> From c03bfb5dc571ee2cfe99c75edcda889160f255a1 Mon Sep 17 00:00:00 2001 > From: Jiebing Li <[email protected]> > Date: Fri, 22 Oct 2010 15:54:50 +0800 > Subject: [PATCH] usb gadget: add support for medfield in composite > gadget. > > This patch for composite gadget adds support for medfield. In > medfield, eem/ecm gadget function will not support.
> It would be far better to pass a struct and to have fields which have > generic information. The struct also wants to be in a header file > where others can use it. > So instead of passing magic moorestown/medfield cookies to the nokia > gadget, pass 'max endpoints supported'. That way the feature is > generically useful. Thanks for the indication. I agree that passing this magic info is not a reasonable solution. However, even passing'max endpoints supported' is not generic enough neither, as function binding API like XXX_bind_config() will not return number of endpoints that is used by each function, and the left available ep number has to be computed in gadget driver. I thought about another way to solve this. That is, if current binding API returns -ENODEV, it means there's no more endpoints/interaces to use, thus the whole binding operations will exit with return 0 to ensure the already bound functions work well. Could you take a look at the modified patch again? Thanks a lot! _______________________________________________ Meego-kernel mailing list [email protected] http://lists.meego.com/listinfo/meego-kernel
