static char* getFileName(MonkState* state, long pFileId) { char* pFile = queryPFileForFileId(state->dbManager, pFileId); if (!pFile) { printf("file not found for pFileId=%ld\n", pFileId); return NULL; } char* pFileName; #ifdef MONK_MULTI_THREAD #pragma omp critical(getFileName) #endif { pFileName = fo_RepMkPath("files", pFile); } if (!pFileName) { printf("file '%s' not found\n", pFile); } free(pFile); return pFileName; }
/** * \brief Get the mimetype for a package * \param argc the number of command line arguments * \param argv the command line arguments * \return 0 on a successful program execution */ int main(int argc, char *argv[]) { int arg; char *Parm = NULL; char *Path = NULL; int c; char *agent_desc = "Determines mimetype for each file"; int pfile_count = 0; int Agent_pk; int ars_pk = 0; int upload_pk = 0; // the upload primary key int user_pk = 0; char *AgentARSName = "mimetype_ars"; int rv; PGresult *result; char sqlbuf[1024]; int CmdlineFlag = 0; ///< run from command line flag, 1 yes, 0 not char *COMMIT_HASH; char *VERSION; char agent_rev[MAXCMD]; /* initialize the scheduler connection */ fo_scheduler_connect(&argc, argv, &pgConn); /* Process command-line */ while((c = getopt(argc,argv,"iCc:hvV")) != -1) { switch(c) { case 'i': PQfinish(pgConn); return(0); case 'c': /* do nothing with this option */ break; case 'C': CmdlineFlag = 1; break; case 'v': agent_verbose++; break; case 'V': printf("%s", BuildVersion); PQfinish(pgConn); return(0); default: Usage(argv[0]); PQfinish(pgConn); exit(-1); } } COMMIT_HASH = fo_sysconfig("mimetype", "COMMIT_HASH"); VERSION = fo_sysconfig("mimetype", "VERSION"); sprintf(agent_rev, "%s.%s", VERSION, COMMIT_HASH); /* Get the Agent Key from the DB */ Agent_pk = fo_GetAgentKey(pgConn, basename(argv[0]), 0, agent_rev, agent_desc); FMimetype = fopen("/etc/mime.types","rb"); if (!FMimetype) { LOG_WARNING("Unable to open /etc/mime.types\n"); } MagicCookie = magic_open(MAGIC_PRESERVE_ATIME|MAGIC_MIME); if (MagicCookie == NULL) { LOG_FATAL("Failed to initialize magic cookie\n"); PQfinish(pgConn); exit(-1); } if (magic_load(MagicCookie,NULL) != 0) { LOG_FATAL("Failed to load magic file: UnMagic\n"); PQfinish(pgConn); exit(-1); } /* Run from the command-line (for testing) */ for(arg=optind; arg < argc; arg++) { Akey = -1; memset(A,'\0',sizeof(A)); strncpy(A,argv[arg],sizeof(A)); DBCheckMime(A); } /* Run from scheduler! */ if (0 == CmdlineFlag) { user_pk = fo_scheduler_userID(); /* get user_pk for user who queued the agent */ while(fo_scheduler_next()) { /* get piece of information, including upload_pk, others */ Parm = fo_scheduler_current(); if (Parm && Parm[0]) { upload_pk = atoi(Parm); /* Check Permissions */ if (GetUploadPerm(pgConn, upload_pk, user_pk) < PERM_WRITE) { LOG_ERROR("You have no update permissions on upload %d", upload_pk); continue; } /* does ars table exist? * If not, create it. */ rv = fo_tableExists(pgConn, AgentARSName); if (!rv) { rv = fo_CreateARSTable(pgConn, AgentARSName); if (!rv) return(0); } /* check ars table if this is duplicate request*/ memset(sqlbuf, 0, sizeof(sqlbuf)); snprintf(sqlbuf, sizeof(sqlbuf), "select ars_pk from mimetype_ars,agent \ where agent_pk=agent_fk and ars_success=true \ and upload_fk='%d' and agent_fk='%d'", upload_pk, Agent_pk); result = PQexec(pgConn, sqlbuf); if (fo_checkPQresult(pgConn, result, sqlbuf, __FILE__, __LINE__)) exit(-1); if (PQntuples(result) > 0) { PQclear(result); LOG_WARNING("Ignoring requested mimetype analysis of upload %d - Results are already in database.\n",upload_pk); continue; } PQclear(result); /* Record analysis start in mimetype_ars, the mimetype audit trail. */ ars_pk = fo_WriteARS(pgConn, ars_pk, upload_pk, Agent_pk, AgentARSName, 0, 0); /* get all pfile ids on a upload record */ memset(sqlbuf, 0, sizeof(sqlbuf)); snprintf(sqlbuf, sizeof(sqlbuf), "SELECT DISTINCT(pfile_pk) as Akey, pfile_sha1 || '.' || pfile_md5 || '.' || pfile_size AS A FROM uploadtree, pfile WHERE uploadtree.pfile_fk = pfile.pfile_pk AND pfile_mimetypefk is NULL AND upload_fk = '%d';", upload_pk); result = PQexec(pgConn, sqlbuf); if (fo_checkPQresult(pgConn, result, sqlbuf, __FILE__, __LINE__)) exit(-1); pfile_count = PQntuples(result); int i; for(i=0; i < pfile_count; i++) { Akey = atoi(PQgetvalue(result, i, 0)); strncpy(A, PQgetvalue(result, i, 1), sizeof(A)); if (Akey <= 0 || A[0]=='\0') { printf("ERROR: Data is in an unknown format.\n"); PQfinish(pgConn); exit(-1); } /* Process the repository file */ /* Find the path */ Path = fo_RepMkPath("files",A); if (Path && fo_RepExist("files",A)) { /* Get the mimetype! */ DBCheckMime(Path); } else { printf("ERROR pfile %d Unable to process.\n",Akey); printf("LOG pfile %d File '%s' not found.\n",Akey,A); PQfinish(pgConn); exit(-1); } /* Clean up Path memory */ if(Path) { free(Path); Path = NULL; } fo_scheduler_heart(1); } PQclear(result); /* Record analysis success in mimetype_ars. */ if (ars_pk) fo_WriteARS(pgConn, ars_pk, upload_pk, Agent_pk, AgentARSName, 0, 1); } } } /* if run from scheduler */ /* Clean up */ if (FMimetype) fclose(FMimetype); magic_close(MagicCookie); if (DBMime) PQclear(DBMime); if (pgConn) PQfinish(pgConn); /* after cleaning up agent, disconnect from the scheduler, this doesn't return */ fo_scheduler_disconnect(0); return(0); } /* main() */
int main(int argc, char *argv[]) { int Pid; int c; int rvExist1=0, rvExist2=0; PGresult *result; char *NewDir="."; char *AgentName = "ununpack"; char *AgentARSName = "ununpack_ars"; char *agent_desc = "Unpacks archives (iso, tar, etc)"; int Recurse=0; int ars_pk = 0; int user_pk = 0; long Pfile_size = 0; char *ListOutName=NULL; char *Fname = NULL; char *FnameCheck = NULL; char *COMMIT_HASH; char *VERSION; char agent_rev[PATH_MAX]; struct stat Stat; /* connect to the scheduler */ fo_scheduler_connect(&argc, argv, &pgConn); while((c = getopt(argc,argv,"ACc:d:FfHhL:m:PQiqRr:T:t:U:VvXx")) != -1) { switch(c) { case 'A': SetContainerArtifact=0; break; case 'C': ForceContinue=1; break; case 'c': break; /* handled by fo_scheduler_connect() */ case 'd': /* if there is a %U in the path, substitute a unique ID */ NewDir=PathCheck(optarg); break; case 'F': UseRepository=1; break; case 'f': ForceDuplicate=1; break; case 'L': ListOutName=optarg; break; case 'm': MaxThread = atoi(optarg); if (MaxThread < 1) MaxThread=1; break; case 'P': PruneFiles=1; break; case 'R': Recurse=-1; break; case 'r': Recurse=atoi(optarg); break; case 'i': if (!IsExe("dpkg-source",Quiet)) LOG_WARNING("dpkg-source is not available on this system. This means that debian source packages will NOT be unpacked."); SafeExit(0); break; /* never reached */ case 'Q': UseRepository=1; user_pk = fo_scheduler_userID(); /* get user_pk for user who queued the agent */ /* Get the upload_pk from the scheduler */ if((Upload_Pk = fo_scheduler_next()) == NULL) SafeExit(0); break; case 'q': Quiet=1; break; case 'T': memset(REP_GOLD,0,sizeof(REP_GOLD)); strncpy(REP_GOLD,optarg,sizeof(REP_GOLD)-1); break; case 't': memset(REP_FILES,0,sizeof(REP_FILES)); strncpy(REP_FILES,optarg,sizeof(REP_FILES)-1); break; case 'U': UseRepository = 1; Recurse = -1; Upload_Pk = optarg; break; case 'V': printf("%s", BuildVersion);SafeExit(0); case 'v': Verbose++; break; case 'X': UnlinkSource=1; break; case 'x': UnlinkAll=1; break; default: Usage(argv[0], BuildVersion); SafeExit(25); } } /* Open DB and Initialize CMD table */ if (UseRepository) { /* Check Permissions */ if (GetUploadPerm(pgConn, atoi(Upload_Pk), user_pk) < PERM_WRITE) { LOG_ERROR("You have no update permissions on upload %s", Upload_Pk); SafeExit(100); } COMMIT_HASH = fo_sysconfig(AgentName, "COMMIT_HASH"); VERSION = fo_sysconfig(AgentName, "VERSION"); sprintf(agent_rev, "%s.%s", VERSION, COMMIT_HASH); /* Get the unpack agent key */ agent_pk = fo_GetAgentKey(pgConn, AgentName, atoi(Upload_Pk), agent_rev,agent_desc); InitCmd(); /* Make sure ars table exists */ if (!fo_CreateARSTable(pgConn, AgentARSName)) SafeExit(0); /* Has this user previously unpacked this upload_pk successfully? * In this case we are done. No new ars record is needed since no * processing is initiated. * The unpack version is ignored. */ snprintf(SQL,MAXSQL, "SELECT ars_pk from %s where upload_fk='%s' and ars_success=TRUE", AgentARSName, Upload_Pk); result = PQexec(pgConn, SQL); if (fo_checkPQresult(pgConn, result, SQL, __FILE__, __LINE__)) SafeExit(101); if (PQntuples(result) > 0) /* if there is a value */ { PQclear(result); LOG_WARNING("Upload_pk %s, has already been unpacked. No further action required", Upload_Pk) SafeExit(0); } PQclear(result); /* write the unpack_ars start record */ ars_pk = fo_WriteARS(pgConn, ars_pk, atoi(Upload_Pk), agent_pk, AgentARSName, 0, 0); /* Get Pfile path and Pfile_Pk, from Upload_Pk */ snprintf(SQL,MAXSQL, "SELECT pfile.pfile_sha1 || '.' || pfile.pfile_md5 || '.' || pfile.pfile_size AS pfile, pfile_fk, pfile_size FROM upload INNER JOIN pfile ON upload.pfile_fk = pfile.pfile_pk WHERE upload.upload_pk = '%s'", Upload_Pk); result = PQexec(pgConn, SQL); if (fo_checkPQresult(pgConn, result, SQL, __FILE__, __LINE__)) SafeExit(102); if (PQntuples(result) > 0) /* if there is a value */ { Pfile = strdup(PQgetvalue(result,0,0)); Pfile_Pk = strdup(PQgetvalue(result,0,1)); Pfile_size = atol(PQgetvalue(result, 0, 2)); if (Pfile_size == 0) { PQclear(result); LOG_WARNING("Uploaded file (Upload_pk %s), is zero length. There is nothing to unpack.", Upload_Pk) SafeExit(0); } PQclear(result); } // Determine if uploadtree records should go into a separate table. // If the input file size is > 500MB, then create a separate uploadtree_{upload_pk} table // that inherits from the master uploadtree table. // Save uploadtree_tablename, it will get written to upload.uploadtree_tablename later. if (Pfile_size > 500000000) { sprintf(uploadtree_tablename, "uploadtree_%s", Upload_Pk); if (!fo_tableExists(pgConn, uploadtree_tablename)) { snprintf(SQL,MAXSQL,"CREATE TABLE %s (LIKE uploadtree INCLUDING DEFAULTS INCLUDING CONSTRAINTS INCLUDING INDEXES); ALTER TABLE %s ADD CONSTRAINT %s CHECK (upload_fk=%s); ALTER TABLE %s INHERIT uploadtree", uploadtree_tablename, uploadtree_tablename, uploadtree_tablename, Upload_Pk, uploadtree_tablename); PQsetNoticeProcessor(pgConn, SQLNoticeProcessor, SQL); // ignore notice about implicit primary key index creation result = PQexec(pgConn, SQL); // Ignore postgres notice about creating an implicit index if (PQresultStatus(result) != PGRES_NONFATAL_ERROR) if (fo_checkPQcommand(pgConn, result, SQL, __FILE__, __LINE__)) SafeExit(103); PQclear(result); } } else strcpy(uploadtree_tablename, "uploadtree_a"); } CheckCommands(Quiet); if (NewDir) MkDir(NewDir); if (Verbose) { fclose(stderr) ; stderr=stdout; } /* don't interlace! */ if (ListOutName != NULL) { if ((ListOutName[0]=='-') && (ListOutName[1]=='\0')) ListOutFile=stdout; else ListOutFile = fopen(ListOutName,"w"); if (!ListOutFile) { LOG_ERROR("pfile %s Unable to write to %s\n",Pfile_Pk,ListOutName) SafeExit(104); } else { /* Start the file */ fputs("<xml tool=\"ununpack\" ",ListOutFile); fputs("version=\"",ListOutFile); fputs(Version,ListOutFile); fputs("\" ",ListOutFile); fputs("compiled_date=\"",ListOutFile); fputs(__DATE__,ListOutFile); fputs(" ",ListOutFile); fputs(__TIME__,ListOutFile); fputs("\"",ListOutFile); fputs(">\n",ListOutFile); } /* Problem: When parallel processing, the XML may be generated out of order. Solution? When using XML, only use 1 thread. */ MaxThread=1; } // Set ReunpackSwitch if the uploadtree records are missing from the database. if (!ReunpackSwitch && UseRepository) { snprintf(SQL,MAXSQL,"SELECT uploadtree_pk FROM uploadtree WHERE upload_fk=%s limit 1;",Upload_Pk); result = PQexec(pgConn, SQL); if (fo_checkPQresult(pgConn, result, SQL, __FILE__, __LINE__)) SafeExit(105); if (PQntuples(result) == 0) ReunpackSwitch=1; PQclear(result); } /*** process files from command line ***/ for( ; optind<argc; optind++) { CksumFile *CF=NULL; Cksum *Sum; int i; if (Fname) { free(Fname); Fname=NULL; } if (ListOutName != NULL) { fprintf(ListOutFile,"<source source=\"%s\" ",argv[optind]); if (UseRepository && !fo_RepExist(REP_FILES,argv[optind])) { /* make sure the source exists in the src repository */ if (fo_RepImport(argv[optind],REP_FILES,argv[optind],1) != 0) { LOG_ERROR("Failed to import '%s' as '%s' into the repository",argv[optind],argv[optind]) SafeExit(106); } } } if (UseRepository) { if (fo_RepExist(REP_FILES,argv[optind])) { Fname=fo_RepMkPath(REP_FILES,argv[optind]); } else if (fo_RepExist(REP_GOLD,argv[optind])) { Fname=fo_RepMkPath(REP_GOLD,argv[optind]); if (fo_RepImport(Fname,REP_FILES,argv[optind],1) != 0) { LOG_ERROR("Failed to import '%s' as '%s' into the repository",Fname,argv[optind]) SafeExit(107); } } if (Fname) { FnameCheck = Fname; CF = SumOpenFile(Fname); } else { LOG_ERROR("NO file unpacked. File %s does not exist either in GOLD or FILES", Pfile); SafeExit(108); } /* else: Fname is NULL and CF is NULL */ } else { FnameCheck = argv[optind]; CF = SumOpenFile(argv[optind]); } /* Check file to unpack. Does it exist? Is it zero length? */ if (stat(FnameCheck,&Stat)) { LOG_ERROR("File to unpack is unavailable: %s, error: %s", Fname, strerror(errno)); SafeExit(109); } else if (Stat.st_size < 1) { LOG_WARNING("File to unpack is empty: %s", Fname); SafeExit(110); } if (ListOutFile) { if (CF) { Sum = SumComputeBuff(CF); SumCloseFile(CF); if (Sum) { fputs("fuid=\"",ListOutFile); for(i=0; i<20; i++) { fprintf(ListOutFile,"%02X",Sum->SHA1digest[i]); } fputs(".",ListOutFile); for(i=0; i<16; i++) { fprintf(ListOutFile,"%02X",Sum->MD5digest[i]); } fputs(".",ListOutFile); fprintf(ListOutFile,"%Lu",(long long unsigned int)Sum->DataLen); fputs("\" ",ListOutFile); free(Sum); } /* if Sum */ } /* if CF */ else /* file too large to mmap (probably) */ { FILE *Fin; Fin = fopen(argv[optind],"rb"); if (Fin) { Sum = SumComputeFile(Fin); if (Sum) { fputs("fuid=\"",ListOutFile); for(i=0; i<20; i++) { fprintf(ListOutFile,"%02X",Sum->SHA1digest[i]); } fputs(".",ListOutFile); for(i=0; i<16; i++) { fprintf(ListOutFile,"%02X",Sum->MD5digest[i]); } fputs(".",ListOutFile); fprintf(ListOutFile,"%Lu",(long long unsigned int)Sum->DataLen); fputs("\" ",ListOutFile); free(Sum); } fclose(Fin); } } /* else no CF */ fprintf(ListOutFile,">\n"); /* end source XML */ } if (Fname) TraverseStart(Fname,"called by main via args",NewDir,Recurse); else TraverseStart(argv[optind],"called by main",NewDir,Recurse); if (ListOutName != NULL) fprintf(ListOutFile,"</source>\n"); } /* end for */ /* free memory */ if (Fname) { free(Fname); Fname=NULL; } /* process pfile from scheduler */ if (Pfile) { if (0 == (rvExist1 = fo_RepExist2(REP_FILES,Pfile))) { Fname=fo_RepMkPath(REP_FILES,Pfile); } else if (0 == (rvExist2 = fo_RepExist2(REP_GOLD,Pfile))) { Fname=fo_RepMkPath(REP_GOLD,Pfile); if (fo_RepImport(Fname,REP_FILES,Pfile,1) != 0) { LOG_ERROR("Failed to import '%s' as '%s' into the repository",Fname,Pfile) SafeExit(111); } } if (Fname) { TraverseStart(Fname,"called by main via env",NewDir,Recurse); free(Fname); Fname=NULL; } else { LOG_ERROR("NO file unpacked!"); if (rvExist1 > 0) { Fname=fo_RepMkPath(REP_FILES, Pfile); LOG_ERROR("Error is %s for %s", strerror(rvExist1), Fname); } if (rvExist2 > 0) { Fname=fo_RepMkPath(REP_GOLD, Pfile); LOG_ERROR("Error is %s for %s", strerror(rvExist2), Fname); } SafeExit(112); } } /* recurse on all the children */ if (Thread > 0) do { Pid = ParentWait(); Thread--; if (Pid >= 0) { if (!Queue[Pid].ChildEnd) { /* copy over data */ if (Recurse > 0) Traverse(Queue[Pid].ChildRecurse,NULL,"called by wait",NULL,Recurse-1,&Queue[Pid].PI); else if (Recurse < 0) Traverse(Queue[Pid].ChildRecurse,NULL,"called by wait",NULL,Recurse,&Queue[Pid].PI); } } } while(Pid >= 0); if (MagicCookie) magic_close(MagicCookie); if (ListOutFile) { fprintf(ListOutFile,"<summary files_regular=\"%d\" files_compressed=\"%d\" artifacts=\"%d\" directories=\"%d\" containers=\"%d\" />\n", TotalFiles,TotalCompressedFiles,TotalArtifacts, TotalDirectories,TotalContainers); fputs("</xml>\n",ListOutFile); } if (pgConn) { /* If it completes, mark it! */ if (Upload_Pk) { snprintf(SQL,MAXSQL,"UPDATE upload SET upload_mode = (upload_mode | (1<<5)), uploadtree_tablename='%s' WHERE upload_pk = '%s';",uploadtree_tablename, Upload_Pk); result = PQexec(pgConn, SQL); /* UPDATE upload */ if (fo_checkPQcommand(pgConn, result, SQL, __FILE__ ,__LINE__)) SafeExit(113); PQclear(result); snprintf(SQL,MAXSQL,"UPDATE %s SET realparent = getItemParent(uploadtree_pk) WHERE upload_fk = '%s'",uploadtree_tablename, Upload_Pk); result = PQexec(pgConn, SQL); /* UPDATE uploadtree */ if (fo_checkPQcommand(pgConn, result, SQL, __FILE__ ,__LINE__)) SafeExit(114); PQclear(result); } if (ars_pk) fo_WriteARS(pgConn, ars_pk, atoi(Upload_Pk), agent_pk, AgentARSName, 0, 1); } if (ListOutFile && (ListOutFile != stdout)) { fclose(ListOutFile); } if (UnlinkAll && MaxThread > 1) { /* Delete temporary files */ if (strcmp(NewDir, ".")) RemoveDir(NewDir); } SafeExit(0); return(0); // never executed but makes the compiler happy }
/** * @brief perform the analysis of a given list of files * * loops over the file_list given and performs the analysis of all the files, * when finished simply check if the results should be entered into the * database, this is indicated by the second element of the pair not being NULL * * @param pgConn the connection to the database * @param copy the copyright instance to use to perform the analysis * @param current_file the file and the pfile_pk that is currently being analyzed * @param agent_pk the primary key for this agent, use to enter info into the database * @param mout a logging file to used for debugging * @param type report_type binary number xyz, 0x1(bit z): statement, 0x2(bit y): url, 0x4(bit x): email */ void perform_analysis(PGconn* pgConn, copyright copy, pair current_file, long agent_pk, FILE* mout, int report_type) { /* locals */ char sql[2048]; // buffer to hold the sql commands char buf[2048]; // buffer to hold string that have been escaped for sql char hash[256]; // holds the hash of the copyright string for entry into database char* file_name; // holds the name of the file to open copyright_iterator finds; // an iterator to access the copyrights FILE* input_fp; // the file that will be analyzed /* initialize memory */ memset(sql, 0, sizeof(sql)); file_name = NULL; finds = NULL; input_fp = NULL; /* find the correct path to the file */ if(*(int*)pair_second(current_file) >= 0) { file_name = fo_RepMkPath("files", (char*)pair_first(current_file)); } else { file_name = (char*)pair_first(current_file); } /* attempt to open the file */ input_fp = fopen(file_name, "rb"); if(!input_fp) { fprintf(cerr, "ERROR %s.%d Failure to open file %s\n", __FILE__, __LINE__, file_name); fprintf(cerr, "ERROR %s\n", strerror(errno)); fflush(cerr); copyright_clear(copy); return; } /* only free temp if running as an agent */ if(*(int*)pair_second(current_file) >= 0) { free(file_name); } /* perform the actual analysis */ copyright_analyze(copy, input_fp, report_type); /* if running command line, print file name */ if(*(int*)pair_second(current_file) < 0) { fprintf(cout, "%s ::\n", (char*)pair_first(current_file)); } /* loop across the found copyrights */ if(copyright_size(copy) > 0) { for(finds = copyright_begin(copy); finds != copyright_end(copy); finds++) { copy_entry entry = *finds; if(verbose) { fprintf(mout, "=== %s ==============================================\n", (char*)pair_first(current_file)); fprintf(mout, "DICT: %s\nNAME: %s\nTEXT[%s]\n", copy_entry_dict(entry), copy_entry_name(entry), copy_entry_text(entry)); } if(*(int*)pair_second(current_file) >= 0) { /* ensure legal sql */ escape_string(pgConn, buf, copy_entry_text(entry), sizeof(buf)); /* get the hash for the string */ sprintf(hash, "0x%lx", hash_string(copy_entry_text(entry))); /* place the copyright in the table */ memset(sql, '\0', sizeof(sql)); snprintf(sql, sizeof(sql), "%ld\t%d\t%d\t%d\t%s\t%s\t%s\n", agent_pk, *(int*)pair_second(current_file), copy_entry_start(entry), copy_entry_end(entry), buf, hash, copy_entry_type(entry)); fo_sqlCopyAdd(sqlcpy, sql); } else { fprintf(cout, "\t[%d:%d:%s] '%s'", copy_entry_start(entry), copy_entry_end(entry), copy_entry_type(entry), copy_entry_text(entry)); if(copy_entry_text(entry)[strlen(copy_entry_text(entry)) - 1] != '\n') { fprintf(cout, "\n"); } } } } fclose(input_fp); fo_scheduler_heart(1); }