From 3c81c94596748810a56171b5701536b57398f1e0 Mon Sep 17 00:00:00 2001
From: Raffi Khatchadourian <raffi.khatchadourian@hunter.cuny.edu>
Date: Mon, 30 Oct 2023 13:51:55 -0400
Subject: [PATCH] Handle global writes in the Python ModRef analysis (#101)

Previously, the Python ModRef analysis did not consider writes to global
variables that used the `global` keyword as a heap modification.
---
 com.ibm.wala.cast.python.test/data/globals.py | 11 ++++
 .../python/test/TestPythonModRefAnalysis.java | 60 +++++++++++++++++++
 .../wala/cast/python/modref/PythonModRef.java | 21 +++++++
 3 files changed, 92 insertions(+)
 create mode 100644 com.ibm.wala.cast.python.test/data/globals.py
 create mode 100644 com.ibm.wala.cast.python.test/source/com/ibm/wala/cast/python/test/TestPythonModRefAnalysis.java

diff --git a/com.ibm.wala.cast.python.test/data/globals.py b/com.ibm.wala.cast.python.test/data/globals.py
new file mode 100644
index 000000000..0fd8368f8
--- /dev/null
+++ b/com.ibm.wala.cast.python.test/data/globals.py
@@ -0,0 +1,11 @@
+a = 10
+
+
+def f():
+    global a
+    a = a + 1
+
+
+assert a == 10
+f()
+assert a == 11
diff --git a/com.ibm.wala.cast.python.test/source/com/ibm/wala/cast/python/test/TestPythonModRefAnalysis.java b/com.ibm.wala.cast.python.test/source/com/ibm/wala/cast/python/test/TestPythonModRefAnalysis.java
new file mode 100644
index 000000000..05ba7481f
--- /dev/null
+++ b/com.ibm.wala.cast.python.test/source/com/ibm/wala/cast/python/test/TestPythonModRefAnalysis.java
@@ -0,0 +1,60 @@
+package com.ibm.wala.cast.python.test;
+
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertTrue;
+
+import com.ibm.wala.cast.python.client.PythonAnalysisEngine;
+import com.ibm.wala.cast.python.modref.PythonModRef;
+import com.ibm.wala.classLoader.IField;
+import com.ibm.wala.core.util.strings.Atom;
+import com.ibm.wala.ipa.callgraph.CGNode;
+import com.ibm.wala.ipa.callgraph.CallGraph;
+import com.ibm.wala.ipa.callgraph.propagation.InstanceKey;
+import com.ibm.wala.ipa.callgraph.propagation.PointerKey;
+import com.ibm.wala.ipa.callgraph.propagation.SSAPropagationCallGraphBuilder;
+import com.ibm.wala.ipa.callgraph.propagation.StaticFieldKey;
+import com.ibm.wala.ipa.cha.ClassHierarchyException;
+import com.ibm.wala.ipa.modref.ModRef;
+import com.ibm.wala.util.CancelException;
+import com.ibm.wala.util.intset.OrdinalSet;
+import java.io.IOException;
+import java.util.Collection;
+import java.util.Map;
+import org.junit.Test;
+
+public class TestPythonModRefAnalysis extends TestPythonCallGraphShape {
+
+  @Test
+  public void testComputeModCallGraphPointerAnalysisOfT()
+      throws ClassHierarchyException, IllegalArgumentException, CancelException, IOException {
+    PythonAnalysisEngine<?> engine = makeEngine("globals.py");
+    SSAPropagationCallGraphBuilder builder =
+        (SSAPropagationCallGraphBuilder) engine.defaultCallGraphBuilder();
+    CallGraph CG = builder.makeCallGraph(builder.getOptions());
+
+    Collection<CGNode> nodes = getNodes(CG, "script globals.py/f");
+    assertEquals(1, nodes.size());
+
+    assertTrue(nodes.iterator().hasNext());
+    CGNode fNode = nodes.iterator().next();
+
+    ModRef<InstanceKey> modRef = new PythonModRef();
+    Map<CGNode, OrdinalSet<PointerKey>> mod = modRef.computeMod(CG, builder.getPointerAnalysis());
+
+    // what heap locations does f() (transitively) modify?
+    OrdinalSet<PointerKey> modSet = mod.get(fNode);
+
+    // should only modify the global.
+    assertEquals(1, modSet.size());
+
+    assertTrue(modSet.iterator().hasNext());
+    PointerKey pointerKey = modSet.iterator().next();
+
+    assertTrue(pointerKey instanceof StaticFieldKey);
+    StaticFieldKey staticFieldKey = (StaticFieldKey) pointerKey;
+
+    IField field = staticFieldKey.getField();
+    Atom name = field.getName();
+    assertEquals(Atom.findOrCreateAsciiAtom("global a"), name);
+  }
+}
diff --git a/com.ibm.wala.cast.python/source/com/ibm/wala/cast/python/modref/PythonModRef.java b/com.ibm.wala.cast.python/source/com/ibm/wala/cast/python/modref/PythonModRef.java
index 98046bfc5..4576225aa 100644
--- a/com.ibm.wala.cast.python/source/com/ibm/wala/cast/python/modref/PythonModRef.java
+++ b/com.ibm.wala.cast.python/source/com/ibm/wala/cast/python/modref/PythonModRef.java
@@ -2,11 +2,15 @@
 
 import com.ibm.wala.cast.ipa.callgraph.AstHeapModel;
 import com.ibm.wala.cast.ipa.modref.AstModRef;
+import com.ibm.wala.cast.ir.ssa.AstGlobalWrite;
 import com.ibm.wala.cast.python.ssa.PythonInstructionVisitor;
+import com.ibm.wala.classLoader.IField;
+import com.ibm.wala.core.util.strings.Atom;
 import com.ibm.wala.ipa.callgraph.CGNode;
 import com.ibm.wala.ipa.callgraph.propagation.InstanceKey;
 import com.ibm.wala.ipa.callgraph.propagation.PointerAnalysis;
 import com.ibm.wala.ipa.callgraph.propagation.PointerKey;
+import com.ibm.wala.ipa.callgraph.propagation.StaticFieldKey;
 import com.ibm.wala.ipa.modref.ExtendedHeapModel;
 import java.util.Collection;
 
@@ -41,6 +45,23 @@ public PythonModVisitor(
         boolean ignoreAllocHeapDefs) {
       super(n, result, (AstHeapModel) h, pa);
     }
+
+    @Override
+    public void visitAstGlobalWrite(AstGlobalWrite instruction) {
+      super.visitAstGlobalWrite(instruction);
+      String globalName = instruction.getGlobalName();
+
+      // find the pointer key corresponding to this global.
+      for (PointerKey pk : this.pa.getPointerKeys()) {
+        if (pk instanceof StaticFieldKey) {
+          StaticFieldKey staticFieldKey = (StaticFieldKey) pk;
+          IField field = staticFieldKey.getField();
+          Atom fieldName = field.getName();
+          if (fieldName.toString().equals(globalName))
+            this.result.add(this.h.getPointerKeyForStaticField(field));
+        }
+      }
+    }
   }
 
   @Override