Merge pull request #13116 from rdmarsh2/rdmarsh2/cpp/cobo-array-vars

C++: include stack-allocated arrays in off-by-one query
2025-12-16 16:53:25 +01:00 · 2023-06-26 15:26:45 -04:00
parent 9c5aff395a d68b0605cd
commit 757f40c0cb
3 changed files with 108 additions and 37 deletions
--- a/cpp/ql/src/experimental/Security/CWE/CWE-193/ConstantSizeArrayOffByOne.ql
+++ b/cpp/ql/src/experimental/Security/CWE/CWE-193/ConstantSizeArrayOffByOne.ql
@@ -14,7 +14,7 @@ import semmle.code.cpp.rangeanalysis.new.internal.semantic.analysis.RangeAnalysi
 import semmle.code.cpp.rangeanalysis.new.internal.semantic.SemanticExprSpecific
 import semmle.code.cpp.ir.IR
 import semmle.code.cpp.ir.dataflow.DataFlow
-import FieldAddressToDerefFlow::PathGraph
+import ArrayAddressToDerefFlow::PathGraph

 pragma[nomagic]
 Instruction getABoundIn(SemBound b, IRFunction func) {
@@ -78,28 +78,45 @@ predicate isInvalidPointerDerefSink2(DataFlow::Node sink, Instruction i, string
  )
 }

+predicate arrayTypeCand(ArrayType arrayType) {
+  any(Variable v).getUnspecifiedType() = arrayType and
+  exists(arrayType.getArraySize())
+}
+
 pragma[nomagic]
 predicate arrayTypeHasSizes(ArrayType arr, int baseTypeSize, int arraySize) {
+  arrayTypeCand(arr) and
  arr.getBaseType().getSize() = baseTypeSize and
  arr.getArraySize() = arraySize
 }

-predicate pointerArithOverflow0(
-  PointerArithmeticInstruction pai, Field f, int size, int bound, int delta
-) {
-  not f.getNamespace() instanceof StdNamespace and
-  arrayTypeHasSizes(f.getUnspecifiedType(), pai.getElementSize(), size) and
-  semBounded(getSemanticExpr(pai.getRight()), any(SemZeroBound b), bound, true, _) and
-  delta = bound - size and
-  delta >= 0 and
-  size != 0 and
-  size != 1
+bindingset[pai]
+pragma[inline_late]
+predicate constantUpperBounded(PointerArithmeticInstruction pai, int delta) {
+  semBounded(getSemanticExpr(pai.getRight()), any(SemZeroBound b), delta, true, _)
+}
+
+bindingset[pai, size]
+predicate pointerArithOverflow0Impl(PointerArithmeticInstruction pai, int size, int delta) {
+  exists(int bound |
+    constantUpperBounded(pai, bound) and
+    delta = bound - size and
+    delta >= 0 and
+    size != 0 and
+    size != 1
+  )
+}
+
+pragma[nomagic]
+predicate pointerArithOverflow0(PointerArithmeticInstruction pai, int delta) {
+  exists(int size |
+    arrayTypeHasSizes(_, pai.getElementSize(), size) and
+    pointerArithOverflow0Impl(pai, size, delta)
+  )
 }

 module PointerArithmeticToDerefConfig implements DataFlow::ConfigSig {
-  predicate isSource(DataFlow::Node source) {
-    pointerArithOverflow0(source.asInstruction(), _, _, _, _)
-  }
+  predicate isSource(DataFlow::Node source) { pointerArithOverflow0(source.asInstruction(), _) }

  predicate isBarrierIn(DataFlow::Node node) { isSource(node) }

@@ -110,25 +127,38 @@ module PointerArithmeticToDerefConfig implements DataFlow::ConfigSig {

 module PointerArithmeticToDerefFlow = DataFlow::Global<PointerArithmeticToDerefConfig>;

-predicate pointerArithOverflow(
-  PointerArithmeticInstruction pai, Field f, int size, int bound, int delta
-) {
-  pointerArithOverflow0(pai, f, size, bound, delta) and
+predicate pointerArithOverflow(PointerArithmeticInstruction pai, int delta) {
+  pointerArithOverflow0(pai, delta) and
  PointerArithmeticToDerefFlow::flow(DataFlow::instructionNode(pai), _)
 }

-module FieldAddressToDerefConfig implements DataFlow::StateConfigSig {
+bindingset[v]
+predicate finalPointerArithOverflow(Variable v, PointerArithmeticInstruction pai, int delta) {
+  exists(int size |
+    arrayTypeHasSizes(pragma[only_bind_out](v.getUnspecifiedType()), pai.getElementSize(), size) and
+    pointerArithOverflow0Impl(pai, size, delta)
+  )
+}
+
+predicate isSourceImpl(DataFlow::Node source, Variable v) {
+  (
+    source.asInstruction().(FieldAddressInstruction).getField() = v
+    or
+    source.asInstruction().(VariableAddressInstruction).getAstVariable() = v
+  ) and
+  arrayTypeCand(v.getUnspecifiedType())
+}
+
+module ArrayAddressToDerefConfig implements DataFlow::StateConfigSig {
  newtype FlowState =
-    additional TArray(Field f) { pointerArithOverflow(_, f, _, _, _) } or
+    additional TArray() or
    additional TOverflowArithmetic(PointerArithmeticInstruction pai) {
-      pointerArithOverflow(pai, _, _, _, _)
+      pointerArithOverflow(pai, _)
    }

  predicate isSource(DataFlow::Node source, FlowState state) {
-    exists(Field f |
-      source.asInstruction().(FieldAddressInstruction).getField() = f and
-      state = TArray(f)
-    )
+    isSourceImpl(source, _) and
+    state = TArray()
  }

  predicate isSink(DataFlow::Node sink, FlowState state) {
@@ -147,27 +177,27 @@ module FieldAddressToDerefConfig implements DataFlow::StateConfigSig {
  predicate isAdditionalFlowStep(
    DataFlow::Node node1, FlowState state1, DataFlow::Node node2, FlowState state2
  ) {
-    exists(PointerArithmeticInstruction pai, Field f |
-      state1 = TArray(f) and
+    exists(PointerArithmeticInstruction pai |
+      state1 = TArray() and
      state2 = TOverflowArithmetic(pai) and
      pai.getLeft() = node1.asInstruction() and
      node2.asInstruction() = pai and
-      pointerArithOverflow(pai, f, _, _, _)
+      pointerArithOverflow(pai, _)
    )
  }
 }

-module FieldAddressToDerefFlow = DataFlow::GlobalWithState<FieldAddressToDerefConfig>;
+module ArrayAddressToDerefFlow = DataFlow::GlobalWithState<ArrayAddressToDerefConfig>;

 from
-  Field f, FieldAddressToDerefFlow::PathNode source, PointerArithmeticInstruction pai,
-  FieldAddressToDerefFlow::PathNode sink, Instruction deref, string operation, int delta
+  Variable v, ArrayAddressToDerefFlow::PathNode source, PointerArithmeticInstruction pai,
+  ArrayAddressToDerefFlow::PathNode sink, Instruction deref, string operation, int delta
 where
-  FieldAddressToDerefFlow::flowPath(source, sink) and
+  ArrayAddressToDerefFlow::flowPath(pragma[only_bind_into](source), pragma[only_bind_into](sink)) and
  isInvalidPointerDerefSink2(sink.getNode(), deref, operation) and
-  source.getState() = FieldAddressToDerefConfig::TArray(f) and
-  sink.getState() = FieldAddressToDerefConfig::TOverflowArithmetic(pai) and
-  pointerArithOverflow(pai, f, _, _, delta)
+  pragma[only_bind_out](sink.getState()) = ArrayAddressToDerefConfig::TOverflowArithmetic(pai) and
+  isSourceImpl(source.getNode(), v) and
+  finalPointerArithOverflow(v, pai, delta)
 select pai, source, sink,
  "This pointer arithmetic may have an off-by-" + (delta + 1) +
-    " error allowing it to overrun $@ at this $@.", f, f.getName(), deref, operation
+    " error allowing it to overrun $@ at this $@.", v, v.getName(), deref, operation
--- a/cpp/ql/test/experimental/query-tests/Security/CWE/CWE-193/constant-size/ConstantSizeArrayOffByOne.expected
+++ b/cpp/ql/test/experimental/query-tests/Security/CWE/CWE-193/constant-size/ConstantSizeArrayOffByOne.expected
@@ -11,6 +11,13 @@ edges
 | test.cpp:77:32:77:34 | buf | test.cpp:77:26:77:44 | & ... |
 | test.cpp:79:27:79:34 | buf | test.cpp:70:33:70:33 | p |
 | test.cpp:79:32:79:34 | buf | test.cpp:79:27:79:34 | buf |
+| test.cpp:85:34:85:36 | buf | test.cpp:87:5:87:31 | access to array |
+| test.cpp:85:34:85:36 | buf | test.cpp:88:5:88:27 | access to array |
+| test.cpp:128:9:128:11 | arr | test.cpp:128:9:128:14 | access to array |
+| test.cpp:134:25:134:27 | arr | test.cpp:136:9:136:16 | ... += ... |
+| test.cpp:136:9:136:16 | ... += ... | test.cpp:138:13:138:15 | arr |
+| test.cpp:143:18:143:21 | asdf | test.cpp:134:25:134:27 | arr |
+| test.cpp:143:18:143:21 | asdf | test.cpp:143:18:143:21 | asdf |
 nodes
 | test.cpp:35:5:35:22 | access to array | semmle.label | access to array |
 | test.cpp:35:10:35:12 | buf | semmle.label | buf |
@@ -33,6 +40,16 @@ nodes
 | test.cpp:77:32:77:34 | buf | semmle.label | buf |
 | test.cpp:79:27:79:34 | buf | semmle.label | buf |
 | test.cpp:79:32:79:34 | buf | semmle.label | buf |
+| test.cpp:85:34:85:36 | buf | semmle.label | buf |
+| test.cpp:87:5:87:31 | access to array | semmle.label | access to array |
+| test.cpp:88:5:88:27 | access to array | semmle.label | access to array |
+| test.cpp:128:9:128:11 | arr | semmle.label | arr |
+| test.cpp:128:9:128:14 | access to array | semmle.label | access to array |
+| test.cpp:134:25:134:27 | arr | semmle.label | arr |
+| test.cpp:136:9:136:16 | ... += ... | semmle.label | ... += ... |
+| test.cpp:138:13:138:15 | arr | semmle.label | arr |
+| test.cpp:143:18:143:21 | asdf | semmle.label | asdf |
+| test.cpp:143:18:143:21 | asdf | semmle.label | asdf |
 subpaths
 #select
 | test.cpp:35:5:35:22 | PointerAdd: access to array | test.cpp:35:10:35:12 | buf | test.cpp:35:5:35:22 | access to array | This pointer arithmetic may have an off-by-1 error allowing it to overrun $@ at this $@. | test.cpp:15:9:15:11 | buf | buf | test.cpp:35:5:35:26 | Store: ... = ... | write |
@@ -44,3 +61,5 @@ subpaths
 | test.cpp:61:9:61:19 | PointerAdd: access to array | test.cpp:61:14:61:16 | buf | test.cpp:61:9:61:19 | access to array | This pointer arithmetic may have an off-by-2 error allowing it to overrun $@ at this $@. | test.cpp:19:9:19:11 | buf | buf | test.cpp:61:9:61:23 | Store: ... = ... | write |
 | test.cpp:72:5:72:15 | PointerAdd: access to array | test.cpp:79:32:79:34 | buf | test.cpp:72:5:72:15 | access to array | This pointer arithmetic may have an off-by-1 error allowing it to overrun $@ at this $@. | test.cpp:15:9:15:11 | buf | buf | test.cpp:72:5:72:19 | Store: ... = ... | write |
 | test.cpp:77:27:77:44 | PointerAdd: access to array | test.cpp:77:32:77:34 | buf | test.cpp:66:32:66:32 | p | This pointer arithmetic may have an off-by-1 error allowing it to overrun $@ at this $@. | test.cpp:15:9:15:11 | buf | buf | test.cpp:67:5:67:10 | Store: ... = ... | write |
+| test.cpp:128:9:128:14 | PointerAdd: access to array | test.cpp:128:9:128:11 | arr | test.cpp:128:9:128:14 | access to array | This pointer arithmetic may have an off-by-1 error allowing it to overrun $@ at this $@. | test.cpp:125:11:125:13 | arr | arr | test.cpp:128:9:128:18 | Store: ... = ... | write |
+| test.cpp:136:9:136:16 | PointerAdd: ... += ... | test.cpp:143:18:143:21 | asdf | test.cpp:138:13:138:15 | arr | This pointer arithmetic may have an off-by-2 error allowing it to overrun $@ at this $@. | test.cpp:142:10:142:13 | asdf | asdf | test.cpp:138:12:138:15 | Load: * ... | read |
--- a/cpp/ql/test/experimental/query-tests/Security/CWE/CWE-193/constant-size/test.cpp
+++ b/cpp/ql/test/experimental/query-tests/Security/CWE/CWE-193/constant-size/test.cpp
@@ -85,7 +85,7 @@ void testCharIndex(BigArray *arr) {
    char *charBuf = (char*) arr->buf;

    charBuf[MAX_SIZE_BYTES - 1] = 0; // GOOD
-    charBuf[MAX_SIZE_BYTES] = 0; // BAD [FALSE NEGATIVE]
+    charBuf[MAX_SIZE_BYTES] = 0; // BAD
 }

 void testEqRefinement() {
@@ -120,3 +120,25 @@ void testEqRefinement2() {
        }
    }
 }
+
+void testStackAllocated() {
+    char *arr[MAX_SIZE];
+
+    for(int i = 0; i <= MAX_SIZE; i++) {
+        arr[i] = 0; // BAD
+    }
+}
+
+int strncmp(const char*, const char*, int);
+
+char testStrncmp2(char *arr) {
+    if(strncmp(arr, "<test>", 6) == 0) {
+        arr += 6;
+    }
+    return *arr; // GOOD [FALSE POSITIVE]
+}
+
+void testStrncmp1() {
+    char asdf[5];
+    testStrncmp2(asdf);
+}