Merge branch 'github:main' into amammad-python-WebAppsConstatntSecretKeys

Author: am0o0

.github/workflows/check-implicit-this.yml

@@ -0,0 +1,29 @@
+name: "Check implicit this warnings"
+
+on:
+  workflow_dispatch:
+  pull_request:
+    paths:
+      - "**qlpack.yml"
+    branches:
+      - main
+      - "rc/*"
+
+jobs:
+  check:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - name: Check that implicit this warnings is enabled for all packs
+        shell: bash
+        run: |
+          EXIT_CODE=0
+          packs="$(find . -iname 'qlpack.yml')"
+          for pack_file in ${packs}; do
+            option="$(yq '.warnOnImplicitThis' ${pack_file})"
+            if [ "${option}" != "true" ]; then
+              echo "::error file=${pack_file}::warnOnImplicitThis property must be set to 'true' for pack ${pack_file}"
+              EXIT_CODE=1
+            fi
+          done
+          exit "${EXIT_CODE}"

.pre-commit-config.yaml

@@ -5,9 +5,9 @@ repos:
     rev: v3.2.0
     hooks:
       - id: trailing-whitespace
-        exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)
+        exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)|.*\.patch
       - id: end-of-file-fixer
-        exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)
+        exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)|.*\.patch
 
   - repo: https://github.com/pre-commit/mirrors-clang-format
     rev: v13.0.1
@@ -21,6 +21,11 @@ repos:
       - id: autopep8
         files: ^misc/codegen/.*\.py
 
+  - repo: https://github.com/warchant/pre-commit-buildifier
+    rev: 0.0.2
+    hooks:
+      - id: buildifier
+
   - repo: local
     hooks:
       - id: codeql-format

cpp/downgrades/qlpack.yml

@@ -2,3 +2,4 @@ name: codeql/cpp-downgrades
 groups: cpp
 downgrades: .
 library: true
+warnOnImplicitThis: true

cpp/ql/examples/qlpack.yml

@@ -4,3 +4,4 @@ groups:
   - examples
 dependencies:
   codeql/cpp-all: ${workspace}
+warnOnImplicitThis: true

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowPrivate.qll

@@ -321,9 +321,11 @@ private class PrimaryArgumentNode extends ArgumentNode, OperandNode {
 
 private class SideEffectArgumentNode extends ArgumentNode, SideEffectOperandNode {
   override predicate argumentOf(DataFlowCall dfCall, ArgumentPosition pos) {
-    this.getCallInstruction() = dfCall and
-    pos.(IndirectionPosition).getArgumentIndex() = this.getArgumentIndex() and
-    super.hasAddressOperandAndIndirectionIndex(_, pos.(IndirectionPosition).getIndirectionIndex())
+    exists(int indirectionIndex |
+      pos = TIndirectionPosition(argumentIndex, pragma[only_bind_into](indirectionIndex)) and
+      this.getCallInstruction() = dfCall and
+      super.hasAddressOperandAndIndirectionIndex(_, pragma[only_bind_into](indirectionIndex))
+    )
   }
 }
 
@@ -651,13 +653,16 @@ predicate jumpStep(Node n1, Node n2) {
  * Holds if data can flow from `node1` to `node2` via an assignment to `f`.
  * Thus, `node2` references an object with a field `f` that contains the
  * value of `node1`.
+ *
+ * The boolean `certain` is true if the destination address does not involve
+ * any pointer arithmetic, and false otherwise.
  */
-predicate storeStep(Node node1, Content c, PostFieldUpdateNode node2) {
+predicate storeStepImpl(Node node1, Content c, PostFieldUpdateNode node2, boolean certain) {
   exists(int indirectionIndex1, int numberOfLoads, StoreInstruction store |
     nodeHasInstruction(node1, store, pragma[only_bind_into](indirectionIndex1)) and
     node2.getIndirectionIndex() = 1 and
     numberOfLoadsFromOperand(node2.getFieldAddress(), store.getDestinationAddressOperand(),
-      numberOfLoads)
+      numberOfLoads, certain)
   |
     exists(FieldContent fc | fc = c |
       fc.getField() = node2.getUpdatedField() and
@@ -671,35 +676,51 @@ predicate storeStep(Node node1, Content c, PostFieldUpdateNode node2) {
   )
 }
 
+/**
+ * Holds if data can flow from `node1` to `node2` via an assignment to `f`.
+ * Thus, `node2` references an object with a field `f` that contains the
+ * value of `node1`.
+ */
+predicate storeStep(Node node1, Content c, PostFieldUpdateNode node2) {
+  storeStepImpl(node1, c, node2, _)
+}
+
 /**
  * Holds if `operandFrom` flows to `operandTo` using a sequence of conversion-like
  * operations and exactly `n` `LoadInstruction` operations.
  */
-private predicate numberOfLoadsFromOperandRec(Operand operandFrom, Operand operandTo, int ind) {
+private predicate numberOfLoadsFromOperandRec(
+  Operand operandFrom, Operand operandTo, int ind, boolean certain
+) {
   exists(Instruction load | Ssa::isDereference(load, operandFrom) |
-    operandTo = operandFrom and ind = 0
+    operandTo = operandFrom and ind = 0 and certain = true
     or
-    numberOfLoadsFromOperand(load.getAUse(), operandTo, ind - 1)
+    numberOfLoadsFromOperand(load.getAUse(), operandTo, ind - 1, certain)
   )
   or
-  exists(Operand op, Instruction instr |
+  exists(Operand op, Instruction instr, boolean isPointerArith, boolean certain0 |
     instr = op.getDef() and
-    conversionFlow(operandFrom, instr, _, _) and
-    numberOfLoadsFromOperand(op, operandTo, ind)
+    conversionFlow(operandFrom, instr, isPointerArith, _) and
+    numberOfLoadsFromOperand(op, operandTo, ind, certain0)
+  |
+    if isPointerArith = true then certain = false else certain = certain0
   )
 }
 
 /**
  * Holds if `operandFrom` flows to `operandTo` using a sequence of conversion-like
  * operations and exactly `n` `LoadInstruction` operations.
  */
-private predicate numberOfLoadsFromOperand(Operand operandFrom, Operand operandTo, int n) {
-  numberOfLoadsFromOperandRec(operandFrom, operandTo, n)
+private predicate numberOfLoadsFromOperand(
+  Operand operandFrom, Operand operandTo, int n, boolean certain
+) {
+  numberOfLoadsFromOperandRec(operandFrom, operandTo, n, certain)
   or
   not Ssa::isDereference(_, operandFrom) and
   not conversionFlow(operandFrom, _, _, _) and
   operandFrom = operandTo and
-  n = 0
+  n = 0 and
+  certain = true
 }
 
 // Needed to join on both an operand and an index at the same time.
@@ -729,7 +750,7 @@ predicate readStep(Node node1, Content c, Node node2) {
     // The `1` here matches the `node2.getIndirectionIndex() = 1` conjunct
     // in `storeStep`.
     nodeHasOperand(node1, fa1.getObjectAddressOperand(), 1) and
-    numberOfLoadsFromOperand(fa1, operand, numberOfLoads)
+    numberOfLoadsFromOperand(fa1, operand, numberOfLoads, _)
   |
     exists(FieldContent fc | fc = c |
       fc.getField() = fa1.getField() and
@@ -747,7 +768,33 @@ predicate readStep(Node node1, Content c, Node node2) {
  * Holds if values stored inside content `c` are cleared at node `n`.
  */
 predicate clearsContent(Node n, Content c) {
-  none() // stub implementation
+  n =
+    any(PostUpdateNode pun, Content d | d.impliesClearOf(c) and storeStepImpl(_, d, pun, true) | pun)
+        .getPreUpdateNode() and
+  (
+    // The crement operations and pointer addition and subtraction self-assign. We do not
+    // want to clear the contents if it is indirectly pointed at by any of these operations,
+    // as part of the contents might still be accessible afterwards. If there is no such
+    // indirection clearing the contents is safe.
+    not exists(Operand op, Cpp::Operation p |
+      n.(IndirectOperand).hasOperandAndIndirectionIndex(op, _) and
+      (
+        p instanceof Cpp::AssignPointerAddExpr or
+        p instanceof Cpp::AssignPointerSubExpr or
+        p instanceof Cpp::CrementOperation
+      )
+    |
+      p.getAnOperand() = op.getUse().getAst()
+    )
+    or
+    forex(PostUpdateNode pun, Content d |
+      pragma[only_bind_into](d).impliesClearOf(pragma[only_bind_into](c)) and
+      storeStepImpl(_, d, pun, true) and
+      pun.getPreUpdateNode() = n
+    |
+      c.getIndirectionIndex() = d.getIndirectionIndex()
+    )
+  )
 }
 
 /**
@@ -809,7 +856,73 @@ class DataFlowCall extends CallInstruction {
   Function getEnclosingCallable() { result = this.getEnclosingFunction() }
 }
 
-predicate isUnreachableInCall(Node n, DataFlowCall call) { none() } // stub implementation
+module IsUnreachableInCall {
+  private import semmle.code.cpp.ir.ValueNumbering
+  private import semmle.code.cpp.controlflow.IRGuards as G
+
+  private class ConstantIntegralTypeArgumentNode extends PrimaryArgumentNode {
+    int value;
+
+    ConstantIntegralTypeArgumentNode() {
+      value = op.getDef().(IntegerConstantInstruction).getValue().toInt()
+    }
+
+    int getValue() { result = value }
+  }
+
+  pragma[nomagic]
+  private predicate ensuresEq(Operand left, Operand right, int k, IRBlock block, boolean areEqual) {
+    any(G::IRGuardCondition guard).ensuresEq(left, right, k, block, areEqual)
+  }
+
+  pragma[nomagic]
+  private predicate ensuresLt(Operand left, Operand right, int k, IRBlock block, boolean areEqual) {
+    any(G::IRGuardCondition guard).ensuresLt(left, right, k, block, areEqual)
+  }
+
+  predicate isUnreachableInCall(Node n, DataFlowCall call) {
+    exists(
+      DirectParameterNode paramNode, ConstantIntegralTypeArgumentNode arg,
+      IntegerConstantInstruction constant, int k, Operand left, Operand right, IRBlock block
+    |
+      // arg flows into `paramNode`
+      DataFlowImplCommon::viableParamArg(call, paramNode, arg) and
+      left = constant.getAUse() and
+      right = valueNumber(paramNode.getInstruction()).getAUse() and
+      block = n.getBasicBlock()
+    |
+      // and there's a guard condition which ensures that the result of `left == right + k` is `areEqual`
+      exists(boolean areEqual |
+        ensuresEq(pragma[only_bind_into](left), pragma[only_bind_into](right),
+          pragma[only_bind_into](k), pragma[only_bind_into](block), areEqual)
+      |
+        // this block ensures that left = right + k, but it holds that `left != right + k`
+        areEqual = true and
+        constant.getValue().toInt() != arg.getValue() + k
+        or
+        // this block ensures that or `left != right + k`, but it holds that `left = right + k`
+        areEqual = false and
+        constant.getValue().toInt() = arg.getValue() + k
+      )
+      or
+      // or there's a guard condition which ensures that the result of `left < right + k` is `isLessThan`
+      exists(boolean isLessThan |
+        ensuresLt(pragma[only_bind_into](left), pragma[only_bind_into](right),
+          pragma[only_bind_into](k), pragma[only_bind_into](block), isLessThan)
+      |
+        isLessThan = true and
+        // this block ensures that `left < right + k`, but it holds that `left >= right + k`
+        constant.getValue().toInt() >= arg.getValue() + k
+        or
+        // this block ensures that `left >= right + k`, but it holds that `left < right + k`
+        isLessThan = false and
+        constant.getValue().toInt() < arg.getValue() + k
+      )
+    )
+  }
+}
+
+import IsUnreachableInCall
 
 int accessPathLimit() { result = 5 }
 

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowUtil.qll

@@ -1832,6 +1832,20 @@ class Content extends TContent {
   predicate hasLocationInfo(string path, int sl, int sc, int el, int ec) {
     path = "" and sl = 0 and sc = 0 and el = 0 and ec = 0
   }
+
+  /** Gets the indirection index of this `Content`. */
+  abstract int getIndirectionIndex();
+
+  /**
+   * INTERNAL: Do not use.
+   *
+   * Holds if a write to this `Content` implies that `c` is
+   * also cleared.
+   *
+   * For example, a write to a field `f` implies that any content of
+   * the form `*f` is also cleared.
+   */
+  abstract predicate impliesClearOf(Content c);
 }
 
 /** A reference through a non-union instance field. */
@@ -1849,10 +1863,21 @@ class FieldContent extends Content, TFieldContent {
 
   Field getField() { result = f }
 
+  /** Gets the indirection index of this `FieldContent`. */
   pragma[inline]
-  int getIndirectionIndex() {
+  override int getIndirectionIndex() {
     pragma[only_bind_into](result) = pragma[only_bind_out](indirectionIndex)
   }
+
+  override predicate impliesClearOf(Content c) {
+    exists(FieldContent fc |
+      fc = c and
+      fc.getField() = f and
+      // If `this` is `f` then `c` is cleared if it's of the
+      // form `*f`, `**f`, etc.
+      fc.getIndirectionIndex() >= indirectionIndex
+    )
+  }
 }
 
 /** A reference through an instance field of a union. */
@@ -1877,9 +1902,21 @@ class UnionContent extends Content, TUnionContent {
 
   /** Gets the indirection index of this `UnionContent`. */
   pragma[inline]
-  int getIndirectionIndex() {
+  override int getIndirectionIndex() {
     pragma[only_bind_into](result) = pragma[only_bind_out](indirectionIndex)
   }
+
+  override predicate impliesClearOf(Content c) {
+    exists(UnionContent uc |
+      uc = c and
+      uc.getUnion() = u and
+      // If `this` is `u` then `c` is cleared if it's of the
+      // form `*u`, `**u`, etc. (and we ignore `bytes` because
+      // we know the entire union is overwritten because it's a
+      // union).
+      uc.getIndirectionIndex() >= indirectionIndex
+    )
+  }
 }
 
 /**

cpp/ql/lib/semmle/code/cpp/rangeanalysis/new/internal/semantic/SemanticSSA.qll

@@ -70,7 +70,7 @@ predicate semBackEdge(SemSsaPhiNode phi, SemSsaVariable inp, SemSsaReadPositionP
   // Conservatively assume that every edge is a back edge if we don't have dominance information.
   (
     phi.getBasicBlock().bbDominates(edge.getOrigBlock()) or
-    irreducibleSccEdge(phi.getBasicBlock(), edge.getOrigBlock()) or
+    irreducibleSccEdge(edge.getOrigBlock(), phi.getBasicBlock()) or
     not edge.getOrigBlock().hasDominanceInformation()
   )
 }

cpp/ql/src/Security/CWE/CWE-119/OverrunWriteProductFlow.ql

@@ -120,6 +120,10 @@ module ValidState {
 
     predicate isBarrier(DataFlow::Node node, FlowState state) { none() }
 
+    predicate isBarrierOut(DataFlow::Node node) {
+      node = any(DataFlow::SsaPhiNode phi).getAnInput(true)
+    }
+
     predicate isAdditionalFlowStep(
       DataFlow::Node node1, FlowState state1, DataFlow::Node node2, FlowState state2
     ) {
@@ -233,7 +237,8 @@ module StringSizeConfig implements ProductFlow::StateConfigSig {
     // we use `state2` to remember that there was an offset (in this case an offset of `1`) added
     // to the size of the allocation. This state is then checked in `isSinkPair`.
     exists(state1) and
-    hasSize(bufSource.asConvertedExpr(), sizeSource, state2)
+    hasSize(bufSource.asConvertedExpr(), sizeSource, state2) and
+    validState(sizeSource, state2)
   }
 
   predicate isSinkPair(

cpp/ql/src/experimental/Security/CWE/CWE-193/ConstantSizeArrayOffByOne.ql

@@ -80,14 +80,14 @@ predicate isInvalidPointerDerefSink2(DataFlow::Node sink, Instruction i, string
 
 predicate arrayTypeCand(ArrayType arrayType) {
   any(Variable v).getUnspecifiedType() = arrayType and
-  exists(arrayType.getArraySize())
+  exists(arrayType.getByteSize())
 }
 
-pragma[nomagic]
-predicate arrayTypeHasSizes(ArrayType arr, int baseTypeSize, int arraySize) {
+bindingset[baseTypeSize]
+pragma[inline_late]
+predicate arrayTypeHasSizes(ArrayType arr, int baseTypeSize, int size) {
   arrayTypeCand(arr) and
-  arr.getBaseType().getSize() = baseTypeSize and
-  arr.getArraySize() = arraySize
+  arr.getByteSize() / baseTypeSize = size
 }
 
 bindingset[pai]