Re: My $0.02 on devd and devfs

Horst von Brand (vonbrand@sleipnir.valparaiso.cl)
Sun, 17 Oct 1999 12:28:09 -0300


David Lang <dlang@diginsite.com> said:
> On Thu, 14 Oct 1999, Horst von Brand wrote:
> > orc@pell.portland.or.us (david parsons) said:
> > > In article <linux.kernel.199910141142.IAA10452@pincoya.inf.utfsm.cl>,
> > > Horst von Brand <vonbrand@inf.utfsm.cl> wrote:

[...]

> > > >The described kernel will notify devd to create the nodes when they show
> > > >up, i.e., on boot or when the device is plugged in.

> > > How?

> > > Until I insmod the parallel port driver, the kernel doesn't know
> > > that there are any parallel ports around.

> > Because it doesn't check before.

> > > >Any dynamic /dev will
> > > >have to include this exact functionality somehow,

> > > Perhaps it would be a good idea if you wasted the five minutes
> > > necessary to look at Richard Gooch's code before you commented
> > > further on it.

> > Look, to find out if the foo device is present, you _have_ to check
> > somehow. No way around that. I somehow seriously doubt devfs does some
> > mindreading trick here.

> Actually, in a way it does. It can be setup so that if you attempt to
> access a parallel port device it then goes and insmods the driver. this
> allows it to not do a scan of everything to start.

It does not. It is just checking when you try to access the device. That is
exactly what happens here now.

But if you factor in hotplugging (this is what all this discussion is
supposed to be about, wasn't it?), the kernel _has_ to know at each instant
what is available and what is not (and any halfway decent PnP hardware
standard will provide notification to the kernel), if only to ask the user
(via some devd as HPA advocates, and which devfs also has) what should be
done with/to the new device.

Don't use dumb hardware (parallel port, serial ports) onto which pluggable
devices (Zip drives, CD writers, parallel IDE, ...) are kludged on for PCs
just because the interface is universally available and the box doesn't
need to be opened for doing this; with real hotplugging, PnP devices (Not
ISA ones. Shame that the concept has been rendered useless by you know
who).

-- 
Horst von Brand                             vonbrand@sleipnir.valparaiso.cl
Casilla 9G, Viņa del Mar, Chile                               +56 32 672616

- 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/