spark-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From vanzin <...@git.apache.org>
Subject [GitHub] spark pull request: SPARK-10771: Implement the shuffle encryption ...
Date Thu, 08 Oct 2015 23:34:57 GMT
Github user vanzin commented on a diff in the pull request:

    https://github.com/apache/spark/pull/8880#discussion_r41583434
  
    --- Diff: core/src/main/scala/org/apache/spark/crypto/CryptoInputStream.scala ---
    @@ -0,0 +1,439 @@
    +/*
    + * 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.spark.crypto
    +
    +import java.io.{IOException, InputStream, FilterInputStream}
    +import java.nio.ByteBuffer
    +import java.nio.channels.ReadableByteChannel
    +import java.security.GeneralSecurityException
    +import java.util.Queue
    +import java.util.concurrent.ConcurrentLinkedQueue
    +
    +import com.google.common.base.Preconditions
    +
    +/**
    + * CryptoInputStream decrypts data. It is not thread-safe. AES CTR mode is
    + * required in order to ensure that the plain text and cipher text have a 1:1
    + * mapping. The decryption is buffer based. The key points of the decryption
    + * are (1) calculating the counter and (2) padding through stream position:
    + * <p/>
    + * counter = base + pos/(algorithm blocksize);
    + * padding = pos%(algorithm blocksize);
    + * <p/>
    + * The underlying stream offset is maintained as state.
    + */
    +class CryptoInputStream(
    +    in: InputStream,
    +    codecVal: CryptoCodec,
    +    bufferSizeVal: Integer,
    +    keyVal: Array[Byte],
    +    ivVal: Array[Byte],
    +    streamOffsetVal: Long)
    +  extends FilterInputStream(in: InputStream) with ReadableByteChannel {
    +  val oneByteBuf = new Array[Byte](1)
    +  val codec = codecVal
    +
    +  val bufferSize = CryptoStreamUtils.checkBufferSize(codecVal, bufferSizeVal)
    +  /**
    +   * Input data buffer. The data starts at inBuffer.position() and ends at
    +   * to inBuffer.limit().
    +   */
    +  val inBuffer: ByteBuffer = ByteBuffer.allocateDirect(bufferSizeVal)
    +
    +  /**
    +   * The decrypted data buffer. The data starts at outBuffer.position() and
    +   * ends at outBuffer.limit()
    +   */
    +  val outBuffer: ByteBuffer = ByteBuffer.allocateDirect(bufferSizeVal)
    +  var streamOffset: Long = streamOffsetVal // Underlying stream offset.
    +
    +  /**
    +   * Whether the underlying stream supports
    +   * {@link org.apache.hadoop.fs.ByteBufferReadable}
    +   */
    +  var usingByteBufferRead: Boolean = false
    +  var usingByteBufferReadInitialized: Boolean = false
    +  /**
    +   * Padding = pos%(algorithm blocksize) Padding is put into {@link #inBuffer}
    +   * before any other data goes in. The purpose of padding is to put the input
    +   * data at proper position.
    +   */
    +  var padding: Byte = '0'
    +  var closed: Boolean = false
    +  var key: Array[Byte] = keyVal.clone()
    +  var initIV: Array[Byte] = ivVal.clone()
    +  var iv: Array[Byte] = ivVal.clone()
    +  var isReadableByteChannel: Boolean = in.isInstanceOf[ReadableByteChannel]
    +
    +  /** DirectBuffer pool */
    +  var bufferPool: Queue[ByteBuffer] = new ConcurrentLinkedQueue[ByteBuffer]()
    +  /** Decryptor pool */
    +  var decryptorPool: Queue[Decryptor] = new ConcurrentLinkedQueue[Decryptor]()
    +
    +  var tmpBuf: Array[Byte] = null
    +  var decryptor: Decryptor = getDecryptor
    +  CryptoStreamUtils.checkCodec(codecVal)
    +  resetStreamOffset(streamOffset)
    +
    +  def this(in: InputStream,
    +      codec: CryptoCodec,
    +      bufferSize: Integer,
    +      key: Array[Byte],
    +      iv: Array[Byte]) {
    +    this(in, codec, bufferSize, key, iv, 0)
    +  }
    +
    +  def this(in: InputStream,
    +      codec: CryptoCodec,
    +      key: Array[Byte],
    +      iv: Array[Byte]) {
    +    this(in, codec, CryptoStreamUtils.getBufferSize, key, iv)
    +  }
    +
    +  def getWrappedStream(): InputStream = in
    +
    +  /**
    +   * Decryption is buffer based.
    +   * If there is data in {@link #outBuffer}, then read it out of this buffer.
    +   * If there is no data in {@link #outBuffer}, then read more from the
    +   * underlying stream and do the decryption.
    +   * @param b the buffer into which the decrypted data is read.
    +   * @param off the buffer offset.
    +   * @param len the maximum number of decrypted data bytes to read.
    +   * @return int the total number of decrypted data bytes read into the buffer.
    +   * @throws IOException
    +   */
    +  override def read(b: Array[Byte], off: Int, len: Int): Int = {
    +    checkStream
    +    if (b == null) {
    +      throw new NullPointerException()
    +    } else if (off < 0 || len < 0 || len > b.length - off) {
    +      throw new IndexOutOfBoundsException()
    +    } else if (len == 0) {
    +      0
    +    } else {
    +      val remaining: Integer = outBuffer.remaining()
    +      if (remaining > 0) {
    +        val n: Integer = Math.min(len, remaining)
    +        outBuffer.get(b, off, n)
    +        n
    +      } else {
    +        var n: Integer = 0
    +        /*
    +         * Check whether the underlying stream is {@link ByteBufferReadable},
    +         * it can avoid bytes copy.
    +         */
    +        if (usingByteBufferReadInitialized == false) {
    +          usingByteBufferReadInitialized = true
    +          if (isReadableByteChannel) {
    +            try {
    +
    +              n = (in.asInstanceOf[ReadableByteChannel]).read(inBuffer)
    +              usingByteBufferRead = true
    +            } catch {
    +              case e: UnsupportedOperationException =>
    +                usingByteBufferRead = false
    +            }
    +          } else {
    +            usingByteBufferRead = false
    +          }
    +          if (!usingByteBufferRead) {
    +            n = readFromUnderlyingStream(inBuffer)
    +          }
    +        } else {
    +          if (usingByteBufferRead) {
    +            n = (in.asInstanceOf[ReadableByteChannel]).read(inBuffer)
    +          } else {
    +            n = readFromUnderlyingStream(inBuffer)
    +          }
    +        }
    +        if (n <= 0) {
    +          n
    +        } else {
    +          streamOffset += n // Read n bytes
    +          decrypt(decryptor, inBuffer, outBuffer, padding)
    +          padding = afterDecryption(decryptor, inBuffer, streamOffset, iv)
    +          n = Math.min(len, outBuffer.remaining())
    +          outBuffer.get(b, off, n)
    +          n
    +        }
    +      }
    +    }
    +  }
    +
    +  /** Read data from underlying stream. */
    +  def readFromUnderlyingStream(inBuffer: ByteBuffer): Int = {
    +    val toRead: Int = inBuffer.remaining
    +    val tmp: Array[Byte] = getTmpBuf
    +    val n: Int = in.read(tmp, 0, toRead)
    +    if (n > 0) {
    +      inBuffer.put(tmp, 0, n)
    +    }
    +    n
    +  }
    +
    +
    +  def getTmpBuf: Array[Byte] = {
    +    if (tmpBuf == null) {
    +      tmpBuf = new Array[Byte](bufferSize)
    +    }
    +    tmpBuf
    +  }
    +
    +  /**
    +   * Do the decryption using inBuffer as input and outBuffer as output.
    +   * Upon return, inBuffer is cleared the decrypted data starts at
    +   * outBuffer.position() and ends at outBuffer.limit()
    +   */
    +  def decrypt(decryptor: Decryptor, inBuffer: ByteBuffer, outBuffer: ByteBuffer, padding:
Byte) {
    +    Preconditions.checkState(inBuffer.position >= padding)
    +    if (inBuffer.position != padding) {
    +      inBuffer.flip
    +      outBuffer.clear
    +      decryptor.decrypt(inBuffer, outBuffer)
    +      inBuffer.clear
    +      outBuffer.flip
    +      if (padding > 0) {
    +        outBuffer.position(padding)
    +      }
    +    }
    +  }
    +
    +  /**
    +   * This method is executed immediately after decryption. Check whether
    +   * decryptor should be updated and recalculate padding if needed.
    +   */
    +  def afterDecryption(decryptor: Decryptor, inBuffer: ByteBuffer, position: Long,
    +                      iv: Array[Byte]): Byte = {
    +    var padding: Byte = 0
    +    if (decryptor.isContextReset) {
    +      updateDecryptor(decryptor, position, iv)
    +      padding = getPadding(position)
    +      inBuffer.position(padding)
    +    }
    +    padding
    +  }
    +
    +  def getCounter(position: Long): Long = {
    +    position / codec.getCipherSuite.algoBlockSize
    +  }
    +
    +  def getPadding(position: Long): Byte = {
    +    (position % codec.getCipherSuite.algoBlockSize).asInstanceOf[Byte]
    +  }
    +
    +  /** Calculate the counter and iv, update the decryptor. */
    +  def updateDecryptor(decryptor: Decryptor, position: Long, iv: Array[Byte]) {
    +    val counter: Long = getCounter(position)
    +    codec.calculateIV(initIV, counter, iv)
    +    decryptor.init(key, iv)
    +  }
    +
    +  /**
    +   * Reset the underlying stream offset clear {@link #inBuffer} and
    +   * {@link #outBuffer}. This Typically happens during {@link #seek(long)}
    +   * or {@link #skip(long)}.
    +   */
    +  def resetStreamOffset(offset: Long) {
    +    streamOffset = offset
    +    inBuffer.clear
    +    outBuffer.clear
    +    outBuffer.limit(0)
    +    updateDecryptor(decryptor, offset, iv)
    +    padding = getPadding(offset)
    +    inBuffer.position(padding)
    +  }
    +
    +  override def close {
    +    if (!closed) {
    +      super.close
    +      freeBuffers
    +      closed = true
    +    }
    +  }
    +
    +  /** Skip n bytes */
    +  override def skip(nVal: Long): Long = {
    +    var n: Long = nVal
    +    Preconditions.checkArgument(n >= 0, "Negative skip length.", new Array[String](1))
    +    checkStream
    +    if (n == 0) {
    +      0
    +    }
    +    else if (n <= outBuffer.remaining) {
    +      val pos: Int = outBuffer.position + n.asInstanceOf[Int]
    +      outBuffer.position(pos)
    +      n
    +    }
    +    else {
    +      n -= outBuffer.remaining
    +      var skipped: Long = in.skip(n)
    +      if (skipped < 0) {
    +        skipped = 0
    +      }
    +      val pos: Long = streamOffset + skipped
    +      skipped += outBuffer.remaining
    +      resetStreamOffset(pos)
    +      skipped
    +    }
    +  }
    +
    +  /** ByteBuffer read. */
    +  override def read(buf: ByteBuffer): Int = {
    +    checkStream
    +    if (isReadableByteChannel) {
    +      val unread: Int = outBuffer.remaining
    +      if (unread > 0) {
    +        val toRead: Int = buf.remaining
    +        if (toRead <= unread) {
    +          val limit: Int = outBuffer.limit
    +          outBuffer.limit(outBuffer.position + toRead)
    +          buf.put(outBuffer)
    +          outBuffer.limit(limit)
    +          toRead
    +        }
    +        else {
    +          buf.put(outBuffer)
    +        }
    +      }
    +      val pos: Int = buf.position
    +      val n: Int = (in.asInstanceOf[ReadableByteChannel]).read(buf)
    +      if (n > 0) {
    +        streamOffset += n
    +        decrypt(buf, n, pos)
    +      }
    +      if (n >= 0) {
    +        unread + n
    +      }
    +      else {
    --- End diff --
    
    nit: move to previous line. Also in lots of other places in the code. I'll just stop pointing
these out.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org


Mime
View raw message