Highly-opinionated (ex-bullshit-free) MTPROTO proxy for Telegram. If you use v1.0 or upgrade broke you proxy, please read the chapter Version 2
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.

firehol.go 6.6KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327
  1. package ipblocklist
  2. import (
  3. "bufio"
  4. "context"
  5. "fmt"
  6. "io"
  7. "io/ioutil"
  8. "net"
  9. "net/http"
  10. "net/url"
  11. "os"
  12. "regexp"
  13. "strings"
  14. "sync"
  15. "time"
  16. "github.com/9seconds/mtg/v2/mtglib"
  17. "github.com/kentik/patricia"
  18. "github.com/kentik/patricia/bool_tree"
  19. "github.com/panjf2000/ants"
  20. )
  21. const (
  22. fireholIPv4DefaultCIDR = 32
  23. fireholIPv6DefaultCIDR = 128
  24. )
  25. var fireholRegexpComment = regexp.MustCompile(`\s*#.*?$`)
  26. type Firehol struct {
  27. logger mtglib.Logger
  28. rwMutex sync.RWMutex
  29. remoteURLs []string
  30. localFiles []string
  31. httpClient *http.Client
  32. workerPool *ants.Pool
  33. treeV4 *bool_tree.TreeV4
  34. treeV6 *bool_tree.TreeV6
  35. }
  36. func (f *Firehol) Contains(ip net.IP) bool {
  37. if ip == nil {
  38. return true
  39. }
  40. ip4 := ip.To4()
  41. f.rwMutex.RLock()
  42. defer f.rwMutex.RUnlock()
  43. if ip4 != nil {
  44. return f.containsIPv4(ip4)
  45. }
  46. return f.containsIPv6(ip.To16())
  47. }
  48. func (f *Firehol) containsIPv4(addr net.IP) bool {
  49. ip := patricia.NewIPv4AddressFromBytes(addr, 32)
  50. if ok, _, err := f.treeV4.FindDeepestTag(ip); ok && err == nil {
  51. return true
  52. }
  53. return false
  54. }
  55. func (f *Firehol) containsIPv6(addr net.IP) bool {
  56. ip := patricia.NewIPv6Address(addr, 128)
  57. if ok, _, err := f.treeV6.FindDeepestTag(ip); ok && err == nil {
  58. return true
  59. }
  60. return false
  61. }
  62. func (f *Firehol) Run(ctx context.Context, updateEach time.Duration) {
  63. ticker := time.NewTicker(updateEach)
  64. defer func() {
  65. ticker.Stop()
  66. select {
  67. case <-ticker.C:
  68. default:
  69. }
  70. }()
  71. if err := f.update(ctx); err != nil {
  72. f.logger.WarningError("cannot update blocklist", err)
  73. }
  74. for {
  75. select {
  76. case <-ctx.Done():
  77. return
  78. case <-ticker.C:
  79. if err := f.update(ctx); err != nil {
  80. f.logger.WarningError("cannot update blocklist", err)
  81. }
  82. }
  83. }
  84. }
  85. func (f *Firehol) update(ctx context.Context) error { // nolint: funlen, cyclop
  86. ctx, cancel := context.WithCancel(ctx)
  87. defer cancel()
  88. wg := &sync.WaitGroup{}
  89. wg.Add(len(f.remoteURLs) + len(f.localFiles))
  90. treeMutex := &sync.Mutex{}
  91. v4tree := bool_tree.NewTreeV4()
  92. v6tree := bool_tree.NewTreeV6()
  93. errorChan := make(chan error, 1)
  94. defer close(errorChan)
  95. for _, v := range f.localFiles {
  96. go func(filename string) {
  97. defer wg.Done()
  98. if err := f.updateLocalFile(ctx, filename, treeMutex, v4tree, v6tree); err != nil {
  99. cancel()
  100. f.logger.BindStr("filename", filename).WarningError("cannot update", err)
  101. select {
  102. case errorChan <- err:
  103. default:
  104. }
  105. }
  106. }(v)
  107. }
  108. for _, v := range f.remoteURLs {
  109. value := v
  110. f.workerPool.Submit(func() { // nolint: errcheck
  111. defer wg.Done()
  112. if err := f.updateRemoteURL(ctx, value, treeMutex, v4tree, v6tree); err != nil {
  113. cancel()
  114. f.logger.BindStr("url", value).WarningError("cannot update", err)
  115. select {
  116. case errorChan <- err:
  117. default:
  118. }
  119. }
  120. })
  121. }
  122. wg.Wait()
  123. select {
  124. case err := <-errorChan:
  125. return fmt.Errorf("cannot update trees: %w", err)
  126. default:
  127. }
  128. f.rwMutex.Lock()
  129. defer f.rwMutex.Unlock()
  130. f.treeV4 = v4tree
  131. f.treeV6 = v6tree
  132. return nil
  133. }
  134. func (f *Firehol) updateLocalFile(ctx context.Context, filename string,
  135. mutex sync.Locker,
  136. v4tree *bool_tree.TreeV4, v6tree *bool_tree.TreeV6) error {
  137. filefp, err := os.Open(filename)
  138. if err != nil {
  139. return fmt.Errorf("cannot open file: %w", err)
  140. }
  141. defer filefp.Close()
  142. return f.updateTrees(ctx, mutex, filefp, v4tree, v6tree)
  143. }
  144. func (f *Firehol) updateRemoteURL(ctx context.Context, url string,
  145. mutex sync.Locker,
  146. v4tree *bool_tree.TreeV4, v6tree *bool_tree.TreeV6) error {
  147. req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
  148. if err != nil {
  149. return fmt.Errorf("cannot build a request: %w", err)
  150. }
  151. resp, err := f.httpClient.Do(req)
  152. if err != nil {
  153. return fmt.Errorf("cannot request a remote URL %s: %w", url, err)
  154. }
  155. defer func() {
  156. io.Copy(ioutil.Discard, resp.Body) // nolint: errcheck
  157. resp.Body.Close()
  158. }()
  159. return f.updateTrees(ctx, mutex, resp.Body, v4tree, v6tree)
  160. }
  161. func (f *Firehol) updateTrees(ctx context.Context,
  162. mutex sync.Locker,
  163. reader io.Reader,
  164. v4tree *bool_tree.TreeV4,
  165. v6tree *bool_tree.TreeV6) error {
  166. scanner := bufio.NewScanner(reader)
  167. for scanner.Scan() {
  168. select {
  169. case <-ctx.Done():
  170. return ctx.Err()
  171. default:
  172. }
  173. text := scanner.Text()
  174. text = fireholRegexpComment.ReplaceAllLiteralString(text, "")
  175. text = strings.TrimSpace(text)
  176. if text == "" {
  177. continue
  178. }
  179. ip, cidr, err := f.updateParseLine(text)
  180. if err != nil {
  181. return fmt.Errorf("cannot parse a line: %w", err)
  182. }
  183. if err := f.updateAddToTrees(ip, cidr, mutex, v4tree, v6tree); err != nil {
  184. return fmt.Errorf("cannot add a node to the tree: %w", err)
  185. }
  186. }
  187. if scanner.Err() != nil {
  188. return fmt.Errorf("cannot parse a response: %w", scanner.Err())
  189. }
  190. return nil
  191. }
  192. func (f *Firehol) updateParseLine(text string) (net.IP, uint, error) {
  193. _, ipnet, err := net.ParseCIDR(text)
  194. if err != nil {
  195. ipaddr := net.ParseIP(text)
  196. if ipaddr == nil {
  197. return nil, 0, fmt.Errorf("incorrect ip address %s", text)
  198. }
  199. ip4 := ipaddr.To4()
  200. if ip4 != nil {
  201. return ip4, fireholIPv4DefaultCIDR, nil
  202. }
  203. return ipaddr.To16(), fireholIPv6DefaultCIDR, nil
  204. }
  205. ones, _ := ipnet.Mask.Size()
  206. return ipnet.IP, uint(ones), nil
  207. }
  208. func (f *Firehol) updateAddToTrees(ip net.IP, cidr uint,
  209. mutex sync.Locker,
  210. v4tree *bool_tree.TreeV4, v6tree *bool_tree.TreeV6) error {
  211. mutex.Lock()
  212. defer mutex.Unlock()
  213. if ip.To4() != nil {
  214. addr := patricia.NewIPv4AddressFromBytes(ip, cidr)
  215. if _, _, err := v4tree.Set(addr, true); err != nil {
  216. return err // nolint: wrapcheck
  217. }
  218. } else {
  219. addr := patricia.NewIPv6Address(ip, cidr)
  220. if _, _, err := v6tree.Set(addr, true); err != nil {
  221. return err // nolint: wrapcheck
  222. }
  223. }
  224. return nil
  225. }
  226. func NewFirehol(logger mtglib.Logger, network mtglib.Network,
  227. downloadConcurrency uint,
  228. remoteURLs []string,
  229. localFiles []string) (*Firehol, error) {
  230. for _, v := range remoteURLs {
  231. parsed, err := url.Parse(v)
  232. if err != nil {
  233. return nil, fmt.Errorf("incorrect url %s: %w", v, err)
  234. }
  235. switch parsed.Scheme {
  236. case "http", "https":
  237. default:
  238. return nil, fmt.Errorf("unsupported url %s", v)
  239. }
  240. }
  241. for _, v := range localFiles {
  242. if stat, err := os.Stat(v); os.IsNotExist(err) || stat.IsDir() || stat.Mode().Perm()&0o400 == 0 {
  243. return nil, fmt.Errorf("%s is not a readable file", v)
  244. }
  245. }
  246. if downloadConcurrency == 0 {
  247. downloadConcurrency = 1
  248. }
  249. workerPool, _ := ants.NewPool(int(downloadConcurrency))
  250. return &Firehol{
  251. logger: logger.Named("firehol"),
  252. httpClient: network.MakeHTTPClient(nil),
  253. treeV4: bool_tree.NewTreeV4(),
  254. treeV6: bool_tree.NewTreeV6(),
  255. workerPool: workerPool,
  256. remoteURLs: remoteURLs,
  257. localFiles: localFiles,
  258. }, nil
  259. }