On Wed, 27 May 2015, Golmer Palmer wrote: > Hi Alan, > > Thank you for the new patch! > However, a comment: > > * At page 74 of the specifications, > section F.3 Boot Keyboard Requirements, > (http://www.usb.org/developers/hidpage/HID1_11.pdf) > is defined that: > "The Boot Keyboard shall, upon reset, return to the non- boot protocol > which is > described in its Report descriptor. That is, the Report descriptor for a > Boot > Keyboard does not necessarily match the boot protocol. The Report > descriptor > for a Boot Keyboard is the non-boot protocol descriptor. " > > This enforces that the default mode (protocol) is REPORT, not BOOT. > So why initialize in the patch to "1" for devices without boot support > and "0" for devices with boot support? > > I suggest to always initialize to "1" (=REPORT mode). > > Remember that the assumption is that the DESCRIPTOR of the BOOT protocol > is IDENTICAL to the descriptor of the REPORT protocol, so they are the > same. This is the key point! > > Moreover, if a device don't supports BOOT mode, then the function > Get_Protocol() needs to be unimplemented. See page 54, section > 7.2.5 Get_Protocol Request. So, I suggest to maintain the old behaviour > ("goto stall" for returning ERROR) in the Get_Protocol() in case of > bInterfaceSubClass != USB_INTERFACE_SUBCLASS_BOOT. This can be in the > same way as you done in the Set_Protocol function. Golmer: I sent you a revised patch incorporating this correction more than a month ago, and I haven't heard back. Have you tested the newest patch? Does it work correctly? Alan Stern -- 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
