Fix Segfault in iperf_err(exit)

[martin-ottens]

• Version of iperf3 (or development branch, such as master or
  3.1-STABLE) to which this pull request applies: all

• Issues fixed (if any): -

• Brief description of code changes (suitable for use as a commit message): Fix Segfault in iperf_err(exit)

In some places iperf_err and iperf_errexit are called with NULL for the parameter struct iperf_test *test. In this case the pthread_mutex_lock call will Segfault. Fixed by adding a NULL-check.

Reproduce this Segfault

Example of a usage of `iperf_errexit` with `NULL` (here)[https://github.com/esnet/iperf/blob/master/src/iperf_util.c#L68]. (Reproduce with a full `/tmp` filesystem, `readentropy` will fail)

Valgrind output:

==2274== Process terminating with default action of signal 11 (SIGSEGV): dumping core
==2274==  Access not within mapped region at address 0x10
==2274==    at 0x49559E4: pthread_mutex_lock@@GLIBC_2.2.5 (pthread_mutex_lock.c:80)
==2274==    by 0x48A94B8: iperf_errexit (iperf_error.c:114)
==2274==    by 0x48A95BF: readentropy (iperf_util.c:74)
==2274==    by 0x48A279D: iperf_new_stream (iperf_api.c:4487)
==2274==    by 0x48AC633: iperf_run_server (iperf_server_api.c:786)
==2274==    by 0x1092FD: run (main.c:169)
==2274==    by 0x109068: main (main.c:124)