001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  You may obtain a copy of the License at
008 *
009 *      https://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017
018package org.apache.commons.xml.secure;
019
020import java.io.InputStream;
021import java.io.Reader;
022import java.lang.invoke.MethodHandle;
023import java.util.Objects;
024
025import javax.xml.stream.EventFilter;
026import javax.xml.stream.FactoryConfigurationError;
027import javax.xml.stream.StreamFilter;
028import javax.xml.stream.XMLEventReader;
029import javax.xml.stream.XMLInputFactory;
030import javax.xml.stream.XMLReporter;
031import javax.xml.stream.XMLResolver;
032import javax.xml.stream.XMLStreamException;
033import javax.xml.stream.XMLStreamReader;
034import javax.xml.stream.util.XMLEventAllocator;
035import javax.xml.transform.Source;
036
037/**
038 * Creates new, secure {@link XMLInputFactory} instances.
039 * <p>
040 * The three universal guarantees on {@link org.apache.commons.xml.secure} apply; StAX exposes no additional vectors beyond them.
041 * </p>
042 * <p>
043 * This class is not itself a {@link XMLInputFactory}, so it inherits none of the static JAXP factory methods. A caller therefore cannot obtain an unsecured
044 * factory through this class by calling a method such as {@code newDefaultFactory()}. The secure factories are instances of a nested, non-public wrapper class.
045 * </p>
046 *
047 * @see org.apache.commons.xml.secure
048 */
049public final class SecureXMLInputFactory {
050
051    /**
052     * {@link XMLInputFactory} wrapper that installs a non-removable {@link FallbackIgnoreXMLResolver} floor on the delegate's entity-resolution hook so the
053     * caller cannot remove it.
054     * <p>
055     * The constructor installs the floor through {@code setXMLResolver}, which every implementation routes external resolution through (Woodstox fans it out to
056     * both its DTD-subset and entity resolvers). Woodstox keeps one hook outside that fan-out, {@value SecureXMLInputFactory#WSTX_UNDECLARED_ENTITY_RESOLVER},
057     * which is deliberately left empty: emptying the external subset leaves any entity it declared undeclared, and Woodstox then rejects the reference. The
058     * rejection is implementation-prescribed and keeps the resource unfetched, like the empty resolution the other implementations produce; a caller who wants
059     * those references resolved can still set the property, and their resolver lands behind a floor like on every other resolver hook.
060     * </p>
061     * <p>
062     * Every resolver-valued entry point ({@link #setXMLResolver(XMLResolver)}, {@code setProperty(XMLInputFactory.RESOLVER, ...)} and the Woodstox
063     * {@code com.ctc.wstx.*Resolver} keys) is routed uniformly: a caller who supplies their own {@link FallbackIgnoreXMLResolver} takes control and it is
064     * passed straight to the delegate; otherwise the caller's resolver is wrapped in a new floor installed on that hook, so the floor cannot be removed by dropping
065     * the resolver. This matters because Woodstox does not chain resolvers: when a resolver returns {@code null},
066     * {@code DefaultInputResolver} falls through to fetching the systemId URL itself, so a caller-set resolver that returns {@code null} must still land behind
067     * the floor. {@link #getXMLResolver()} and {@code getProperty} report the caller's resolver unwrapped.
068     * </p>
069     * <p>
070     * The floor on a hook is replaced rather than mutated, which is what keeps each hook and each reader independent. Woodstox routes {@code setXMLResolver} to
071     * both its DTD-subset and entity hooks, so one floor object sits on several of them, and it copies that reference into every reader it creates; setting a
072     * delegate on the object in place would therefore also answer hooks the caller never named, and would change the resolution policy of readers already
073     * created, including ones parsing on another thread. Installing a new floor leaves both untouched: a hook keeps whatever floor it was given, and a reader
074     * keeps the one it captured when it was created.
075     * </p>
076     */
077    private static final class Wrapper extends XMLInputFactory {
078
079        private static boolean isResolverProperty(final String name) {
080            return XMLInputFactory.RESOLVER.equals(name)
081                    || WSTX_DTD_RESOLVER.equals(name)
082                    || WSTX_ENTITY_RESOLVER.equals(name)
083                    || WSTX_UNDECLARED_ENTITY_RESOLVER.equals(name);
084        }
085
086
087        private static XMLResolver unwrap(final XMLResolver resolver) {
088            return resolver instanceof FallbackIgnoreXMLResolver ? ((FallbackIgnoreXMLResolver) resolver).getDelegate() : resolver;
089        }
090
091        private final XMLInputFactory delegate;
092
093        /**
094         * Constructs a new instance.
095         *
096         * @param delegate The delegate to wrap; must not be {@code null}.
097         * @throws NullPointerException Thrown if {@code delegate} is {@code null}.
098         */
099        private Wrapper(final XMLInputFactory delegate) {
100            this.delegate = Objects.requireNonNull(delegate, "delegate");
101            delegate.setXMLResolver(new FallbackIgnoreXMLResolver(null));
102        }
103
104        @Override
105        public XMLEventReader createFilteredReader(final XMLEventReader reader, final EventFilter filter) throws XMLStreamException {
106            return delegate.createFilteredReader(reader, filter);
107        }
108
109        @Override
110        public XMLStreamReader createFilteredReader(final XMLStreamReader reader, final StreamFilter filter) throws XMLStreamException {
111            return delegate.createFilteredReader(reader, filter);
112        }
113
114        @Override
115        public XMLEventReader createXMLEventReader(final InputStream stream) throws XMLStreamException {
116            return delegate.createXMLEventReader(stream);
117        }
118
119        @Override
120        public XMLEventReader createXMLEventReader(final InputStream stream, final String encoding) throws XMLStreamException {
121            return delegate.createXMLEventReader(stream, encoding);
122        }
123
124        @Override
125        public XMLEventReader createXMLEventReader(final Reader reader) throws XMLStreamException {
126            return delegate.createXMLEventReader(reader);
127        }
128
129        @Override
130        public XMLEventReader createXMLEventReader(final Source source) throws XMLStreamException {
131            return delegate.createXMLEventReader(source);
132        }
133
134        @Override
135        public XMLEventReader createXMLEventReader(final String systemId, final InputStream stream) throws XMLStreamException {
136            return delegate.createXMLEventReader(systemId, stream);
137        }
138
139        @Override
140        public XMLEventReader createXMLEventReader(final String systemId, final Reader reader) throws XMLStreamException {
141            return delegate.createXMLEventReader(systemId, reader);
142        }
143
144        @Override
145        public XMLEventReader createXMLEventReader(final XMLStreamReader reader) throws XMLStreamException {
146            return delegate.createXMLEventReader(reader);
147        }
148
149        @Override
150        public XMLStreamReader createXMLStreamReader(final InputStream stream) throws XMLStreamException {
151            return delegate.createXMLStreamReader(stream);
152        }
153
154        @Override
155        public XMLStreamReader createXMLStreamReader(final InputStream stream, final String encoding) throws XMLStreamException {
156            return delegate.createXMLStreamReader(stream, encoding);
157        }
158
159        @Override
160        public XMLStreamReader createXMLStreamReader(final Reader reader) throws XMLStreamException {
161            return delegate.createXMLStreamReader(reader);
162        }
163
164        @Override
165        public XMLStreamReader createXMLStreamReader(final Source source) throws XMLStreamException {
166            return delegate.createXMLStreamReader(source);
167        }
168
169        @Override
170        public XMLStreamReader createXMLStreamReader(final String systemId, final InputStream stream) throws XMLStreamException {
171            return delegate.createXMLStreamReader(systemId, stream);
172        }
173
174        @Override
175        public XMLStreamReader createXMLStreamReader(final String systemId, final Reader reader) throws XMLStreamException {
176            return delegate.createXMLStreamReader(systemId, reader);
177        }
178
179
180        @Override
181        public XMLEventAllocator getEventAllocator() {
182            return delegate.getEventAllocator();
183        }
184
185        @Override
186        public Object getProperty(final String name) {
187            if (isResolverProperty(name)) {
188                return unwrap((XMLResolver) delegate.getProperty(name));
189            }
190            return delegate.getProperty(name);
191        }
192
193        @Override
194        public XMLReporter getXMLReporter() {
195            return delegate.getXMLReporter();
196        }
197
198        @Override
199        public XMLResolver getXMLResolver() {
200            return unwrap(delegate.getXMLResolver());
201        }
202
203        @Override
204        public boolean isPropertySupported(final String name) {
205            return delegate.isPropertySupported(name);
206        }
207
208        @Override
209        public void setEventAllocator(final XMLEventAllocator allocator) {
210            delegate.setEventAllocator(allocator);
211        }
212
213        @Override
214        public void setProperty(final String name, final Object value) {
215            // If a resolver property has a value of the wrong type, pass it to the delegate to generate an appropriate exception.
216            if (isResolverProperty(name) && (value == null || value instanceof XMLResolver)) {
217                setResolverProperty(name, (XMLResolver) value);
218            } else {
219                delegate.setProperty(name, value);
220            }
221        }
222
223        /**
224         * Routes a caller-set resolver for the property {@code name} behind the floor currently installed on that hook.
225         *
226         * @param name     The resolver-valued property being set.
227         * @param resolver The caller's resolver, or their own {@link FallbackIgnoreXMLResolver} to take control.
228         */
229        private void setResolverProperty(final String name, final XMLResolver resolver) {
230            if (resolver instanceof FallbackIgnoreXMLResolver) {
231                // The caller supplies their own floor: hand it to the delegate as-is.
232                delegate.setProperty(name, resolver);
233            } else {
234                // A fresh floor for this hook rather than a new delegate on the floor already there: Woodstox puts one floor object on several hooks and copies
235                // the reference into every reader it creates, so mutating it would reach hooks, and readers already parsing, that this call never named.
236                delegate.setProperty(name, new FallbackIgnoreXMLResolver(resolver));
237            }
238        }
239
240        @Override
241        public void setXMLReporter(final XMLReporter reporter) {
242            delegate.setXMLReporter(reporter);
243        }
244
245        @Override
246        public void setXMLResolver(final XMLResolver resolver) {
247            setResolverProperty(XMLInputFactory.RESOLVER, resolver);
248        }
249    }
250    /** Woodstox property: resolver consulted for the external DTD subset. */
251    static final String WSTX_DTD_RESOLVER = "com.ctc.wstx.dtdResolver";
252    /** Woodstox property: resolver consulted for declared external general entities. */
253    static final String WSTX_ENTITY_RESOLVER = "com.ctc.wstx.entityResolver";
254    /** Woodstox property: resolver consulted for undeclared entity references. */
255    static final String WSTX_UNDECLARED_ENTITY_RESOLVER = "com.ctc.wstx.undeclaredEntityResolver";
256
257    /** Class name of the JDK's built-in default implementation, the Java 8 fallback for {@link #newDefaultFactory()}. */
258    private static final String JDK_XML_INPUT_FACTORY = "com.sun.xml.internal.stream.XMLInputFactoryImpl";
259
260    private static final MethodHandle MH_newDefaultInstance = MethodHandleFactory.findStatic(XMLInputFactory.class, "newDefaultFactory");
261
262    /**
263     * Returns a new, secure {@link XMLInputFactory} of the system-default implementation.
264     * <p>
265     * Obtained from {@code XMLInputFactory.newDefaultFactory()} where the platform provides it (Java 9 or later), and by instantiating the JDK's built-in
266     * implementation directly on Java 8.
267     * </p>
268     *
269     * @return A secure factory.
270     * @throws IllegalStateException     Thrown if a required secure setting cannot be applied to the underlying implementation.
271     * @throws FactoryConfigurationError Thrown if the running platform provides neither {@code newDefaultFactory()} nor the JDK's built-in implementation
272     *                                   (for example Android).
273     */
274    public static XMLInputFactory newDefaultFactory() {
275        if (MH_newDefaultInstance != null) {
276            return secure(MethodHandleFactory.invokeExact(() -> (XMLInputFactory) MH_newDefaultInstance.invokeExact(), FactoryConfigurationError.class));
277        }
278        try {
279            // Java 8: the method does not exist, and XMLInputFactory has no class-name-taking lookup; instantiate the JDK's built-in default directly.
280            return secure((XMLInputFactory) Class.forName(JDK_XML_INPUT_FACTORY).getConstructor().newInstance());
281        } catch (final ReflectiveOperationException e) {
282            // Where the class does not exist either (for example Android), report the miss like any StAX factory lookup: with FactoryConfigurationError.
283            throw new FactoryConfigurationError(e, "Neither XMLInputFactory.newDefaultFactory() nor " + JDK_XML_INPUT_FACTORY + " is available");
284        }
285    }
286
287    /**
288     * Returns a new, secure {@link XMLInputFactory}, like {@link XMLInputFactory#newFactory()}.
289     *
290     * @return A secure factory.
291     * @throws IllegalStateException     Thrown if a required secure setting cannot be applied to the underlying implementation.
292     * @throws FactoryConfigurationError Thrown if an instance of this factory cannot be loaded.
293     */
294    public static XMLInputFactory newFactory() {
295        // XMLInputFactory.newInstance, not newFactory: the same specified lookup, but Android's StAX API predates newFactory.
296        return secure(XMLInputFactory.newInstance());
297    }
298
299    /**
300     * Returns a new, secure {@link XMLInputFactory} resolved from the given factory id.
301     *
302     * @param factoryId   The name of the factory to find; a system property or service id to look up, not the class name of the implementation.
303     * @param classLoader The class loader used in the lookup; {@code null} means the current thread's context class loader.
304     * @return A secure factory.
305     * @throws IllegalStateException     Thrown if a required secure setting cannot be applied to the underlying implementation.
306     * @throws FactoryConfigurationError Thrown in case of a service configuration error or if the implementation is not available or cannot be instantiated.
307     * @throws NullPointerException      Thrown if {@code factoryId} is {@code null}.
308     */
309    public static XMLInputFactory newFactory(final String factoryId, final ClassLoader classLoader) {
310        return secure(XMLInputFactory.newFactory(factoryId, classLoader));
311    }
312
313    /**
314     * Returns a new, secure {@link XMLInputFactory}.
315     *
316     * @return A secure factory.
317     * @throws IllegalStateException     Thrown if a required secure setting cannot be applied to the underlying implementation.
318     * @throws FactoryConfigurationError Thrown if an instance of this factory cannot be loaded.
319     */
320    public static XMLInputFactory newInstance() {
321        return secure(XMLInputFactory.newInstance());
322    }
323
324    /**
325     * Capability-driven secure for any {@link XMLInputFactory} (StAX) on the classpath.
326     *
327     * <p>One recipe covers both the JDK Zephyr and Woodstox: the wrapper installs a non-removable {@link FallbackIgnoreXMLResolver} floor on
328     * every entity-resolution hook, leaving the standard {@code SUPPORT_DTD} / {@code IS_SUPPORTING_EXTERNAL_ENTITIES} defaults untouched; see the wrapper's
329     * Javadoc for the per-implementation hook routing.</p>
330     *
331     * @param factory The factory to secure; never {@code null}.
332     * @return a secure factory.
333     */
334    static XMLInputFactory secure(final XMLInputFactory factory) {
335        // The wrapper installs the non-removable ignore-all resolver floor that resolves every external DTD and entity to empty content.
336        return new Wrapper(factory);
337    }
338
339    private SecureXMLInputFactory() {
340        // static only
341    }
342}