> Now it seems to me currently it is not possible using > Zadig/libwdi to achieve this.
I agree. As far as I understand, Zadig/libwdi simply offers what it gets from the system - i.e. the "subdevices" generated by usbccgp.sys (one subdevice only for the IAD). I assume no inf-hacking would help here for the same reason...? (referring to your other e-mail) >> Brief comparison with libusb-win32 seems to indicate that this lib has >> no problem with such setup (usbccgp.sys on parent, libusb0.sys on IF 0), >> but of course there the enumeration process is completely different. > > Could you please share more details about this? Supposedly > libusb-win32 may have the same issue as well. I have not much to share in the moment. I did just a quick cross-check some time ago and libusb-win32 seemed to allow to claim/use the other interfaces even without touching the composite parent. But I didn't get into details since I don't use and don't intend (currently) to use libusb-win32. Just noticed that libusb-win32 is enumerating the devices through its own driver rather than through the SetupAPI as libusbx, therefore it doesn't get unaligned. >> Finally, during my earlier experiments, I tried a trivial >> "hack"-modification of libusbx, assigning the same backend (subAPI) to >> the associated interfaces as to the first one - and after this >> modification everything "just worked", therefore I was hoping there >> would be a solution. However, due to my lack of knowledge of libusbx >> internals, I'm of course not sure if such approach might have some >> unexpected side effects. > > Could you post your hack? Sure, but please understand it's really just a hack, no attempt for proper solution ;-) In windows_usb.c, function set_composite_interface(), after lines priv->usb_interface[interface_number].path = dev_interface_path; priv->usb_interface[interface_number].apib = &usb_api_backend[api]; priv->usb_interface[interface_number].sub_api = sub_api; I have added something like char *copy1, *copy2; copy1 = malloc (1024); strcpy (copy1, dev_interface_path); copy2 = (char*) malloc (1024); strcpy (copy2, dev_interface_path); priv->usb_interface[interface_number+1].path = copy1; priv->usb_interface[interface_number+1].apib = &usb_api_backend[api]; priv->usb_interface[interface_number+1].sub_api = sub_api; priv->usb_interface[interface_number+2].path = copy2; priv->usb_interface[interface_number+2].apib = &usb_api_backend[api]; priv->usb_interface[interface_number+2].sub_api = sub_api; Note that this hack does not do any IAD parsing, it's hardcoded for the single device of interest with "known" device, therefore it blindly sets the backend of IF 0 also to IF 1&2. The copies of dev_interface_path strings seemed to be necessary because the lib frees them somewhere during cleanup. I did not even try to check further details around this, assuming that the proper solution would need to be anyway designed by somebody with good knowledge of libusbx internals... > If there is a solution inside libwdi/Zadig, I would prefer to sort > the issue there. I don't think so (see above). Jan ------------------------------------------------------------------------------ Get 100% visibility into Java/.NET code with AppDynamics Lite! It's a free troubleshooting tool designed for production. Get down to code-level detail for bottlenecks, with <2% overhead. Download for free and get started troubleshooting in minutes. http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk _______________________________________________ libusbx-devel mailing list libusbx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libusbx-devel
