The two 'usbd' implementations I provided are both on my
jusb.sourceforge.net page (click the 'usbd' link at left).
Then just download the prototype, install in /etc/usb,
and have a go at it.
- the Java one is pretty fully functional, and I expect
it's a good learning excercise for someone to make it
into a C daemon that leverages Johannes' patches to
expose all USB descriptors and the driver binding;
- the usbdevfs one uses a C module that should be easily
turned into a C version, but needs a bit more smarts
as should be described in the comments. Since it needed
the usbdevfs/devfs merge patch that's not going into 2.4,
its main role is to be such a model.
As noted, they both use the same shell scripts to drive things,
and I'm seeing those as more critical than those implementations.
For some reason, not everyone uses Java as much as me ... ;-)
Meanwhile, the Java daemon has seemed to work just fine when
driving those scripts to load (and to some extent configure)
the modules, and it's painless to get a JDK 1.2 package to
drive everything for such debugging.
Current status: they will "modprobe" all the drivers I can
test, and should work for some of the others. There's a CPIA
issue, at lest in my config, in that it doesn't show up in the
/proc/bus/usb/drivers file after loading.
If the driver is already loaded, it won't be reloaded, and if
the modprobe fails, things stop there.
The trickier issues seem to revolve around HID support. It's
easy to autoload the HID driver, but then something smart needs
to be provided to figure out whether to load keyboard, mouse,
joystick, and so on.
What I'd really like to see happening is: someone takes those
shell scripts, checks them into the linux-usb sourceforge CVS,
and they start to evolve from there both (a) to get all the
other kernel drivers loaded, and (b) to handle any other admin
issues that crop up, in particular stuff like configuring network
adapters.
The goal IMHO should be to have a standard set of shell scripts
that most distributions can use as-is. The stumbling block I
know of will be, sadly, network config -- but PCMCIA is clearly
the model to follow there.
- Dave
p.s. I should be back to regular network access next week.
----- Original Message -----
From: Dunlap, Randy <[EMAIL PROTECTED]>
To: 'Miles Lane' <[EMAIL PROTECTED]>; David Brownell <[EMAIL PROTECTED]>
Cc: Johannes Erdfelt <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Friday, May 26, 2000 8:32 PM
Subject: RE: [linux-usb] [exp. patch] userspace driver binding (and
followingpatches)
> Would someone provide me and the list with the
> proposed/alternative/out-there USB driver autoloading/binding
> solutions, where they are (URLs), status, etc.?
>
> I know that some are listed on www.linux-usb.org under
> the heading "USB /proc & /dev tools". Are there some
> others to consider?
>
> Thanks,
> ~Randy
>
> > -----Original Message-----
> > From: Miles Lane [mailto:[EMAIL PROTECTED]]
> > Sent: Friday, May 26, 2000 12:23 PM
> > To: David Brownell
> > Cc: Johannes Erdfelt; Dunlap, Randy; [EMAIL PROTECTED]
> > Subject: Re: [linux-usb] [exp. patch] userspace driver binding (and
> > follow ingpatches)
> >
> >
> > David,
> >
> > I will provide feedback if I can. You may need to help
> > me get things in a testable state. I'm not a developer,
> > but maybe that can be a good thing. I'm probably more
> > like your target audience.
> >
> > I haven't provided feedback up 'til now because I have been
> > focused on ORB drive testing. Plus, I thought you were mostly
> > looking for technical feedback from the rest of the USB
> > development team.
> >
> > Let me know what I can do to help. If you can provide two
> > sets of patches for your alternate implementations that I
> > can download from somewhere (or e-mail them to me), that
> > would help.
> >
> > I'll be in Washington D.C. for the National Spelling Bee this
> > next week (my wife is on the Word List Committee) and since
> > I am on childcare duty part of the time, my availability
> > will be sporadic until the first Friday in June.
> >
> > BTW: Speaking as a user, I am greatly anticipating seeing
> > your work come to fruition! I think it's really essential
> > stuff.
> >
> > Thanks,
> > Miles
> >
> >
> > On Fri, 26 May 2000, David Brownell wrote:
> >
> > > Date: Fri, 26 May 2000 11:47:17 +0100
> > > From: David Brownell <[EMAIL PROTECTED]>
> > > To: Johannes Erdfelt <[EMAIL PROTECTED]>,
> > > "Dunlap, Randy" <[EMAIL PROTECTED]>
> > > Cc: [EMAIL PROTECTED]
> > > Subject: Re: [linux-usb] [exp. patch] userspace driver binding (and
> > > follow ingpatches)
> > >
> > > > > Johannes's devfs patches for USB have been especially difficult
> > > > > for me to decide and no doubt I have taken a long time on it.
> > > > > I really couldn't decide.
> > > > >
> > > > > Johannes and David Brownell had some good discussion
> > and some good
> > > > > progress with JE's patch, and my thanks to David Brownell for
> > > > > experimenting with this patch.
> > > >
> > > > David was the only person to give me feedback. Thanks
> > David. I appreciate
> > > > it.
> > >
> > > For the record, despite all the discussion about wanting a
> > USB Daemon
> > > to autoload drivers, I've gotten no feedback about the two different
> > > implementations I provided ... or, more significantly, the
> > shell scripts
> > > that they share to handle policies associated with driver
> > loading and
> > > subsequent configuration.
> > >
> > > If folk are serious about dynamically loading USB modules,
> > I'd expect
> > > to see some more discussion on standards for such
> > administrative policies.
> > >
> > > My tentative conclusion is that not many people have gotten
> > beyond the "yay,
> > > my device works!!" stage to the "Wow, 2.4 does USB
> > Plug'n'Play !!" stage.
> > >
> > > That's a matter of time of course, but I'll encourage more
> > folk to put
> > > that higher on their priority lists ...
> > >
> > >
> > > > I'll work on splitting out the unrelated features of my
> > previous patches
> > > > and send them to Randy and the list ASAP. These are
> > unrelated to anything
> > > > with devfs. I expect this will be sent today.
> > >
> > > Good. I'd pointed out to Randy that I wanted to see two of them
> > > (the ioctl to expose driver-to-interface binding, and seeing all
> > > the device and config descriptors) in 2.4 regardless.
> > >
> > > - Dave
> > >
> > >
> > >
> > >
> > >
> > >
> > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > For additional commands, e-mail: [EMAIL PROTECTED]
> > >
> > >
> >
> >
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
