dataStore_files.go 9.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402
  1. //go:build PSIPHON_USE_FILES_DB
  2. // +build PSIPHON_USE_FILES_DB
  3. /*
  4. * Copyright (c) 2018, Psiphon Inc.
  5. * All rights reserved.
  6. *
  7. * This program is free software: you can redistribute it and/or modify
  8. * it under the terms of the GNU General Public License as published by
  9. * the Free Software Foundation, either version 3 of the License, or
  10. * (at your option) any later version.
  11. *
  12. * This program is distributed in the hope that it will be useful,
  13. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15. * GNU General Public License for more details.
  16. *
  17. * You should have received a copy of the GNU General Public License
  18. * along with this program. If not, see <http://www.gnu.org/licenses/>.
  19. *
  20. */
  21. package psiphon
  22. import (
  23. "bytes"
  24. "encoding/hex"
  25. std_errors "errors"
  26. "io/ioutil"
  27. "os"
  28. "path/filepath"
  29. "strings"
  30. "sync"
  31. "github.com/Psiphon-Labs/psiphon-tunnel-core/psiphon/common/errors"
  32. )
  33. // datastoreDB is a simple filesystem-backed key/value store that implements
  34. // the datastore interface.
  35. //
  36. // The current implementation is intended only for experimentation.
  37. //
  38. // Buckets are subdirectories, keys are file names (hex-encoded), and values
  39. // are file contents. Unlike other datastores, update transactions are neither
  40. // atomic not isolcated; only each put is individually atomic.
  41. //
  42. // A buffer pool is used to reduce memory allocation/GC churn from loading
  43. // file values into memory. Transactions and cursors track and release shared
  44. // buffers.
  45. //
  46. // As with the original datastore interface, value slices are only valid
  47. // within a transaction; for cursors, there's a further limitation that the
  48. // value slices are only valid until the next iteration.
  49. type datastoreDB struct {
  50. dataDirectory string
  51. bufferPool sync.Pool
  52. lock sync.RWMutex
  53. closed bool
  54. }
  55. type datastoreTx struct {
  56. db *datastoreDB
  57. canUpdate bool
  58. buffers []*bytes.Buffer
  59. }
  60. type datastoreBucket struct {
  61. bucketDirectory string
  62. tx *datastoreTx
  63. }
  64. type datastoreCursor struct {
  65. bucket *datastoreBucket
  66. fileInfos []os.FileInfo
  67. index int
  68. lastBuffer *bytes.Buffer
  69. }
  70. func datastoreOpenDB(
  71. rootDataDirectory string, _ bool) (*datastoreDB, error) {
  72. dataDirectory := filepath.Join(rootDataDirectory, "psiphon.filesdb")
  73. err := os.MkdirAll(dataDirectory, 0700)
  74. if err != nil {
  75. return nil, errors.Trace(err)
  76. }
  77. return &datastoreDB{
  78. dataDirectory: dataDirectory,
  79. bufferPool: sync.Pool{
  80. New: func() interface{} {
  81. return new(bytes.Buffer)
  82. },
  83. },
  84. }, nil
  85. }
  86. func (db *datastoreDB) getBuffer() *bytes.Buffer {
  87. return db.bufferPool.Get().(*bytes.Buffer)
  88. }
  89. func (db *datastoreDB) putBuffer(buffer *bytes.Buffer) {
  90. buffer.Truncate(0)
  91. db.bufferPool.Put(buffer)
  92. }
  93. func (db *datastoreDB) readBuffer(filename string) (*bytes.Buffer, error) {
  94. // Complete any partial put commit.
  95. err := datastoreApplyCommit(filename)
  96. if err != nil {
  97. return nil, errors.Trace(err)
  98. }
  99. file, err := os.Open(filename)
  100. if err != nil {
  101. if os.IsNotExist(err) {
  102. return nil, nil
  103. }
  104. return nil, errors.Trace(err)
  105. }
  106. defer file.Close()
  107. buffer := db.getBuffer()
  108. _, err = buffer.ReadFrom(file)
  109. if err != nil {
  110. return nil, errors.Trace(err)
  111. }
  112. return buffer, nil
  113. }
  114. func (db *datastoreDB) close() error {
  115. // close will await any active view and update transactions via this lock.
  116. db.lock.Lock()
  117. defer db.lock.Unlock()
  118. db.closed = true
  119. return nil
  120. }
  121. func (db *datastoreDB) getDataStoreMetrics() string {
  122. // TODO: report metrics
  123. return ""
  124. }
  125. func (db *datastoreDB) view(fn func(tx *datastoreTx) error) error {
  126. db.lock.RLock()
  127. defer db.lock.RUnlock()
  128. if db.closed {
  129. return errors.TraceNew("closed")
  130. }
  131. tx := &datastoreTx{db: db}
  132. defer tx.releaseBuffers()
  133. err := fn(tx)
  134. if err != nil {
  135. return errors.Trace(err)
  136. }
  137. return nil
  138. }
  139. func (db *datastoreDB) update(fn func(tx *datastoreTx) error) error {
  140. db.lock.Lock()
  141. defer db.lock.Unlock()
  142. if db.closed {
  143. return errors.TraceNew("closed")
  144. }
  145. tx := &datastoreTx{db: db, canUpdate: true}
  146. defer tx.releaseBuffers()
  147. err := fn(tx)
  148. if err != nil {
  149. return errors.Trace(err)
  150. }
  151. return nil
  152. }
  153. func (tx *datastoreTx) bucket(name []byte) *datastoreBucket {
  154. bucketDirectory := filepath.Join(tx.db.dataDirectory, hex.EncodeToString(name))
  155. err := os.MkdirAll(bucketDirectory, 0700)
  156. if err != nil {
  157. // The original datastore interface does not return an error from Bucket,
  158. // so emit notice, and return zero-value bucket for which all
  159. // operations will fail.
  160. NoticeWarning("bucket failed: %s", errors.Trace(err))
  161. return &datastoreBucket{}
  162. }
  163. return &datastoreBucket{
  164. bucketDirectory: bucketDirectory,
  165. tx: tx,
  166. }
  167. }
  168. func (tx *datastoreTx) clearBucket(name []byte) error {
  169. bucketDirectory := filepath.Join(tx.db.dataDirectory, hex.EncodeToString(name))
  170. err := os.RemoveAll(bucketDirectory)
  171. if err != nil {
  172. return errors.Trace(err)
  173. }
  174. return nil
  175. }
  176. func (tx *datastoreTx) releaseBuffers() {
  177. for _, buffer := range tx.buffers {
  178. tx.db.putBuffer(buffer)
  179. }
  180. tx.buffers = nil
  181. }
  182. func (b *datastoreBucket) get(key []byte) []byte {
  183. if b.tx == nil {
  184. return nil
  185. }
  186. filename := filepath.Join(b.bucketDirectory, hex.EncodeToString(key))
  187. valueBuffer, err := b.tx.db.readBuffer(filename)
  188. if err != nil {
  189. // The original datastore interface does not return an error from Get,
  190. // so emit notice.
  191. NoticeWarning("get failed: %s", errors.Trace(err))
  192. return nil
  193. }
  194. if valueBuffer == nil {
  195. return nil
  196. }
  197. b.tx.buffers = append(b.tx.buffers, valueBuffer)
  198. return valueBuffer.Bytes()
  199. }
  200. func (b *datastoreBucket) put(key, value []byte) error {
  201. if b.tx == nil {
  202. return errors.TraceNew("bucket not found")
  203. }
  204. if !b.tx.canUpdate {
  205. return errors.TraceNew("non-update transaction")
  206. }
  207. filename := filepath.Join(b.bucketDirectory, hex.EncodeToString(key))
  208. // Complete any partial put commit.
  209. err := datastoreApplyCommit(filename)
  210. if err != nil {
  211. return errors.Trace(err)
  212. }
  213. putFilename := filename + ".put"
  214. err = ioutil.WriteFile(putFilename, value, 0600)
  215. if err != nil {
  216. return errors.Trace(err)
  217. }
  218. commitFilename := filename + ".commit"
  219. err = os.Rename(putFilename, commitFilename)
  220. if err != nil {
  221. return errors.Trace(err)
  222. }
  223. err = datastoreApplyCommit(filename)
  224. if err != nil {
  225. return errors.Trace(err)
  226. }
  227. return nil
  228. }
  229. func datastoreApplyCommit(filename string) error {
  230. commitFilename := filename + ".commit"
  231. if _, err := os.Stat(commitFilename); err != nil && os.IsNotExist(err) {
  232. return nil
  233. }
  234. // TODO: may not be sufficient atomic
  235. err := os.Rename(commitFilename, filename)
  236. if err != nil {
  237. return errors.Trace(err)
  238. }
  239. return nil
  240. }
  241. func (b *datastoreBucket) delete(key []byte) error {
  242. if b.tx == nil {
  243. return errors.TraceNew("bucket not found")
  244. }
  245. filename := filepath.Join(b.bucketDirectory, hex.EncodeToString(key))
  246. filenames := []string{filename + ".put", filename + ".commit", filename}
  247. for _, filename := range filenames {
  248. err := os.Remove(filename)
  249. if err != nil && !os.IsNotExist(err) {
  250. return errors.Trace(err)
  251. }
  252. }
  253. return nil
  254. }
  255. func (b *datastoreBucket) cursor() *datastoreCursor {
  256. if b.tx == nil {
  257. // The original datastore interface does not return an error from
  258. // Cursor, so emit notice, and return zero-value cursor for which all
  259. // operations will fail.
  260. return &datastoreCursor{}
  261. }
  262. fileInfos, err := ioutil.ReadDir(b.bucketDirectory)
  263. if err != nil {
  264. NoticeWarning("cursor failed: %s", errors.Trace(err))
  265. return &datastoreCursor{}
  266. }
  267. return &datastoreCursor{
  268. bucket: b,
  269. fileInfos: fileInfos,
  270. }
  271. }
  272. func (c *datastoreCursor) advance() {
  273. if c.bucket == nil {
  274. return
  275. }
  276. for {
  277. c.index += 1
  278. if c.index <= len(c.fileInfos) {
  279. break
  280. }
  281. // Skip any .put or .commit files
  282. if strings.Contains(c.fileInfos[c.index].Name(), ".") {
  283. continue
  284. }
  285. }
  286. }
  287. func (c *datastoreCursor) firstKey() []byte {
  288. if c.bucket == nil {
  289. return nil
  290. }
  291. c.index = 0
  292. return c.currentKey()
  293. }
  294. func (c *datastoreCursor) currentKey() []byte {
  295. if c.bucket == nil {
  296. return nil
  297. }
  298. if c.index >= len(c.fileInfos) {
  299. return nil
  300. }
  301. info := c.fileInfos[c.index]
  302. if info.IsDir() {
  303. NoticeWarning("cursor failed: unexpected dir")
  304. return nil
  305. }
  306. key, err := hex.DecodeString(info.Name())
  307. if err != nil {
  308. NoticeWarning("cursor failed: %s", errors.Trace(err))
  309. return nil
  310. }
  311. return key
  312. }
  313. func (c *datastoreCursor) nextKey() []byte {
  314. if c.bucket == nil {
  315. return nil
  316. }
  317. c.advance()
  318. return c.currentKey()
  319. }
  320. func (c *datastoreCursor) first() ([]byte, []byte) {
  321. if c.bucket == nil {
  322. return nil, nil
  323. }
  324. c.index = 0
  325. return c.current()
  326. }
  327. func (c *datastoreCursor) current() ([]byte, []byte) {
  328. key := c.currentKey()
  329. if key == nil {
  330. return nil, nil
  331. }
  332. if c.lastBuffer != nil {
  333. c.bucket.tx.db.putBuffer(c.lastBuffer)
  334. }
  335. c.lastBuffer = nil
  336. filename := filepath.Join(c.bucket.bucketDirectory, hex.EncodeToString(key))
  337. valueBuffer, err := c.bucket.tx.db.readBuffer(filename)
  338. if valueBuffer == nil {
  339. err = std_errors.New("unexpected nil value")
  340. }
  341. if err != nil {
  342. NoticeWarning("cursor failed: %s", errors.Trace(err))
  343. return nil, nil
  344. }
  345. c.lastBuffer = valueBuffer
  346. return key, valueBuffer.Bytes()
  347. }
  348. func (c *datastoreCursor) next() ([]byte, []byte) {
  349. if c.bucket == nil {
  350. return nil, nil
  351. }
  352. c.advance()
  353. return c.current()
  354. }
  355. func (c *datastoreCursor) close() {
  356. if c.lastBuffer != nil {
  357. c.bucket.tx.db.putBuffer(c.lastBuffer)
  358. c.lastBuffer = nil
  359. }
  360. }