At 09:01 31/07/2003 -0700, you wrote:
Hm, any reason you took this off of the list?

It should have been on there, my mistake



On Thu, Jul 31, 2003 at 09:43:35AM +0200, Koen Van Renterghem wrote:
> At 14:16 30/07/2003 -0700, you wrote:
> >On Tue, Jul 29, 2003 at 04:03:15PM +0200, Koen Van Renterghem wrote:
> >> Hello,
> >>
> >> I noticed that usb_register_dev() has changed in recent kernels. The
> >'old'
> >> function (2.5.69) could allocate a continious chunk of 16 minors, but the
> >> new version only allocates a single minor. I would like to have several
> >> minors available to my driver, how can this be implemented?
> >
> >You still have many minors available to your driver, just ask for them
> >when your device is connected.
>
> Could you elaborate a bit more on how this should be handled? In the probe
> function I used to allocate 16 minors with usb_register_dev(). I want
> to control a single piece of hardware through that minor range. How do
> I ask for additional minors after calling usb_register_dev() in the
> probe function?


You want more than one minor per USB device?  During 2.5,
usb_register_dev() could handle that, but no one ever used it that way,
so it was changed to only dole out 1 minor at a time per device.

It's a good thing to keep the code as clean as possible, but for me this
limits the possibilities. I can imagine that other developers will run into
similar trouble in the future. Maybe it should be reconsidered to allow
allocation of a block of minors, I don't think that it adds a lot of complexity
to the code.
I also consider it a good thing to keep the API as versatile as possible,
but thats just one persons opinion of course.


If you
want your driver to go into the main kernel tree, I can reconsider
changing the function back.

Since I'am working on something that only exists in the lab as a first prototype there
is no point in including it in the main kernel tree right now I guess.



> > What driver are you writing, and have
> >you reserved a minor number range for it?
>
> No I don't, but this isn't a real problem since I'am working on custom
> hardware based on the Cypress FX2. Everything is currently in the
> research phase.

Any reason why you have to write a kernel driver and can't do everything
from libusb/usbfs?

I'am working on something that can log data very fast. With 2.5.59 I was able to transfer 40Mb/s sustained
from the cypress FX2, over usb, to harddisk. Using libusb or usbfs would introduce unwanted latency,
and I really need to eliminate all possible overhead.



thanks,

greg k-h

Thanks for your reply, I really appreciate it, Koen

___________________________________________________________________________
ir. Koen Van Renterghem
Ghent University
Department of Information technology / INTEC-design
Sint-Pietersnieuwstraat 41
B-9000 GENT
Belgium, Europe
___________________________________________________________________________



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to