8190711: Assert in G1MMUTracker due to concurrent modification
Reviewed-by: tschatzl, rehn, sangheki
/*
* Copyright (c) 2009, 2010, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation. Oracle designates this
* particular file as subject to the "Classpath" exception as provided
* by Oracle in the LICENSE file that accompanied this code.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
package $PACKAGE$;
import java.io.*;
import java.nio.CharBuffer;
import java.nio.ByteBuffer;
import java.nio.charset.Charset;
import java.nio.charset.CharsetDecoder;
import java.nio.charset.CharsetEncoder;
import java.nio.charset.CoderResult;
import java.util.Arrays;
import sun.nio.cs.HistoricallyNamedCharset;
import static sun.nio.cs.CharsetMapping.*;
public class EUC_TW extends Charset implements HistoricallyNamedCharset
{
private static final int SS2 = 0x8E;
/*
(1) EUC_TW
Second byte of EUC_TW for cs2 is in range of
0xA1-0xB0 for plane 1-16. According to CJKV /163,
plane1 is coded in both cs1 and cs2. This impl
however does not decode the codepoints of plane1
in cs2, so only p2-p7 and p15 are supported in cs2.
Plane2 0xA2;
Plane3 0xA3;
Plane4 0xA4;
Plane5 0xA5;
Plane6 0xA6;
Plane7 0xA7;
Plane15 0xAF;
(2) Mapping
The fact that all supplementary characters encoded in EUC_TW are
in 0x2xxxx range gives us the room to optimize the data tables.
Decoding:
(1) save the lower 16-bit value of all codepoints of b->c mapping
in a String array table String[plane] b2c.
(2) save "codepoint is supplementary" info (one bit) in a
byte[] b2cIsSupp, so 8 codepoints (same codepoint value, different
plane No) share one byte.
Encoding:
(1)c->b mappings are stored in
char[]c2b/char[]c2bIndex
char[]c2bSupp/char[]c2bIndexsupp (indexed by lower 16-bit
(2)byte[] c2bPlane stores the "plane info" of each euc-tw codepoints,
BMP and Supp share the low/high 4 bits of one byte.
Mapping tables are stored separated in EUC_TWMapping, which
is generated by tool.
*/
public EUC_TW() {
super("x-EUC-TW", $ALIASES$);
}
public String historicalName() {
return "EUC_TW";
}
public boolean contains(Charset cs) {
return ((cs.name().equals("US-ASCII"))
|| (cs instanceof EUC_TW));
}
public CharsetDecoder newDecoder() {
return new Decoder(this);
}
public CharsetEncoder newEncoder() {
return new Encoder(this);
}
public static class Decoder extends CharsetDecoder {
public Decoder(Charset cs) {
super(cs, 2.0f, 2.0f);
}
char[] c1 = new char[1];
char[] c2 = new char[2];
public char[] toUnicode(int b1, int b2, int p) {
return decode(b1, b2, p, c1, c2);
}
static final String[] b2c = EUC_TWMapping.b2c;
static final int b1Min = EUC_TWMapping.b1Min;
static final int b1Max = EUC_TWMapping.b1Max;
static final int b2Min = EUC_TWMapping.b2Min;
static final int b2Max = EUC_TWMapping.b2Max;
static final int dbSegSize = b2Max - b2Min + 1;
static final byte[] b2cIsSupp;
// adjust from cns planeNo to the plane index of b2c
static final byte[] cnspToIndex = new byte[0x100];
static {
Arrays.fill(cnspToIndex, (byte)-1);
cnspToIndex[0xa2] = 1; cnspToIndex[0xa3] = 2; cnspToIndex[0xa4] = 3;
cnspToIndex[0xa5] = 4; cnspToIndex[0xa6] = 5; cnspToIndex[0xa7] = 6;
cnspToIndex[0xaf] = 7;
}
//static final BitSet b2cIsSupp;
static {
String b2cIsSuppStr = EUC_TWMapping.b2cIsSuppStr;
// work on a local copy is much faster than operate
// directly on b2cIsSupp
byte[] flag = new byte[b2cIsSuppStr.length() << 1];
int off = 0;
for (int i = 0; i < b2cIsSuppStr.length(); i++) {
char c = b2cIsSuppStr.charAt(i);
flag[off++] = (byte)(c >> 8);
flag[off++] = (byte)(c & 0xff);
}
b2cIsSupp = flag;
}
static boolean isLegalDB(int b) {
return b >= b1Min && b <= b1Max;
}
static char[] decode(int b1, int b2, int p, char[] c1, char[] c2)
{
if (b1 < b1Min || b1 > b1Max || b2 < b2Min || b2 > b2Max)
return null;
int index = (b1 - b1Min) * dbSegSize + b2 - b2Min;
char c = b2c[p].charAt(index);
if (c == UNMAPPABLE_DECODING)
return null;
if ((b2cIsSupp[index] & (1 << p)) == 0) {
c1[0] = c;
return c1;
} else {
c2[0] = Character.highSurrogate(0x20000 + c);
c2[1] = Character.lowSurrogate(0x20000 + c);
return c2;
}
}
private CoderResult decodeArrayLoop(ByteBuffer src,
CharBuffer dst)
{
byte[] sa = src.array();
int sp = src.arrayOffset() + src.position();
int sl = src.arrayOffset() + src.limit();
char[] da = dst.array();
int dp = dst.arrayOffset() + dst.position();
int dl = dst.arrayOffset() + dst.limit();
try {
while (sp < sl) {
int byte1 = sa[sp] & 0xff;
if (byte1 == SS2) { // Codeset 2 G2
if ( sl - sp < 4)
return CoderResult.UNDERFLOW;
int cnsPlane = cnspToIndex[sa[sp + 1] & 0xff];
if (cnsPlane < 0)
return CoderResult.malformedForLength(2);
byte1 = sa[sp + 2] & 0xff;
int byte2 = sa[sp + 3] & 0xff;
char[] cc = toUnicode(byte1, byte2, cnsPlane);
if (cc == null) {
if (!isLegalDB(byte1) || !isLegalDB(byte2))
return CoderResult.malformedForLength(4);
return CoderResult.unmappableForLength(4);
}
if (dl - dp < cc.length)
return CoderResult.OVERFLOW;
if (cc.length == 1) {
da[dp++] = cc[0];
} else {
da[dp++] = cc[0];
da[dp++] = cc[1];
}
sp += 4;
} else if (byte1 < 0x80) { // ASCII G0
if (dl - dp < 1)
return CoderResult.OVERFLOW;
da[dp++] = (char) byte1;
sp++;
} else { // Codeset 1 G1
if ( sl - sp < 2)
return CoderResult.UNDERFLOW;
int byte2 = sa[sp + 1] & 0xff;
char[] cc = toUnicode(byte1, byte2, 0);
if (cc == null) {
if (!isLegalDB(byte1) || !isLegalDB(byte2))
return CoderResult.malformedForLength(1);
return CoderResult.unmappableForLength(2);
}
if (dl - dp < 1)
return CoderResult.OVERFLOW;
da[dp++] = cc[0];
sp += 2;
}
}
return CoderResult.UNDERFLOW;
} finally {
src.position(sp - src.arrayOffset());
dst.position(dp - dst.arrayOffset());
}
}
private CoderResult decodeBufferLoop(ByteBuffer src,
CharBuffer dst)
{
int mark = src.position();
try {
while (src.hasRemaining()) {
int byte1 = src.get() & 0xff;
if (byte1 == SS2) { // Codeset 2 G2
if ( src.remaining() < 3)
return CoderResult.UNDERFLOW;
int cnsPlane = cnspToIndex[src.get() & 0xff];
if (cnsPlane < 0)
return CoderResult.malformedForLength(2);
byte1 = src.get() & 0xff;
int byte2 = src.get() & 0xff;
char[] cc = toUnicode(byte1, byte2, cnsPlane);
if (cc == null) {
if (!isLegalDB(byte1) || !isLegalDB(byte2))
return CoderResult.malformedForLength(4);
return CoderResult.unmappableForLength(4);
}
if (dst.remaining() < cc.length)
return CoderResult.OVERFLOW;
if (cc.length == 1) {
dst.put(cc[0]);
} else {
dst.put(cc[0]);
dst.put(cc[1]);
}
mark += 4;
} else if (byte1 < 0x80) { // ASCII G0
if (!dst.hasRemaining())
return CoderResult.OVERFLOW;
dst.put((char) byte1);
mark++;
} else { // Codeset 1 G1
if (!src.hasRemaining())
return CoderResult.UNDERFLOW;
int byte2 = src.get() & 0xff;
char[] cc = toUnicode(byte1, byte2, 0);
if (cc == null) {
if (!isLegalDB(byte1) || !isLegalDB(byte2))
return CoderResult.malformedForLength(1);
return CoderResult.unmappableForLength(2);
}
if (!dst.hasRemaining())
return CoderResult.OVERFLOW;
dst.put(cc[0]);
mark +=2;
}
}
return CoderResult.UNDERFLOW;
} finally {
src.position(mark);
}
}
protected CoderResult decodeLoop(ByteBuffer src, CharBuffer dst)
{
if (src.hasArray() && dst.hasArray())
return decodeArrayLoop(src, dst);
else
return decodeBufferLoop(src, dst);
}
}
public static class Encoder extends CharsetEncoder {
private byte[] bb = new byte[4];
public Encoder(Charset cs) {
super(cs, 4.0f, 4.0f);
}
public boolean canEncode(char c) {
return (c <= '\u007f' || toEUC(c, bb) != -1);
}
public boolean canEncode(CharSequence cs) {
int i = 0;
while (i < cs.length()) {
char c = cs.charAt(i++);
if (Character.isHighSurrogate(c)) {
if (i == cs.length())
return false;
char low = cs.charAt(i++);
if (!Character.isLowSurrogate(low) || toEUC(c, low, bb) == -1)
return false;
} else if (!canEncode(c)) {
return false;
}
}
return true;
}
public int toEUC(char hi, char low, byte[] bb) {
return encode(hi, low, bb);
}
public int toEUC(char c, byte[] bb) {
return encode(c, bb);
}
private CoderResult encodeArrayLoop(CharBuffer src,
ByteBuffer dst)
{
char[] sa = src.array();
int sp = src.arrayOffset() + src.position();
int sl = src.arrayOffset() + src.limit();
byte[] da = dst.array();
int dp = dst.arrayOffset() + dst.position();
int dl = dst.arrayOffset() + dst.limit();
int inSize;
int outSize;
try {
while (sp < sl) {
char c = sa[sp];
inSize = 1;
if (c < 0x80) { // ASCII
bb[0] = (byte)c;
outSize = 1;
} else {
outSize = toEUC(c, bb);
if (outSize == -1) {
// to check surrogates only after BMP failed
// has the benefit of improving the BMP encoding
// 10% faster, with the price of the slowdown of
// supplementary character encoding. given the use
// of supplementary characters is really rare, this
// is something worth doing.
if (Character.isHighSurrogate(c)) {
if ((sp + 1) == sl)
return CoderResult.UNDERFLOW;
if (!Character.isLowSurrogate(sa[sp + 1]))
return CoderResult.malformedForLength(1);
outSize = toEUC(c, sa[sp+1], bb);
inSize = 2;
} else if (Character.isLowSurrogate(c)) {
return CoderResult.malformedForLength(1);
}
}
}
if (outSize == -1)
return CoderResult.unmappableForLength(inSize);
if ( dl - dp < outSize)
return CoderResult.OVERFLOW;
for (int i = 0; i < outSize; i++)
da[dp++] = bb[i];
sp += inSize;
}
return CoderResult.UNDERFLOW;
} finally {
src.position(sp - src.arrayOffset());
dst.position(dp - dst.arrayOffset());
}
}
private CoderResult encodeBufferLoop(CharBuffer src,
ByteBuffer dst)
{
int outSize;
int inSize;
int mark = src.position();
try {
while (src.hasRemaining()) {
inSize = 1;
char c = src.get();
if (c < 0x80) { // ASCII
outSize = 1;
bb[0] = (byte)c;
} else {
outSize = toEUC(c, bb);
if (outSize == -1) {
if (Character.isHighSurrogate(c)) {
if (!src.hasRemaining())
return CoderResult.UNDERFLOW;
char c2 = src.get();
if (!Character.isLowSurrogate(c2))
return CoderResult.malformedForLength(1);
outSize = toEUC(c, c2, bb);
inSize = 2;
} else if (Character.isLowSurrogate(c)) {
return CoderResult.malformedForLength(1);
}
}
}
if (outSize == -1)
return CoderResult.unmappableForLength(inSize);
if (dst.remaining() < outSize)
return CoderResult.OVERFLOW;
for (int i = 0; i < outSize; i++)
dst.put(bb[i]);
mark += inSize;
}
return CoderResult.UNDERFLOW;
} finally {
src.position(mark);
}
}
protected CoderResult encodeLoop(CharBuffer src, ByteBuffer dst)
{
if (src.hasArray() && dst.hasArray())
return encodeArrayLoop(src, dst);
else
return encodeBufferLoop(src, dst);
}
static int encode(char hi, char low, byte[] bb) {
int c = Character.toCodePoint(hi, low);
if ((c & 0xf0000) != 0x20000)
return -1;
c -= 0x20000;
int index = c2bSuppIndex[c >> 8];
if (index == UNMAPPABLE_ENCODING)
return -1;
index = index + (c & 0xff);
int db = c2bSupp[index];
if (db == UNMAPPABLE_ENCODING)
return -1;
int p = (c2bPlane[index] >> 4) & 0xf;
bb[0] = (byte)SS2;
bb[1] = (byte)(0xa0 | p);
bb[2] = (byte)(db >> 8);
bb[3] = (byte)db;
return 4;
}
static int encode(char c, byte[] bb) {
int index = c2bIndex[c >> 8];
if (index == UNMAPPABLE_ENCODING)
return -1;
index = index + (c & 0xff);
int db = c2b[index];
if (db == UNMAPPABLE_ENCODING)
return -1;
int p = c2bPlane[index] & 0xf;
if (p == 0) {
bb[0] = (byte)(db >> 8);
bb[1] = (byte)db;
return 2;
} else {
bb[0] = (byte)SS2;
bb[1] = (byte)(0xa0 | p);
bb[2] = (byte)(db >> 8);
bb[3] = (byte)db;
return 4;
}
}
static final char[] c2b;
static final char[] c2bIndex;
static final char[] c2bSupp;
static final char[] c2bSuppIndex;
static final byte[] c2bPlane;
static {
int b1Min = Decoder.b1Min;
int b1Max = Decoder.b1Max;
int b2Min = Decoder.b2Min;
int b2Max = Decoder.b2Max;
int dbSegSize = Decoder.dbSegSize;
String[] b2c = Decoder.b2c;
byte[] b2cIsSupp = Decoder.b2cIsSupp;
c2bIndex = EUC_TWMapping.c2bIndex;
c2bSuppIndex = EUC_TWMapping.c2bSuppIndex;
char[] c2b0 = new char[EUC_TWMapping.C2BSIZE];
char[] c2bSupp0 = new char[EUC_TWMapping.C2BSUPPSIZE];
byte[] c2bPlane0 = new byte[Math.max(EUC_TWMapping.C2BSIZE,
EUC_TWMapping.C2BSUPPSIZE)];
Arrays.fill(c2b0, (char)UNMAPPABLE_ENCODING);
Arrays.fill(c2bSupp0, (char)UNMAPPABLE_ENCODING);
for (int p = 0; p < b2c.length; p++) {
String db = b2c[p];
/*
adjust the "plane" from 0..7 to 0, 2, 3, 4, 5, 6, 7, 0xf,
which helps balance between footprint (to save the plane
info in 4 bits) and runtime performance (to require only
one operation "0xa0 | plane" to encode the plane byte)
*/
int plane = p;
if (plane == 7)
plane = 0xf;
else if (plane != 0)
plane = p + 1;
int off = 0;
for (int b1 = b1Min; b1 <= b1Max; b1++) {
for (int b2 = b2Min; b2 <= b2Max; b2++) {
char c = db.charAt(off);
if (c != UNMAPPABLE_DECODING) {
if ((b2cIsSupp[off] & (1 << p)) != 0) {
int index = c2bSuppIndex[c >> 8] + (c&0xff);
c2bSupp0[index] = (char)((b1 << 8) + b2);
c2bPlane0[index] |= (byte)(plane << 4);
} else {
int index = c2bIndex[c >> 8] + (c&0xff);
c2b0[index] = (char)((b1 << 8) + b2);
c2bPlane0[index] |= (byte)plane;
}
}
off++;
}
}
}
c2b = c2b0;
c2bSupp = c2bSupp0;
c2bPlane = c2bPlane0;
}
}
}