Java: Make support for generated as a part of kind.

Author: michaelnebel

java/ql/lib/semmle/code/java/dataflow/ExternalFlow.qll

@@ -404,10 +404,17 @@ private predicate summaryModel(string row) {
   any(SummaryModelCsv s).row(row)
 }
 
+bindingset[input]
+private predicate getKind(string input, string kind, boolean generated) {
+  input.splitAt(":", 0) = "generated" and kind = input.splitAt(":", 1) and generated = true
+  or
+  not input.matches("%:%") and kind = input and generated = false
+}
+
 /** Holds if a source model exists for the given parameters. */
 predicate sourceModel(
   string namespace, string type, boolean subtypes, string name, string signature, string ext,
-  string output, string kind
+  string output, string kind, boolean generated
 ) {
   exists(string row |
     sourceModel(row) and
@@ -419,14 +426,14 @@ predicate sourceModel(
     row.splitAt(";", 4) = signature and
     row.splitAt(";", 5) = ext and
     row.splitAt(";", 6) = output and
-    row.splitAt(";", 7) = kind
+    exists(string k | row.splitAt(";", 7) = k and getKind(k, kind, generated))
   )
 }
 
 /** Holds if a sink model exists for the given parameters. */
 predicate sinkModel(
   string namespace, string type, boolean subtypes, string name, string signature, string ext,
-  string input, string kind
+  string input, string kind, boolean generated
 ) {
   exists(string row |
     sinkModel(row) and
@@ -438,22 +445,22 @@ predicate sinkModel(
     row.splitAt(";", 4) = signature and
     row.splitAt(";", 5) = ext and
     row.splitAt(";", 6) = input and
-    row.splitAt(";", 7) = kind
+    exists(string k | row.splitAt(";", 7) = k and getKind(k, kind, generated))
   )
 }
 
 /** Holds if a summary model exists for the given parameters. */
 predicate summaryModel(
   string namespace, string type, boolean subtypes, string name, string signature, string ext,
-  string input, string output, string kind
+  string input, string output, string kind, boolean generated
 ) {
-  summaryModel(namespace, type, subtypes, name, signature, ext, input, output, kind, _)
+  summaryModel(namespace, type, subtypes, name, signature, ext, input, output, kind, generated, _)
 }
 
 /** Holds if a summary model `row` exists for the given parameters. */
 predicate summaryModel(
   string namespace, string type, boolean subtypes, string name, string signature, string ext,
-  string input, string output, string kind, string row
+  string input, string output, string kind, boolean generated, string row
 ) {
   summaryModel(row) and
   row.splitAt(";", 0) = namespace and
@@ -465,13 +472,13 @@ predicate summaryModel(
   row.splitAt(";", 5) = ext and
   row.splitAt(";", 6) = input and
   row.splitAt(";", 7) = output and
-  row.splitAt(";", 8) = kind
+  exists(string k | row.splitAt(";", 8) = k and getKind(k, kind, generated))
 }
 
 private predicate relevantPackage(string package) {
-  sourceModel(package, _, _, _, _, _, _, _) or
-  sinkModel(package, _, _, _, _, _, _, _) or
-  summaryModel(package, _, _, _, _, _, _, _, _)
+  sourceModel(package, _, _, _, _, _, _, _, _) or
+  sinkModel(package, _, _, _, _, _, _, _, _) or
+  summaryModel(package, _, _, _, _, _, _, _, _, _, _)
 }
 
 private predicate packageLink(string shortpkg, string longpkg) {
@@ -499,25 +506,25 @@ predicate modelCoverage(string package, int pkgs, string kind, string part, int
     part = "source" and
     n =
       strictcount(string subpkg, string type, boolean subtypes, string name, string signature,
-        string ext, string output |
+        string ext, string output, boolean generated |
         canonicalPkgLink(package, subpkg) and
-        sourceModel(subpkg, type, subtypes, name, signature, ext, output, kind)
+        sourceModel(subpkg, type, subtypes, name, signature, ext, output, kind, generated)
       )
     or
     part = "sink" and
     n =
       strictcount(string subpkg, string type, boolean subtypes, string name, string signature,
-        string ext, string input |
+        string ext, string input, boolean generated |
         canonicalPkgLink(package, subpkg) and
-        sinkModel(subpkg, type, subtypes, name, signature, ext, input, kind)
+        sinkModel(subpkg, type, subtypes, name, signature, ext, input, kind, generated)
       )
     or
     part = "summary" and
     n =
       strictcount(string subpkg, string type, boolean subtypes, string name, string signature,
-        string ext, string input, string output |
+        string ext, string input, string output, boolean generated |
         canonicalPkgLink(package, subpkg) and
-        summaryModel(subpkg, type, subtypes, name, signature, ext, input, output, kind)
+        summaryModel(subpkg, type, subtypes, name, signature, ext, input, output, kind, generated)
       )
   )
 }
@@ -527,11 +534,11 @@ module CsvValidation {
   /** Holds if some row in a CSV-based flow model appears to contain typos. */
   query predicate invalidModelRow(string msg) {
     exists(string pred, string namespace, string type, string name, string signature, string ext |
-      sourceModel(namespace, type, _, name, signature, ext, _, _) and pred = "source"
+      sourceModel(namespace, type, _, name, signature, ext, _, _, _) and pred = "source"
       or
-      sinkModel(namespace, type, _, name, signature, ext, _, _) and pred = "sink"
+      sinkModel(namespace, type, _, name, signature, ext, _, _, _) and pred = "sink"
       or
-      summaryModel(namespace, type, _, name, signature, ext, _, _, _) and pred = "summary"
+      summaryModel(namespace, type, _, name, signature, ext, _, _, _, _) and pred = "summary"
     |
       not namespace.regexpMatch("[a-zA-Z0-9_\\.]+") and
       msg = "Dubious namespace \"" + namespace + "\" in " + pred + " model."
@@ -550,9 +557,9 @@ module CsvValidation {
     )
     or
     exists(string pred, string input, string part |
-      sinkModel(_, _, _, _, _, _, input, _) and pred = "sink"
+      sinkModel(_, _, _, _, _, _, input, _, _) and pred = "sink"
       or
-      summaryModel(_, _, _, _, _, _, input, _, _) and pred = "summary"
+      summaryModel(_, _, _, _, _, _, input, _, _, _) and pred = "summary"
     |
       (
         invalidSpecComponent(input, part) and
@@ -567,9 +574,9 @@ module CsvValidation {
     )
     or
     exists(string pred, string output, string part |
-      sourceModel(_, _, _, _, _, _, output, _) and pred = "source"
+      sourceModel(_, _, _, _, _, _, output, _, _) and pred = "source"
       or
-      summaryModel(_, _, _, _, _, _, _, output, _) and pred = "summary"
+      summaryModel(_, _, _, _, _, _, _, output, _, _) and pred = "summary"
     |
       invalidSpecComponent(output, part) and
       not part = "" and
@@ -598,16 +605,22 @@ module CsvValidation {
         msg = "Invalid boolean \"" + b + "\" in " + pred + " model."
       )
     )
+    or
+    exists(string row, string kind | summaryModel(row) |
+      kind = row.splitAt(";", 8) and
+      not kind = ["taint", "value", "generated:taint", "generated:value"] and
+      msg = "Invalid kind \"" + kind + "\" in summary model."
+    )
   }
 }
 
 pragma[nomagic]
 private predicate elementSpec(
   string namespace, string type, boolean subtypes, string name, string signature, string ext
 ) {
-  sourceModel(namespace, type, subtypes, name, signature, ext, _, _) or
-  sinkModel(namespace, type, subtypes, name, signature, ext, _, _) or
-  summaryModel(namespace, type, subtypes, name, signature, ext, _, _, _)
+  sourceModel(namespace, type, subtypes, name, signature, ext, _, _, _) or
+  sinkModel(namespace, type, subtypes, name, signature, ext, _, _, _) or
+  summaryModel(namespace, type, subtypes, name, signature, ext, _, _, _, _)
 }
 
 private string paramsStringPart(Callable c, int i) {

java/ql/lib/semmle/code/java/dataflow/internal/FlowSummaryImplSpecific.qll

@@ -62,9 +62,8 @@ predicate summaryElement(
   exists(
     string namespace, string type, boolean subtypes, string name, string signature, string ext
   |
-    summaryModel(namespace, type, subtypes, name, signature, ext, input, output, kind) and
-    c.asCallable() = interpretElement(namespace, type, subtypes, name, signature, ext) and
-    generated = false
+    summaryModel(namespace, type, subtypes, name, signature, ext, input, output, kind, generated) and
+    c.asCallable() = interpretElement(namespace, type, subtypes, name, signature, ext)
   )
 }
 
@@ -123,9 +122,8 @@ predicate sourceElement(SourceOrSinkElement e, string output, string kind, boole
   exists(
     string namespace, string type, boolean subtypes, string name, string signature, string ext
   |
-    sourceModel(namespace, type, subtypes, name, signature, ext, output, kind) and
-    e = interpretElement(namespace, type, subtypes, name, signature, ext) and
-    generated = false
+    sourceModel(namespace, type, subtypes, name, signature, ext, output, kind, generated) and
+    e = interpretElement(namespace, type, subtypes, name, signature, ext)
   )
 }
 
@@ -138,9 +136,8 @@ predicate sinkElement(SourceOrSinkElement e, string input, string kind, boolean
   exists(
     string namespace, string type, boolean subtypes, string name, string signature, string ext
   |
-    sinkModel(namespace, type, subtypes, name, signature, ext, input, kind) and
-    e = interpretElement(namespace, type, subtypes, name, signature, ext) and
-    generated = false
+    sinkModel(namespace, type, subtypes, name, signature, ext, input, kind, generated) and
+    e = interpretElement(namespace, type, subtypes, name, signature, ext)
   )
 }
 

java/ql/src/utils/flowtestcasegenerator/FlowTestCase.qll

@@ -41,7 +41,7 @@ private class CallableToTest extends Callable {
     exists(
       string namespace, string type, boolean subtypes, string name, string signature, string ext
     |
-      summaryModel(namespace, type, subtypes, name, signature, ext, _, _, _) and
+      summaryModel(namespace, type, subtypes, name, signature, ext, _, _, _, _) and
       this = interpretElement(namespace, type, subtypes, name, signature, ext) and
       this.isPublic() and
       getRootType(this.getDeclaringType()).(RefType).isPublic()
@@ -64,7 +64,8 @@ private newtype TTestCase =
       string inputSpec, string outputSpec
     |
       any(TargetSummaryModelCsv tsmc).row(row) and
-      summaryModel(namespace, type, subtypes, name, signature, ext, inputSpec, outputSpec, kind, row) and
+      summaryModel(namespace, type, subtypes, name, signature, ext, inputSpec, outputSpec, kind,
+        false, row) and
       callable = interpretElement(namespace, type, subtypes, name, signature, ext) and
       Private::External::interpretSpec(inputSpec, input) and
       Private::External::interpretSpec(outputSpec, output)

java/ql/src/utils/flowtestcasegenerator/GenerateFlowTestCase.qll

@@ -21,25 +21,25 @@ query string getAParseFailure(string reason) {
   any(TargetSummaryModelCsv target).row(result) and
   any(SummaryModelCsv model).row(result) and
   (
-    not summaryModel(_, _, _, _, _, _, _, _, _, result) and
+    not summaryModel(_, _, _, _, _, _, _, _, _, _, result) and
     reason = "row could not be parsed"
     or
     exists(
       string namespace, string type, boolean subtypes, string name, string signature, string ext
     |
-      summaryModel(namespace, type, subtypes, name, signature, ext, _, _, _, result) and
+      summaryModel(namespace, type, subtypes, name, signature, ext, _, _, _, _, result) and
       not interpretElement(namespace, type, subtypes, name, signature, ext) instanceof Callable and
       reason = "callable could not be resolved"
     )
     or
     exists(string inputSpec |
-      summaryModel(_, _, _, _, _, _, inputSpec, _, _, result) and
+      summaryModel(_, _, _, _, _, _, inputSpec, _, _, _, result) and
       not Private::External::interpretSpec(inputSpec, _) and
       reason = "input spec could not be parsed"
     )
     or
     exists(string outputSpec |
-      summaryModel(_, _, _, _, _, _, _, outputSpec, _, result) and
+      summaryModel(_, _, _, _, _, _, _, outputSpec, _, _, result) and
       not Private::External::interpretSpec(outputSpec, _) and
       reason = "output spec could not be parsed"
     )