dataStore_files.go 9.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397
  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) view(fn func(tx *datastoreTx) error) error {
  122. db.lock.RLock()
  123. defer db.lock.RUnlock()
  124. if db.closed {
  125. return errors.TraceNew("closed")
  126. }
  127. tx := &datastoreTx{db: db}
  128. defer tx.releaseBuffers()
  129. err := fn(tx)
  130. if err != nil {
  131. return errors.Trace(err)
  132. }
  133. return nil
  134. }
  135. func (db *datastoreDB) update(fn func(tx *datastoreTx) error) error {
  136. db.lock.Lock()
  137. defer db.lock.Unlock()
  138. if db.closed {
  139. return errors.TraceNew("closed")
  140. }
  141. tx := &datastoreTx{db: db, canUpdate: true}
  142. defer tx.releaseBuffers()
  143. err := fn(tx)
  144. if err != nil {
  145. return errors.Trace(err)
  146. }
  147. return nil
  148. }
  149. func (tx *datastoreTx) bucket(name []byte) *datastoreBucket {
  150. bucketDirectory := filepath.Join(tx.db.dataDirectory, hex.EncodeToString(name))
  151. err := os.MkdirAll(bucketDirectory, 0700)
  152. if err != nil {
  153. // The original datastore interface does not return an error from Bucket,
  154. // so emit notice, and return zero-value bucket for which all
  155. // operations will fail.
  156. NoticeWarning("bucket failed: %s", errors.Trace(err))
  157. return &datastoreBucket{}
  158. }
  159. return &datastoreBucket{
  160. bucketDirectory: bucketDirectory,
  161. tx: tx,
  162. }
  163. }
  164. func (tx *datastoreTx) clearBucket(name []byte) error {
  165. bucketDirectory := filepath.Join(tx.db.dataDirectory, hex.EncodeToString(name))
  166. err := os.RemoveAll(bucketDirectory)
  167. if err != nil {
  168. return errors.Trace(err)
  169. }
  170. return nil
  171. }
  172. func (tx *datastoreTx) releaseBuffers() {
  173. for _, buffer := range tx.buffers {
  174. tx.db.putBuffer(buffer)
  175. }
  176. tx.buffers = nil
  177. }
  178. func (b *datastoreBucket) get(key []byte) []byte {
  179. if b.tx == nil {
  180. return nil
  181. }
  182. filename := filepath.Join(b.bucketDirectory, hex.EncodeToString(key))
  183. valueBuffer, err := b.tx.db.readBuffer(filename)
  184. if err != nil {
  185. // The original datastore interface does not return an error from Get,
  186. // so emit notice.
  187. NoticeWarning("get failed: %s", errors.Trace(err))
  188. return nil
  189. }
  190. if valueBuffer == nil {
  191. return nil
  192. }
  193. b.tx.buffers = append(b.tx.buffers, valueBuffer)
  194. return valueBuffer.Bytes()
  195. }
  196. func (b *datastoreBucket) put(key, value []byte) error {
  197. if b.tx == nil {
  198. return errors.TraceNew("bucket not found")
  199. }
  200. if !b.tx.canUpdate {
  201. return errors.TraceNew("non-update transaction")
  202. }
  203. filename := filepath.Join(b.bucketDirectory, hex.EncodeToString(key))
  204. // Complete any partial put commit.
  205. err := datastoreApplyCommit(filename)
  206. if err != nil {
  207. return errors.Trace(err)
  208. }
  209. putFilename := filename + ".put"
  210. err = ioutil.WriteFile(putFilename, value, 0600)
  211. if err != nil {
  212. return errors.Trace(err)
  213. }
  214. commitFilename := filename + ".commit"
  215. err = os.Rename(putFilename, commitFilename)
  216. if err != nil {
  217. return errors.Trace(err)
  218. }
  219. err = datastoreApplyCommit(filename)
  220. if err != nil {
  221. return errors.Trace(err)
  222. }
  223. return nil
  224. }
  225. func datastoreApplyCommit(filename string) error {
  226. commitFilename := filename + ".commit"
  227. if _, err := os.Stat(commitFilename); err != nil && os.IsNotExist(err) {
  228. return nil
  229. }
  230. // TODO: may not be sufficient atomic
  231. err := os.Rename(commitFilename, filename)
  232. if err != nil {
  233. return errors.Trace(err)
  234. }
  235. return nil
  236. }
  237. func (b *datastoreBucket) delete(key []byte) error {
  238. if b.tx == nil {
  239. return errors.TraceNew("bucket not found")
  240. }
  241. filename := filepath.Join(b.bucketDirectory, hex.EncodeToString(key))
  242. filenames := []string{filename + ".put", filename + ".commit", filename}
  243. for _, filename := range filenames {
  244. err := os.Remove(filename)
  245. if err != nil && !os.IsNotExist(err) {
  246. return errors.Trace(err)
  247. }
  248. }
  249. return nil
  250. }
  251. func (b *datastoreBucket) cursor() *datastoreCursor {
  252. if b.tx == nil {
  253. // The original datastore interface does not return an error from
  254. // Cursor, so emit notice, and return zero-value cursor for which all
  255. // operations will fail.
  256. return &datastoreCursor{}
  257. }
  258. fileInfos, err := ioutil.ReadDir(b.bucketDirectory)
  259. if err != nil {
  260. NoticeWarning("cursor failed: %s", errors.Trace(err))
  261. return &datastoreCursor{}
  262. }
  263. return &datastoreCursor{
  264. bucket: b,
  265. fileInfos: fileInfos,
  266. }
  267. }
  268. func (c *datastoreCursor) advance() {
  269. if c.bucket == nil {
  270. return
  271. }
  272. for {
  273. c.index += 1
  274. if c.index <= len(c.fileInfos) {
  275. break
  276. }
  277. // Skip any .put or .commit files
  278. if strings.Contains(c.fileInfos[c.index].Name(), ".") {
  279. continue
  280. }
  281. }
  282. }
  283. func (c *datastoreCursor) firstKey() []byte {
  284. if c.bucket == nil {
  285. return nil
  286. }
  287. c.index = 0
  288. return c.currentKey()
  289. }
  290. func (c *datastoreCursor) currentKey() []byte {
  291. if c.bucket == nil {
  292. return nil
  293. }
  294. if c.index >= len(c.fileInfos) {
  295. return nil
  296. }
  297. info := c.fileInfos[c.index]
  298. if info.IsDir() {
  299. NoticeWarning("cursor failed: unexpected dir")
  300. return nil
  301. }
  302. key, err := hex.DecodeString(info.Name())
  303. if err != nil {
  304. NoticeWarning("cursor failed: %s", errors.Trace(err))
  305. return nil
  306. }
  307. return key
  308. }
  309. func (c *datastoreCursor) nextKey() []byte {
  310. if c.bucket == nil {
  311. return nil
  312. }
  313. c.advance()
  314. return c.currentKey()
  315. }
  316. func (c *datastoreCursor) first() ([]byte, []byte) {
  317. if c.bucket == nil {
  318. return nil, nil
  319. }
  320. c.index = 0
  321. return c.current()
  322. }
  323. func (c *datastoreCursor) current() ([]byte, []byte) {
  324. key := c.currentKey()
  325. if key == nil {
  326. return nil, nil
  327. }
  328. if c.lastBuffer != nil {
  329. c.bucket.tx.db.putBuffer(c.lastBuffer)
  330. }
  331. c.lastBuffer = nil
  332. filename := filepath.Join(c.bucket.bucketDirectory, hex.EncodeToString(key))
  333. valueBuffer, err := c.bucket.tx.db.readBuffer(filename)
  334. if valueBuffer == nil {
  335. err = std_errors.New("unexpected nil value")
  336. }
  337. if err != nil {
  338. NoticeWarning("cursor failed: %s", errors.Trace(err))
  339. return nil, nil
  340. }
  341. c.lastBuffer = valueBuffer
  342. return key, valueBuffer.Bytes()
  343. }
  344. func (c *datastoreCursor) next() ([]byte, []byte) {
  345. if c.bucket == nil {
  346. return nil, nil
  347. }
  348. c.advance()
  349. return c.current()
  350. }
  351. func (c *datastoreCursor) close() {
  352. if c.lastBuffer != nil {
  353. c.bucket.tx.db.putBuffer(c.lastBuffer)
  354. c.lastBuffer = nil
  355. }
  356. }