seek_data.c

/* Reproduce incorrect lseek(fd, 0, SEEK_DATA) results on freshly written files. */

#define _GNU_SOURCE
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
#include <errno.h>

int rand_fd;
char buf[1000000];

void make_random_file(const char *filename, size_t size) {
    // it must be a new file
    unlink(filename);

    int fd = open(filename, O_CREAT|O_WRONLY, 0644);
    if (fd < 0) {
        perror("open");
        exit(1);
    }

    size_t len = read(rand_fd, buf, size);
    if (len < 0) {
        perror("read /dev/urandom");
        exit(1);
    }

    write(fd, buf, len);
    close(fd);
}

int main(int argc, char **argv) {
    int nfiles = 1000;

    if (argc != 2) {
        printf("usage: %s workdir\n", argv[0]);
        return 1;
    }

    rand_fd = open("/dev/urandom", O_RDONLY);
    if (rand_fd < 0) {
        perror("open /dev/urandom");
        exit(1);
    }

    char srcname[PATH_MAX];
    for (int i=0; i<nfiles; i++) {
        sprintf(srcname, "%s/test%05x", argv[1], i);

        int size = (i & 1) ? 10000 : 1000000;
        make_random_file(srcname, size);
    }

    close(rand_fd);

    int bad = 0;

    for (int i=0; i<nfiles; i++) {
        sprintf(srcname, "%s/test%05x", argv[1], i);

        int fd = open(srcname, O_RDONLY);
        struct stat st;
        fstat(fd, &st);

        off_t offset = lseek(fd, 0, SEEK_DATA);
        if (offset < 0) {
            if (errno == ENXIO)
                printf("%s: SEEK_DATA says no data, true size %ld\n", srcname, st.st_size);
            else
                printf("%s: unexpected error??\n", srcname);
            bad = 1;
        }
    }

    return bad;
}