Userspace alternative to gpstime kernel module

This code can be used when models run in userspace and without GPS hardware. It removes the dependency on the special device file /dev/gpstime. GPS time is obtained in userspace from the system clock. Enable at compile time by defining USE_GPSCLOCK. When building a model, add userspacegps=1 to the CDS parameters.

Reviewing.

@christopher.wipf Thanks for the MR.

I like the idea, it would be good to allow things to run w/o custom kernel drivers when possible. There are a few things I would like changed to help with future maintenance/support and packaging.

1. We can do all of this at run time, so I don't think we need to do this via a configuration time check. This will make it easier for support as we will have less packages and variants to work with.

   a. The code can try to open /dev/gpstime, and if that fails call the gpsclock calls.

2. For readability and maintenance down the line I would like to reduce the number of ifdefs. Typically consolidate them at the front of a file, including small static inline functions/macros that are used later in the code to make the code paths read the same.

3. You found some bits of dead/unused code that we should just drop.

I took the liberty of implementing these as a MR against this MR, and it has been submitted for your review.

Thanks.

added 2 commits

• 58acce0a - Updates to the gpsclock.
• 85017637 - Merge branch 'gpsclock' into 'gpsclock'

Compare with previous version

@rolf.bork Can you sign off on the RCG side of the changes. I'm ready to merge this if it looks good to you.

Specifically:

• timing changes in controllerAppUser.c, controllerIopUser.c, there are supporting changes to feCodeGen.pl, Parameters.pm, and createUserMakefile.pm

merged

mentioned in commit 5b9c93d2

Is there a reason you store the original value in an env var here? I don't understand the point of that.

The timezone setting is saved and restored later to avoid any unintended side effects, in case programs call this function and subsequently need to do other date/time calculations of their own.

@christopher.wipf I didn't mean why is it saved and restored, I meant why is it saved to an environment variable, rather than just in the savetz variable (GPSCLOCK_TZ_ORIG)? Seems like writing it to the environment is an unnecessary step.

savetz is a pointer and I don't know if it remains valid after the environment variable TZ is reset. The value is copied to make sure it remains available, and I use another environment variable so I don't have to worry about how much space to allocate.