Adding volatile for memory mapped IO, adding function for kernel space timespec retrieval and adding test for the gpstime module

Fixing permissions on scripts that I changed

See merge request cds/advligorts!299

Cleaning up findSharedMemory usage, fixing error checks on the returned...

See merge request cds/advligorts!297

Remove shared files causing collisions when generating the code for two models at the same time. Also speed up generation time by 10s per model.

See merge request cds/advligorts!296

Reduce foot print of cycle1 -> cycle2 dma size change for fast adcs.

See merge request cds/advligorts!298

Old code for changing DMA size from the ( possibly smaller ) first cycle
was too expensive in some sense, causing momentary channel hopping on the
ADC when run at 65k and time_shift=0.

The new code just changes the size value of an already existing DMA
module instead of totally recreating it.

Cleaning up findSharedMemory usage, fixing error checks on the returned pointer, consolidating code for findSharedMemory() and findSharedMemorySize() so we don't have duplicate, removing unused sprintf in both functions

The feCodeGen.pl script was using ./diags.txt and ./diags2.txt to build the <model>_partConnectionList.txt for each model. This meant that multiple models could not be built at the same time bacause they were sharing the diags files. This commit replaces those file with ones specific to the model being built. We also clean up the code a bit and remove some sleeps that slow down builing the models by 10 s each.

Shared memory cleanup 1

See merge request cds/advligorts!295

 * typeo in one of the commits left it as the invalid 'lingo_ioremap.h' instead of 'ioremap_selection.h'

Remove all references to ioremap_nocache and replace with the LIGO_IOREMAP macro.

See merge request cds/advligorts!292

 * ligo_iomap.h becomes iomap_selection.h to denote what it actually does.
 * LIGO_IOREMAP becomes simply IOREMAP to note that it isn't LIGO specific.

 * This allows the code to call the appropriate version if ioremap/ioremap_nocache depending on the kernel version.

Adding the use of the rts-cpu-isolator.h header, and calling the...

Closes #337

See merge request cds/advligorts!291

Fixing set_fe_code_idle( func ), func prototype needs to match the expected void (*)( void ) and NOT void * (*) (void*)

Adapt the rcg build process to use rts-cpu-isolator symvers file (if present)

See merge request cds/advligorts!290

Adding the use of the rts-cpu-isolator.h header, and calling the swappable/better named functions rts_isolator_exec()/rts_isolator_cleanup() instead of cpu up/down directly

Adding functions with more general real-time/isolation language because this functionality extends to more general core isolation use-cases.

Closes #334

See merge request cds/advligorts!289

 * work towards #335
 * modify find_module_symvers to user argparser and have a quiet option to make missing symvers files empty strings
 * update the src/epics/util/Makefile.in to pass the location of the rts cpu isolator symver file to feCodeGen.pl
 * update createKernelMakefile.pm to add the rts cpu isolator symvers file to the KBUILD_EXTRA_SYMBOLS for the module

Adding funtions with more general real-time/isolation language because this functionality extends to more general core isolation usecases.

 * rename from ligo_rt_* to rts_cpu_isolator_*

Close issue #333 rename ligo_rt to rts-cpu-isolator

Closes #333

See merge request cds/advligorts!288

Fix for 332, standalone_edcu does not reflect all of its created values into EPICS.

Closes #332

See merge request cds/advligorts!287

Six channels are created by the EDCU.  Only three were sent to both EPICS and the data stream.
 * Update the standalone_edcu to export the UPTIME, GPS, and DATA_RATE_KB_PER_S values in EPICS as well.
 * Minor formatting changes as per clang-format run.

Closes #332