# for maintainability's sake:
#   keep this filename in one place
our $database_file = 'dbfile.pdb';

# self-explanatory name
sub add_person {
    my ($name,$surname,$phone) = @_;

    # a fully qualified call to open
    open my $dbfile, '+<' , $database_file
         or die "Error opening '$database_file': $!";

    flock $dbfile,LOCK_SH #share this file for now
         or die "Error flocking '$database_file': $!";

    # $dbfile is what was DBFILE before, but lexikals are
    #   less dangerous, because the handle is not global
    #   and you needn't locale()ize the *GLOB

    # using a while loop here
    #  note explicit localization of $_!

    # you had two loops around the handle, which was
    #  nonsense. This here says: while you can read data 
    #  from $dbfile

    # a flag indicating a duplicate was found;
    my $dupe = 0;

    while(defined(local $_ = <$dbfile> )){

        # source of errors: don't forget to chomp!
        chomp;

        # use usefull variable names
        my ($check_name,$check_surname) = split /,/;

        if( $name    eq $check_name and
            $surname eq $check_surname
           ){ # huston, we have a duplicate
           ++ $dupe;
           # so we can stope here
           last;
        }
    } # that's it

    if( $dupe ){
        close $dbfile;
        return; #false, this means: the name was already in
                # the database
    }
    else {
        # now we'll write to the file, so let's lock it
        flock $dbfile,LOCK_EX
              or die "Couldn't get excl. lock ... : $!";

        # go to the end of the file
        seek $dbfile , 0 , 2;

        # and add the new dataset
        print $dbfile "$name,$surname,$phone\n";

        close $dbfile
              or die "Write error in '$database_file': $!";

        return 1; #success
    }
}