Merge branch '5.3'

Author: cowtowncoder

release-notes/VERSION

@@ -110,12 +110,14 @@ Project: woodstox
 #78: Shade MSV dependency
 - Add `module-info.class` for Java 9+ module system
 
-5.3.1 (not yet released)
+5.4.0 (not yet released)
 
 #104: `NullPointerException` in `DTDValidator.validateElementEnd()`
   for element undefined in DTD
  (reported by ChrisTrenkamp@github)
 #105: W3CSchemaFactory constructor incorrectly references relaxng
+#160: Add limit and configuration setting for maximum nesting for DTD subsets
+ (contributed by @pjfanning)
 
 5.3.0 (15-Jul-2019)
 

src/main/java/com/ctc/wstx/api/ReaderConfig.java

@@ -46,6 +46,9 @@ public final class ReaderConfig
     public final static int DEFAULT_MAX_ENTITY_DEPTH = 500;
     public final static int DEFAULT_MAX_ENTITY_COUNT = 100 * 1000;
 
+    // @since 5.4/6.4
+    public final static int DEFAULT_MAX_DTD_DEPTH = 500;
+
     /*
     ///////////////////////////////////////////////////////////////////////
     // Constants for reader properties:
@@ -133,7 +136,9 @@ public final class ReaderConfig
     final static int PROP_MAX_TEXT_LENGTH = 66;
     final static int PROP_MAX_ENTITY_COUNT = 67;
     final static int PROP_MAX_ENTITY_DEPTH = 68;
-    
+
+    final static int PROP_MAX_DTD_DEPTH = 69;
+
     /*
     ////////////////////////////////////////////////
     // Limits for numeric properties
@@ -341,6 +346,10 @@ public final class ReaderConfig
                  PROP_MAX_ENTITY_DEPTH);
         sProperties.put(WstxInputProperties.P_MAX_ENTITY_COUNT,
                  PROP_MAX_ENTITY_COUNT);
+        // since 5.4/6.4
+        sProperties.put(WstxInputProperties.P_MAX_DTD_DEPTH,
+                PROP_MAX_DTD_DEPTH);
+
         sProperties.put(WstxInputProperties.P_MAX_CHARACTERS, PROP_MAX_CHARACTERS);
         sProperties.put(WstxInputProperties.P_CUSTOM_INTERNAL_ENTITIES,
                 Integer.valueOf(PROP_CUSTOM_INTERNAL_ENTITIES));
@@ -400,6 +409,9 @@ public final class ReaderConfig
 
     protected int mMaxEntityDepth = DEFAULT_MAX_ENTITY_DEPTH;
     protected long mMaxEntityCount = DEFAULT_MAX_ENTITY_COUNT;
+
+    // since 5.4/6.4
+    protected int mMaxDtdDepth = DEFAULT_MAX_DTD_DEPTH;
 
     /**
      * Base URL to use as the resolution context for relative entity
@@ -506,6 +518,7 @@ private ReaderConfig(ReaderConfig base,
             mMaxTextLength = base.mMaxTextLength;
             mMaxEntityDepth = base.mMaxEntityDepth;
             mMaxEntityCount = base.mMaxEntityCount;
+            mMaxDtdDepth = base.mMaxDtdDepth;
         }
 
         /* Ok, let's then see if we can find a buffer recycler. Since they
@@ -569,6 +582,7 @@ public ReaderConfig createNonShared(SymbolTable sym)
         rc.mMaxElementDepth = mMaxElementDepth;
         rc.mMaxEntityDepth = mMaxEntityDepth;
         rc.mMaxEntityCount = mMaxEntityCount;
+        rc.mMaxDtdDepth = mMaxDtdDepth;
         if (mSpecialProperties != null) {
             int len = mSpecialProperties.length;
             Object[] specProps = new Object[len];
@@ -735,6 +749,8 @@ public boolean willAllowXml11EscapedCharsInXml10() {
     public int getMaxEntityDepth() { return mMaxEntityDepth; }
     public long getMaxEntityCount() { return mMaxEntityCount; }
 
+    public int getMaxDtdDepth() { return mMaxDtdDepth; }
+
     public long getMaxCharacters() { return mMaxCharacters; }
     public long getMaxTextLength() { return mMaxTextLength; }
 
@@ -1000,6 +1016,10 @@ public void setMaxEntityDepth(int value) {
     public void setMaxEntityCount(long value) {
         mMaxEntityCount = value;
     }
+    // @since 5.4/6.4
+    public void setMaxDtdDepth(int value) {
+        mMaxDtdDepth = value;
+    }
 
     public void setCustomInternalEntities(Map<String,?> m)
     {
@@ -1498,6 +1518,8 @@ public Object getProperty(int id)
             return getMaxEntityDepth();
         case PROP_MAX_ENTITY_COUNT:
             return getMaxEntityCount();
+        case PROP_MAX_DTD_DEPTH:
+            return getMaxDtdDepth();
 
         case PROP_MIN_TEXT_SEGMENT:
             return getShortestReportedTextSegment();
@@ -1686,6 +1708,9 @@ public boolean setProperty(String propName, int id, Object value)
         case PROP_MAX_ENTITY_COUNT:
             setMaxEntityCount(ArgUtil.convertToLong(propName, value, 1));
             break;
+        case PROP_MAX_DTD_DEPTH:
+            setMaxDtdDepth(ArgUtil.convertToInt(propName, value, 1));
+            break;
 
         case PROP_MIN_TEXT_SEGMENT:
             setShortestReportedTextSegment(ArgUtil.convertToInt(propName, value, 1));

src/main/java/com/ctc/wstx/api/WstxInputProperties.java

@@ -184,7 +184,6 @@ public final class WstxInputProperties
 
     // // // Constraints on sizes of text segments parsed:
 
-
     /**
      * Property to specify shortest non-complete text segment (part of
      * CDATA section or text content) that parser is allowed to return,
@@ -252,6 +251,15 @@ public final class WstxInputProperties
      */
     public final static String P_MAX_ENTITY_DEPTH = "com.ctc.wstx.maxEntityDepth";
 
+    // and yet more size constraints (4.3+)
+
+    /**
+     * Maximum level of nesting of XML elements, starting with root element.
+     *
+     * @since 5.4 / 6.4
+     */
+    public final static String P_MAX_DTD_DEPTH = "com.ctc.wstx.maxDtdDepth";
+    
     // // // Entity handling
 
     /**

src/main/java/com/ctc/wstx/dtd/FullDTDReader.java

@@ -2266,7 +2266,7 @@ private void handleElementDecl()
                 vldContent = XMLValidator.CONTENT_ALLOW_ANY_TEXT; // checked against DTD
             } else {
                 --mInputPtr; // let's push it back...
-                ContentSpec spec = readContentSpec(elemName, true, mCfgFullyValidating);
+                ContentSpec spec = readContentSpec(elemName, mCfgFullyValidating, 0);
                 val = spec.getSimpleValidator();
                 if (val == null) {
                     val = new DFAValidator(DFAState.constructDFA(spec));
@@ -3044,13 +3044,14 @@ private StructValidator readMixedSpec(PrefixedName elemName, boolean construct)
         return val;
     }
 
-    /**
-	 * @param mainLevel Whether this is the main-level content specification or nested 
-	 */
-    private ContentSpec readContentSpec(PrefixedName elemName, boolean mainLevel,
-                                        boolean construct)
+    private ContentSpec readContentSpec(final PrefixedName elemName, final boolean construct,
+            final int recursionDepth)
         throws XMLStreamException
     {
+        verifyLimit("Maximum DTD nesting depth (WstxInputProperties.P_MAX_DTD_DEPTH)",
+                mConfig.getMaxDtdDepth(),
+                recursionDepth);
+
         ArrayList<ContentSpec> subSpecs = new ArrayList<ContentSpec>();
         boolean isChoice = false; // default to sequence
         boolean choiceSet = false;
@@ -3082,7 +3083,7 @@ private ContentSpec readContentSpec(PrefixedName elemName, boolean mainLevel,
                 }
             }
             if (c == '(') {
-                ContentSpec cs = readContentSpec(elemName, false, construct);
+                ContentSpec cs = readContentSpec(elemName, construct, recursionDepth + 1);
                 subSpecs.add(cs);
                 continue;
             }

src/test/java/wstxtest/BaseWstxTest.java

@@ -494,6 +494,29 @@ protected static String getAndVerifyText(XMLStreamReader sr)
         return text;
     }
 
+    protected byte[] readResource(String ref)
+    {
+        ByteArrayOutputStream bytes = new ByteArrayOutputStream();
+        final byte[] buf = new byte[4000];
+
+        InputStream in = getClass().getResourceAsStream(ref);
+        if (in != null) {
+            try {
+                int len;
+                while ((len = in.read(buf)) > 0) {
+                    bytes.write(buf, 0, len);
+                }
+                in.close();
+            } catch (IOException e) {
+                throw new RuntimeException("Failed to read resource '"+ref+"': "+e);
+            }
+        }
+        if (bytes.size() == 0) {
+            throw new IllegalArgumentException("Failed to read resource '"+ref+"': empty resource?");
+        }
+        return bytes.toByteArray();
+    }
+
     /*
     //////////////////////////////////////////////////
     // Misc other helpers

src/test/java/wstxtest/fuzz/Fuzz_DTDReadTest.java

@@ -0,0 +1,80 @@
+package wstxtest.fuzz;
+
+import java.io.ByteArrayInputStream;
+import java.io.InputStreamReader;
+import java.io.Reader;
+
+import javax.xml.stream.XMLStreamReader;
+
+import com.ctc.wstx.api.WstxInputProperties;
+import com.ctc.wstx.exc.WstxLazyException;
+import com.ctc.wstx.stax.WstxInputFactory;
+import org.codehaus.stax2.io.Stax2ByteArraySource;
+import wstxtest.stream.BaseStreamTest;
+
+public class Fuzz_DTDReadTest extends BaseStreamTest
+{
+    private final byte[] DOC = readResource("/fuzz/clusterfuzz-testcase-modified-XmlFuzzer-5219006592450560.txt");
+
+    private final WstxInputFactory STAX_F = getWstxInputFactory();
+
+    public void testIssueInputStream() throws Exception
+    {
+        XMLStreamReader sr = STAX_F.createXMLStreamReader(new ByteArrayInputStream(DOC));
+        try {
+            streamThrough(sr);
+            fail("Should not pass");
+        } catch (WstxLazyException e) {
+            verifyException(e, "Maximum DTD nesting depth");
+            verifyException(e, "500");
+        }
+        sr.close();
+    }
+
+    public void testIssueInputStreamHigherRecursionLimit() throws Exception
+    {
+        final WstxInputFactory staxF = getWstxInputFactory();
+        staxF.setProperty(WstxInputProperties.P_MAX_DTD_DEPTH, 1100);
+
+        XMLStreamReader sr = staxF.createXMLStreamReader(new ByteArrayInputStream(DOC));
+        try {
+            streamThrough(sr);
+            fail("Should not pass");
+        } catch (WstxLazyException e) {
+            verifyException(e, "Maximum DTD nesting depth");
+            verifyException(e, "1100");
+        }
+        sr.close();
+    }
+    
+    public void testIssueReader() throws Exception
+    {
+        Reader r = new InputStreamReader(new ByteArrayInputStream(DOC),
+                "UTF-8");
+        XMLStreamReader sr = STAX_F.createXMLStreamReader(r);
+        try {
+            streamThrough(sr);
+            fail("Should not pass");
+        } catch (WstxLazyException e) {
+            verifyException(e, "Maximum DTD nesting depth");
+            verifyException(e, "500");
+        }
+        sr.close();
+    }
+
+    public void testIssueStax2ByteArray() throws Exception
+    {
+        // Then "native" Byte array
+        Stax2ByteArraySource src = new Stax2ByteArraySource(DOC, 0, DOC.length);
+        XMLStreamReader sr = STAX_F.createXMLStreamReader(src);
+        try {
+            streamThrough(sr);
+            fail("Should not pass");
+        } catch (WstxLazyException e) {
+            verifyException(e, "Maximum DTD nesting depth");
+            verifyException(e, "500");
+        }
+        sr.close();
+    }
+}
+