You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

ChunkIndex.java 11KB

Store Git on any DHT jgit.storage.dht is a storage provider implementation for JGit that permits storing the Git repository in a distributed hashtable, NoSQL system, or other database. The actual underlying storage system is undefined, and can be plugged in by implementing 7 small interfaces: * Database * RepositoryIndexTable * RepositoryTable * RefTable * ChunkTable * ObjectIndexTable * WriteBuffer The storage provider interface tries to assume very little about the underlying storage system, and requires only three key features: * key -> value lookup (a hashtable is suitable) * atomic updates on single rows * asynchronous operations (Java's ExecutorService is easy to use) Most NoSQL database products offer all 3 of these features in their clients, and so does any decent network based cache system like the open source memcache product. Relying only on key equality for data retrevial makes it simple for the storage engine to distribute across multiple machines. Traditional SQL systems could also be used with a JDBC based spi implementation. Before submitting this change I have implemented six storage systems for the spi layer: * Apache HBase[1] * Apache Cassandra[2] * Google Bigtable[3] * an in-memory implementation for unit testing * a JDBC implementation for SQL * a generic cache provider that can ride on top of memcache All six systems came in with an spi layer around 1000 lines of code to implement the above 7 interfaces. This is a huge reduction in size compared to prior attempts to implement a new JGit storage layer. As this package shows, a complete JGit storage implementation is more than 17,000 lines of fairly complex code. A simple cache is provided in storage.dht.spi.cache. Implementers can use CacheDatabase to wrap any other type of Database and perform fast reads against a network based cache service, such as the open source memcached[4]. An implementation of CacheService must be provided to glue this spi onto the network cache. [1] https://github.com/spearce/jgit_hbase [2] https://github.com/spearce/jgit_cassandra [3] http://labs.google.com/papers/bigtable.html [4] http://memcached.org/ Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295 Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
13 years ago
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428
  1. /*
  2. * Copyright (C) 2011, Google Inc.
  3. * and other copyright owners as documented in the project's IP log.
  4. *
  5. * This program and the accompanying materials are made available
  6. * under the terms of the Eclipse Distribution License v1.0 which
  7. * accompanies this distribution, is reproduced below, and is
  8. * available at http://www.eclipse.org/org/documents/edl-v10.php
  9. *
  10. * All rights reserved.
  11. *
  12. * Redistribution and use in source and binary forms, with or
  13. * without modification, are permitted provided that the following
  14. * conditions are met:
  15. *
  16. * - Redistributions of source code must retain the above copyright
  17. * notice, this list of conditions and the following disclaimer.
  18. *
  19. * - Redistributions in binary form must reproduce the above
  20. * copyright notice, this list of conditions and the following
  21. * disclaimer in the documentation and/or other materials provided
  22. * with the distribution.
  23. *
  24. * - Neither the name of the Eclipse Foundation, Inc. nor the
  25. * names of its contributors may be used to endorse or promote
  26. * products derived from this software without specific prior
  27. * written permission.
  28. *
  29. * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
  30. * CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
  31. * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
  32. * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  33. * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
  34. * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  35. * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
  36. * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  37. * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
  38. * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
  39. * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  40. * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
  41. * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  42. */
  43. package org.eclipse.jgit.storage.dht;
  44. import java.text.MessageFormat;
  45. import java.util.Collections;
  46. import java.util.List;
  47. import org.eclipse.jgit.lib.AnyObjectId;
  48. import static org.eclipse.jgit.lib.Constants.*;
  49. import org.eclipse.jgit.lib.ObjectId;
  50. import org.eclipse.jgit.transport.PackedObjectInfo;
  51. import org.eclipse.jgit.util.NB;
  52. /** Index into a {@link PackChunk}. */
  53. public abstract class ChunkIndex {
  54. private static final int V1 = 0x01;
  55. static ChunkIndex fromBytes(ChunkKey key, byte[] index, int ptr, int len)
  56. throws DhtException {
  57. int v = index[ptr] & 0xff;
  58. switch (v) {
  59. case V1: {
  60. final int offsetFormat = index[ptr + 1] & 7;
  61. switch (offsetFormat) {
  62. case 1:
  63. return new Offset1(index, ptr, len, key);
  64. case 2:
  65. return new Offset2(index, ptr, len, key);
  66. case 3:
  67. return new Offset3(index, ptr, len, key);
  68. case 4:
  69. return new Offset4(index, ptr, len, key);
  70. default:
  71. throw new DhtException(MessageFormat.format(
  72. DhtText.get().unsupportedChunkIndex,
  73. Integer.toHexString(NB.decodeUInt16(index, ptr)), key));
  74. }
  75. }
  76. default:
  77. throw new DhtException(MessageFormat.format(
  78. DhtText.get().unsupportedChunkIndex,
  79. Integer.toHexString(v), key));
  80. }
  81. }
  82. /**
  83. * Format the chunk index and return its binary representation.
  84. *
  85. * @param list
  86. * the list of objects that appear in the chunk. This list will
  87. * be sorted in-place if it has more than 1 element.
  88. * @return binary representation of the chunk's objects and their starting
  89. * offsets. The format is private to this class.
  90. */
  91. @SuppressWarnings("null")
  92. static byte[] create(List<? extends PackedObjectInfo> list) {
  93. int cnt = list.size();
  94. sortObjectList(list);
  95. int fanoutFormat = 0;
  96. int[] buckets = null;
  97. if (64 < cnt) {
  98. buckets = new int[256];
  99. for (PackedObjectInfo oe : list)
  100. buckets[oe.getFirstByte()]++;
  101. fanoutFormat = selectFanoutFormat(buckets);
  102. }
  103. int offsetFormat = selectOffsetFormat(list);
  104. byte[] index = new byte[2 // header
  105. + 256 * fanoutFormat // (optional) fanout
  106. + cnt * OBJECT_ID_LENGTH // ids
  107. + cnt * offsetFormat // offsets
  108. ];
  109. index[0] = V1;
  110. index[1] = (byte) ((fanoutFormat << 3) | offsetFormat);
  111. int ptr = 2;
  112. switch (fanoutFormat) {
  113. case 0:
  114. break;
  115. case 1:
  116. for (int i = 0; i < 256; i++, ptr++)
  117. index[ptr] = (byte) buckets[i];
  118. break;
  119. case 2:
  120. for (int i = 0; i < 256; i++, ptr += 2)
  121. NB.encodeInt16(index, ptr, buckets[i]);
  122. break;
  123. case 3:
  124. for (int i = 0; i < 256; i++, ptr += 3)
  125. encodeUInt24(index, ptr, buckets[i]);
  126. break;
  127. case 4:
  128. for (int i = 0; i < 256; i++, ptr += 4)
  129. NB.encodeInt32(index, ptr, buckets[i]);
  130. break;
  131. }
  132. for (PackedObjectInfo oe : list) {
  133. oe.copyRawTo(index, ptr);
  134. ptr += OBJECT_ID_LENGTH;
  135. }
  136. switch (offsetFormat) {
  137. case 1:
  138. for (PackedObjectInfo oe : list)
  139. index[ptr++] = (byte) oe.getOffset();
  140. break;
  141. case 2:
  142. for (PackedObjectInfo oe : list) {
  143. NB.encodeInt16(index, ptr, (int) oe.getOffset());
  144. ptr += 2;
  145. }
  146. break;
  147. case 3:
  148. for (PackedObjectInfo oe : list) {
  149. encodeUInt24(index, ptr, (int) oe.getOffset());
  150. ptr += 3;
  151. }
  152. break;
  153. case 4:
  154. for (PackedObjectInfo oe : list) {
  155. NB.encodeInt32(index, ptr, (int) oe.getOffset());
  156. ptr += 4;
  157. }
  158. break;
  159. }
  160. return index;
  161. }
  162. private static int selectFanoutFormat(int[] buckets) {
  163. int fmt = 1;
  164. int max = 1 << (8 * fmt);
  165. for (int cnt : buckets) {
  166. while (max <= cnt && fmt < 4) {
  167. if (++fmt == 4)
  168. return fmt;
  169. max = 1 << (8 * fmt);
  170. }
  171. }
  172. return fmt;
  173. }
  174. private static int selectOffsetFormat(List<? extends PackedObjectInfo> list) {
  175. int fmt = 1;
  176. int max = 1 << (8 * fmt);
  177. for (PackedObjectInfo oe : list) {
  178. while (max <= oe.getOffset() && fmt < 4) {
  179. if (++fmt == 4)
  180. return fmt;
  181. max = 1 << (8 * fmt);
  182. }
  183. }
  184. return fmt;
  185. }
  186. private static void sortObjectList(List<? extends PackedObjectInfo> list) {
  187. Collections.sort(list);
  188. }
  189. private final byte[] indexBuf;
  190. private final int indexPtr;
  191. private final int indexLen;
  192. private final int[] fanout;
  193. private final int idTable;
  194. private final int offsetTable;
  195. private final int count;
  196. ChunkIndex(byte[] indexBuf, int ptr, int len, ChunkKey key)
  197. throws DhtException {
  198. final int ctl = indexBuf[ptr + 1];
  199. final int fanoutFormat = (ctl >>> 3) & 7;
  200. final int offsetFormat = ctl & 7;
  201. switch (fanoutFormat) {
  202. case 0:
  203. fanout = null; // no fanout, too small
  204. break;
  205. case 1: {
  206. int last = 0;
  207. fanout = new int[256];
  208. for (int i = 0; i < 256; i++) {
  209. last += indexBuf[ptr + 2 + i] & 0xff;
  210. fanout[i] = last;
  211. }
  212. break;
  213. }
  214. case 2: {
  215. int last = 0;
  216. fanout = new int[256];
  217. for (int i = 0; i < 256; i++) {
  218. last += NB.decodeUInt16(indexBuf, ptr + 2 + i * 2);
  219. fanout[i] = last;
  220. }
  221. break;
  222. }
  223. case 3: {
  224. int last = 0;
  225. fanout = new int[256];
  226. for (int i = 0; i < 256; i++) {
  227. last += decodeUInt24(indexBuf, ptr + 2 + i * 3);
  228. fanout[i] = last;
  229. }
  230. break;
  231. }
  232. case 4: {
  233. int last = 0;
  234. fanout = new int[256];
  235. for (int i = 0; i < 256; i++) {
  236. last += NB.decodeInt32(indexBuf, ptr + 2 + i * 4);
  237. fanout[i] = last;
  238. }
  239. break;
  240. }
  241. default:
  242. throw new DhtException(MessageFormat.format(
  243. DhtText.get().unsupportedChunkIndex,
  244. Integer.toHexString(NB.decodeUInt16(indexBuf, ptr)), key));
  245. }
  246. this.indexBuf = indexBuf;
  247. this.indexPtr = ptr;
  248. this.indexLen = len;
  249. this.idTable = indexPtr + 2 + 256 * fanoutFormat;
  250. int recsz = OBJECT_ID_LENGTH + offsetFormat;
  251. this.count = (indexLen - (idTable - indexPtr)) / recsz;
  252. this.offsetTable = idTable + count * OBJECT_ID_LENGTH;
  253. }
  254. /**
  255. * Get the total number of objects described by this index.
  256. *
  257. * @return number of objects in this index and its associated chunk.
  258. */
  259. public final int getObjectCount() {
  260. return count;
  261. }
  262. /**
  263. * Get an ObjectId from this index.
  264. *
  265. * @param nth
  266. * the object to return. Must be in range [0, getObjectCount).
  267. * @return the object id.
  268. */
  269. public final ObjectId getObjectId(int nth) {
  270. return ObjectId.fromRaw(indexBuf, idPosition(nth));
  271. }
  272. /**
  273. * Get the offset of an object in the chunk.
  274. *
  275. * @param nth
  276. * offset to return. Must be in range [0, getObjectCount).
  277. * @return the offset.
  278. */
  279. public final int getOffset(int nth) {
  280. return getOffset(indexBuf, offsetTable, nth);
  281. }
  282. /** @return the size of this index, in bytes. */
  283. int getIndexSize() {
  284. int sz = indexBuf.length;
  285. if (fanout != null)
  286. sz += 12 + 256 * 4;
  287. return sz;
  288. }
  289. /**
  290. * Search for an object in the index.
  291. *
  292. * @param objId
  293. * the object to locate.
  294. * @return offset of the object in the corresponding chunk; -1 if not found.
  295. */
  296. final int findOffset(AnyObjectId objId) {
  297. int hi, lo;
  298. if (fanout != null) {
  299. int fb = objId.getFirstByte();
  300. lo = fb == 0 ? 0 : fanout[fb - 1];
  301. hi = fanout[fb];
  302. } else {
  303. lo = 0;
  304. hi = count;
  305. }
  306. while (lo < hi) {
  307. final int mid = (lo + hi) >>> 1;
  308. final int cmp = objId.compareTo(indexBuf, idPosition(mid));
  309. if (cmp < 0)
  310. hi = mid;
  311. else if (cmp == 0)
  312. return getOffset(mid);
  313. else
  314. lo = mid + 1;
  315. }
  316. return -1;
  317. }
  318. abstract int getOffset(byte[] indexArray, int offsetTableStart, int nth);
  319. private int idPosition(int nth) {
  320. return idTable + (nth * OBJECT_ID_LENGTH);
  321. }
  322. private static class Offset1 extends ChunkIndex {
  323. Offset1(byte[] index, int ptr, int len, ChunkKey key)
  324. throws DhtException {
  325. super(index, ptr, len, key);
  326. }
  327. int getOffset(byte[] index, int offsetTable, int nth) {
  328. return index[offsetTable + nth] & 0xff;
  329. }
  330. }
  331. private static class Offset2 extends ChunkIndex {
  332. Offset2(byte[] index, int ptr, int len, ChunkKey key)
  333. throws DhtException {
  334. super(index, ptr, len, key);
  335. }
  336. int getOffset(byte[] index, int offsetTable, int nth) {
  337. return NB.decodeUInt16(index, offsetTable + (nth * 2));
  338. }
  339. }
  340. private static class Offset3 extends ChunkIndex {
  341. Offset3(byte[] index, int ptr, int len, ChunkKey key)
  342. throws DhtException {
  343. super(index, ptr, len, key);
  344. }
  345. int getOffset(byte[] index, int offsetTable, int nth) {
  346. return decodeUInt24(index, offsetTable + (nth * 3));
  347. }
  348. }
  349. private static class Offset4 extends ChunkIndex {
  350. Offset4(byte[] index, int ptr, int len, ChunkKey key)
  351. throws DhtException {
  352. super(index, ptr, len, key);
  353. }
  354. int getOffset(byte[] index, int offsetTable, int nth) {
  355. return NB.decodeInt32(index, offsetTable + (nth * 4));
  356. }
  357. }
  358. private static void encodeUInt24(byte[] intbuf, int offset, int v) {
  359. intbuf[offset + 2] = (byte) v;
  360. v >>>= 8;
  361. intbuf[offset + 1] = (byte) v;
  362. v >>>= 8;
  363. intbuf[offset] = (byte) v;
  364. }
  365. private static int decodeUInt24(byte[] intbuf, int offset) {
  366. int r = (intbuf[offset] & 0xff) << 8;
  367. r |= intbuf[offset + 1] & 0xff;
  368. r <<= 8;
  369. r |= intbuf[offset + 2] & 0xff;
  370. return r;
  371. }
  372. }