'Llo All,
This patch switches to the second (and so on, and so on ...) configuration
when there is no driver available for this configuration. It is made against
the ("old") 2.4.20. I hope I don't make anyone furious about it ;) I can
make it against the 2.6 too,
For 2.6, patches have already been circulated to completely export this policy decision to userspace, through a writable syfs file.
So for example a hotplug script would be able to notice a Brand-X Model FOO, and the set the configuration to match the drivers that were installed on that system. (I think I sent them around against the 2.5.67 kernel. Recent test2 patches have changed some of the affected code, and I'll submit some of the updates.)
I don't think it's right to have the kernel make the particular policy decision you're proposing. As Oliver noted, even the way it's done is troublesome for a typical case: where the drivers can't have been modprobed yet by the time your patch concluded that there were no drivers for that configuration.
There are a lot of other updates needed, within usbcore, to make changing configurations safe; since usbcore (and the HCDs) keep lots of state around, and only the more recent 2.6 kernels synchronize it properly with config change events (including set_interface). I'd be cautious with changing 2.4 kernels too.
- Dave
------------------------------------------------------- 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
