> [Gerard Roudier <groudier@club-internet.fr>]
> >
> > The right thing to do is:
> >
> > 1) Do all system intialisations and driver initialisations with interrupts
> > disabled.
> > 2) For all hardware that can be generically detected (or guessed) and that
> > no driver claims handling, warn user about.
> > 3) Enable interrupts now.
> >
> > As a result, loading a driver for hardware after system initialisation is
> > a great idiocy, in my opinion. No need to say that I donnot use modules
> > for drivers that deal with the hardware and that I will _never_ do so with
> > Free O/Ses that allow to use linked driver modules. "ld" does the
> > appropriate loading job for me when I am concerned with drivers that deal
> > with the hardware.
>
> All well and good until you have to deal with PCMCIA, hot-swap PCI,
> and other dynamic busses. You don't have the luxury of having a single
> time where you can enumerate everything -- devices could be added later
> that are not present at startup, etc.
I dislike surgery and especially without anesthesia, and frankly, I didn't
even have had a single look at the specs of this kind of stuff. May-be,
the protocols are not too unsafe. I will read hot-swap PCI as soon as I
will have time for (I just downloaded all the latest PCI specs for free
some days ago, thanks to the stupidity of some system administrator that
made readable the directory that hosted the spec. documents :)
> I used to be very skeptical about modules for this sort of thing, but
> after working with an OS that only loads drivers dynamically, I can say
> there are some nice things about it.
A thing designed to fail does not seem nice to me.
> Handling strange conditions is also nice for driver developers -- you
> can get a bit more debugging time if the system is graceful under odd
> conditions.
Having to deal with nasty situations due to useless complexity does not
seem nice to me.
Dynamic loading of kernel modules is for sure a facility for packaging and
distributing commercial O/Ses, drivers and other kinds of kernel objects.
Technically, it is just unsafe bloat and useless complexity, in my
opinion.
Gérard.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/