice in these Makefiles is to indent subsequent lines to be to the right of the '='.

lib/Target/WebAssembly/WebAssemblyInstrAtomics.td

12

For full doxygen happiness, these should get triple-slash comments, as is done in other files.

lib/Target/WebAssembly/WebAssemblyInstrCall.td

1

If we have extra characters available in the top-line comment, it's nice to leave in some of the '='s that LLVM typically uses at the beginning and end.

lib/Target/WebAssembly/WebAssemblyInstrConv.td

1

"Conv" in closely-related contexts is short for Convention, as in CallingConv. I'm ok with the filename here, but can you make the top-level comment clarify that this is about conversions?

lib/Target/WebAssembly/WebAssemblyInstrInfo.td

40

Is there a reason for reordering these? Operands and DAG Nodes and DAG Node Types are all inter-related so it's nice to have them together.

50

If comma and conditional end up being included in WebAssembly, they'll have embedded control flow, so we won't be able to represent them explicitly in either SelectionDAG or MachineInstrs.

lib/Target/WebAssembly/WebAssemblyRegisterInfo.cpp

49

This should also set FP32/FP64 as reserved, at least when TFI->hasFP(), but maybe always, since these are virtual registers anyway.

56

report_fatal_error is for legitimate conditions that codegen can't handle, not for reporting bugs in codegen. llvm_unreachable is more appropriate here because it's just a "bug" that this function isn't implemented yet.

61

Using unsigned instead of unsigned char for holding register values.

65

We should static_cast this to WebAssemblyFrameLowering so that the hasFP call can be easily devirtualized.

lib/Target/WebAssembly/WebAssemblyRegisterInfo.td

13

More triple-slash comments.

30

WebAssembly won't support it right away, and it's not yet decided whether it will support it in the future. But actually, we need separate registers anyway so that we can put them in separate register classes.

43

This code comes from NVPTXRegisterInfo.td, and we are doing something similar here, but I'm curious what it's really needed for. Could we try leaving it out and seeing what breaks?

[jfb]

Rebase.

[jfb]

• Address sunfish's comments.

[jfb]

Updated, PTAL.

lib/Target/WebAssembly/Makefile

16

Those are tabs (same as for other Makefiles), they just don't appear nicely in phab.

lib/Target/WebAssembly/WebAssemblyInstrInfo.td

50

Good point, I removed them.

lib/Target/WebAssembly/WebAssemblyRegisterInfo.cpp

49

Oops, I added FP later and forgot to add it here too. Changed to a range-based for loop too.

61

Imagine all the bits we're wasting! ;-)

65

I'll send a separate PR to do the same in other backends.

lib/Target/WebAssembly/WebAssemblyRegisterInfo.td

30

Are you agreeing, or suggesting something?

43

I'll add a todo instead to try it out later, if needed? I'd rather follow what others have done, and then try to change it, seems like we'll have less WTF moments that way.

[sunfish]

Looks good.

lib/Target/WebAssembly/WebAssemblyRegisterInfo.td

30

I think I'm just suggesting the comment say something like "may someday support" instead of "supports".

[jfb]

• Address sunfish's comments.
• Use the shiny new devirtualized getFrameLowering from D11093.
• 'May someday support'.

[jfb]

Clarified the comment, and updated the patch to use devirtualized frame lowering.

lib/Target/WebAssembly/CMakeLists.txt

 set(LLVM_TARGET_DEFINITIONS WebAssembly.td)
 
+tablegen(LLVM WebAssemblyGenRegisterInfo.inc -gen-register-info)
 tablegen(LLVM WebAssemblyGenMCCodeEmitter.inc -gen-emitter)
 tablegen(LLVM WebAssemblyGenSubtargetInfo.inc -gen-subtarget)
 add_public_tablegen_target(WebAssemblyCommonTableGen)
 
 add_llvm_target(WebAssemblyCodeGen
   WebAssemblyFrameLowering.cpp
   WebAssemblyInstrInfo.cpp
   WebAssemblyISelDAGToDAG.cpp

lib/Target/WebAssembly/MCTargetDesc/WebAssemblyMCTargetDesc.h

MCAsmBackend *createWebAssemblyAsmBackend(const Target &T,
                                           const MCRegisterInfo &MRI,
                                           StringRef TT, StringRef CPU);
 
 } // end namespace llvm
 
 // Defines symbolic names for WebAssembly registers. This defines a mapping from
 // register name to register number.
 //
+#define GET_REGINFO_ENUM
+#include "WebAssemblyGenRegisterInfo.inc"
+
 #define GET_SUBTARGETINFO_ENUM
 #include "WebAssemblyGenSubtargetInfo.inc"
 
 #endif

lib/Target/WebAssembly/MCTargetDesc/WebAssemblyMCTargetDesc.cpp

 #include "llvm/Support/TargetRegistry.h"
 using namespace llvm;
 
 #define DEBUG_TYPE "wasm-mc-target-desc"
 
 #define GET_SUBTARGETINFO_MC_DESC
 #include "WebAssemblyGenSubtargetInfo.inc"
 
+#define GET_REGINFO_MC_DESC
+#include "WebAssemblyGenRegisterInfo.inc"
+
 static MCAsmInfo *createWebAssemblyMCAsmInfo(const MCRegisterInfo &MRI,
                                              const Triple &TT) {
   MCAsmInfo *MAI = new WebAssemblyMCAsmInfo(TT);
   return MAI;
 }
 
 static MCInstPrinter *
 createWebAssemblyMCInstPrinter(const Triple &T, unsigned SyntaxVariant,

lib/Target/WebAssembly/Makefile

 ##===- lib/Target/WebAssembly/Makefile ---------------------*- Makefile -*-===##
 #
 #                     The LLVM Compiler Infrastructure
 #
 # This file is distributed under the University of Illinois Open Source
 # License. See LICENSE.TXT for details.
 #
 ##===----------------------------------------------------------------------===##
 
 LEVEL = ../../..
 LIBRARYNAME = LLVMWebAssemblyCodeGen
 TARGET = WebAssembly
 
 # Make sure that tblgen is run, first thing.
-BUILT_SOURCES = WebAssemblyGenSubtargetInfo.inc WebAssemblyGenMCCodeEmitter.inc
+BUILT_SOURCES = WebAssemblyGenRegisterInfo.inc WebAssemblyGenSubtargetInfo.inc \
+    WebAssemblyGenMCCodeEmitter.inc
 
 DIRS = InstPrinter TargetInfo MCTargetDesc
 
 include $(LEVEL)/Makefile.common

lib/Target/WebAssembly/WebAssemblyInstrAtomics.td

 // WebAssemblyInstrAtomics.td-WebAssembly Atomic codegen support-*- tablegen -*-
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
-// WebAssembly Atomic operand code-gen constructs.
+// \file
+// \brief WebAssembly Atomic operand code-gen constructs.
 //
 //===----------------------------------------------------------------------===//
 
 // TODO: Implement atomic instructions.
 
 //===----------------------------------------------------------------------===//
 // Atomic fences
 //===----------------------------------------------------------------------===//
 

lib/Target/WebAssembly/WebAssemblyInstrCall.td

-// WebAssemblyInstrSIMD.td - WebAssembly SIMD codegen support -*- tablegen -*-//
+// WebAssemblyInstrCall.td-WebAssembly Call codegen support -----*- tablegen -*-
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
-// WebAssembly SIMD operand code-gen constructs.
+// \file
+// \brief WebAssembly Call operand code-gen constructs.
 //
 //===----------------------------------------------------------------------===//
 
-// TODO: Implement SIMD instructions.
-// Note: use Requires<[HasSIMD128]>.
+/*
+ * TODO(jfb): Add the following.
+ *
+ * call_direct: call function directly
+ * call_indirect: call function indirectly
+ * addressof: obtain a function pointer value for a given function
+ */

lib/Target/WebAssembly/WebAssemblyInstrConv.td

+// WebAssemblyInstrConv.td-WebAssembly Conv codegen support -----*- tablegen -*-
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// \file
+// \brief WebAssembly datatype conversions, truncations, reinterpretations, promotions,
+// and demotions operand code-gen constructs.
+//
+//===----------------------------------------------------------------------===//
+
+/*
+ * TODO(jfb): Add the following.
+ *
+ * int32.wrap[int64]: wrap a 64-bit integer to a 32-bit integer
+ * int32.trunc_signed[float32]: truncate a 32-bit float to a signed 32-bit integer
+ * int32.trunc_signed[float64]: truncate a 64-bit float to a signed 32-bit integer
+ * int32.trunc_unsigned[float32]: truncate a 32-bit float to an unsigned 32-bit integer
+ * int32.trunc_unsigned[float64]: truncate a 64-bit float to an unsigned 32-bit integer
+ * int32.reinterpret[float32]: reinterpret the bits of a 32-bit float as a 32-bit integer
+ * int64.extend_signed[int32]: extend a signed 32-bit integer to a 64-bit integer
+ * int64.extend_unsigned[int32]: extend an unsigned 32-bit integer to a 64-bit integer
+ * int64.trunc_signed[float32]: truncate a 32-bit float to a signed 64-bit integer
+ * int64.trunc_signed[float64]: truncate a 64-bit float to a signed 64-bit integer
+ * int64.trunc_unsigned[float32]: truncate a 32-bit float to an unsigned 64-bit integer
+ * int64.trunc_unsigned[float64]: truncate a 64-bit float to an unsigned 64-bit integer
+ * int64.reinterpret[float64]: reinterpret the bits of a 64-bit float as a 64-bit integer
+ * float32.demote[float64]: demote a 64-bit float to a 32-bit float
+ * float32.cvt_signed[int32]: convert a signed 32-bit integer to a 32-bit float
+ * float32.cvt_signed[int64]: convert a signed 64-bit integer to a 32-bit float
+ * float32.cvt_unsigned[int32]: convert an unsigned 32-bit integer to a 32-bit float
+ * float32.cvt_unsigned[int64]: convert an unsigned 64-bit integer to a 32-bit float
+ * float32.reinterpret[int32]: reinterpret the bits of a 32-bit integer as a 32-bit float
+ * float64.promote[float32]: promote a 32-bit float to a 64-bit float
+ * float64.cvt_signed[int32]: convert a signed 32-bit integer to a 64-bit float
+ * float64.cvt_signed[int64]: convert a signed 64-bit integer to a 64-bit float
+ * float64.cvt_unsigned[int32]: convert an unsigned 32-bit integer to a 64-bit float
+ * float64.cvt_unsigned[int64]: convert an unsigned 64-bit integer to a 64-bit float
+ * float64.reinterpret[int64]: reinterpret the bits of a 64-bit integer as a 64-bit float
+ */

lib/Target/WebAssembly/WebAssemblyInstrFloat.td

+// WebAssemblyInstrFloat.td-WebAssembly Float codegen support ---*- tablegen -*-
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// \file
+// \brief WebAssembly Floating-point operand code-gen constructs.
+//
+//===----------------------------------------------------------------------===//
+
+/*
+ * TODO(jfb): Add the following for 32-bit and 64-bit.
+ *
+ * float32.add: addition
+ * float32.sub: subtraction
+ * float32.mul: multiplication
+ * float32.div: division
+ * float32.abs: absolute value
+ * float32.neg: negation
+ * float32.copysign: copysign
+ * float32.ceil: ceiling operation
+ * float32.floor: floor operation
+ * float32.trunc: round to nearest integer towards zero
+ * float32.nearestint: round to nearest integer, ties to even
+ * float32.eq: compare equal
+ * float32.lt: less than
+ * float32.le: less than or equal
+ * float32.gt: greater than
+ * float32.ge: greater than or equal
+ * float32.sqrt: square root
+ * float32.min: minimum (binary operator); if either operand is NaN, returns NaN
+ * float32.max: maximum (binary operator); if either operand is NaN, returns NaN
+ */

lib/Target/WebAssembly/WebAssemblyInstrFormats.td

 // WebAssemblyInstrFormats.td - WebAssembly Instruction Formats -*- tblgen -*-//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
-// WebAssembly instruction format definitions.
+// \file
+// \brief WebAssembly instruction format definitions.
 //
 //===----------------------------------------------------------------------===//
 
 // WebAssembly Instruction Format
 class WebAssemblyInst<string cstr> : Instruction {
   field bits<0> Inst; // Instruction encoding.
   let Namespace   = "WebAssembly";
   let Pattern     = [];

lib/Target/WebAssembly/WebAssemblyInstrInfo.td

 // WebAssemblyInstrInfo.td-Describe the WebAssembly Instructions-*- tablegen -*-
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
-// WebAssembly Instruction definitions.
+// \file
+// \brief WebAssembly Instruction definitions.
 //
 //===----------------------------------------------------------------------===//
 
 //===----------------------------------------------------------------------===//
 // WebAssembly Instruction Predicate Definitions.
 //===----------------------------------------------------------------------===//
 
 def HasAddr32 : Predicate<"!Subtarget->hasAddr64()">;
 def HasAddr64 : Predicate<"Subtarget->hasAddr64()">;
 def HasSIMD128 : Predicate<"Subtarget->hasSIMD128()">,
                            AssemblerPredicate<"FeatureSIMD128", "simd128">;
 
 //===----------------------------------------------------------------------===//
 // WebAssembly-specific DAG Node Types.
 //===----------------------------------------------------------------------===//
 
 //===----------------------------------------------------------------------===//
 // WebAssembly-specific DAG Nodes.
 //===----------------------------------------------------------------------===//
 
 //===----------------------------------------------------------------------===//
-// WebAssembly-specific Operands.
+// WebAssembly Instruction Format Definitions.
 //===----------------------------------------------------------------------===//
 
+include "WebAssemblyInstrFormats.td"
+
 //===----------------------------------------------------------------------===//
-// WebAssembly Instruction Format Definitions.
+// WebAssembly-specific Operands.
 //===----------------------------------------------------------------------===//
 
-include "WebAssemblyInstrFormats.td"
+/*
+ * TODO(jfb): Add the following.
+ *
+ * get_local: read the current value of a local variable
+ * set_local: set the current value of a local variable
+ *
+ * comma: evaluate and ignore the result of the first operand, evaluate and
+ *        return the second operand
+ * conditional: basically ternary ?: operator
+*/
 
 //===----------------------------------------------------------------------===//
 // Additional sets of instructions.
 //===----------------------------------------------------------------------===//
 
+include "WebAssemblyInstrMemory.td"
+include "WebAssemblyInstrCall.td"
+include "WebAssemblyInstrInteger.td"
+include "WebAssemblyInstrFloat.td"
+include "WebAssemblyInstrConv.td"
 include "WebAssemblyInstrAtomics.td"
 include "WebAssemblyInstrSIMD.td"

lib/Target/WebAssembly/WebAssemblyInstrInteger.td

+// WebAssemblyInstrInteger.td-WebAssembly Integer codegen -------*- tablegen -*-
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// \file
+// \brief WebAssembly Integer operand code-gen constructs.
+//
+//===----------------------------------------------------------------------===//
+
+/*
+ * TODO(jfb): Add the following for 32-bit and 64-bit.
+ *
+ * int32.add: signed-less addition
+ * int32.sub: signed-less subtraction
+ * int32.mul: signed-less multiplication (lower 32-bits)
+ * int32.sdiv: signed division
+ * int32.udiv: unsigned division
+ * int32.srem: signed remainder
+ * int32.urem: unsigned remainder
+ * int32.and: signed-less logical and
+ * int32.ior: signed-less inclusive or
+ * int32.xor: signed-less exclusive or
+ * int32.shl: signed-less shift left
+ * int32.shr: signed-less logical shift right
+ * int32.sar: signed-less arithmetic shift right
+ * int32.eq: signed-less compare equal
+ * int32.slt: signed less than
+ * int32.sle: signed less than or equal
+ * int32.ult: unsigned less than
+ * int32.ule: unsigned less than or equal
+ * int32.sgt: signed greater than
+ * int32.sge: signed greater than or equal
+ * int32.ugt: unsigned greater than
+ * int32.uge: unsigned greater than or equal
+ * int32.clz: count leading zeroes (defined for all values, including zero)
+ * int32.ctz: count trailing zeroes (defined for all values, including zero)
+ * int32.popcnt: count number of ones
+ */

lib/Target/WebAssembly/WebAssemblyInstrMemory.td

+// WebAssemblyInstrMemory.td-WebAssembly Memory codegen support -*- tablegen -*-
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// \file
+// \brief WebAssembly Memory operand code-gen constructs.
+//
+//===----------------------------------------------------------------------===//
+
+/*
+ * TODO(jfb): Add the following.
+ * Each has optional alignment and immediate byte offset.
+ *
+ * int32.load_sx[int8]: sign-extend to int32
+ * int32.load_sx[int16]: sign-extend to int32
+ * int32.load_zx[int8]: zero-extend to int32
+ * int32.load_zx[int16]: zero-extend to int32
+ * int32.load[int32]: (no conversion)
+ * int64.load_sx[int8]: sign-extend to int64
+ * int64.load_sx[int16]: sign-extend to int64
+ * int64.load_sx[int32]: sign-extend to int64
+ * int64.load_zx[int8]: zero-extend to int64
+ * int64.load_zx[int16]: zero-extend to int64
+ * int64.load_zx[int32]: zero-extend to int64
+ * int64.load[int64]: (no conversion)
+ * float32.load[float32]: (no conversion)
+ * float64.load[float64]: (no conversion)
+ * 
+ * int32.store[int8]: wrap int32 to int8
+ * int32.store[int16]: wrap int32 to int16
+ * int32.store[int32]: (no conversion)
+ * int64.store[int8]: wrap int64 to int8
+ * int64.store[int16]: wrap int64 to int16
+ * int64.store[int32]: wrap int64 to int32
+ * int64.store[int64]: (no conversion)
+ * float32.store[float32]: (no conversion)
+ * float64.store[float64]: (no conversion)
+ * 
+ * load_global: load the value of a given global variable
+ * store_global: store a given value to a given global variable
+ */

lib/Target/WebAssembly/WebAssemblyInstrSIMD.td

 // WebAssemblyInstrSIMD.td - WebAssembly SIMD codegen support -*- tablegen -*-//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
-// WebAssembly SIMD operand code-gen constructs.
+// \file
+// \brief WebAssembly SIMD operand code-gen constructs.
 //
 //===----------------------------------------------------------------------===//
 
 // TODO: Implement SIMD instructions.
 // Note: use Requires<[HasSIMD128]>.

lib/Target/WebAssembly/WebAssemblyRegisterInfo.h

 /// \brief This file contains the WebAssembly implementation of the
 /// WebAssemblyRegisterInfo class.
 ///
 //===----------------------------------------------------------------------===//
 
 #ifndef LLVM_LIB_TARGET_WEBASSEMBLY_WEBASSEMBLYREGISTERINFO_H
 #define LLVM_LIB_TARGET_WEBASSEMBLY_WEBASSEMBLYREGISTERINFO_H
 
+#define GET_REGINFO_HEADER
+#include "WebAssemblyGenRegisterInfo.inc"
+
 namespace llvm {
 
 class MachineFunction;
 class RegScavenger;
 class TargetRegisterClass;
 class Triple;
 
-class WebAssemblyRegisterInfo final {
+class WebAssemblyRegisterInfo final : public WebAssemblyGenRegisterInfo {
   const Triple &TT;
 
 public:
   explicit WebAssemblyRegisterInfo(const Triple &TT);
+
+  // Code Generation virtual methods.
+  const MCPhysReg *getCalleeSavedRegs(const MachineFunction *MF) const override;
+  BitVector getReservedRegs(const MachineFunction &MF) const override;
+  void eliminateFrameIndex(MachineBasicBlock::iterator MI, int SPAdj,
+                           unsigned FIOperandNum,
+                           RegScavenger *RS = nullptr) const override;
+
+  // Debug information queries.
+  unsigned getFrameRegister(const MachineFunction &MF) const override;
+
+  // Base pointer (stack realignment) support.
+  bool canRealignStack(const MachineFunction &MF) const;
+  bool needsStackRealignment(const MachineFunction &MF) const override;
 };
 
 } // end namespace llvm
 
 #endif

lib/Target/WebAssembly/WebAssemblyRegisterInfo.cpp

 #include "llvm/IR/Function.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Target/TargetFrameLowering.h"
 #include "llvm/Target/TargetOptions.h"
 using namespace llvm;
 
 #define DEBUG_TYPE "wasm-reg-info"
 
-WebAssemblyRegisterInfo::WebAssemblyRegisterInfo(const Triple &TT) : TT(TT) {}
+#define GET_REGINFO_TARGET_DESC
+#include "WebAssemblyGenRegisterInfo.inc"
+
+WebAssemblyRegisterInfo::WebAssemblyRegisterInfo(const Triple &TT)
+    : WebAssemblyGenRegisterInfo(0), TT(TT) {}
+
+const MCPhysReg *
+WebAssemblyRegisterInfo::getCalleeSavedRegs(const MachineFunction *) const {
+  static const MCPhysReg CalleeSavedRegs[] = {0};
+  return CalleeSavedRegs;
+}
+
+BitVector
+WebAssemblyRegisterInfo::getReservedRegs(const MachineFunction &MF) const {
+  BitVector Reserved(getNumRegs());
+  Reserved.set(WebAssembly::SP32);
+  Reserved.set(WebAssembly::SP64);
+  return Reserved;
+}
+
+void WebAssemblyRegisterInfo::eliminateFrameIndex(
+    MachineBasicBlock::iterator II, int SPAdj, unsigned FIOperandNum,
+    RegScavenger *RS) const {
+  report_fatal_error("WebAssemblyRegisterInfo::eliminateFrameIndex"); // FIXME
+}
+
+unsigned
+WebAssemblyRegisterInfo::getFrameRegister(const MachineFunction &MF) const {
+  static const unsigned char Regs[2][2] = {
+      /*            !isArch64Bit       isArch64Bit      */
+      /* !hasFP */ {WebAssembly::SP32, WebAssembly::SP64},
+      /*  hasFP */ {WebAssembly::FP32, WebAssembly::FP64}};
+  const TargetFrameLowering *TFI = MF.getSubtarget().getFrameLowering();
+  return Regs[TFI->hasFP(MF)][TT.isArch64Bit()];
+}
+
+bool WebAssemblyRegisterInfo::canRealignStack(const MachineFunction &MF) const {
+  return !MF.getFunction()->hasFnAttribute("no-realign-stack");
+}
+
+// FIXME: share this with other backends with identical implementation?
+bool WebAssemblyRegisterInfo::needsStackRealignment(
+    const MachineFunction &MF) const {
+  const MachineFrameInfo *MFI = MF.getFrameInfo();
+  const Function *F = MF.getFunction();
+  unsigned StackAlign = MF.getTarget()
+                            .getSubtargetImpl(*MF.getFunction())
+                            ->getFrameLowering()
+                            ->getStackAlignment();
+  bool requiresRealignment =
+      ((MFI->getMaxAlignment() > StackAlign) ||
+       F->getAttributes().hasAttribute(AttributeSet::FunctionIndex,
+                                       Attribute::StackAlignment));
+
+  return requiresRealignment && canRealignStack(MF);
+}

lib/Target/WebAssembly/WebAssemblyRegisterInfo.td

 //WebAssemblyRegisterInfo.td-Describe the WebAssembly Registers -*- tablegen -*-
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
-// This file describes the WebAssembly register classes and some nominal
+// \file
+// \brief This file describes the WebAssembly register classes and some nominal
 // physical registers.
 //
 //===----------------------------------------------------------------------===//
 
 class WebAssemblyReg<string n> : Register<n> {
   let Namespace = "WebAssembly";
 }
 
 class WebAssemblyRegClass<list<ValueType> regTypes, int alignment, dag regList>
      : RegisterClass<"WebAssembly", regTypes, alignment, regList>;
 
 //===----------------------------------------------------------------------===//
 // Registers
 //===----------------------------------------------------------------------===//
 
+// Special registers used as the frame and stack pointer.
+//
+// WebAssembly supports mixed 32-bit and 64-bit heaps in the same application,
+// which requires separate width FP and SP.
+def FP32 : WebAssemblyReg<"%FP32">;
+def FP64 : WebAssemblyReg<"%FP64">;
+def SP32 : WebAssemblyReg<"%SP32">;
+def SP64 : WebAssemblyReg<"%SP64">;
+
+// WebAssembly uses virtual registers, but the backend defines a few physical
+// registers here to keep SDAG and the MachineInstr layers happy.
+foreach i = 0-4 in {
+  def I#i : WebAssemblyReg<"%i."#i>; // i32
+  def L#i : WebAssemblyReg<"%l."#i>; // i64
+  def F#i : WebAssemblyReg<"%f."#i>; // f32
+  def D#i : WebAssemblyReg<"%d."#i>; // f64
+}
+
 //===----------------------------------------------------------------------===//
 //  Register classes
 //===----------------------------------------------------------------------===//
+
+def Int32 : WebAssemblyRegClass<[i32], 32, (add (sequence "I%u", 0, 4), SP32)>;
+def Int64 : WebAssemblyRegClass<[i64], 64, (add (sequence "L%u", 0, 4), SP64)>;
+def Float32 : WebAssemblyRegClass<[f32], 32, (add (sequence "F%u", 0, 4))>;
+def Float64 : WebAssemblyRegClass<[f64], 64, (add (sequence "D%u", 0, 4))>;