As I said, you perhaps could try the latest "dev snapshots", for example
minix3_2_1_ide_20131118_f24634c.iso.bz2 to see if it helps.
I do not remember all the details, but I am pretty sure the area of
powering down the system was revised, several times, including after
official 3.2.1 release. And if anything is to change in this area, it
will change away from the current state of things.
> To be sure about the different behaviour of shutdown and poweroff, I just
> double checked and the behaviour is still the same: I call shutdown and the
> "self installed" machine freezes after "Minix will now be shut down"
> console output,
As the online help and the man page explains, raw (no parameters)
shutdown(8) performs the "HALT" action (-h), which in 3.2.1 stops the
operating system, prints a message, and freezes (halts) the machine.
Pretty much what you described.
It is true that before 3.2.x, the course of action was a bit different,
since control was driven back to the 'boot monitor'; this is no longer
possible since the current boot manager is not any more available after
the system is boot strapped (same behaviour as with any other systems
these days.)
> I call poweroff, the machine is powered down.
Since poweroff(8) calls shutdown(8) with the -p (POWEROFF) argument, it
also appears to behaves as expected to me.
I should add that I believe this (powering down) is the area which had
the major changes since 3.2.1; if plain 3.2.1 works OK for you for
`shutdown -p`, I'll consider it a good thing rather than a problem.
> As my aim is to learn about the system, I am asking to get some hints about
> where to look for the error or how to proceed.
Can you please explain what you are considering an *error*?
Documentation? Or lack of it? Mismatching documentation? Missing
directions on the wiki page?
> As this thread was already existing, I guess I am not the only one
> encountering this problem.
As I said, this area experimented changes; furthermore, you took a
thread from 2012, and one might guess that "the problem" might have been
resolved since then; particularly since in that case, the poster to
which you responded, "hijacked" in November a thread from March (subject
was "NetBSD Boot Monitor", basically about AHCI), and he was asked in
November to provide details, which he did not; as a result, nobody can
know if it still was a problem after pikpik's answer on November 18th, 2012.
Antoine