Description of problem:
When the server binary is checked for its selinux context, its path is not full, and if the binary is located in a nonstandard location, getfilecon() returns -1 and xinetd reports that changing process context failed.

In set_context_from_socket():
   exepath = SC_SERVER_ARGV( scp )[0];
   if (getfilecon(exepath, &exec_context) < 0)
     goto fail;


Version-Release number of selected component (if applicable):
xinetd-2.3.15-5


How reproducible:
always


Steps to Reproduce:
1. Have a testing server

2. Configure a xinetd service
service test_labeled_ipv4
{
        id              = test_labeled_ipv4
        type            = UNLISTED
        flags           = REUSE IPv4 LABELED
        wait            = no
        user            = root
        disable         = no
        instances       = 1
        socket_type     = stream
        protocol        = tcp
        port            = 4001
        server          = /usr/local/non-standard/path/test_labeled_ipv4
}

3. Connect to the server

4. Observe an error


Actual results:
13/6/25@15:44:01: ERROR: 8373 {exec_server} Changing process context failed for test_labeled_ipv4


Expected results:
Connection succeeds if all the selinux stuff is right.