HDFS-3246: pRead equivalent for direct read path (apache#597)

HDFS-3246: pRead equivalent for direct read path

Contributed by Sahil Takiar

Author: sahilTakiar

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/crypto/CryptoInputStream.java

@@ -0,0 +1,66 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.hadoop.fs;
+
+import java.io.IOException;
+import java.nio.ByteBuffer;
+
+import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceStability;
+
+/**
+ * Implementers of this interface provide a positioned read API that writes to a
+ * {@link ByteBuffer} rather than a {@code byte[]}.
+ *
+ * @see PositionedReadable
+ * @see ByteBufferReadable
+ */
+@InterfaceAudience.Public
+@InterfaceStability.Evolving
+public interface ByteBufferPositionedReadable {
+  /**
+   * Reads up to {@code buf.remaining()} bytes into buf from a given position
+   * in the file and returns the number of bytes read. Callers should use
+   * {@code buf.limit(...)} to control the size of the desired read and
+   * {@code buf.position(...)} to control the offset into the buffer the data
+   * should be written to.
+   * <p>
+   * After a successful call, {@code buf.position()} will be advanced by the
+   * number of bytes read and {@code buf.limit()} will be unchanged.
+   * <p>
+   * In the case of an exception, the state of the buffer (the contents of the
+   * buffer, the {@code buf.position()}, the {@code buf.limit()}, etc.) is
+   * undefined, and callers should be prepared to recover from this
+   * eventuality.
+   * <p>
+   * Callers should use {@link StreamCapabilities#hasCapability(String)} with
+   * {@link StreamCapabilities#PREADBYTEBUFFER} to check if the underlying
+   * stream supports this interface, otherwise they might get a
+   * {@link UnsupportedOperationException}.
+   * <p>
+   * Implementations should treat 0-length requests as legitimate, and must not
+   * signal an error upon their receipt.
+   *
+   * @param position position within file
+   * @param buf the ByteBuffer to receive the results of the read operation.
+   * @return the number of bytes read, possibly zero, or -1 if reached
+   *         end-of-stream
+   * @throws IOException if there is some error performing the read
+   */
+  int read(long position, ByteBuffer buf) throws IOException;
+}

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/ByteBufferPositionedReadable.java

@@ -19,6 +19,7 @@
 
 import java.io.IOException;
 import java.nio.ByteBuffer;
+
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
 
@@ -32,18 +33,20 @@ public interface ByteBufferReadable {
   /**
    * Reads up to buf.remaining() bytes into buf. Callers should use
    * buf.limit(..) to control the size of the desired read.
-   * <p/>
-   * After a successful call, buf.position() will be advanced by the number 
-   * of bytes read and buf.limit() should be unchanged.
-   * <p/>
-   * In the case of an exception, the values of buf.position() and buf.limit()
-   * are undefined, and callers should be prepared to recover from this
+   * <p>
+   * After a successful call, {@code buf.position()} will be advanced by the
+   * number of bytes read and {@code buf.limit()} will be unchanged.
+   * <p>
+   * In the case of an exception, the state of the buffer (the contents of the
+   * buffer, the {@code buf.position()}, the {@code buf.limit()}, etc.) is
+   * undefined, and callers should be prepared to recover from this
    * eventuality.
-   * <p/>
-   * Many implementations will throw {@link UnsupportedOperationException}, so
-   * callers that are not confident in support for this method from the
-   * underlying filesystem should be prepared to handle that exception.
-   * <p/>
+   * <p>
+   * Callers should use {@link StreamCapabilities#hasCapability(String)} with
+   * {@link StreamCapabilities#READBYTEBUFFER} to check if the underlying
+   * stream supports this interface, otherwise they might get a
+   * {@link UnsupportedOperationException}.
+   * <p>
    * Implementations should treat 0-length requests as legitimate, and must not
    * signal an error upon their receipt.
    *

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/ByteBufferReadable.java

@@ -38,7 +38,8 @@
 public class FSDataInputStream extends DataInputStream
     implements Seekable, PositionedReadable, 
       ByteBufferReadable, HasFileDescriptor, CanSetDropBehind, CanSetReadahead,
-      HasEnhancedByteBufferAccess, CanUnbuffer, StreamCapabilities {
+      HasEnhancedByteBufferAccess, CanUnbuffer, StreamCapabilities,
+      ByteBufferPositionedReadable {
   /**
    * Map ByteBuffers that we have handed out to readers to ByteBufferPool 
    * objects
@@ -147,7 +148,8 @@ public int read(ByteBuffer buf) throws IOException {
       return ((ByteBufferReadable)in).read(buf);
     }
 
-    throw new UnsupportedOperationException("Byte-buffer read unsupported by input stream");
+    throw new UnsupportedOperationException("Byte-buffer read unsupported " +
+            "by input stream");
   }
 
   @Override
@@ -246,4 +248,13 @@ public boolean hasCapability(String capability) {
   public String toString() {
     return super.toString() + ": " + in;
   }
+
+  @Override
+  public int read(long position, ByteBuffer buf) throws IOException {
+    if (in instanceof ByteBufferPositionedReadable) {
+      return ((ByteBufferPositionedReadable) in).read(position, buf);
+    }
+    throw new UnsupportedOperationException("Byte-buffer pread unsupported " +
+        "by input stream");
+  }
 }

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/FSDataInputStream.java

@@ -59,6 +59,18 @@ public interface StreamCapabilities {
    */
   String UNBUFFER = "in:unbuffer";
 
+  /**
+   * Stream read(ByteBuffer) capability implemented by
+   * {@link ByteBufferReadable#read(java.nio.ByteBuffer)}.
+   */
+  String READBYTEBUFFER = "in:readbytebuffer";
+
+  /**
+   * Stream read(long, ByteBuffer) capability implemented by
+   * {@link ByteBufferPositionedReadable#read(long, java.nio.ByteBuffer)}.
+   */
+  String PREADBYTEBUFFER = "in:preadbytebuffer";
+
   /**
    * Capabilities that a stream can support and be queried for.
    */