Index: lib/Target/WebAssembly/WebAssemblyAsmPrinter.cpp
===================================================================
--- lib/Target/WebAssembly/WebAssemblyAsmPrinter.cpp
+++ lib/Target/WebAssembly/WebAssemblyAsmPrinter.cpp
@@ -211,6 +211,10 @@
       OutStreamer->AddBlankLine();
     }
     break;
+  case WebAssembly::COMPILER_FENCE:
+    // This is a compiler barrier that prevents instruction reordering during
+    // backend compilation, and should not be emitted.
+    break;
   default: {
     WebAssemblyMCInstLower MCInstLowering(OutContext, *this);
     MCInst TmpInst;
Index: lib/Target/WebAssembly/WebAssemblyISelDAGToDAG.cpp
===================================================================
--- lib/Target/WebAssembly/WebAssemblyISelDAGToDAG.cpp
+++ lib/Target/WebAssembly/WebAssemblyISelDAGToDAG.cpp
@@ -75,14 +75,95 @@
     return;
   }
 
-  // Few custom selection stuff. If we need WebAssembly-specific selection,
-  // uncomment this block add corresponding case statements.
-  /*
+  // Few custom selection stuff.
+  SDLoc DL(Node);
+  MachineFunction &MF = CurDAG->getMachineFunction();
   switch (Node->getOpcode()) {
+  case ISD::ATOMIC_FENCE: {
+    if (!MF.getSubtarget<WebAssemblySubtarget>().hasAtomics())
+      break;
+
+    uint64_t SyncScopeID =
+        cast<ConstantSDNode>(Node->getOperand(2).getNode())->getZExtValue();
+    switch (SyncScopeID) {
+    case SyncScope::SingleThread: {
+      // We lower a single-thread fence to a pseudo compiler barrier instruction
+      // preventing instruction reordering. This will not be emitted in final
+      // binary.
+      MachineSDNode *Fence =
+          CurDAG->getMachineNode(WebAssembly::COMPILER_FENCE,
+                                 DL,                 // debug loc
+                                 MVT::Other,         // outchain type
+                                 Node->getOperand(0) // inchain
+          );
+      ReplaceNode(Node, Fence);
+      CurDAG->RemoveDeadNode(Node);
+      return;
+    }
+
+    case SyncScope::System: {
+      // Wasm does not have a fence instruction, but because all atomic
+      // instructions in wasm are sequentially consistent, we translate a fence
+      // to an idempotent atomic RMW instruction to a linear memory address.
+      // All atomic instructions in wasm are sequentially consistent, but this
+      // is to ensure a fence also prevents reordering of non-atomic
+      // instructions in the VM. Even though LLVM IR's fence instruction does
+      // not say anything about its relationship with non-atomic instructions,
+      // we think this is more user-friendly.
+      //
+      // While any address can work, here we use a value stored in
+      // __stack_pointer wasm global because there's high chance that area is in
+      // cache.
+      //
+      // So the selected instructions will be in the form of:
+      //   %addr = get_global $__stack_pointer
+      //   i32.atomic.rmw.or %addr, 0
+      SDValue StackPtrSym = CurDAG->getTargetExternalSymbol(
+          "__stack_pointer", TLI->getPointerTy(CurDAG->getDataLayout()),
+          WebAssemblyII::MO_SYMBOL_GLOBAL);
+      MachineSDNode *GetGlobal =
+          CurDAG->getMachineNode(WebAssembly::GET_GLOBAL_I32, // opcode
+                                 DL,                          // debug loc
+                                 MVT::i32,                    // result type
+                                 StackPtrSym // __stack_pointer symbol
+          );
+
+      SDValue Zero = CurDAG->getTargetConstant(0, DL, MVT::i32);
+      auto *MMO = MF.getMachineMemOperand(
+          MachinePointerInfo::getUnknownStack(MF),
+          // FIXME Volatile isn't really correct, but currently all LLVM atomic
+          // instructions are treated as volatiles in the backend, so we should
+          // be consistent.
+          MachineMemOperand::MOVolatile | MachineMemOperand::MOLoad |
+              MachineMemOperand::MOStore,
+          4, 4, AAMDNodes(), nullptr, SyncScope::System,
+          AtomicOrdering::SequentiallyConsistent);
+      MachineSDNode *AtomicRMW =
+          CurDAG->getMachineNode(WebAssembly::ATOMIC_RMW_OR_I32, // opcode
+                                 DL,                             // debug loc
+                                 MVT::i32,                       // result type
+                                 MVT::Other, // outchain type
+                                 {
+                                     Zero,                  // alignment
+                                     Zero,                  // offset
+                                     SDValue(GetGlobal, 0), // __stack_pointer
+                                     Zero, // OR with 0 to make it idempotent
+                                     Node->getOperand(0) // inchain
+                                 });
+
+      CurDAG->setNodeMemRefs(AtomicRMW, {MMO});
+      ReplaceUses(SDValue(Node, 0), SDValue(AtomicRMW, 1));
+      CurDAG->RemoveDeadNode(Node);
+      return;
+    }
+    default:
+      llvm_unreachable("Unknown scope!");
+    }
+  }
+
   default:
     break;
   }
-  */
 
   // Select the default instruction.
   SelectCode(Node);
Index: lib/Target/WebAssembly/WebAssemblyInstrAtomics.td
===================================================================
--- lib/Target/WebAssembly/WebAssemblyInstrAtomics.td
+++ lib/Target/WebAssembly/WebAssemblyInstrAtomics.td
@@ -1025,3 +1025,13 @@
 def : WaitPatExternSymOffOnly<i32, int_wasm_atomic_wait_i32, ATOMIC_WAIT_I32>;
 def : WaitPatExternSymOffOnly<i64, int_wasm_atomic_wait_i64, ATOMIC_WAIT_I64>;
 } // Predicates = [HasAtomics]
+
+//===----------------------------------------------------------------------===//
+// Atomic fences
+//===----------------------------------------------------------------------===//
+
+// A compiler fence instruction that prevents reordering of instructions.
+let Defs = [ARGUMENTS] in {
+let isPseudo = 1, hasSideEffects = 1 in
+defm COMPILER_FENCE : NRI<(outs), (ins), [], "compiler_fence">;
+} // Defs = [ARGUMENTS]
Index: test/CodeGen/WebAssembly/atomic-fence.ll
===================================================================
--- /dev/null
+++ test/CodeGen/WebAssembly/atomic-fence.ll
@@ -0,0 +1,41 @@
+; RUN: not llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt
+; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -disable-wasm-explicit-locals -mattr=+atomics,+sign-ext | FileCheck %s
+
+target datalayout = "e-m:e-p:32:32-i64:64-n32:64-S128"
+target triple = "wasm32-unknown-unknown"
+
+; A multithread fence turns into 'get_global $__stack_pointer' followed by an
+; idempotent atomicrmw instruction.
+; CHECK-LABEL: multithread_fence:
+; CHECK:  get_global  $push0=, __stack_pointer
+; CHECK-NEXT:  i32.atomic.rmw.or  $drop=, 0($pop0), 0
+define void @multithread_fence() {
+  fence seq_cst
+  ret void
+}
+
+; Fences with weaker memory orderings than seq_cst should be treated the same
+; because atomic memory access in wasm are sequentially consistent.
+; CHECK-LABEL: multithread_weak_fence:
+; CHECK:  get_global  $push{{.+}}=, __stack_pointer
+; CHECK:  i32.atomic.rmw.or
+; CHECK:  i32.atomic.rmw.or
+; CHECK:  i32.atomic.rmw.or
+define void @multithread_weak_fence() {
+  fence acquire
+  fence release
+  fence acq_rel
+  ret void
+}
+
+; A singlethread fence becomes compiler_fence instruction, a pseudo instruction
+; that acts as a compiler barrier. The barrier should not be emitted to .s file.
+; CHECK-LABEL: singlethread_fence:
+; CHECK-NOT:  compiler_fence
+define void @singlethread_fence() {
+  fence syncscope("singlethread") seq_cst
+  fence syncscope("singlethread") acquire
+  fence syncscope("singlethread") release
+  fence syncscope("singlethread") acq_rel
+  ret void
+}