Hi, > Are you able to check what does Windows host actually see in bcd_version if > it is set as "0x0001" at the FunctionFS gadget side? I'm sorry, It is hard for me to check how the bcd_version value is handled internally because I don't have a Windows development environment and I simply don't know enough about Windows. The least I can do is try to manipulate something and watch the behavior from the outside. (e.g., if the driver was autoloaded by passing an Extended Compat ID Descriptor with bcd_version=0x0001, this descriptor would work, etc.) I forgot to write about the background in my last post. Before trying FunctionFS, I had tried GadgetFS with a configuration that included OS descriptor, and found that bcd_version=0x0100 worked (the Windows host recognized it). For this reason, I thought the same descriptor would work in FunctionFS, but was actually rejected by the kernel. > And then, whether it does > not complain because it sees what it expects or because it does not care even > though it sees something it doesn't expect? Since the Windows host appears to recognize both the value 0x0001 and 0x0100, I guess the Windows host "does not care even though it sees something it doesn't expect." I don't know which value the Windows host actually expects, but since the MSFT document says bcd_version=0x0100, it probably expects 0x100. Regards, Yuta Hayama
