udfd.c 43.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/*
 * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
 *
 * This program is free software: you can use, redistribute, and/or modify
 * it under the terms of the GNU Affero General Public License, version 3
 * or later ("AGPL"), as published by the Free Software Foundation.
 *
 * This program 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.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
 */
dengyihao's avatar
dengyihao 已提交
15 16

// clang-format off
17 18
#include "uv.h"
#include "os.h"
S
slzhou 已提交
19 20
#include "fnLog.h"
#include "thash.h"
S
shenglian zhou 已提交
21

22 23 24
#include "tudf.h"
#include "tudfInt.h"

25
#include "tdatablock.h"
26 27 28 29
#include "tdataformat.h"
#include "tglobal.h"
#include "tmsg.h"
#include "trpc.h"
30
#include "tmisce.h"
31
// clang-format on
32

S
slzhou 已提交
33 34 35
#define UDFD_MAX_SCRIPT_PLUGINS 64
#define UDFD_MAX_SCRIPT_TYPE    1
#define UDFD_MAX_PLUGIN_FUNCS   9
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

typedef struct SUdfCPluginCtx {
  uv_lib_t lib;

  TUdfScalarProcFunc scalarProcFunc;

  TUdfAggStartFunc   aggStartFunc;
  TUdfAggProcessFunc aggProcFunc;
  TUdfAggFinishFunc  aggFinishFunc;
  TUdfAggMergeFunc   aggMergeFunc;

  TUdfInitFunc    initFunc;
  TUdfDestroyFunc destroyFunc;
} SUdfCPluginCtx;

51
int32_t udfdCPluginOpen(SScriptUdfEnvItem *items, int numItems) { return 0; }
52

53
int32_t udfdCPluginClose() { return 0; }
54

S
slzhou 已提交
55
int32_t udfdCPluginUdfInit(SScriptUdfInfo *udf, void **pUdfCtx) {
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
  int32_t         err = 0;
  SUdfCPluginCtx *udfCtx = taosMemoryCalloc(1, sizeof(SUdfCPluginCtx));
  err = uv_dlopen(udf->path, &udfCtx->lib);
  if (err != 0) {
    fnError("can not load library %s. error: %s", udf->path, uv_strerror(err));
    return TSDB_CODE_UDF_LOAD_UDF_FAILURE;
  }
  const char *udfName = udf->name;
  char        initFuncName[TSDB_FUNC_NAME_LEN + 5] = {0};
  char       *initSuffix = "_init";
  strcpy(initFuncName, udfName);
  strncat(initFuncName, initSuffix, strlen(initSuffix));
  uv_dlsym(&udfCtx->lib, initFuncName, (void **)(&udfCtx->initFunc));

  char  destroyFuncName[TSDB_FUNC_NAME_LEN + 5] = {0};
  char *destroySuffix = "_destroy";
  strcpy(destroyFuncName, udfName);
  strncat(destroyFuncName, destroySuffix, strlen(destroySuffix));
  uv_dlsym(&udfCtx->lib, destroyFuncName, (void **)(&udfCtx->destroyFunc));

S
slzhou 已提交
76
  if (udf->funcType == UDF_FUNC_TYPE_SCALAR) {
77 78 79
    char processFuncName[TSDB_FUNC_NAME_LEN] = {0};
    strcpy(processFuncName, udfName);
    uv_dlsym(&udfCtx->lib, processFuncName, (void **)(&udfCtx->scalarProcFunc));
S
slzhou 已提交
80
  } else if (udf->funcType == UDF_FUNC_TYPE_AGG) {
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
    char processFuncName[TSDB_FUNC_NAME_LEN] = {0};
    strcpy(processFuncName, udfName);
    uv_dlsym(&udfCtx->lib, processFuncName, (void **)(&udfCtx->aggProcFunc));
    char  startFuncName[TSDB_FUNC_NAME_LEN + 6] = {0};
    char *startSuffix = "_start";
    strncpy(startFuncName, processFuncName, sizeof(startFuncName));
    strncat(startFuncName, startSuffix, strlen(startSuffix));
    uv_dlsym(&udfCtx->lib, startFuncName, (void **)(&udfCtx->aggStartFunc));
    char  finishFuncName[TSDB_FUNC_NAME_LEN + 7] = {0};
    char *finishSuffix = "_finish";
    strncpy(finishFuncName, processFuncName, sizeof(finishFuncName));
    strncat(finishFuncName, finishSuffix, strlen(finishSuffix));
    uv_dlsym(&udfCtx->lib, finishFuncName, (void **)(&udfCtx->aggFinishFunc));
    char  mergeFuncName[TSDB_FUNC_NAME_LEN + 6] = {0};
    char *mergeSuffix = "_merge";
    strncpy(mergeFuncName, processFuncName, sizeof(mergeFuncName));
    strncat(mergeFuncName, mergeSuffix, strlen(mergeSuffix));
    uv_dlsym(&udfCtx->lib, mergeFuncName, (void **)(&udfCtx->aggMergeFunc));
  }
S
slzhou 已提交
100
  int32_t code = 0;
101
  if (udfCtx->initFunc) {
S
slzhou 已提交
102 103 104 105 106 107 108
    // TODO: handle init call return error
    code = (udfCtx->initFunc)();
    if (code != 0) {
      uv_dlclose(&udfCtx->lib);
      taosMemoryFree(udfCtx);
      return code;
    }
109 110 111 112 113 114 115
  }
  *pUdfCtx = udfCtx;
  return 0;
}

int32_t udfdCPluginUdfDestroy(void *udfCtx) {
  SUdfCPluginCtx *ctx = udfCtx;
S
slzhou 已提交
116
  int32_t         code = 0;
117
  if (ctx->destroyFunc) {
S
slzhou 已提交
118
    code = (ctx->destroyFunc)();
119 120 121
  }
  uv_dlclose(&ctx->lib);
  taosMemoryFree(ctx);
S
slzhou 已提交
122
  return code;
123 124 125 126 127
}

int32_t udfdCPluginUdfScalarProc(SUdfDataBlock *block, SUdfColumn *resultCol, void *udfCtx) {
  SUdfCPluginCtx *ctx = udfCtx;
  if (ctx->scalarProcFunc) {
S
slzhou 已提交
128 129
    return ctx->scalarProcFunc(block, resultCol);
  } else {
S
slzhou 已提交
130 131
    fnError("udfd c plugin scalar proc not implemented");
    return TSDB_CODE_UDF_FUNC_EXEC_FAILURE;
132 133 134 135 136 137
  }
}

int32_t udfdCPluginUdfAggStart(SUdfInterBuf *buf, void *udfCtx) {
  SUdfCPluginCtx *ctx = udfCtx;
  if (ctx->aggStartFunc) {
S
slzhou 已提交
138 139
    return ctx->aggStartFunc(buf);
  } else {
S
slzhou 已提交
140 141
    fnError("udfd c plugin aggregation start not implemented");
    return TSDB_CODE_UDF_FUNC_EXEC_FAILURE;
142 143 144 145 146 147 148
  }
  return 0;
}

int32_t udfdCPluginUdfAggProc(SUdfDataBlock *block, SUdfInterBuf *interBuf, SUdfInterBuf *newInterBuf, void *udfCtx) {
  SUdfCPluginCtx *ctx = udfCtx;
  if (ctx->aggProcFunc) {
S
slzhou 已提交
149 150
    return ctx->aggProcFunc(block, interBuf, newInterBuf);
  } else {
S
slzhou 已提交
151 152
    fnError("udfd c plugin aggregation process not implemented");
    return TSDB_CODE_UDF_FUNC_EXEC_FAILURE;
153 154 155 156 157 158 159
  }
}

int32_t udfdCPluginUdfAggMerge(SUdfInterBuf *inputBuf1, SUdfInterBuf *inputBuf2, SUdfInterBuf *outputBuf,
                               void *udfCtx) {
  SUdfCPluginCtx *ctx = udfCtx;
  if (ctx->aggMergeFunc) {
S
slzhou 已提交
160 161
    return ctx->aggMergeFunc(inputBuf1, inputBuf2, outputBuf);
  } else {
S
slzhou 已提交
162 163
    fnError("udfd c plugin aggregation merge not implemented");
    return TSDB_CODE_UDF_FUNC_EXEC_FAILURE;
164 165 166 167 168 169
  }
}

int32_t udfdCPluginUdfAggFinish(SUdfInterBuf *buf, SUdfInterBuf *resultData, void *udfCtx) {
  SUdfCPluginCtx *ctx = udfCtx;
  if (ctx->aggFinishFunc) {
S
slzhou 已提交
170 171
    return ctx->aggFinishFunc(buf, resultData);
  } else {
S
slzhou 已提交
172 173
    fnError("udfd c plugin aggregation finish not implemented");
    return TSDB_CODE_UDF_FUNC_EXEC_FAILURE;
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199
  }
  return 0;
}

// for c, the function pointer are filled directly and libloaded = true;
// for others, dlopen/dlsym to find function pointers
typedef struct SUdfScriptPlugin {
  int8_t scriptType;

  char     libPath[PATH_MAX];
  bool     libLoaded;
  uv_lib_t lib;

  TScriptUdfScalarProcFunc udfScalarProcFunc;
  TScriptUdfAggStartFunc   udfAggStartFunc;
  TScriptUdfAggProcessFunc udfAggProcFunc;
  TScriptUdfAggMergeFunc   udfAggMergeFunc;
  TScriptUdfAggFinishFunc  udfAggFinishFunc;

  TScriptUdfInitFunc    udfInitFunc;
  TScriptUdfDestoryFunc udfDestroyFunc;

  TScriptOpenFunc  openFunc;
  TScriptCloseFunc closeFunc;
} SUdfScriptPlugin;

S
slzhou 已提交
200
typedef struct SUdfdContext {
201
  uv_loop_t  *loop;
S
slzhou 已提交
202
  uv_pipe_t   ctrlPipe;
203
  uv_signal_t intrSignal;
204
  char        listenPipeName[PATH_MAX + UDF_LISTEN_PIPE_NAME_LEN + 2];
S
slzhou 已提交
205
  uv_pipe_t   listeningPipe;
S
slzhou 已提交
206

207 208 209
  void     *clientRpc;
  SCorEpSet mgmtEp;

S
slzhou 已提交
210
  uv_mutex_t udfsMutex;
211
  SHashObj  *udfsHash;
S
slzhou 已提交
212

213
  uv_mutex_t        scriptPluginsMutex;
S
slzhou 已提交
214
  SUdfScriptPlugin *scriptPlugins[UDFD_MAX_SCRIPT_PLUGINS];
215

216
  SArray *residentFuncs;
217

S
slzhou 已提交
218 219 220 221
  bool printVersion;
} SUdfdContext;

SUdfdContext global;
222

223 224 225
struct SUdfdUvConn;
struct SUvUdfWork;

226
typedef struct SUdfdUvConn {
S
slzhou 已提交
227
  uv_stream_t *client;
228
  char        *inputBuf;
S
slzhou 已提交
229 230 231
  int32_t      inputLen;
  int32_t      inputCap;
  int32_t      inputTotal;
232 233

  struct SUvUdfWork *pWorkList;  // head of work list
234 235 236
} SUdfdUvConn;

typedef struct SUvUdfWork {
237
  SUdfdUvConn *conn;
S
slzhou 已提交
238 239
  uv_buf_t     input;
  uv_buf_t     output;
240 241

  struct SUvUdfWork *pWorkNext;
242 243
} SUvUdfWork;

244
typedef enum { UDF_STATE_INIT = 0, UDF_STATE_LOADING, UDF_STATE_READY } EUdfState;
245

S
slzhou 已提交
246
typedef struct SUdf {
247
  char name[TSDB_FUNC_NAME_LEN + 1];
S
slzhou 已提交
248

dengyihao's avatar
dengyihao 已提交
249 250 251
  int8_t  funcType;
  int8_t  scriptType;
  int8_t  outputType;
S
slzhou 已提交
252 253 254
  int32_t outputLen;
  int32_t bufSize;

dengyihao's avatar
dengyihao 已提交
255
  char path[PATH_MAX];
S
slzhou 已提交
256

257 258 259 260 261
  int32_t    refCount;
  EUdfState  state;
  uv_mutex_t lock;
  uv_cond_t  condReady;
  bool       resident;
262

263 264
  SUdfScriptPlugin *scriptPlugin;
  void             *scriptUdfCtx;
265 266
} SUdf;

S
slzhou 已提交
267
// TODO: add private udf structure.
268
typedef struct SUdfcFuncHandle {
S
slzhou 已提交
269
  SUdf *udf;
270
} SUdfcFuncHandle;
271

272 273 274 275 276 277 278
typedef enum EUdfdRpcReqRspType {
  UDFD_RPC_MNODE_CONNECT = 0,
  UDFD_RPC_RETRIVE_FUNC,
} EUdfdRpcReqRspType;

typedef struct SUdfdRpcSendRecvInfo {
  EUdfdRpcReqRspType rpcType;
dengyihao's avatar
dengyihao 已提交
279
  int32_t            code;
280
  void              *param;
dengyihao's avatar
dengyihao 已提交
281
  uv_sem_t           resultSem;
282 283
} SUdfdRpcSendRecvInfo;

dengyihao's avatar
dengyihao 已提交
284
static void    udfdProcessRpcRsp(void *parent, SRpcMsg *pMsg, SEpSet *pEpSet);
S
shenglian zhou 已提交
285 286
static int32_t udfdFillUdfInfoFromMNode(void *clientRpc, char *udfName, SUdf *udf);
static int32_t udfdConnectToMnode();
dengyihao's avatar
dengyihao 已提交
287
static bool    udfdRpcRfp(int32_t code, tmsg_t msgType);
dengyihao's avatar
dengyihao 已提交
288
static int     initEpSetFromCfg(const char *firstEp, const char *secondEp, SCorEpSet *pEpSet);
S
shenglian zhou 已提交
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305
static int32_t udfdOpenClientRpc();
static int32_t udfdCloseClientRpc();

static void udfdProcessSetupRequest(SUvUdfWork *uvUdf, SUdfRequest *request);
static void udfdProcessCallRequest(SUvUdfWork *uvUdf, SUdfRequest *request);
static void udfdProcessTeardownRequest(SUvUdfWork *uvUdf, SUdfRequest *request);
static void udfdProcessRequest(uv_work_t *req);
static void udfdOnWrite(uv_write_t *req, int status);
static void udfdSendResponse(uv_work_t *work, int status);
static void udfdAllocBuffer(uv_handle_t *handle, size_t suggestedSize, uv_buf_t *buf);
static bool isUdfdUvMsgComplete(SUdfdUvConn *pipe);
static void udfdHandleRequest(SUdfdUvConn *conn);
static void udfdPipeCloseCb(uv_handle_t *pipe);
static void udfdUvHandleError(SUdfdUvConn *conn) { uv_close((uv_handle_t *)conn->client, udfdPipeCloseCb); }
static void udfdPipeRead(uv_stream_t *client, ssize_t nread, const uv_buf_t *buf);
static void udfdOnNewConnection(uv_stream_t *server, int status);

dengyihao's avatar
dengyihao 已提交
306
static void    udfdIntrSignalHandler(uv_signal_t *handle, int signum);
S
shenglian zhou 已提交
307 308
static int32_t removeListeningPipe();

dengyihao's avatar
dengyihao 已提交
309
static void    udfdPrintVersion();
S
shenglian zhou 已提交
310 311 312
static int32_t udfdParseArgs(int32_t argc, char *argv[]);
static int32_t udfdInitLog();

dengyihao's avatar
dengyihao 已提交
313 314
static void    udfdCtrlAllocBufCb(uv_handle_t *handle, size_t suggested_size, uv_buf_t *buf);
static void    udfdCtrlReadCb(uv_stream_t *q, ssize_t nread, const uv_buf_t *buf);
S
shenglian zhou 已提交
315
static int32_t udfdUvInit();
dengyihao's avatar
dengyihao 已提交
316
static void    udfdCloseWalkCb(uv_handle_t *handle, void *arg);
S
shenglian zhou 已提交
317
static int32_t udfdRun();
dengyihao's avatar
dengyihao 已提交
318
static void    udfdConnectMnodeThreadFunc(void *args);
S
shenglian zhou 已提交
319

320 321 322 323 324 325 326 327 328 329 330
void udfdInitializeCPlugin(SUdfScriptPlugin *plugin) {
  plugin->scriptType = TSDB_FUNC_SCRIPT_BIN_LIB;
  plugin->openFunc = udfdCPluginOpen;
  plugin->closeFunc = udfdCPluginClose;
  plugin->udfInitFunc = udfdCPluginUdfInit;
  plugin->udfDestroyFunc = udfdCPluginUdfDestroy;
  plugin->udfScalarProcFunc = udfdCPluginUdfScalarProc;
  plugin->udfAggStartFunc = udfdCPluginUdfAggStart;
  plugin->udfAggProcFunc = udfdCPluginUdfAggProc;
  plugin->udfAggMergeFunc = udfdCPluginUdfAggMerge;
  plugin->udfAggFinishFunc = udfdCPluginUdfAggFinish;
331

S
slzhou 已提交
332 333
  SScriptUdfEnvItem items[1] = {{"LD_LIBRARY_PATH", tsUdfdLdLibPath}};
  plugin->openFunc(items, 1);
334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
  return;
}

int32_t udfdLoadSharedLib(char *libPath, uv_lib_t *pLib, const char *funcName[], void **func[], int numOfFuncs) {
  int err = uv_dlopen(libPath, pLib);
  if (err != 0) {
    fnError("can not load library %s. error: %s", libPath, uv_strerror(err));
    return TSDB_CODE_UDF_LOAD_UDF_FAILURE;
  }

  for (int i = 0; i < numOfFuncs; ++i) {
    err = uv_dlsym(pLib, funcName[i], func[i]);
    if (err != 0) {
      fnError("load library function failed. lib %s function %s", libPath, funcName[i]);
    }
  }
  return 0;
}

353
int32_t udfdInitializePythonPlugin(SUdfScriptPlugin *plugin) {
354
  plugin->scriptType = TSDB_FUNC_SCRIPT_PYTHON;
S
slzhou 已提交
355
  // todo: windows support
S
slzhou 已提交
356
  sprintf(plugin->libPath, "%s", "libtaospyudf.so");
357
  plugin->libLoaded = false;
S
slzhou 已提交
358 359 360 361
  const char *funcName[UDFD_MAX_PLUGIN_FUNCS] = {"pyOpen",         "pyClose",         "pyUdfInit",
                                                 "pyUdfDestroy",   "pyUdfScalarProc", "pyUdfAggStart",
                                                 "pyUdfAggFinish", "pyUdfAggProc",    "pyUdfAggMerge"};
  void      **funcs[UDFD_MAX_PLUGIN_FUNCS] = {
362 363 364
      (void **)&plugin->openFunc,         (void **)&plugin->closeFunc,         (void **)&plugin->udfInitFunc,
      (void **)&plugin->udfDestroyFunc,   (void **)&plugin->udfScalarProcFunc, (void **)&plugin->udfAggStartFunc,
      (void **)&plugin->udfAggFinishFunc, (void **)&plugin->udfAggProcFunc,    (void **)&plugin->udfAggMergeFunc};
S
slzhou 已提交
365
  int32_t err = udfdLoadSharedLib(plugin->libPath, &plugin->lib, funcName, funcs, UDFD_MAX_PLUGIN_FUNCS);
366 367
  if (err != 0) {
    fnError("can not load python plugin. lib path %s", plugin->libPath);
368
    return err;
369
  }
370

371
  if (plugin->openFunc) {
372 373 374 375 376 377 378
    int16_t lenPythonPath = strlen(tsUdfdLdLibPath) + strlen(tsTempDir) + 1 + 1;  // tsTempDir:tsUdfdLdLibPath
    char   *pythonPath = taosMemoryMalloc(lenPythonPath);
#ifdef WINDOWS
    snprintf(pythonPath, lenPythonPath, "%s;%s", tsTempDir, tsUdfdLdLibPath);
#else
    snprintf(pythonPath, lenPythonPath, "%s:%s", tsTempDir, tsUdfdLdLibPath);
#endif
S
slzhou 已提交
379
    SScriptUdfEnvItem items[] = {{"PYTHONPATH", pythonPath}, {"LOGDIR", tsLogDir}};
S
slzhou 已提交
380
    err = plugin->openFunc(items, 2);
381
    taosMemoryFree(pythonPath);
382
  }
383 384 385 386 387
  if (err != 0) {
    fnError("udf script python plugin open func failed. error: %d", err);
    uv_dlclose(&plugin->lib);
    return err;
  }
388
  plugin->libLoaded = true;
389 390

  return 0;
391 392
}

393
void udfdDeinitCPlugin(SUdfScriptPlugin *plugin) {
394 395 396 397 398 399 400 401 402 403 404
  if (plugin->closeFunc) {
    plugin->closeFunc();
  }
  plugin->openFunc = NULL;
  plugin->closeFunc = NULL;
  plugin->udfInitFunc = NULL;
  plugin->udfDestroyFunc = NULL;
  plugin->udfScalarProcFunc = NULL;
  plugin->udfAggStartFunc = NULL;
  plugin->udfAggProcFunc = NULL;
  plugin->udfAggMergeFunc = NULL;
405
  plugin->udfAggFinishFunc = NULL;
406 407 408 409
  return;
}

void udfdDeinitPythonPlugin(SUdfScriptPlugin *plugin) {
410 411 412
  if (plugin->closeFunc) {
    plugin->closeFunc();
  }
S
slzhou 已提交
413
  uv_dlclose(&plugin->lib);
414 415 416
  if (plugin->libLoaded) {
    plugin->libLoaded = false;
  }
417 418 419 420 421 422 423 424 425
  plugin->openFunc = NULL;
  plugin->closeFunc = NULL;
  plugin->udfInitFunc = NULL;
  plugin->udfDestroyFunc = NULL;
  plugin->udfScalarProcFunc = NULL;
  plugin->udfAggStartFunc = NULL;
  plugin->udfAggProcFunc = NULL;
  plugin->udfAggMergeFunc = NULL;
  plugin->udfAggFinishFunc = NULL;
426 427
}

S
slzhou 已提交
428 429
int32_t udfdInitScriptPlugin(int8_t scriptType) {
  SUdfScriptPlugin *plugin = taosMemoryCalloc(1, sizeof(SUdfScriptPlugin));
430

S
slzhou 已提交
431 432 433 434
  switch (scriptType) {
    case TSDB_FUNC_SCRIPT_BIN_LIB:
      udfdInitializeCPlugin(plugin);
      break;
435 436 437 438 439 440
    case TSDB_FUNC_SCRIPT_PYTHON: {
      int32_t err = udfdInitializePythonPlugin(plugin);
      if (err != 0) {
        taosMemoryFree(plugin);
        return err;
      }
S
slzhou 已提交
441
      break;
442
    }
S
slzhou 已提交
443 444 445 446 447
    default:
      fnError("udf script type %d not supported", scriptType);
      taosMemoryFree(plugin);
      return TSDB_CODE_UDF_SCRIPT_NOT_SUPPORTED;
  }
448

S
slzhou 已提交
449 450
  global.scriptPlugins[scriptType] = plugin;
  return TSDB_CODE_SUCCESS;
451 452 453
}

void udfdDeinitScriptPlugins() {
454 455
  SUdfScriptPlugin *plugin = NULL;
  plugin = global.scriptPlugins[TSDB_FUNC_SCRIPT_PYTHON];
S
slzhou 已提交
456 457 458 459
  if (plugin != NULL) {
    udfdDeinitPythonPlugin(plugin);
    taosMemoryFree(plugin);
  }
460 461

  plugin = global.scriptPlugins[TSDB_FUNC_SCRIPT_BIN_LIB];
S
slzhou 已提交
462 463 464 465
  if (plugin != NULL) {
    udfdDeinitCPlugin(plugin);
    taosMemoryFree(plugin);
  }
466 467 468
  return;
}

S
shenglian zhou 已提交
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493
void udfdProcessRequest(uv_work_t *req) {
  SUvUdfWork *uvUdf = (SUvUdfWork *)(req->data);
  SUdfRequest request = {0};
  decodeUdfRequest(uvUdf->input.base, &request);

  switch (request.type) {
    case UDF_TASK_SETUP: {
      udfdProcessSetupRequest(uvUdf, &request);
      break;
    }

    case UDF_TASK_CALL: {
      udfdProcessCallRequest(uvUdf, &request);
      break;
    }
    case UDF_TASK_TEARDOWN: {
      udfdProcessTeardownRequest(uvUdf, &request);
      break;
    }
    default: {
      break;
    }
  }
}

S
slzhou 已提交
494
void convertUdf2UdfInfo(SUdf *udf, SScriptUdfInfo *udfInfo) {
495
  udfInfo->bufSize = udf->bufSize;
S
slzhou 已提交
496 497 498 499 500
  if (udf->funcType == TSDB_FUNC_TYPE_AGGREGATE) {
    udfInfo->funcType = UDF_FUNC_TYPE_AGG;
  } else if (udf->funcType == TSDB_FUNC_TYPE_SCALAR) {
    udfInfo->funcType = UDF_FUNC_TYPE_SCALAR;
  }
S
slzhou 已提交
501
  udfInfo->name = udf->name;
502 503
  udfInfo->outputLen = udf->outputLen;
  udfInfo->outputType = udf->outputType;
S
slzhou 已提交
504
  udfInfo->path = udf->path;
505 506 507
  udfInfo->scriptType = udf->scriptType;
}

S
slzhou 已提交
508 509 510 511 512 513 514 515 516
int32_t udfdRenameUdfFile(SUdf *udf) {
  char newPath[PATH_MAX];
  if (udf->scriptType == TSDB_FUNC_SCRIPT_BIN_LIB) {
    snprintf(newPath, PATH_MAX, "%s/lib%s.so", tsTempDir, udf->name);
  } else if (udf->scriptType == TSDB_FUNC_SCRIPT_PYTHON) {
    snprintf(newPath, PATH_MAX, "%s/%s.py", tsTempDir, udf->name);
  } else {
    return TSDB_CODE_UDF_SCRIPT_NOT_SUPPORTED;
  }
517 518 519 520
  int32_t code = taosRenameFile(udf->path, newPath);
  if (code == 0) {
    sprintf(udf->path, "%s", newPath);
  }
S
slzhou 已提交
521 522 523
  return 0;
}

524 525 526 527 528 529 530
int32_t udfdInitUdf(char *udfName, SUdf *udf) {
  int32_t err = 0;
  err = udfdFillUdfInfoFromMNode(global.clientRpc, udfName, udf);
  if (err != 0) {
    fnError("can not retrieve udf from mnode. udf name %s", udfName);
    return TSDB_CODE_UDF_LOAD_UDF_FAILURE;
  }
S
slzhou 已提交
531 532 533 534 535
  if (udf->scriptType > UDFD_MAX_SCRIPT_TYPE) {
    fnError("udf name %s script type %d not supported", udfName, udf->scriptType);
    return TSDB_CODE_UDF_SCRIPT_NOT_SUPPORTED;
  }

536 537
  uv_mutex_lock(&global.scriptPluginsMutex);
  SUdfScriptPlugin *scriptPlugin = global.scriptPlugins[udf->scriptType];
S
slzhou 已提交
538
  if (scriptPlugin == NULL) {
S
slzhou 已提交
539 540 541 542 543
    err = udfdInitScriptPlugin(udf->scriptType);
    if (err != 0) {
      uv_mutex_unlock(&global.scriptPluginsMutex);
      return err;
    }
544 545
  }
  uv_mutex_unlock(&global.scriptPluginsMutex);
S
slzhou 已提交
546 547 548 549
  udf->scriptPlugin = global.scriptPlugins[udf->scriptType];

  udfdRenameUdfFile(udf);

S
slzhou 已提交
550
  SScriptUdfInfo info = {0};
551
  convertUdf2UdfInfo(udf, &info);
S
slzhou 已提交
552 553 554 555 556
  err = udf->scriptPlugin->udfInitFunc(&info, &udf->scriptUdfCtx);
  if (err != 0) {
    fnError("udf name %s init failed. error %d", udfName, err);
    return err;
  }
557 558

  fnInfo("udf init succeeded. name %s type %d context %p", udf->name, udf->scriptType, (void*)udf->scriptUdfCtx);
559 560 561 562 563
  return 0;
}

SUdf *udfdGetOrCreateUdf(const char *udfName) {
  SUdf *udf = NULL;
S
shenglian zhou 已提交
564
  uv_mutex_lock(&global.udfsMutex);
565
  SUdf **udfInHash = taosHashGet(global.udfsHash, udfName, strlen(udfName));
S
shenglian zhou 已提交
566 567 568 569 570 571 572
  if (udfInHash) {
    ++(*udfInHash)->refCount;
    udf = *udfInHash;
    uv_mutex_unlock(&global.udfsMutex);
  } else {
    SUdf *udfNew = taosMemoryCalloc(1, sizeof(SUdf));
    udfNew->refCount = 1;
573 574
    strncpy(udfNew->name, udfName, TSDB_FUNC_NAME_LEN);

S
shenglian zhou 已提交
575 576 577
    udfNew->state = UDF_STATE_INIT;
    uv_mutex_init(&udfNew->lock);
    uv_cond_init(&udfNew->condReady);
S
slzhou 已提交
578

S
shenglian zhou 已提交
579
    udf = udfNew;
580 581 582 583 584 585 586 587
    udf->resident = false;
    for (int32_t i = 0; i < taosArrayGetSize(global.residentFuncs); ++i) {
      char *funcName = taosArrayGet(global.residentFuncs, i);
      if (strcmp(udfName, funcName) == 0) {
        udf->resident = true;
        break;
      }
    }
588
    SUdf **pUdf = &udf;
589
    taosHashPut(global.udfsHash, udfName, strlen(udfName), pUdf, POINTER_BYTES);
S
shenglian zhou 已提交
590 591
    uv_mutex_unlock(&global.udfsMutex);
  }
592 593 594 595 596 597 598 599 600
  return udf;
}

void udfdProcessSetupRequest(SUvUdfWork *uvUdf, SUdfRequest *request) {
  // TODO: tracable id from client. connect, setup, call, teardown
  fnInfo("setup request. seq num: %" PRId64 ", udf name: %s", request->seqNum, request->setup.udfName);

  SUdfSetupRequest *setup = &request->setup;
  int32_t           code = TSDB_CODE_SUCCESS;
601
  SUdf             *udf = NULL;
602 603

  udf = udfdGetOrCreateUdf(setup->udfName);
S
shenglian zhou 已提交
604 605 606 607

  uv_mutex_lock(&udf->lock);
  if (udf->state == UDF_STATE_INIT) {
    udf->state = UDF_STATE_LOADING;
608
    code = udfdInitUdf(setup->udfName, udf);
609 610 611 612
    if (code == 0) {
      udf->state = UDF_STATE_READY;
    } else {
      udf->state = UDF_STATE_INIT;
S
slzhou 已提交
613
    }
S
shenglian zhou 已提交
614 615 616
    uv_cond_broadcast(&udf->condReady);
    uv_mutex_unlock(&udf->lock);
  } else {
617
    while (udf->state == UDF_STATE_LOADING) {
S
shenglian zhou 已提交
618 619 620 621 622 623 624 625 626 627
      uv_cond_wait(&udf->condReady, &udf->lock);
    }
    uv_mutex_unlock(&udf->lock);
  }
  SUdfcFuncHandle *handle = taosMemoryMalloc(sizeof(SUdfcFuncHandle));
  handle->udf = udf;

  SUdfResponse rsp;
  rsp.seqNum = request->seqNum;
  rsp.type = request->type;
S
slzhou 已提交
628
  rsp.code = (code != 0) ? TSDB_CODE_UDF_FUNC_EXEC_FAILURE : 0;
S
shenglian zhou 已提交
629 630
  rsp.setupRsp.udfHandle = (int64_t)(handle);
  rsp.setupRsp.outputType = udf->outputType;
S
slzhou 已提交
631
  rsp.setupRsp.bytes = udf->outputLen;
S
shenglian zhou 已提交
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647
  rsp.setupRsp.bufSize = udf->bufSize;

  int32_t len = encodeUdfResponse(NULL, &rsp);
  rsp.msgLen = len;
  void *bufBegin = taosMemoryMalloc(len);
  void *buf = bufBegin;
  encodeUdfResponse(&buf, &rsp);

  uvUdf->output = uv_buf_init(bufBegin, len);

  taosMemoryFree(uvUdf->input.base);
  return;
}

void udfdProcessCallRequest(SUvUdfWork *uvUdf, SUdfRequest *request) {
  SUdfCallRequest *call = &request->call;
648 649 650 651
  fnDebug("call request. call type %d, handle: %" PRIx64 ", seq num %" PRId64, call->callType, call->udfHandle,
          request->seqNum);
  SUdfcFuncHandle  *handle = (SUdfcFuncHandle *)(call->udfHandle);
  SUdf             *udf = handle->udf;
S
shenglian zhou 已提交
652
  SUdfResponse      response = {0};
653
  SUdfResponse     *rsp = &response;
S
shenglian zhou 已提交
654 655 656 657 658 659
  SUdfCallResponse *subRsp = &rsp->callRsp;

  int32_t code = TSDB_CODE_SUCCESS;
  switch (call->callType) {
    case TSDB_UDF_CALL_SCALA_PROC: {
      SUdfColumn output = {0};
S
slzhou 已提交
660
      output.colMeta.bytes = udf->outputLen;
661 662 663
      output.colMeta.type = udf->outputType;
      output.colMeta.precision = 0;
      output.colMeta.scale = 0;
S
shenglian zhou 已提交
664 665
      SUdfDataBlock input = {0};
      convertDataBlockToUdfDataBlock(&call->block, &input);
666
      code = udf->scriptPlugin->udfScalarProcFunc(&input, &output, udf->scriptUdfCtx);
S
shenglian zhou 已提交
667 668 669 670 671 672 673
      freeUdfDataDataBlock(&input);
      convertUdfColumnToDataBlock(&output, &response.callRsp.resultData);
      freeUdfColumn(&output);
      break;
    }
    case TSDB_UDF_CALL_AGG_INIT: {
      SUdfInterBuf outBuf = {.buf = taosMemoryMalloc(udf->bufSize), .bufLen = udf->bufSize, .numOfResult = 0};
674
      code = udf->scriptPlugin->udfAggStartFunc(&outBuf, udf->scriptUdfCtx);
S
shenglian zhou 已提交
675 676 677 678 679 680 681
      subRsp->resultBuf = outBuf;
      break;
    }
    case TSDB_UDF_CALL_AGG_PROC: {
      SUdfDataBlock input = {0};
      convertDataBlockToUdfDataBlock(&call->block, &input);
      SUdfInterBuf outBuf = {.buf = taosMemoryMalloc(udf->bufSize), .bufLen = udf->bufSize, .numOfResult = 0};
682
      code = udf->scriptPlugin->udfAggProcFunc(&input, &call->interBuf, &outBuf, udf->scriptUdfCtx);
S
shenglian zhou 已提交
683 684 685 686 687 688
      freeUdfInterBuf(&call->interBuf);
      freeUdfDataDataBlock(&input);
      subRsp->resultBuf = outBuf;

      break;
    }
S
slzhou 已提交
689 690
    case TSDB_UDF_CALL_AGG_MERGE: {
      SUdfInterBuf outBuf = {.buf = taosMemoryMalloc(udf->bufSize), .bufLen = udf->bufSize, .numOfResult = 0};
691
      code = udf->scriptPlugin->udfAggMergeFunc(&call->interBuf, &call->interBuf2, &outBuf, udf->scriptUdfCtx);
S
slzhou 已提交
692 693 694 695 696 697
      freeUdfInterBuf(&call->interBuf);
      freeUdfInterBuf(&call->interBuf2);
      subRsp->resultBuf = outBuf;

      break;
    }
S
shenglian zhou 已提交
698 699
    case TSDB_UDF_CALL_AGG_FIN: {
      SUdfInterBuf outBuf = {.buf = taosMemoryMalloc(udf->bufSize), .bufLen = udf->bufSize, .numOfResult = 0};
700
      code = udf->scriptPlugin->udfAggFinishFunc(&call->interBuf, &outBuf, udf->scriptUdfCtx);
S
shenglian zhou 已提交
701 702 703 704 705 706 707 708 709 710
      freeUdfInterBuf(&call->interBuf);
      subRsp->resultBuf = outBuf;
      break;
    }
    default:
      break;
  }

  rsp->seqNum = request->seqNum;
  rsp->type = request->type;
S
slzhou 已提交
711
  rsp->code = (code != 0) ? TSDB_CODE_UDF_FUNC_EXEC_FAILURE : 0;
S
shenglian zhou 已提交
712 713 714 715 716 717 718 719 720 721 722
  subRsp->callType = call->callType;

  int32_t len = encodeUdfResponse(NULL, rsp);
  rsp->msgLen = len;
  void *bufBegin = taosMemoryMalloc(len);
  void *buf = bufBegin;
  encodeUdfResponse(&buf, rsp);
  uvUdf->output = uv_buf_init(bufBegin, len);

  switch (call->callType) {
    case TSDB_UDF_CALL_SCALA_PROC: {
723 724
      blockDataFreeRes(&call->block);
      blockDataFreeRes(&subRsp->resultData);
S
shenglian zhou 已提交
725 726 727 728 729 730 731
      break;
    }
    case TSDB_UDF_CALL_AGG_INIT: {
      freeUdfInterBuf(&subRsp->resultBuf);
      break;
    }
    case TSDB_UDF_CALL_AGG_PROC: {
732
      blockDataFreeRes(&call->block);
S
shenglian zhou 已提交
733 734 735
      freeUdfInterBuf(&subRsp->resultBuf);
      break;
    }
S
slzhou 已提交
736 737 738 739
    case TSDB_UDF_CALL_AGG_MERGE: {
      freeUdfInterBuf(&subRsp->resultBuf);
      break;
    }
S
shenglian zhou 已提交
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
    case TSDB_UDF_CALL_AGG_FIN: {
      freeUdfInterBuf(&subRsp->resultBuf);
      break;
    }
    default:
      break;
  }

  taosMemoryFree(uvUdf->input.base);
  return;
}

void udfdProcessTeardownRequest(SUvUdfWork *uvUdf, SUdfRequest *request) {
  SUdfTeardownRequest *teardown = &request->teardown;
  fnInfo("teardown. seq number: %" PRId64 ", handle:%" PRIx64, request->seqNum, teardown->udfHandle);
  SUdfcFuncHandle *handle = (SUdfcFuncHandle *)(teardown->udfHandle);
756
  SUdf            *udf = handle->udf;
S
shenglian zhou 已提交
757 758 759 760 761
  bool             unloadUdf = false;
  int32_t          code = TSDB_CODE_SUCCESS;

  uv_mutex_lock(&global.udfsMutex);
  udf->refCount--;
S
slzhou 已提交
762
  if (udf->refCount == 0 && !udf->resident) {
S
shenglian zhou 已提交
763 764 765 766 767
    unloadUdf = true;
    taosHashRemove(global.udfsHash, udf->name, strlen(udf->name));
  }
  uv_mutex_unlock(&global.udfsMutex);
  if (unloadUdf) {
768
    fnInfo("udf teardown. udf name: %s type %d: context %p", udf->name, udf->scriptType, (void*)(udf->scriptUdfCtx));
S
shenglian zhou 已提交
769 770
    uv_cond_destroy(&udf->condReady);
    uv_mutex_destroy(&udf->lock);
S
slzhou 已提交
771 772
    code = udf->scriptPlugin->udfDestroyFunc(udf->scriptUdfCtx);
    fnDebug("udfd destroy function returns %d", code);
S
slzhou 已提交
773
    taosMemoryFree(udf);
S
shenglian zhou 已提交
774 775 776
  }
  taosMemoryFree(handle);

S
slzhou 已提交
777
  SUdfResponse  response = {0};
S
shenglian zhou 已提交
778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
  SUdfResponse *rsp = &response;
  rsp->seqNum = request->seqNum;
  rsp->type = request->type;
  rsp->code = code;
  int32_t len = encodeUdfResponse(NULL, rsp);
  rsp->msgLen = len;
  void *bufBegin = taosMemoryMalloc(len);
  void *buf = bufBegin;
  encodeUdfResponse(&buf, rsp);
  uvUdf->output = uv_buf_init(bufBegin, len);

  taosMemoryFree(uvUdf->input.base);
  return;
}

793
void udfdProcessRpcRsp(void *parent, SRpcMsg *pMsg, SEpSet *pEpSet) {
794
  SUdfdRpcSendRecvInfo *msgInfo = (SUdfdRpcSendRecvInfo *)pMsg->info.ahandle;
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810

  if (pEpSet) {
    if (!isEpsetEqual(&global.mgmtEp.epSet, pEpSet)) {
      updateEpSet_s(&global.mgmtEp, pEpSet);
    }
  }

  if (pMsg->code != TSDB_CODE_SUCCESS) {
    fnError("udfd rpc error. code: %s", tstrerror(pMsg->code));
    msgInfo->code = pMsg->code;
    goto _return;
  }

  if (msgInfo->rpcType == UDFD_RPC_MNODE_CONNECT) {
    SConnectRsp connectRsp = {0};
    tDeserializeSConnectRsp(pMsg->pCont, pMsg->contLen, &connectRsp);
811

dengyihao's avatar
dengyihao 已提交
812 813 814 815 816 817
    int32_t now = taosGetTimestampSec();
    int32_t delta = abs(now - connectRsp.svrTimestamp);
    if (delta > 900) {
      msgInfo->code = TSDB_CODE_TIME_UNSYNCED;
      goto _return;
    }
818

819
    if (connectRsp.epSet.numOfEps == 0) {
S
Shengliang Guan 已提交
820
      msgInfo->code = TSDB_CODE_APP_ERROR;
821 822 823 824 825 826 827 828 829 830
      goto _return;
    }

    if (connectRsp.dnodeNum > 1 && !isEpsetEqual(&global.mgmtEp.epSet, &connectRsp.epSet)) {
      updateEpSet_s(&global.mgmtEp, &connectRsp.epSet);
    }
    msgInfo->code = 0;
  } else if (msgInfo->rpcType == UDFD_RPC_RETRIVE_FUNC) {
    SRetrieveFuncRsp retrieveRsp = {0};
    tDeserializeSRetrieveFuncRsp(pMsg->pCont, pMsg->contLen, &retrieveRsp);
S
slzhou 已提交
831 832 833
    if (retrieveRsp.pFuncInfos == NULL) {
      goto _return;
    }
834
    SFuncInfo *pFuncInfo = (SFuncInfo *)taosArrayGet(retrieveRsp.pFuncInfos, 0);
835 836
    // SUdf      *udf = msgInfo->param;
    SUdf *udf = msgInfo->param;
837 838
    udf->funcType = pFuncInfo->funcType;
    udf->scriptType = pFuncInfo->scriptType;
839
    udf->outputType = pFuncInfo->outputType;
840 841
    udf->outputLen = pFuncInfo->outputLen;
    udf->bufSize = pFuncInfo->bufSize;
S
slzhou 已提交
842

wafwerar's avatar
wafwerar 已提交
843 844 845 846 847 848 849
    if (!osTempSpaceAvailable()) {
      terrno = TSDB_CODE_NO_AVAIL_DISK;
      msgInfo->code = terrno;
      fnError("udfd create shared library failed since %s", terrstr(terrno));
      goto _return;
    }

850
    char path[PATH_MAX] = {0};
851
#ifdef WINDOWS
S
slzhou 已提交
852
    snprintf(path, sizeof(path), "%s%s", tsTempDir, pFuncInfo->name);
853
#else
S
slzhou 已提交
854
    snprintf(path, sizeof(path), "%s/%s", tsTempDir, pFuncInfo->name);
855
#endif
856
    TdFilePtr file = taosOpenFile(path, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_READ | TD_FILE_TRUNC);
wafwerar's avatar
wafwerar 已提交
857 858 859
    if (file == NULL) {
      fnError("udfd write udf shared library: %s failed, error: %d %s", path, errno, strerror(errno));
      msgInfo->code = TSDB_CODE_FILE_CORRUPTED;
wafwerar's avatar
wafwerar 已提交
860
      goto _return;
wafwerar's avatar
wafwerar 已提交
861
    }
862

S
shenglian zhou 已提交
863 864 865 866
    int64_t count = taosWriteFile(file, pFuncInfo->pCode, pFuncInfo->codeSize);
    if (count != pFuncInfo->codeSize) {
      fnError("udfd write udf shared library failed");
      msgInfo->code = TSDB_CODE_FILE_CORRUPTED;
wafwerar's avatar
wafwerar 已提交
867
      goto _return;
S
shenglian zhou 已提交
868
    }
869
    taosCloseFile(&file);
870
    strncpy(udf->path, path, PATH_MAX);
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888
    tFreeSFuncInfo(pFuncInfo);
    taosArrayDestroy(retrieveRsp.pFuncInfos);
    msgInfo->code = 0;
  }

_return:
  rpcFreeCont(pMsg->pCont);
  uv_sem_post(&msgInfo->resultSem);
  return;
}

int32_t udfdFillUdfInfoFromMNode(void *clientRpc, char *udfName, SUdf *udf) {
  SRetrieveFuncReq retrieveReq = {0};
  retrieveReq.numOfFuncs = 1;
  retrieveReq.pFuncNames = taosArrayInit(1, TSDB_FUNC_NAME_LEN);
  taosArrayPush(retrieveReq.pFuncNames, udfName);

  int32_t contLen = tSerializeSRetrieveFuncReq(NULL, 0, &retrieveReq);
889
  void   *pReq = rpcMallocCont(contLen);
890 891 892
  tSerializeSRetrieveFuncReq(pReq, contLen, &retrieveReq);
  taosArrayDestroy(retrieveReq.pFuncNames);

dengyihao's avatar
dengyihao 已提交
893
  SUdfdRpcSendRecvInfo *msgInfo = taosMemoryCalloc(1, sizeof(SUdfdRpcSendRecvInfo));
894 895 896 897 898 899 900 901
  msgInfo->rpcType = UDFD_RPC_RETRIVE_FUNC;
  msgInfo->param = udf;
  uv_sem_init(&msgInfo->resultSem, 0);

  SRpcMsg rpcMsg = {0};
  rpcMsg.pCont = pReq;
  rpcMsg.contLen = contLen;
  rpcMsg.msgType = TDMT_MND_RETRIEVE_FUNC;
902
  rpcMsg.info.ahandle = msgInfo;
903 904 905 906 907 908 909 910 911 912 913 914
  rpcSendRequest(clientRpc, &global.mgmtEp.epSet, &rpcMsg, NULL);

  uv_sem_wait(&msgInfo->resultSem);
  uv_sem_destroy(&msgInfo->resultSem);
  int32_t code = msgInfo->code;
  taosMemoryFree(msgInfo);
  return code;
}

int32_t udfdConnectToMnode() {
  SConnectReq connReq = {0};
  connReq.connType = CONN_TYPE__UDFD;
dengyihao's avatar
dengyihao 已提交
915
  tstrncpy(connReq.app, "udfd", sizeof(connReq.app));
916 917 918 919
  tstrncpy(connReq.user, TSDB_DEFAULT_USER, sizeof(connReq.user));
  char pass[TSDB_PASSWORD_LEN + 1] = {0};
  taosEncryptPass_c((uint8_t *)(TSDB_DEFAULT_PASS), strlen(TSDB_DEFAULT_PASS), pass);
  tstrncpy(connReq.passwd, pass, sizeof(connReq.passwd));
920 921
  connReq.pid = taosGetPId();
  connReq.startTime = taosGetTimestampMs();
922 923

  int32_t contLen = tSerializeSConnectReq(NULL, 0, &connReq);
924
  void   *pReq = rpcMallocCont(contLen);
925 926 927 928 929 930 931 932 933 934
  tSerializeSConnectReq(pReq, contLen, &connReq);

  SUdfdRpcSendRecvInfo *msgInfo = taosMemoryCalloc(1, sizeof(SUdfdRpcSendRecvInfo));
  msgInfo->rpcType = UDFD_RPC_MNODE_CONNECT;
  uv_sem_init(&msgInfo->resultSem, 0);

  SRpcMsg rpcMsg = {0};
  rpcMsg.msgType = TDMT_MND_CONNECT;
  rpcMsg.pCont = pReq;
  rpcMsg.contLen = contLen;
935
  rpcMsg.info.ahandle = msgInfo;
936 937 938 939 940 941 942 943
  rpcSendRequest(global.clientRpc, &global.mgmtEp.epSet, &rpcMsg, NULL);

  uv_sem_wait(&msgInfo->resultSem);
  int32_t code = msgInfo->code;
  uv_sem_destroy(&msgInfo->resultSem);
  taosMemoryFree(msgInfo);
  return code;
}
944

dengyihao's avatar
dengyihao 已提交
945
static bool udfdRpcRfp(int32_t code, tmsg_t msgType) {
946
  if (code == TSDB_CODE_RPC_NETWORK_UNAVAIL || code == TSDB_CODE_RPC_BROKEN_LINK || code == TSDB_CODE_SYN_NOT_LEADER ||
947 948
      code == TSDB_CODE_RPC_SOMENODE_NOT_CONNECTED || code == TSDB_CODE_SYN_RESTORING ||
      code == TSDB_CODE_MNODE_NOT_FOUND || code == TSDB_CODE_APP_IS_STARTING || code == TSDB_CODE_APP_IS_STOPPING) {
949 950
    if (msgType == TDMT_SCH_QUERY || msgType == TDMT_SCH_MERGE_QUERY || msgType == TDMT_SCH_FETCH ||
        msgType == TDMT_SCH_MERGE_FETCH) {
dengyihao's avatar
dengyihao 已提交
951
      return false;
952
    }
S
shenglian zhou 已提交
953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010
    return true;
  } else {
    return false;
  }
}

int initEpSetFromCfg(const char *firstEp, const char *secondEp, SCorEpSet *pEpSet) {
  pEpSet->version = 0;

  // init mnode ip set
  SEpSet *mgmtEpSet = &(pEpSet->epSet);
  mgmtEpSet->numOfEps = 0;
  mgmtEpSet->inUse = 0;

  if (firstEp && firstEp[0] != 0) {
    if (strlen(firstEp) >= TSDB_EP_LEN) {
      terrno = TSDB_CODE_TSC_INVALID_FQDN;
      return -1;
    }

    int32_t code = taosGetFqdnPortFromEp(firstEp, &mgmtEpSet->eps[0]);
    if (code != TSDB_CODE_SUCCESS) {
      terrno = TSDB_CODE_TSC_INVALID_FQDN;
      return terrno;
    }

    mgmtEpSet->numOfEps++;
  }

  if (secondEp && secondEp[0] != 0) {
    if (strlen(secondEp) >= TSDB_EP_LEN) {
      terrno = TSDB_CODE_TSC_INVALID_FQDN;
      return -1;
    }

    taosGetFqdnPortFromEp(secondEp, &mgmtEpSet->eps[mgmtEpSet->numOfEps]);
    mgmtEpSet->numOfEps++;
  }

  if (mgmtEpSet->numOfEps == 0) {
    terrno = TSDB_CODE_TSC_INVALID_FQDN;
    return -1;
  }

  return 0;
}

int32_t udfdOpenClientRpc() {
  SRpcInit rpcInit = {0};
  rpcInit.label = "UDFD";
  rpcInit.numOfThreads = 1;
  rpcInit.cfp = (RpcCfp)udfdProcessRpcRsp;
  rpcInit.sessions = 1024;
  rpcInit.connType = TAOS_CONN_CLIENT;
  rpcInit.idleTime = tsShellActivityTimer * 1000;
  rpcInit.user = TSDB_DEFAULT_USER;
  rpcInit.parent = &global;
  rpcInit.rfp = udfdRpcRfp;
dengyihao's avatar
dengyihao 已提交
1011
  rpcInit.compressSize = tsCompressMsgSize;
1012

dengyihao's avatar
dengyihao 已提交
1013 1014 1015 1016 1017 1018
  int32_t connLimitNum = tsNumOfRpcSessions / (tsNumOfRpcThreads * 3);
  connLimitNum = TMAX(connLimitNum, 10);
  connLimitNum = TMIN(connLimitNum, 500);
  rpcInit.connLimitNum = connLimitNum;
  rpcInit.timeToGetConn = tsTimeToGetAvailableConn;

S
shenglian zhou 已提交
1019 1020 1021 1022 1023 1024 1025 1026 1027
  global.clientRpc = rpcOpen(&rpcInit);
  if (global.clientRpc == NULL) {
    fnError("failed to init dnode rpc client");
    return -1;
  }
  return 0;
}

int32_t udfdCloseClientRpc() {
1028
  fnInfo("udfd begin closing rpc");
S
shenglian zhou 已提交
1029
  rpcClose(global.clientRpc);
1030
  fnInfo("udfd finish closing rpc");
S
shenglian zhou 已提交
1031 1032 1033
  return 0;
}

1034
void udfdOnWrite(uv_write_t *req, int status) {
S
slzhou 已提交
1035 1036
  SUvUdfWork *work = (SUvUdfWork *)req->data;
  if (status < 0) {
S
slzhou 已提交
1037
    fnError("udfd send response error, length: %zu code: %s", work->output.len, uv_err_name(status));
S
slzhou 已提交
1038
  }
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
  // remove work from the connection work list
  if (work->conn != NULL) {
    SUvUdfWork **ppWork;
    for (ppWork = &work->conn->pWorkList; *ppWork && (*ppWork != work); ppWork = &((*ppWork)->pWorkNext)) {
    }
    if (*ppWork == work) {
      *ppWork = work->pWorkNext;
    } else {
      fnError("work not in conn any more");
    }
  }
S
slzhou 已提交
1050 1051 1052
  taosMemoryFree(work->output.base);
  taosMemoryFree(work);
  taosMemoryFree(req);
1053 1054 1055
}

void udfdSendResponse(uv_work_t *work, int status) {
S
slzhou 已提交
1056
  SUvUdfWork *udfWork = (SUvUdfWork *)(work->data);
1057

1058 1059 1060 1061 1062
  if (udfWork->conn != NULL) {
    uv_write_t *write_req = taosMemoryMalloc(sizeof(uv_write_t));
    write_req->data = udfWork;
    uv_write(write_req, udfWork->conn->client, &udfWork->output, 1, udfdOnWrite);
  }
S
slzhou 已提交
1063
  taosMemoryFree(work);
1064 1065 1066
}

void udfdAllocBuffer(uv_handle_t *handle, size_t suggestedSize, uv_buf_t *buf) {
S
slzhou 已提交
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
  SUdfdUvConn *ctx = handle->data;
  int32_t      msgHeadSize = sizeof(int32_t) + sizeof(int64_t);
  if (ctx->inputCap == 0) {
    ctx->inputBuf = taosMemoryMalloc(msgHeadSize);
    if (ctx->inputBuf) {
      ctx->inputLen = 0;
      ctx->inputCap = msgHeadSize;
      ctx->inputTotal = -1;

      buf->base = ctx->inputBuf;
      buf->len = ctx->inputCap;
1078
    } else {
dengyihao's avatar
dengyihao 已提交
1079
      fnError("udfd can not allocate enough memory") buf->base = NULL;
S
slzhou 已提交
1080
      buf->len = 0;
1081
    }
1082
  } else if (ctx->inputTotal == -1 && ctx->inputLen < msgHeadSize) {
1083 1084
    buf->base = ctx->inputBuf + ctx->inputLen;
    buf->len = msgHeadSize - ctx->inputLen;
S
slzhou 已提交
1085 1086 1087 1088 1089 1090 1091 1092
  } else {
    ctx->inputCap = ctx->inputTotal > ctx->inputCap ? ctx->inputTotal : ctx->inputCap;
    void *inputBuf = taosMemoryRealloc(ctx->inputBuf, ctx->inputCap);
    if (inputBuf) {
      ctx->inputBuf = inputBuf;
      buf->base = ctx->inputBuf + ctx->inputLen;
      buf->len = ctx->inputCap - ctx->inputLen;
    } else {
dengyihao's avatar
dengyihao 已提交
1093
      fnError("udfd can not allocate enough memory") buf->base = NULL;
S
slzhou 已提交
1094 1095 1096
      buf->len = 0;
    }
  }
1097 1098 1099
}

bool isUdfdUvMsgComplete(SUdfdUvConn *pipe) {
S
slzhou 已提交
1100 1101 1102 1103 1104 1105 1106 1107
  if (pipe->inputTotal == -1 && pipe->inputLen >= sizeof(int32_t)) {
    pipe->inputTotal = *(int32_t *)(pipe->inputBuf);
  }
  if (pipe->inputLen == pipe->inputCap && pipe->inputTotal == pipe->inputCap) {
    fnDebug("receive request complete. length %d", pipe->inputLen);
    return true;
  }
  return false;
1108 1109 1110
}

void udfdHandleRequest(SUdfdUvConn *conn) {
1111
  char   *inputBuf = conn->inputBuf;
S
shenglian zhou 已提交
1112 1113
  int32_t inputLen = conn->inputLen;

1114
  uv_work_t  *work = taosMemoryMalloc(sizeof(uv_work_t));
S
slzhou 已提交
1115
  SUvUdfWork *udfWork = taosMemoryMalloc(sizeof(SUvUdfWork));
1116 1117 1118
  udfWork->conn = conn;
  udfWork->pWorkNext = conn->pWorkList;
  conn->pWorkList = udfWork;
S
shenglian zhou 已提交
1119
  udfWork->input = uv_buf_init(inputBuf, inputLen);
S
slzhou 已提交
1120 1121 1122 1123 1124 1125
  conn->inputBuf = NULL;
  conn->inputLen = 0;
  conn->inputCap = 0;
  conn->inputTotal = -1;
  work->data = udfWork;
  uv_queue_work(global.loop, work, udfdProcessRequest, udfdSendResponse);
1126 1127 1128
}

void udfdPipeCloseCb(uv_handle_t *pipe) {
S
slzhou 已提交
1129
  SUdfdUvConn *conn = pipe->data;
1130
  SUvUdfWork  *pWork = conn->pWorkList;
1131 1132 1133 1134 1135
  while (pWork != NULL) {
    pWork->conn = NULL;
    pWork = pWork->pWorkNext;
  }

S
slzhou 已提交
1136 1137 1138
  taosMemoryFree(conn->client);
  taosMemoryFree(conn->inputBuf);
  taosMemoryFree(conn);
1139 1140 1141
}

void udfdPipeRead(uv_stream_t *client, ssize_t nread, const uv_buf_t *buf) {
1142
  fnDebug("udfd read %zd bytes from client", nread);
S
slzhou 已提交
1143
  if (nread == 0) return;
1144

S
slzhou 已提交
1145
  SUdfdUvConn *conn = client->data;
1146

S
slzhou 已提交
1147 1148 1149 1150 1151 1152
  if (nread > 0) {
    conn->inputLen += nread;
    if (isUdfdUvMsgComplete(conn)) {
      udfdHandleRequest(conn);
    } else {
      // log error or continue;
1153
    }
S
slzhou 已提交
1154 1155
    return;
  }
1156

S
slzhou 已提交
1157 1158
  if (nread < 0) {
    if (nread == UV_EOF) {
S
slzhou 已提交
1159
      fnInfo("udfd pipe read EOF");
S
slzhou 已提交
1160
    } else {
1161
      fnError("Receive error %s", uv_err_name(nread));
1162
    }
S
slzhou 已提交
1163 1164
    udfdUvHandleError(conn);
  }
1165 1166 1167
}

void udfdOnNewConnection(uv_stream_t *server, int status) {
S
slzhou 已提交
1168
  if (status < 0) {
S
slzhou 已提交
1169
    fnError("udfd new connection error. code: %s", uv_strerror(status));
S
slzhou 已提交
1170 1171
    return;
  }
1172

S
slzhou 已提交
1173 1174 1175 1176
  uv_pipe_t *client = (uv_pipe_t *)taosMemoryMalloc(sizeof(uv_pipe_t));
  uv_pipe_init(global.loop, client, 0);
  if (uv_accept(server, (uv_stream_t *)client) == 0) {
    SUdfdUvConn *ctx = taosMemoryMalloc(sizeof(SUdfdUvConn));
1177
    ctx->pWorkList = NULL;
S
slzhou 已提交
1178 1179 1180 1181 1182 1183 1184 1185 1186 1187
    ctx->client = (uv_stream_t *)client;
    ctx->inputBuf = 0;
    ctx->inputLen = 0;
    ctx->inputCap = 0;
    client->data = ctx;
    ctx->client = (uv_stream_t *)client;
    uv_read_start((uv_stream_t *)client, udfdAllocBuffer, udfdPipeRead);
  } else {
    uv_close((uv_handle_t *)client, NULL);
  }
1188 1189
}

1190 1191
void udfdIntrSignalHandler(uv_signal_t *handle, int signum) {
  fnInfo("udfd signal received: %d\n", signum);
S
slzhou 已提交
1192
  uv_fs_t req;
1193 1194 1195
  uv_fs_unlink(global.loop, &req, global.listenPipeName, NULL);
  uv_signal_stop(handle);
  uv_stop(global.loop);
1196 1197
}

S
slzhou 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
static int32_t udfdParseArgs(int32_t argc, char *argv[]) {
  for (int32_t i = 1; i < argc; ++i) {
    if (strcmp(argv[i], "-c") == 0) {
      if (i < argc - 1) {
        if (strlen(argv[++i]) >= PATH_MAX) {
          printf("config file path overflow");
          return -1;
        }
        tstrncpy(configDir, argv[i], PATH_MAX);
      } else {
        printf("'-c' requires a parameter, default is %s\n", configDir);
        return -1;
      }
    } else if (strcmp(argv[i], "-V") == 0) {
      global.printVersion = true;
    } else {
    }
  }
1216 1217 1218 1219

  return 0;
}

S
shenglian zhou 已提交
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
static void udfdPrintVersion() {
#ifdef TD_ENTERPRISE
  char *releaseName = "enterprise";
#else
  char *releaseName = "community";
#endif
  printf("%s version: %s compatible_version: %s\n", releaseName, version, compatible_version);
  printf("gitinfo: %s\n", gitinfo);
  printf("buildInfo: %s\n", buildinfo);
}

S
slzhou 已提交
1231 1232 1233
static int32_t udfdInitLog() {
  char logName[12] = {0};
  snprintf(logName, sizeof(logName), "%slog", "udfd");
wafwerar's avatar
wafwerar 已提交
1234
  return taosCreateLog(logName, 1, configDir, NULL, NULL, NULL, NULL, 0);
S
slzhou 已提交
1235
}
1236

1237 1238 1239 1240 1241 1242 1243 1244
void udfdCtrlAllocBufCb(uv_handle_t *handle, size_t suggested_size, uv_buf_t *buf) {
  buf->base = taosMemoryMalloc(suggested_size);
  buf->len = suggested_size;
}

void udfdCtrlReadCb(uv_stream_t *q, ssize_t nread, const uv_buf_t *buf) {
  if (nread < 0) {
    fnError("udfd ctrl pipe read error. %s", uv_err_name(nread));
1245
    taosMemoryFree(buf->base);
S
slzhou 已提交
1246
    uv_close((uv_handle_t *)q, NULL);
1247 1248 1249 1250 1251 1252 1253 1254 1255
    uv_stop(global.loop);
    return;
  }
  fnError("udfd ctrl pipe read %zu bytes", nread);
  taosMemoryFree(buf->base);
}

static int32_t removeListeningPipe() {
  uv_fs_t req;
S
slzhou 已提交
1256
  int     err = uv_fs_unlink(global.loop, &req, global.listenPipeName, NULL);
1257 1258 1259 1260
  uv_fs_req_cleanup(&req);
  return err;
}

S
slzhou 已提交
1261
static int32_t udfdUvInit() {
S
slzhou 已提交
1262
  uv_loop_t *loop = taosMemoryMalloc(sizeof(uv_loop_t));
S
slzhou 已提交
1263 1264
  if (loop) {
    uv_loop_init(loop);
S
slzhou 已提交
1265 1266
  } else {
    return -1;
S
slzhou 已提交
1267 1268
  }
  global.loop = loop;
1269

1270
  if (tsStartUdfd) {  // udfd is started by taosd, which shall exit when taosd exit
1271 1272 1273 1274
    uv_pipe_init(global.loop, &global.ctrlPipe, 1);
    uv_pipe_open(&global.ctrlPipe, 0);
    uv_read_start((uv_stream_t *)&global.ctrlPipe, udfdCtrlAllocBufCb, udfdCtrlReadCb);
  }
1275
  getUdfdPipeName(global.listenPipeName, sizeof(global.listenPipeName));
S
slzhou 已提交
1276

1277
  removeListeningPipe();
S
slzhou 已提交
1278

1279
  uv_pipe_init(global.loop, &global.listeningPipe, 0);
S
slzhou 已提交
1280

1281 1282
  uv_signal_init(global.loop, &global.intrSignal);
  uv_signal_start(&global.intrSignal, udfdIntrSignalHandler, SIGINT);
S
slzhou 已提交
1283 1284 1285

  int r;
  fnInfo("bind to pipe %s", global.listenPipeName);
1286
  if ((r = uv_pipe_bind(&global.listeningPipe, global.listenPipeName))) {
S
slzhou 已提交
1287
    fnError("Bind error %s", uv_err_name(r));
1288
    removeListeningPipe();
S
slzhou 已提交
1289
    return -2;
S
slzhou 已提交
1290
  }
1291
  if ((r = uv_listen((uv_stream_t *)&global.listeningPipe, 128, udfdOnNewConnection))) {
S
slzhou 已提交
1292
    fnError("Listen error %s", uv_err_name(r));
1293
    removeListeningPipe();
S
slzhou 已提交
1294
    return -3;
S
slzhou 已提交
1295
  }
1296 1297
  return 0;
}
1298

dengyihao's avatar
dengyihao 已提交
1299
static void udfdCloseWalkCb(uv_handle_t *handle, void *arg) {
S
slzhou 已提交
1300 1301 1302 1303 1304
  if (!uv_is_closing(handle)) {
    uv_close(handle, NULL);
  }
}

S
slzhou 已提交
1305
static int32_t udfdRun() {
1306 1307
  uv_mutex_init(&global.scriptPluginsMutex);

S
slzhou 已提交
1308 1309
  global.udfsHash = taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, HASH_NO_LOCK);
  uv_mutex_init(&global.udfsMutex);
1310

S
slzhou 已提交
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
  fnInfo("start udfd event loop");
  uv_run(global.loop, UV_RUN_DEFAULT);
  fnInfo("udfd event loop stopped.");

  uv_loop_close(global.loop);

  uv_walk(global.loop, udfdCloseWalkCb, NULL);
  uv_run(global.loop, UV_RUN_DEFAULT);
  uv_loop_close(global.loop);

S
slzhou 已提交
1321
  return 0;
S
slzhou 已提交
1322
}
1323

dengyihao's avatar
dengyihao 已提交
1324
void udfdConnectMnodeThreadFunc(void *args) {
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
  int32_t retryMnodeTimes = 0;
  int32_t code = 0;
  while (retryMnodeTimes++ <= TSDB_MAX_REPLICA) {
    uv_sleep(100 * (1 << retryMnodeTimes));
    code = udfdConnectToMnode();
    if (code == 0) {
      break;
    }
    fnError("udfd can not connect to mnode, code: %s. retry", tstrerror(code));
  }

  if (code != 0) {
    fnError("udfd can not connect to mnode");
  }
}

1341
int32_t udfdInitResidentFuncs() {
S
slzhou 已提交
1342 1343 1344 1345
  if (strlen(tsUdfdResFuncs) == 0) {
    return TSDB_CODE_SUCCESS;
  }

1346
  global.residentFuncs = taosArrayInit(2, TSDB_FUNC_NAME_LEN);
1347 1348
  char *pSave = tsUdfdResFuncs;
  char *token;
S
slzhou 已提交
1349
  while ((token = strtok_r(pSave, ",", &pSave)) != NULL) {
1350
    char func[TSDB_FUNC_NAME_LEN + 1] = {0};
S
shenglian zhou 已提交
1351
    strncpy(func, token, TSDB_FUNC_NAME_LEN);
1352
    fnInfo("udfd add resident function %s", func);
S
slzhou 已提交
1353 1354 1355
    taosArrayPush(global.residentFuncs, func);
  }

1356 1357 1358 1359
  return TSDB_CODE_SUCCESS;
}

int32_t udfdDeinitResidentFuncs() {
1360
  for (int32_t i = 0; i < taosArrayGetSize(global.residentFuncs); ++i) {
1361 1362
    char  *funcName = taosArrayGet(global.residentFuncs, i);
    SUdf **udfInHash = taosHashGet(global.udfsHash, funcName, strlen(funcName));
S
slzhou 已提交
1363
    if (udfInHash) {
1364
      SUdf   *udf = *udfInHash;
S
slzhou 已提交
1365 1366
      int32_t code = udf->scriptPlugin->udfDestroyFunc(udf->scriptUdfCtx);
      fnDebug("udfd destroy function returns %d", code);
S
slzhou 已提交
1367
      taosHashRemove(global.udfsHash, funcName, strlen(funcName));
S
slzhou 已提交
1368
      taosMemoryFree(udf);
S
slzhou 已提交
1369 1370
    }
  }
1371
  taosArrayDestroy(global.residentFuncs);
1372 1373 1374
  return TSDB_CODE_SUCCESS;
}

1375 1376 1377 1378 1379 1380
int32_t udfdCleanup() {
  uv_mutex_destroy(&global.udfsMutex);
  taosHashCleanup(global.udfsHash);
  return 0;
}

S
slzhou 已提交
1381
int main(int argc, char *argv[]) {
D
dapan1121 已提交
1382 1383
  if (!taosCheckSystemIsLittleEnd()) {
    printf("failed to start since on non-little-end machines\n");
S
slzhou 已提交
1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
    return -1;
  }

  if (udfdParseArgs(argc, argv) != 0) {
    printf("failed to start since parse args error\n");
    return -1;
  }

  if (global.printVersion) {
    udfdPrintVersion();
    return 0;
  }

  if (udfdInitLog() != 0) {
A
Alex Duan 已提交
1398
    // ignore create log failed, because this error no matter
S
slzhou 已提交
1399 1400 1401
    printf("failed to start since init log error\n");
  }

wafwerar's avatar
wafwerar 已提交
1402
  if (taosInitCfg(configDir, NULL, NULL, NULL, NULL, 0) != 0) {
S
slzhou 已提交
1403
    fnError("failed to start since read config error");
1404
    return -2;
S
slzhou 已提交
1405 1406
  }

1407
  initEpSetFromCfg(tsFirst, tsSecond, &global.mgmtEp);
1408 1409 1410 1411 1412
  if (udfdOpenClientRpc() != 0) {
    fnError("open rpc connection to mnode failure");
    return -3;
  }

S
slzhou 已提交
1413 1414 1415 1416 1417
  if (udfdUvInit() != 0) {
    fnError("uv init failure");
    return -5;
  }

1418 1419
  udfdInitResidentFuncs();

1420 1421 1422
  uv_thread_t mnodeConnectThread;
  uv_thread_create(&mnodeConnectThread, udfdConnectMnodeThreadFunc, NULL);

1423
  udfdRun();
dengyihao's avatar
dengyihao 已提交
1424

S
slzhou 已提交
1425 1426
  removeListeningPipe();
  udfdCloseClientRpc();
S
shenglian zhou 已提交
1427

1428
  udfdDeinitResidentFuncs();
1429 1430 1431

  udfdDeinitScriptPlugins();

1432
  udfdCleanup();
S
shenglian zhou 已提交
1433
  return 0;
1434
}