Re: devfs compatibility

Richard Gooch (Richard.Gooch@atnf.CSIRO.AU)
Tue, 4 Aug 1998 19:17:34 +1000

Albert D. Cahalan writes:
> Richard Gooch writes:
> > Albert D. Cahalan writes:
>
> >> I know you will hate this, but better now than later...
> >>
> >> UnixWare 7 has 5-part device names with a c0b0t0d0p0 format.
> >> It is just like the new Linux ones, except for the fourth letter.
> >>
> >> Values of p are:
> >>
> >> 0 whole disk
> >> 1 first PC partition
> >> 2 second PC partition
> >> 3 third PC partition
> >> 4 fourth PC partition
> >>
> >> Although SCO stops there, notice how much their system is like the
> >> new Linux system. It would be great to eliminate the inconsistency
> >> while we still can.
> >
> > We went through this at the beginning of the year. IIRC the format
> > you suggest was considered, but rejected for some reasons I don't
> > fully recall just now.
>
> At the beginning of the year, there was no example with a 5-part name.
> We had Solaris and something else, both with 4-part names. Since we
> needed the 5-part name anyway, we didn't worry about compatibility.

But the reasons I gave in my last message still hold: I think 'u' is
more logical than 'd'.

> > Also, there are now too many people using it for these kinds of
> > changes to be "painless". SCO will just have to change their
> > system: we have the bigger market share ;-)
>
> SCO is already shipping. How about a transitional config option?
> It would let existing devfs users convert at their leisure,
> and would be removed before the 2.4 kernel is released.

I'm really reluctant to change this to something that is less
logical. Even a transitional change with a backwards compatibility
config option is a PITA, in my view. The reasons for change have to be
pretty strong. Frankly, being compatible with SCO doesn't seem like a
good enough reason.
If there is widespread support for changing it, and some of the kernel
gurus support the change (with no dissenters), I'll yield.

Regards,

Richard....

-
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.altern.org/andrebalsa/doc/lkml-faq.html