On Sun, 07 Mar 2004 15:23:56 -0800 David Brownell <[EMAIL PROTECTED]> wrote:
Could you export some form of "controller-type" number (1 for net2280, 3 for pxa, etc.) to simplify the bcdDevice assignment for gadget drivers that use the same scheme as gadget-zero?
I'd rather hold that off. Managing a number namespace for all developers is extra administrative overhead. I don't want to be maintaining such a list, or expect LANA to do so.
This is to avoid that ugly series of if (gadged_is_xxx) to compute bcdDevice field. Even if someone is developing top secret UDC drivers, he could find a second to send a single-line #define diff to, say, this list, to reserve some number for his UDC.
Well, if someone else feels strongly enough about this to set up such a mechanism, more power to you! Let me get an updated version of this out first.
That's also just a bit out of scope for cleanup that's focussing on _endpoint_ autoconfiguration. :)
FWIW I'd say the current policy for how the gadget drivers use the 00.00 to 99.99 is something like
01.23 || |+-- chip ID; 9 is NOT ENOUGH || +--- 0 = 2.6, 1 = 2.4 (from g_file_storage) |+----- gadget driver's version id (1, 2, ..) +------ unused/mbz
It's never been written down, so maybe nobody much noticed how broken it is... I'd expect products to override whatever scheme a "stock" kernel will come up with.
Vendor, product, and version ids should probably be gadget driver module parameters. Much like the manufacturer, product, and serial number strings ... :)
- Dave
------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
