Comments fixes after reading the thread creation and suspension paths.

[kib]

Summary

KSE terminology in the comments for cpu_set_upcall_kse() confuses much more than explains.

Remove two XXX in kern_thread.c. One is probably rotten as is, another one required better formulation.

Remove code duplication in weed_inhib().

The question: does it make sense to rename cpu_set_upcall_kse() ? I stopped there. Good name could be e.g. cpu_set_userret().

[jhb]

Let's please rename the function, but I think cpu_set_upcall() is a better name for it. :(

For fork() we have cpu_fork() and cpu_set_fork_handler(). Perhaps cpu_set_upcall() should be something like 'cpu_thread_create()' and 'cpu_set_upcall_kse()' should be 'cpu_user_upcall()' (named similar to cpu_user_tls()).

The only thing that is somewhat unfortunate is that kthread_add() uses cpu_set_fork_handler(). If we called it 'cpu_set_kthread_handler()' perhaps that would be less bad.

sys/i386/i386/vm_machdep.c

465

This comment has some typos and confusing language as well. :-/

sys/kern/kern_thread.c

673

Perhaps commit this separately?

[kib]

sys/kern/kern_thread.c

673

Sure.

[kib]

Rename:
cpu_set_upcall()->cpu_copy_thread()
cpu_set_upcall_kse()->cpu_set_upcall()
cpu_set_fork_handler()->cpu_fork_krthread_handler()

Update comments.

If you have other preferences, I am willing to change this.