Rust: Fix bad join

hvitved · hvitved · commit aabcc108ddfd · 2024-12-16T10:20:30.000+01:00
```
[2024-12-16 10:10:36] (247s) Tuple counts for DataFlowImpl::RustDataFlow::storeStep/3#98e80e57/3@0618fdm6 after 3m8s:
                      33711       ~0%        {3} r1 = SCAN `DataFlowImpl::VariableCapture::storeStep/3#cb0fdcf6` OUTPUT In.1, In.0 'node1', In.2 'node2'
                      33711       ~6%        {3}    | JOIN WITH DataFlowImpl::TSingletonContentSet#9b15eaba ON FIRST 1 OUTPUT Lhs.1 'node1', Rhs.1 'cs', Lhs.2 'node2'

                      0           ~0%        {3} r2 = JOIN `FlowSummaryImpl::Private::Steps::summaryStoreStep/3#2c853d0d` WITH DataFlowImpl::TFlowSummaryNode#2b28ecb7 ON FIRST 1 OUTPUT Lhs.2, Lhs.1 'cs', Rhs.1 'node1'
                      0           ~0%        {3}    | JOIN WITH DataFlowImpl::TFlowSummaryNode#2b28ecb7 ON FIRST 1 OUTPUT Lhs.2 'node1', Lhs.1 'cs', Rhs.1 'node1'

                      1554        ~0%        {3} r3 = JOIN _DataFlowImpl::TExprNode#83a34c2e__DataFlowImpl::TArrayElement#b9fb9b7b_DataFlowImpl::TSingletonCont__#shared WITH `CfgNodes::ArrayRepeatExprCfgNode.getRepeatOperand/0#dispred#b264e402_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2 'node1'
                      1554        ~0%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Lhs.2 'node1', Lhs.1 'cs', Rhs.1 'node2'

                      870         ~2%        {3} r4 = SCAN `DataFlowImpl::RustDataFlow::tupleAssignment/3#bf3c8690` OUTPUT In.2, In.0 'node1', In.1
                      870         ~0%        {3}    | JOIN WITH DataFlowImpl::TSingletonContentSet#9b15eaba ON FIRST 1 OUTPUT Lhs.2, Rhs.1 'cs', Lhs.1 'node1'
                      870         ~0%        {3}    | JOIN WITH `DataFlowImpl::Node::PostUpdateNode.getPreUpdateNode/0#dispred#53daedc2_10#join_rhs` ON FIRST 1 OUTPUT Lhs.2 'node1', Lhs.1 'cs', Rhs.1 'node2'

                      40037       ~4%        {3} r5 = JOIN _DataFlowImpl::TExprNode#83a34c2e__DataFlowImpl::TArrayElement#b9fb9b7b_DataFlowImpl::TSingletonCont__#shared WITH `CfgNodes::ArrayExprCfgNode.getAnExpr/0#dispred#9d00a6f1_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2 'node1'
                      36929       ~4%        {3}    | JOIN WITH CfgNodes::ArrayListExprCfgNode#07eee614 ON FIRST 1 OUTPUT Lhs.0, Lhs.1 'cs', Lhs.2 'node1'
                      36929       ~0%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Lhs.2 'node1', Lhs.1 'cs', Rhs.1 'node2'

                      14          ~0%        {2} r6 = JOIN DataFlowImpl::TTuplePositionContent#f1d90606_10#join_rhs WITH DataFlowImpl::TSingletonContentSet#9b15eaba ON FIRST 1 OUTPUT Lhs.1, Rhs.1 'cs'
                      47949       ~0%        {3}    | JOIN WITH `CfgNodes::TupleExprCfgNode.getField/1#dispred#9f7c9c63_102#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Rhs.2
                      47949       ~0%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Lhs.2, Lhs.1 'cs', Rhs.1 'node2'
                      47949       ~2%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Rhs.1 'node2', Lhs.1 'cs', Lhs.2 'node2'

                      59801       ~0%        {3} r7 = JOIN _DataFlowImpl::TSingletonContentSet#9b15eaba_DataFlowImpl::TVariantPositionContent#ca6baca0_201#join__#shared WITH `DataFlowImpl::RustDataFlow::tupleVariantConstruction/2#10613c55_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2
                      45509       ~0%        {3}    | JOIN WITH CfgNodes::CallExprCfgNode#9c2a4686_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2
                      45509       ~2%        {4}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Lhs.0, Lhs.2, Lhs.1 'cs', Rhs.1 'node2'
                      45509       ~0%        {3}    | JOIN WITH `CfgNodes::CallExprBaseCfgNode.getArgument/1#dispred#9ebb27c0` ON FIRST 2 OUTPUT Rhs.2, Lhs.2 'cs', Lhs.3 'node2'
                      45509       ~0%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Rhs.1 'node2', Lhs.1 'cs', Lhs.2 'node2'

                      75147       ~1%        {3} r8 = JOIN _DataFlowImpl::TSingletonContentSet#9b15eaba_DataFlowImpl::TStructFieldContent#1d6d7b05_201#join_rhs#shared WITH `DataFlowImpl::RustDataFlow::structConstruction/2#a9656db0_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2
                      59186       ~3%        {3}    | JOIN WITH `CfgNodes::RecordExprCfgNode.getRecordExpr/0#dispred#659ad1af_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2

                      5641        ~2%        {3} r9 = JOIN _DataFlowImpl::TSingletonContentSet#9b15eaba_DataFlowImpl::TVariantFieldContent#4e05bcf1_201#join_rh__#shared WITH `DataFlowImpl::RustDataFlow::recordVariantConstruction/2#34b016f6_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2
                      5268        ~0%        {3}    | JOIN WITH `CfgNodes::RecordExprCfgNode.getRecordExpr/0#dispred#659ad1af_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2

                      64454       ~1%        {3} r10 = r8 UNION r9
                      64454       ~0%        {4}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Lhs.0, Lhs.2, Lhs.1 'cs', Rhs.1 'node2'
                      25923       ~0%        {3}    | JOIN WITH `CfgNodes::RecordExprCfgNode.getFieldExpr/1#d72dca6e` ON FIRST 2 OUTPUT Rhs.2, Lhs.2 'cs', Lhs.3 'node2'
                      25923       ~0%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Rhs.1 'node2', Lhs.1 'cs', Lhs.2 'node2'

                      67759289500 ~251%      {4} r11 = JOIN DataFlowImpl::TSingletonContentSet#9b15eaba WITH DataFlowImpl::TExprNode#83a34c2e CARTESIAN PRODUCT OUTPUT Lhs.0, Lhs.1 'cs', Rhs.0, Rhs.1 'node2'
                      3568000     ~1488%     {3}    | JOIN WITH DataFlowImpl::TArrayElement#b9fb9b7b ON FIRST 1 OUTPUT Lhs.3, Lhs.1 'cs', Lhs.2
                      1223000     ~1291%     {3}    | JOIN WITH `DataFlowImpl::Node::PostUpdateNode.getPreUpdateNode/0#dispred#53daedc2_10#join_rhs` ON FIRST 1 OUTPUT Lhs.2, Lhs.1 'cs', Rhs.1 'node2'
                      11500       ~0%        {3}    | JOIN WITH `CfgNodes::IndexExprCfgNode.getBase/0#dispred#19aba7d8_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2 'node2'
                      1000        ~3%        {3}    | JOIN WITH `CfgNodes::BinaryExprCfgNode.getLhs/0#dispred#bd1c02e7_10#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2 'node2'
                      500         ~3%        {3}    | JOIN WITH CfgNodes::AssignmentExprCfgNode#a9a5c022 ON FIRST 1 OUTPUT Lhs.0, Lhs.1 'cs', Lhs.2 'node2'
                      0           ~0%        {3}    | JOIN WITH `CfgNodes::BinaryExprCfgNode.getRhs/0#dispred#4a1146e4` ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'cs', Lhs.2 'node2'
                      0           ~0%        {3}    | JOIN WITH DataFlowImpl::TExprNode#83a34c2e ON FIRST 1 OUTPUT Rhs.1 'node2', Lhs.1 'cs', Lhs.2 'node2'

                      192445      ~1%        {3} r12 = r1 UNION r2 UNION r3 UNION r4 UNION r5 UNION r6 UNION r7 UNION r10 UNION r11
                                             return r12
```
diff --git a/rust/ql/lib/codeql/rust/dataflow/internal/DataFlowImpl.qll b/rust/ql/lib/codeql/rust/dataflow/internal/DataFlowImpl.qll
@@ -1074,57 +1074,59 @@ module RustDataFlow implements InputSig<Location> {
     )
   }
 
+  pragma[nomagic]
+  private predicate storeContentStep(Node node1, Content c, Node node2) {
+    exists(CallExprCfgNode call, int pos |
+      tupleVariantConstruction(call.getCallExpr(),
+        c.(VariantPositionContent).getVariantCanonicalPath(pos)) and
+      node1.asExpr() = call.getArgument(pos) and
+      node2.asExpr() = call
+    )
+    or
+    exists(RecordExprCfgNode re, string field |
+      (
+        // Expression is for a struct-like enum variant.
+        recordVariantConstruction(re.getRecordExpr(),
+          c.(VariantFieldContent).getVariantCanonicalPath(field))
+        or
+        // Expression is for a struct.
+        structConstruction(re.getRecordExpr(), c.(StructFieldContent).getStructCanonicalPath(field))
+      ) and
+      node1.asExpr() = re.getFieldExpr(field) and
+      node2.asExpr() = re
+    )
+    or
+    exists(TupleExprCfgNode tuple |
+      node1.asExpr() = tuple.getField(c.(TuplePositionContent).getPosition()) and
+      node2.asExpr() = tuple
+    )
+    or
+    c instanceof ArrayElementContent and
+    node1.asExpr() =
+      [
+        node2.asExpr().(ArrayRepeatExprCfgNode).getRepeatOperand(),
+        node2.asExpr().(ArrayListExprCfgNode).getAnExpr()
+      ]
+    or
+    tupleAssignment(node1, node2.(PostUpdateNode).getPreUpdateNode(), c)
+    or
+    exists(AssignmentExprCfgNode assignment, IndexExprCfgNode index |
+      c instanceof ArrayElementContent and
+      assignment.getLhs() = index and
+      node1.asExpr() = assignment.getRhs() and
+      node2.(PostUpdateNode).getPreUpdateNode().asExpr() = index.getBase()
+    )
+    or
+    VariableCapture::storeStep(node1, c, node2)
+  }
+
   /**
    * Holds if data can flow from `node1` to `node2` via a store into `c`.  Thus,
    * `node2` references an object with a content `c.getAStoreContent()` that
    * contains the value of `node1`.
    */
   predicate storeStep(Node node1, ContentSet cs, Node node2) {
-    exists(Content c | c = cs.(SingletonContentSet).getContent() |
-      exists(CallExprCfgNode call, int pos |
-        tupleVariantConstruction(call.getCallExpr(),
-          c.(VariantPositionContent).getVariantCanonicalPath(pos)) and
-        node1.asExpr() = call.getArgument(pos) and
-        node2.asExpr() = call
-      )
-      or
-      exists(RecordExprCfgNode re, string field |
-        (
-          // Expression is for a struct-like enum variant.
-          recordVariantConstruction(re.getRecordExpr(),
-            c.(VariantFieldContent).getVariantCanonicalPath(field))
-          or
-          // Expression is for a struct.
-          structConstruction(re.getRecordExpr(),
-            c.(StructFieldContent).getStructCanonicalPath(field))
-        ) and
-        node1.asExpr() = re.getFieldExpr(field) and
-        node2.asExpr() = re
-      )
-      or
-      exists(TupleExprCfgNode tuple |
-        node1.asExpr() = tuple.getField(c.(TuplePositionContent).getPosition()) and
-        node2.asExpr() = tuple
-      )
-      or
-      c instanceof ArrayElementContent and
-      node1.asExpr() =
-        [
-          node2.asExpr().(ArrayRepeatExprCfgNode).getRepeatOperand(),
-          node2.asExpr().(ArrayListExprCfgNode).getAnExpr()
-        ]
-      or
-      tupleAssignment(node1, node2.(PostUpdateNode).getPreUpdateNode(), c)
-      or
-      exists(AssignmentExprCfgNode assignment, IndexExprCfgNode index |
-        c instanceof ArrayElementContent and
-        assignment.getLhs() = index and
-        node1.asExpr() = assignment.getRhs() and
-        node2.(PostUpdateNode).getPreUpdateNode().asExpr() = index.getBase()
-      )
-      or
-      VariableCapture::storeStep(node1, c, node2)
-    )
+    storeContentStep(node1, cs.(SingletonContentSet).getContent(), node2)
     or
     FlowSummaryImpl::Private::Steps::summaryStoreStep(node1.(Node::FlowSummaryNode).getSummaryNode(),
       cs, node2.(Node::FlowSummaryNode).getSummaryNode())
