Don't change global settings which affect the overall process or the floating point control word of the CPU on preemptive systems

[MBlesken]

Just like the specification the IG should make a remark such as "To prepare the FMU code to run reliably on preemptive systems FMI functions must not change global settings." in the https://modelica.github.io/fmi-guides/main/fmi-guide/#_global_state section

[pmai]

That is already what that section says, and that is true regardless of preemption...

[andreas-junghanns]

@MBlesken : can you close this issue?

[MBlesken]

@MBlesken : can you close this issue?

@andreas-junghanns Hm, I don't think that @pmai and I have the same opinion on that section (> That is already what that section says, and that is true regardless of preemption...) "The functions the FMU provides must not change global settings which affect the overall process and/or thread environment" makes sense to me but "FMU funtions can, on the other hand, change the floating point control word of the CPU during their execution,..." seems to me like an exception that should have a remark on preemption. What do you think?

[pmai]

Not really, because any sane preemptive scheduling system will save CPU/FPU state in context switches (i.e. this is by definition per-thread/task info). I.e. can you point to a preemptive scheduling system that screws this up nowadays? And if there are, they are going to be restricted in many other ways too (i.e. no MMU, no FPU, ...), so specific care would be needed to target them anyways.