Re: devfs again, (was RE: USB device allocation)

danielt@digi.com
Fri, 8 Oct 1999 09:30:46 -0500 (CDT)


On 8 Oct 1999, Jes Sorensen wrote:

> >>>>> "Dan" == Dan Hollis <goemon@sasami.anime.net> writes:
>
> Dan> On Thu, 7 Oct 1999 danielt@digi.com wrote:
> >> 90% of the objections to having devfs in the kernel are easily
> >> solved with "well don't use it then".
>
> Dan> Yup its "i dont want to use it, so i dont want *you* to use it
> Dan> either". Now what if this were applied to other optional kernel
> Dan> options like ipv6, isdn, serial, etc...
>
> Reading all your extremely well documented arguments on this subject
> is hilarious. Have you considered that there are techincal and design
> issues with devfs that people are complaining over?
>
Yes, the remaining 10% that I mentioned in the partially
quoted e-mail above. As I mentioned there, most of those have
one (or both) of two characteristics (I do not recall any that don't,
but my memory is not perfect):
1. They are problems of _any_ dynamic /dev
2. They have been or can be fixed in the framework
of devfs.

> I don't know how much actually changed in devfs since Ted wrote an
> excellent summary about what was wrong with it about a year ago, but I
> haven't seen anything from the devfs author claiming that he actually
> addressed those issues.
>
ISTR that many of those were type 90% issues.

NOTE: in the absence of devfs proper it will continue
to be implemented (badly and piecemeal) under /proc.
If the functionality is not desirable, why do so many
device driver writers implement their own versions?

-- 
Daniel Taylor      Senior Test Engineer     Digi International
danielt@digi.com                             Open systems win.

- 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/