notation change: AST => IR (#1245)

The IR is indeed a tree, but not an "abstract syntax tree" since there is no language for which it is the syntax (except in the most trivial and meaningless sense).

1 files changed, 278 insertions, 0 deletions

diff --git a/src/ir/effects.h b/src/ir/effects.h
new file mode 100644
index 000000000..98911d451
--- /dev/null
+++ b/src/ir/effects.h
@@ -0,0 +1,278 @@
+/*
+ * Copyright 2017 WebAssembly Community Group participants
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef wasm_ir_effects_h
+#define wasm_ir_effects_h
+
+namespace wasm {
+
+// Look for side effects, including control flow
+// TODO: optimize
+
+struct EffectAnalyzer : public PostWalker<EffectAnalyzer> {
+  EffectAnalyzer(PassOptions& passOptions, Expression *ast = nullptr) {
+    ignoreImplicitTraps = passOptions.ignoreImplicitTraps;
+    debugInfo = passOptions.debugInfo;
+    if (ast) analyze(ast);
+  }
+
+  bool ignoreImplicitTraps;
+  bool debugInfo;
+
+  void analyze(Expression *ast) {
+    breakNames.clear();
+    walk(ast);
+    // if we are left with breaks, they are external
+    if (breakNames.size() > 0) branches = true;
+  }
+
+  bool branches = false; // branches out of this expression, returns, infinite loops, etc
+  bool calls = false;
+  std::set<Index> localsRead;
+  std::set<Index> localsWritten;
+  std::set<Name> globalsRead;
+  std::set<Name> globalsWritten;
+  bool readsMemory = false;
+  bool writesMemory = false;
+  bool implicitTrap = false; // a load or div/rem, which may trap. we ignore trap
+                             // differences, so it is ok to reorder these, but we can't
+                             // remove them, as they count as side effects, and we
+                             // can't move them in a way that would cause other noticeable
+                             // (global) side effects
+  bool isAtomic = false; // An atomic load/store/RMW/Cmpxchg or an operator that
+                         // has a defined ordering wrt atomics (e.g. grow_memory)
+
+  bool accessesLocal() { return localsRead.size() + localsWritten.size() > 0; }
+  bool accessesGlobal() { return globalsRead.size() + globalsWritten.size() > 0; }
+  bool accessesMemory() { return calls || readsMemory || writesMemory; }
+  bool hasGlobalSideEffects() { return calls || globalsWritten.size() > 0 || writesMemory || isAtomic; }
+  bool hasSideEffects() { return hasGlobalSideEffects() || localsWritten.size() > 0 || branches || implicitTrap; }
+  bool hasAnything() { return branches || calls || accessesLocal() || readsMemory || writesMemory || accessesGlobal() || implicitTrap || isAtomic; }
+
+  // checks if these effects would invalidate another set (e.g., if we write, we invalidate someone that reads, they can't be moved past us)
+  bool invalidates(EffectAnalyzer& other) {
+    if (branches || other.branches
+                 || ((writesMemory || calls) && other.accessesMemory())
+                 || (accessesMemory() && (other.writesMemory || other.calls))) {
+      return true;
+    }
+    // All atomics are sequentially consistent for now, and ordered wrt other
+    // memory references.
+    if ((isAtomic && other.accessesMemory()) ||
+        (other.isAtomic && accessesMemory())) {
+      return true;
+    }
+    for (auto local : localsWritten) {
+      if (other.localsWritten.count(local) || other.localsRead.count(local)) {
+        return true;
+      }
+    }
+    for (auto local : localsRead) {
+      if (other.localsWritten.count(local)) return true;
+    }
+    if ((accessesGlobal() && other.calls) || (other.accessesGlobal() && calls)) {
+      return true;
+    }
+    for (auto global : globalsWritten) {
+      if (other.globalsWritten.count(global) || other.globalsRead.count(global)) {
+        return true;
+      }
+    }
+    for (auto global : globalsRead) {
+      if (other.globalsWritten.count(global)) return true;
+    }
+    // we are ok to reorder implicit traps, but not conditionalize them
+    if ((implicitTrap && other.branches) || (other.implicitTrap && branches)) {
+      return true;
+    }
+    // we can't reorder an implicit trap in a way that alters global state
+    if ((implicitTrap && other.hasGlobalSideEffects()) || (other.implicitTrap && hasGlobalSideEffects())) {
+      return true;
+    }
+    return false;
+  }
+
+  void mergeIn(EffectAnalyzer& other) {
+    branches = branches || other.branches;
+    calls = calls || other.calls;
+    readsMemory = readsMemory || other.readsMemory;
+    writesMemory = writesMemory || other.writesMemory;
+    for (auto i : other.localsRead) localsRead.insert(i);
+    for (auto i : other.localsWritten) localsWritten.insert(i);
+    for (auto i : other.globalsRead) globalsRead.insert(i);
+    for (auto i : other.globalsWritten) globalsWritten.insert(i);
+  }
+
+  // the checks above happen after the node's children were processed, in the order of execution
+  // we must also check for control flow that happens before the children, i.e., loops
+  bool checkPre(Expression* curr) {
+    if (curr->is<Loop>()) {
+      branches = true;
+      return true;
+    }
+    return false;
+  }
+
+  bool checkPost(Expression* curr) {
+    visit(curr);
+    if (curr->is<Loop>()) {
+      branches = true;
+    }
+    return hasAnything();
+  }
+
+  std::set<Name> breakNames;
+
+  void visitBreak(Break *curr) {
+    breakNames.insert(curr->name);
+  }
+  void visitSwitch(Switch *curr) {
+    for (auto name : curr->targets) {
+      breakNames.insert(name);
+    }
+    breakNames.insert(curr->default_);
+  }
+  void visitBlock(Block* curr) {
+    if (curr->name.is()) breakNames.erase(curr->name); // these were internal breaks
+  }
+  void visitLoop(Loop* curr) {
+    if (curr->name.is()) breakNames.erase(curr->name); // these were internal breaks
+    // if the loop is unreachable, then there is branching control flow:
+    //  (1) if the body is unreachable because of a (return), uncaught (br) etc., then we
+    //      already noted branching, so it is ok to mark it again  (if we have *caught*
+    //      (br)s, then they did not lead to the loop body being unreachable).
+    //      (same logic applies to blocks)
+    //  (2) if the loop is unreachable because it only has branches up to the loop
+    //      top, but no way to get out, then it is an infinite loop, and we consider
+    //      that a branching side effect (note how the same logic does not apply to
+    //      blocks).
+    if (curr->type == unreachable) {
+      branches = true;
+    }
+  }
+
+  void visitCall(Call *curr) { calls = true; }
+  void visitCallImport(CallImport *curr) {
+    calls = true;
+    if (debugInfo) {
+      // debugInfo call imports must be preserved very strongly, do not
+      // move code around them
+      branches = true; // !
+    }
+  }
+  void visitCallIndirect(CallIndirect *curr) { calls = true; }
+  void visitGetLocal(GetLocal *curr) {
+    localsRead.insert(curr->index);
+  }
+  void visitSetLocal(SetLocal *curr) {
+    localsWritten.insert(curr->index);
+  }
+  void visitGetGlobal(GetGlobal *curr) {
+    globalsRead.insert(curr->name);
+  }
+  void visitSetGlobal(SetGlobal *curr) {
+    globalsWritten.insert(curr->name);
+  }
+  void visitLoad(Load *curr) {
+    readsMemory = true;
+    isAtomic |= curr->isAtomic;
+    if (!ignoreImplicitTraps) implicitTrap = true;
+  }
+  void visitStore(Store *curr) {
+    writesMemory = true;
+    isAtomic |= curr->isAtomic;
+    if (!ignoreImplicitTraps) implicitTrap = true;
+  }
+  void visitAtomicRMW(AtomicRMW* curr) {
+    readsMemory = true;
+    writesMemory = true;
+    isAtomic = true;
+    if (!ignoreImplicitTraps) implicitTrap = true;
+  }
+  void visitAtomicCmpxchg(AtomicCmpxchg* curr) {
+    readsMemory = true;
+    writesMemory = true;
+    isAtomic = true;
+    if (!ignoreImplicitTraps) implicitTrap = true;
+  }
+  void visitAtomicWait(AtomicWait* curr) {
+    readsMemory = true;
+    // AtomicWait doesn't strictly write memory, but it does modify the waiters
+    // list associated with the specified address, which we can think of as a
+    // write.
+    writesMemory = true;
+    isAtomic = true;
+    if (!ignoreImplicitTraps) implicitTrap = true;
+  }
+  void visitAtomicWake(AtomicWake* curr) {
+    // AtomicWake doesn't strictly write memory, but it does modify the waiters
+    // list associated with the specified address, which we can think of as a
+    // write.
+    readsMemory = true;
+    writesMemory = true;
+    isAtomic = true;
+    if (!ignoreImplicitTraps) implicitTrap = true;
+  };
+  void visitUnary(Unary *curr) {
+    if (!ignoreImplicitTraps) {
+      switch (curr->op) {
+        case TruncSFloat32ToInt32:
+        case TruncSFloat32ToInt64:
+        case TruncUFloat32ToInt32:
+        case TruncUFloat32ToInt64:
+        case TruncSFloat64ToInt32:
+        case TruncSFloat64ToInt64:
+        case TruncUFloat64ToInt32:
+        case TruncUFloat64ToInt64: {
+          implicitTrap = true;
+          break;
+        }
+        default: {}
+      }
+    }
+  }
+  void visitBinary(Binary *curr) {
+    if (!ignoreImplicitTraps) {
+      switch (curr->op) {
+        case DivSInt32:
+        case DivUInt32:
+        case RemSInt32:
+        case RemUInt32:
+        case DivSInt64:
+        case DivUInt64:
+        case RemSInt64:
+        case RemUInt64: {
+          implicitTrap = true;
+          break;
+        }
+        default: {}
+      }
+    }
+  }
+  void visitReturn(Return *curr) { branches = true; }
+  void visitHost(Host *curr) {
+    calls = true;
+    // grow_memory modifies the set of valid addresses, and thus can be modeled as modifying memory
+    writesMemory = true;
+    // Atomics are also sequentially consistent with grow_memory.
+    isAtomic = true;
+  }
+  void visitUnreachable(Unreachable *curr) { branches = true; }
+};
+
+} // namespace wasm
+
+#endif // wasm_ir_effects_h