This 0/1 explanation mail would better have been part of the patch itself, which would also help thread discussions better. --- On Mon, 11/1/10, Li, Jiebing <jiebing.li@xxxxxxxxx> wrote: > > The original composite gadget tries > to bind several functions one by > one, and the final result is decided by the last function. > That is, > if binding of the last function (ecm or eem) is ok, return > value is 0. Last function need not be ECM o EEM of course. In general, if ANY function can't bind to the composite gadget, then I'd say a hard fail is the appropriate response. After all, the gadget is defined as having a fixed set of functions, and when any are missing, then the gadget is malfunctioning. So there might be a bug lurking here, but if there is, this patch doesn't seem to fix it. > Or return value will be non-zero and the whole gadget will > not be functional. ... which is very appropriate, when it can't provide its full functionality, and instead acts like a different (less functional) gadget > For example, number of endpoints is not enough to support > all the functions > on some platform, That's a reasonable class of problem to worry about, and the original way to handle it is the same as non-composite gadgets: configure only gadgets that can work on that hardware. > This patch is used to handle this more flexibly. I want to see predictability and consistency, which this patch doesn't seem to promote. As long as > one of the > functions is bound succesfully, return value will be 0. NAK. When the composite gadget is defined to require three functions (for example), all of them should bind successfully for the gadget to be initialized. The notion of an "optional" function fights the notion of a fully defined USB peripheral, for which a host driver (e.g. Windows INF file) can be written and supported. Think how those support calls will stretch out (and cost). - Dave -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
