netty5/codec-stomp/src/main/java/io/netty/handler/codec/stomp/StompSubframeDecoder.java
Henning Rohlfs 27ff15319c Reduce memory allocations in StompSubframeDecoder.readHeaders
Motivation:
When decoding stomp frames a lot of unnecessary character arrays are created when parsing headers.
For every header, an array is created to read the line into and then more when splitting the line at the colon.

Modifications:
Parse key and value of a header while reading the line instead of afterwards.
Reuse a single AppendableCharSequence.
Reduce initial size of AppendableCharSequence when reading the command as it is expected to be short.

Result:
Allocations when parsing stomp frames have dropped significantly.
2018-01-23 15:27:57 +01:00

334 lines
13 KiB
Java

/*
* Copyright 2014 The Netty Project
*
* The Netty Project 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 io.netty.handler.codec.stomp;
import java.util.List;
import java.util.Locale;
import io.netty.buffer.ByteBuf;
import io.netty.buffer.Unpooled;
import io.netty.channel.ChannelHandlerContext;
import io.netty.handler.codec.DecoderException;
import io.netty.handler.codec.DecoderResult;
import io.netty.handler.codec.ReplayingDecoder;
import io.netty.handler.codec.TooLongFrameException;
import io.netty.handler.codec.stomp.StompSubframeDecoder.State;
import io.netty.util.internal.AppendableCharSequence;
import static io.netty.buffer.ByteBufUtil.indexOf;
import static io.netty.buffer.ByteBufUtil.readBytes;
/**
* Decodes {@link ByteBuf}s into {@link StompHeadersSubframe}s and
* {@link StompContentSubframe}s.
*
* <h3>Parameters to control memory consumption: </h3>
* {@code maxLineLength} the maximum length of line -
* restricts length of command and header lines
* If the length of the initial line exceeds this value, a
* {@link TooLongFrameException} will be raised.
* <br>
* {@code maxChunkSize}
* The maximum length of the content or each chunk. If the content length
* (or the length of each chunk) exceeds this value, the content or chunk
* ill be split into multiple {@link StompContentSubframe}s whose length is
* {@code maxChunkSize} at maximum.
*
* <h3>Chunked Content</h3>
*
* If the content of a stomp message is greater than {@code maxChunkSize}
* the transfer encoding of the HTTP message is 'chunked', this decoder
* generates multiple {@link StompContentSubframe} instances to avoid excessive memory
* consumption. Note, that every message, even with no content decodes with
* {@link LastStompContentSubframe} at the end to simplify upstream message parsing.
*/
public class StompSubframeDecoder extends ReplayingDecoder<State> {
private static final int DEFAULT_CHUNK_SIZE = 8132;
private static final int DEFAULT_MAX_LINE_LENGTH = 1024;
enum State {
SKIP_CONTROL_CHARACTERS,
READ_HEADERS,
READ_CONTENT,
FINALIZE_FRAME_READ,
BAD_FRAME,
INVALID_CHUNK
}
private final int maxLineLength;
private final int maxChunkSize;
private final boolean validateHeaders;
private int alreadyReadChunkSize;
private LastStompContentSubframe lastContent;
private long contentLength = -1;
public StompSubframeDecoder() {
this(DEFAULT_MAX_LINE_LENGTH, DEFAULT_CHUNK_SIZE);
}
public StompSubframeDecoder(boolean validateHeaders) {
this(DEFAULT_MAX_LINE_LENGTH, DEFAULT_CHUNK_SIZE, validateHeaders);
}
public StompSubframeDecoder(int maxLineLength, int maxChunkSize) {
this(maxLineLength, maxChunkSize, false);
}
public StompSubframeDecoder(int maxLineLength, int maxChunkSize, boolean validateHeaders) {
super(State.SKIP_CONTROL_CHARACTERS);
if (maxLineLength <= 0) {
throw new IllegalArgumentException(
"maxLineLength must be a positive integer: " +
maxLineLength);
}
if (maxChunkSize <= 0) {
throw new IllegalArgumentException(
"maxChunkSize must be a positive integer: " +
maxChunkSize);
}
this.maxChunkSize = maxChunkSize;
this.maxLineLength = maxLineLength;
this.validateHeaders = validateHeaders;
}
@Override
protected void decode(ChannelHandlerContext ctx, ByteBuf in, List<Object> out) throws Exception {
switch (state()) {
case SKIP_CONTROL_CHARACTERS:
skipControlCharacters(in);
checkpoint(State.READ_HEADERS);
// Fall through.
case READ_HEADERS:
StompCommand command = StompCommand.UNKNOWN;
StompHeadersSubframe frame = null;
try {
command = readCommand(in);
frame = new DefaultStompHeadersSubframe(command);
checkpoint(readHeaders(in, frame.headers()));
out.add(frame);
} catch (Exception e) {
if (frame == null) {
frame = new DefaultStompHeadersSubframe(command);
}
frame.setDecoderResult(DecoderResult.failure(e));
out.add(frame);
checkpoint(State.BAD_FRAME);
return;
}
break;
case BAD_FRAME:
in.skipBytes(actualReadableBytes());
return;
}
try {
switch (state()) {
case READ_CONTENT:
int toRead = in.readableBytes();
if (toRead == 0) {
return;
}
if (toRead > maxChunkSize) {
toRead = maxChunkSize;
}
if (contentLength >= 0) {
int remainingLength = (int) (contentLength - alreadyReadChunkSize);
if (toRead > remainingLength) {
toRead = remainingLength;
}
ByteBuf chunkBuffer = readBytes(ctx.alloc(), in, toRead);
if ((alreadyReadChunkSize += toRead) >= contentLength) {
lastContent = new DefaultLastStompContentSubframe(chunkBuffer);
checkpoint(State.FINALIZE_FRAME_READ);
} else {
out.add(new DefaultStompContentSubframe(chunkBuffer));
return;
}
} else {
int nulIndex = indexOf(in, in.readerIndex(), in.writerIndex(), StompConstants.NUL);
if (nulIndex == in.readerIndex()) {
checkpoint(State.FINALIZE_FRAME_READ);
} else {
if (nulIndex > 0) {
toRead = nulIndex - in.readerIndex();
} else {
toRead = in.writerIndex() - in.readerIndex();
}
ByteBuf chunkBuffer = readBytes(ctx.alloc(), in, toRead);
alreadyReadChunkSize += toRead;
if (nulIndex > 0) {
lastContent = new DefaultLastStompContentSubframe(chunkBuffer);
checkpoint(State.FINALIZE_FRAME_READ);
} else {
out.add(new DefaultStompContentSubframe(chunkBuffer));
return;
}
}
}
// Fall through.
case FINALIZE_FRAME_READ:
skipNullCharacter(in);
if (lastContent == null) {
lastContent = LastStompContentSubframe.EMPTY_LAST_CONTENT;
}
out.add(lastContent);
resetDecoder();
}
} catch (Exception e) {
StompContentSubframe errorContent = new DefaultLastStompContentSubframe(Unpooled.EMPTY_BUFFER);
errorContent.setDecoderResult(DecoderResult.failure(e));
out.add(errorContent);
checkpoint(State.BAD_FRAME);
}
}
private StompCommand readCommand(ByteBuf in) {
String commandStr = readLine(in, 16);
StompCommand command = null;
try {
command = StompCommand.valueOf(commandStr);
} catch (IllegalArgumentException iae) {
//do nothing
}
if (command == null) {
commandStr = commandStr.toUpperCase(Locale.US);
try {
command = StompCommand.valueOf(commandStr);
} catch (IllegalArgumentException iae) {
//do nothing
}
}
if (command == null) {
throw new DecoderException("failed to read command from channel");
}
return command;
}
private State readHeaders(ByteBuf buffer, StompHeaders headers) {
AppendableCharSequence buf = new AppendableCharSequence(128);
for (;;) {
boolean headerRead = readHeader(headers, buf, buffer);
if (!headerRead) {
if (headers.contains(StompHeaders.CONTENT_LENGTH)) {
contentLength = getContentLength(headers, 0);
if (contentLength == 0) {
return State.FINALIZE_FRAME_READ;
}
}
return State.READ_CONTENT;
}
}
}
private static long getContentLength(StompHeaders headers, long defaultValue) {
long contentLength = headers.getLong(StompHeaders.CONTENT_LENGTH, defaultValue);
if (contentLength < 0) {
throw new DecoderException(StompHeaders.CONTENT_LENGTH + " must be non-negative");
}
return contentLength;
}
private static void skipNullCharacter(ByteBuf buffer) {
byte b = buffer.readByte();
if (b != StompConstants.NUL) {
throw new IllegalStateException("unexpected byte in buffer " + b + " while expecting NULL byte");
}
}
private static void skipControlCharacters(ByteBuf buffer) {
byte b;
for (;;) {
b = buffer.readByte();
if (b != StompConstants.CR && b != StompConstants.LF) {
buffer.readerIndex(buffer.readerIndex() - 1);
break;
}
}
}
private String readLine(ByteBuf buffer, int initialBufferSize) {
AppendableCharSequence buf = new AppendableCharSequence(initialBufferSize);
int lineLength = 0;
for (;;) {
byte nextByte = buffer.readByte();
if (nextByte == StompConstants.CR) {
//do nothing
} else if (nextByte == StompConstants.LF) {
return buf.toString();
} else {
if (lineLength >= maxLineLength) {
invalidLineLength();
}
lineLength ++;
buf.append((char) nextByte);
}
}
}
private boolean readHeader(StompHeaders headers, AppendableCharSequence buf, ByteBuf buffer) {
buf.reset();
int lineLength = 0;
String key = null;
boolean valid = false;
for (;;) {
byte nextByte = buffer.readByte();
if (nextByte == StompConstants.COLON && key == null) {
key = buf.toString();
valid = true;
buf.reset();
} else if (nextByte == StompConstants.CR) {
//do nothing
} else if (nextByte == StompConstants.LF) {
if (key == null && lineLength == 0) {
return false;
} else if (valid) {
headers.add(key, buf.toString());
} else if (validateHeaders) {
invalidHeader(key, buf.toString());
}
return true;
} else {
if (lineLength >= maxLineLength) {
invalidLineLength();
}
if (nextByte == StompConstants.COLON && key != null) {
valid = false;
}
lineLength ++;
buf.append((char) nextByte);
}
}
}
private void invalidHeader(String key, String value) {
String line = key != null ? key + ":" + value : value;
throw new IllegalArgumentException("a header value or name contains a prohibited character ':'"
+ ", " + line);
}
private void invalidLineLength() {
throw new TooLongFrameException("An STOMP line is larger than " + maxLineLength + " bytes.");
}
private void resetDecoder() {
checkpoint(State.SKIP_CONTROL_CHARACTERS);
contentLength = -1;
alreadyReadChunkSize = 0;
lastContent = null;
}
}