Fix heap buffer overflow in nc_def_var of openfast-cpp

[marchdf]

Feature or improvement description Fixes a heap-buffer-overflow error. nc_def_var size should take the value of the size of the dim array. The size of ptRefDataDims is 1 so the nc_def_var should take a 1. Feels like all of these nc_def_var should really be using *Dims.size() But that would lead to a larger diff.

Diagnosis

(lldb) bt
* thread #1, name = 'naluX', stop reason = Heap buffer overflow
  * frame #0: 0x00005555563e2dc0 naluX`::AsanDie() at asan_rtl.cpp:44
    frame #1: 0x00005555563fc79c naluX`__sanitizer::Die() at sanitizer_termination.cpp:55:7
    frame #2: 0x00005555563dda6f naluX`::~ScopedInErrorReport() at asan_report.cpp:192:7
    frame #3: 0x00005555563e0abd naluX`::ReportGenericError() at asan_report.cpp:497:1
    frame #4: 0x00005555563d7df2 naluX`::___interceptor_memcpy() at sanitizer_common_interceptors_memintrinsics.inc:115:5
    frame #5: 0x00007ffff7c61903 libnetcdf.so.19`new_NC_var + 259
    frame #6: 0x00007ffff7c61746 libnetcdf.so.19`NC3_def_var + 214
    frame #7: 0x00007ffff7c12657 libnetcdf.so.19`nc_def_var + 71
    frame #8: 0x00007ffff4bdf089 libopenfastcpplib.so`fast::OpenFAST::prepareOutputFile(this=0x000052000000d650, iTurbLoc=0) at OpenFAST.cpp:455:16
    frame #9: 0x00007ffff4beca22 libopenfastcpplib.so`fast::OpenFAST::solution0(this=0x000052000000d650, writeFiles=true) at OpenFAST.cpp:954:17
    frame #10: 0x0000555557cde9e6 naluX`sierra::nalu::ActuatorBulkFAST::interpolate_velocities_to_fast()::$_0::operator()(this=0x00007fffe796dea0) const at ActuatorBulkFAST.C:271:44
    frame #11: 0x0000555557cde995 naluX`void std::__invoke_impl<void, sierra::nalu::ActuatorBulkFAST::interpolate_velocities_to_fast()::$_0&>((null)=__invoke_other @ 0x00007fffffff237f, __f=0x00007fffe796dea0) at invoke.h:61:14
    frame #12: 0x0000555557cde945 naluX`std::enable_if<is_invocable_r_v<void, sierra::nalu::ActuatorBulkFAST::interpolate_velocities_to_fast()::$_0&>, void>::type std::__invoke_r<void, sierra::nalu::ActuatorBulkFAST::interpolate_velocities_to_fast()::$_0&>(__fn=0x00007fffe796dea0) at invoke.h:111:2
    frame #13: 0x0000555557cde80d naluX`std::_Function_handler<void (), sierra::nalu::ActuatorBulkFAST::interpolate_velocities_to_fast()::$_0>::_M_invoke(__functor=0x00007fffe796dea0) at std_function.h:290:9
    frame #14: 0x0000555557ce95a3 naluX`std::function<void ()>::operator()(this=0x00007fffe796dea0) const at std_function.h:591:9
    frame #15: 0x0000555557ce0c9b naluX`sierra::nalu::squash_fast_output(func=function<void ()> @ 0x00007fffe796dea0) at ActuatorBulkFAST.h:82:3
    frame #16: 0x0000555557cdbe20 naluX`sierra::nalu::ActuatorBulkFAST::interpolate_velocities_to_fast(this=0x000052000000d080) at ActuatorBulkFAST.C:271:7
    frame #17: 0x0000555557d1788e naluX`sierra::nalu::ActuatorLineFastNGP::operator()(this=0x00005070000833b0) at ActuatorExecutorsFASTNgp.C:45:12
    frame #18: 0x0000555557bf9f6a naluX`sierra::nalu::ActuatorModel::execute(this=0x00005070000608f0, timer=0x000051c000003c90) at ActuatorModel.C:169:13
    frame #19: 0x0000555557bf6347 naluX`sierra::nalu::AeroContainer::execute(this=0x00005070000608f0, actTimer=0x000051c000003c90) at AeroContainer.C:118:20
    frame #20: 0x0000555556c02794 naluX`sierra::nalu::Realm::advance_time_step(this=0x000051c000003880) at Realm.C:1955:18
    frame #21: 0x000055555644d280 naluX`sierra::nalu::TimeIntegrator::integrate_realm(this=0x000050f000017bf0) at TimeIntegrator.C:389:16
    frame #22: 0x000055555644596f naluX`sierra::nalu::Simulation::run(this=0x00007fffe7f00fa0) at Simulation.C:216:20
    frame #23: 0x000055555641ae23 naluX`main(argc=5, argv=0x00007fffffff3688) at nalu.C:198:9
    frame #24: 0x00007fffeca2e7e5 libc.so.6`__libc_start_main + 229
    frame #25: 0x000055555633c92e naluX`_start + 46
(lldb) frame s 8
frame #8: 0x00007ffff4bdf089 libopenfastcpplib.so`fast::OpenFAST::prepareOutputFile(this=0x000052000000d650, iTurbLoc=0) at OpenFAST.cpp:455:16
   452          ncOutVarIDs_["bld_ld"] = tmpVarID;
   453          ierr = nc_def_var(ncid, "bld_ld_loc", NC_DOUBLE, 4, bldDataDims.data(), &tmpVarID);
   454          ncOutVarIDs_["bld_ld_loc"] = tmpVarID;
-> 455          ierr = nc_def_var(ncid, "hub_ref_pos", NC_DOUBLE, 2, ptRefDataDims.data(), &tmpVarID);
   456          ncOutVarIDs_["hub_ref_pos"] = tmpVarID;
   457          ierr = nc_def_var(ncid, "hub_disp", NC_DOUBLE, 2, ptDataDims.data(), &tmpVarID);
   458          ncOutVarIDs_["hub_disp"] = tmpVarID;

Related issue, if one exists

Impacted areas of the software

openfast-cpp

[marchdf]

It should be noted that the turbineData[iTurbLoc].sType == EXTLOADS pathway had the right sizes for those vars. So it was only failing for turbineData[iTurbLoc].sType == EXTINFLOW path.

[andrew-platt]

At first glance, there appears to be some inconsistencies in this section of code. I would excpect hub_disp (line 457), hub_vel (line 459), and hubrotvel (line 461) will also need this change. Similarly for the nacelle section below.

[marchdf]

I am not sure why the test is failing...

[andrew-platt]

There is a segmentation fault with the 5MW_Land_DLL_WTurb_cpp case.

[marchdf]

My simulation ran to completion so I guess this fixes all the remaining issues with this pathway (at least for this configuration). Thanks for you help. This can be merged whenever you want from my standpoint.