/* file_iterate: Reliable, fast monitoring of files * Author: Dave Taht * Date: 22 Nov 2015 * Copyright (C) 2015-6 Michael David Taht * Copyright (C) 2015-6 Toke Høiland-Jørgensen * This program is free software: you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation, either version 3 of the License, or * (at your option) any later version. * * 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. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program. If not, see . */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* openwrt's ash shell does not have usleep, nor nanosecond time in the date command. for i in $(seq $count); do cat $filename date '+Time: %s.%N' echo "---" sleep $interval done So this C program (which is more accurate and lighter weight than the shell script) needs to be used on things like openwrt, when high resolution data is needed. */ #define BUFFERSIZE (1024*1024) #define NSEC_PER_SEC (1000000000.0) struct arg { int count; struct timespec interval; double finterval; char *filename; char *command; char buffer; }; typedef struct arg args; static const struct option long_options[] = { { "count" , required_argument , NULL , 'c' } , { "interval" , required_argument , NULL , 'I' } , { "help" , no_argument , NULL , 'h' } , { "buffer" , no_argument , NULL , 'b' } , }; void usage (char *err) { if(err) fprintf(stderr,"%s\n",err); printf("file_iterate [options]\n"); printf( "\t-h --help \n" "\t-b --buffer \n" "\t-c --count [number of iterations]\n" "\t-I --interval [fractional number of seconds]\n" "\t-f --filename [filename]\n"); exit(-1); } static void defaults(args *a) { a->filename = NULL; a->finterval=.2; a->count=10; a->interval.tv_nsec = 0; a->interval.tv_sec = 0; a->buffer = 0; } #define QSTRING "c:I:f:hb" int process_options(int argc, char **argv, args *o) { int option_index = 0; int opt = 0; optind = 1; while(1) { opt = getopt_long(argc, argv, QSTRING, long_options, &option_index); if(opt == -1) break; switch (opt) { case 'c': o->count = strtoul(optarg,NULL,10); break; case 'I': o->finterval = strtod(optarg,NULL); break; case 'f': o->filename = optarg; break; case 'b': o->buffer = 1; break; case '?': case 'h': usage(NULL); break; default: usage(NULL); } } o->interval.tv_sec = floor(o->finterval); o->interval.tv_nsec = (long long) ((o->finterval - o->interval.tv_sec) * NSEC_PER_SEC); return 0; } static int read_once(char *filename, char *buffer, int bufsize) { int fp = open(filename, O_RDONLY); ssize_t len = 0, r = 0; if(!fp) { perror("Unable to open file"); return -1; } while((r = read(fp, buffer, bufsize-len)) > 0) len += r; return len; } static void result(int out, int size, int bufsize, char *buffer) { struct timespec cur_time; if(bufsize - size > 40) { clock_gettime(CLOCK_REALTIME, &cur_time); int added = sprintf(&buffer[size],"Time: %ld.%09ld\n---\n", cur_time.tv_sec,cur_time.tv_nsec); write(out,buffer,size+added); } else { write(2,"Buffer Overrun\n",sizeof("Buffer Overrun\n")); } } // Since this is linux only we can use timerfd for an isochronous clock int run(args *a) { char tmpfile[] = "/tmp/file_iterateXXXXXX"; int out = a->buffer ? mkstemp(tmpfile) : STDOUT_FILENO; if(a->buffer && !out) { perror("Unable to create tmpfile"); exit(-1); } if (!a->filename) usage("Must specify filename"); struct itimerspec new_value = {0}; int timer = timerfd_create(CLOCK_REALTIME, 0); //new_value.it_interval.tv_nsec = 100000000; // 10ms interval //new_value.it_interval.tv_sec = 0; new_value.it_interval = a->interval; new_value.it_value = a->interval; /* better method would be to poll for input (since writes from the tool could block or return no output for some reason), timestamp the input, and if the difference is less than half, skip this round. this would absorb non-completing stuff */ char buffer[BUFFERSIZE]; if (read_once(a->filename, buffer, sizeof(buffer)) <= 0) { perror("Unable to read file"); exit(-1); } int size = 0; int ctr = 0; timerfd_settime(timer,0,&new_value,NULL); // relative timer do { long long fired; if(read(timer,&fired,sizeof(fired))!=8) perror("reading timer"); ctr+=fired; if((size = read_once(a->filename,buffer,sizeof(buffer))) > 0) { result(out,size,BUFFERSIZE,buffer); } else { result(out,0,BUFFERSIZE,buffer); perror("reading file"); } } while (ctr < a->count); if(a->buffer) { lseek(out, 0, SEEK_SET); while((size = read(out, buffer, sizeof(buffer))) > 0) write(STDOUT_FILENO,buffer,size); unlink(tmpfile); } close(out); wait(NULL); return 0; } int main(int argc,char **argv) { args a; int status = 0; defaults(&a); process_options(argc, argv, &a); run(&a); return status; }