CentOS 6.X SegFault in Control Port

[mbr0wn]

I receive a segmentation fault when I attempt to run a flow graph that exports any properties over control port on CentOS 6.X. This issue does not effect Ubuntu. I have tested GNU Radio tag v3.7.8.2 as well as the current head of master (tag# bd69c31). It crashes when the first property attempts to register with the rpc manager.

Backtrace:
(gdb) bt
#0 0x00007ffff176ca4d in rpcbasic_register_get<gr::blocks::ctrlport_probe2_f, std::vector<float, std::allocator > >::rpcbasic_register_get (this=0x165a660, block_alias="probe2_f0",

functionbase=0x164f878 "samples", function=&virtual table offset 0, min=..., max=..., def=..., units_=0x7ffff17967e4 "volts", desc_=0x164eaa8 "Sample Points", minpriv_=RPC_PRIVLVL_MIN, display_=1)
at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/gnuradio-runtime/include/gnuradio/rpcregisterhelpers.h:1042

#1 0x00007ffff176995b in gr::blocks::ctrlport_probe2_f_impl::setup_rpc (this=0x1650120) at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/gr-blocks/lib/ctrlport_probe2_f_impl.cc:134
#2 0x00007ffff129a608 in gr::hier_block2_detail::flatten_aux (this=0x16226c0, sfg=...) at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/gnuradio-runtime/lib/hier_block2_detail.cc:621
#3 0x00007ffff1293779 in gr::hier_block2::flatten (this=) at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/gnuradio-runtime/lib/hier_block2.cc:157
#4 0x00007ffff12bd1b0 in gr::top_block_impl::start (this=0x163e940, max_noutput_items=) at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/gnuradio-runtime/lib/top_block_impl.cc:109
#5 0x00007ffff12b5b3c in gr::top_block::start (this=0x15115b0, max_noutput_items=) at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/gnuradio-runtime/lib/top_block.cc:62
#6 0x00007fffece1401c in top_block_start_unlocked (r=..., max_noutput_items=10000000) at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/build/gnuradio-runtime/swig/runtime_swigPYTHON_wrap.cxx:6434
#7 0x00007fffece9cad5 in _wrap_top_block_start_unlocked (args=, kwargs=)

at /home/jrm/workspace/FOSS/pybombs/src/gnuradio/build/gnuradio-runtime/swig/runtime_swigPYTHON_wrap.cxx:49375

#8 0x000000358f6d55b6 in PyEval_EvalFrameEx () from /usr/lib64/libpython2.6.so.1.0
#9 0x000000358f6d7647 in PyEval_EvalCodeEx () from /usr/lib64/libpython2.6.so.1.0
#10 0x000000358f6d5a94 in PyEval_EvalFrameEx () from /usr/lib64/libpython2.6.so.1.0
#11 0x000000358f6d7647 in PyEval_EvalCodeEx () from /usr/lib64/libpython2.6.so.1.0
#12 0x000000358f6d5a94 in PyEval_EvalFrameEx () from /usr/lib64/libpython2.6.so.1.0
#13 0x000000358f6d7647 in PyEval_EvalCodeEx () from /usr/lib64/libpython2.6.so.1.0
#14 0x000000358f6d5a94 in PyEval_EvalFrameEx () from /usr/lib64/libpython2.6.so.1.0
#15 0x000000358f6d7647 in PyEval_EvalCodeEx () from /usr/lib64/libpython2.6.so.1.0
#16 0x000000358f6d7722 in PyEval_EvalCode () from /usr/lib64/libpython2.6.so.1.0
#17 0x000000358f6f1b9c in ?? () from /usr/lib64/libpython2.6.so.1.0
#18 0x000000358f6f1c70 in PyRun_FileExFlags () from /usr/lib64/libpython2.6.so.1.0
#19 0x000000358f6f315c in PyRun_SimpleFileExFlags () from /usr/lib64/libpython2.6.so.1.0
#20 0x000000358f6ff892 in Py_Main () from /usr/lib64/libpython2.6.so.1.0
#21 0x000000358221ed5d in __libc_start_main () from /lib64/libc.so.6
#22 0x0000000000400649 in _start ()

The line that causes the seg fault is:
rpcmanager::get()->i()->registerQueryCallback(d_id, inserter);
It appears that the "rpcmanager::get()" call does not return a handle to the Thrift rpc process b/c the process dies when it attempts to call "i()" on the object.

Thrift is installed and GNU Radio recognizes it as a valid control port component:
$ gnuradio-config-info --enabled-components
python-support;testing-support;volk;gnuradio-runtime;gr-ctrlport;* thrift;gr-blocks;gnuradio-companion;gr-fec;gr-fft;gr-filter;gr-analog;gr-digital;gr-dtv;gr-atsc;gr-audio;* alsa;* oss;gr-channels;gr-noaa;gr-pager;gr-qtgui;gr-trellis;gr-uhd;gr-utils;gr-vocoder;gr-fcd;gr-wavelet;gr-wxgui

I can run gr-ctrlport-monitor from CentOS build and attach using Thrift to another flow graph running on an Ubuntu machine.

Attached is a sample flow graph that causes the seg fault.

[marcusmueller]

CentOS6 no longer supported, I'll say it's unlikely we'll fix this. Closing.