On Fri, 31 Aug 2012, Dmitry Torokhov wrote: > > > + /* Send a "check active" packet. The response will be read > > > + * later and ignored. */ > > > + ret = usb_control_msg(udev, usb_rcvctrlpipe(udev, 0), > > > + 0, > > > + USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE, > > > + 0, 0, "\x0A\x01A", 0, > > > > You probably can't send data from the .const section (as well as off the > > stack) -- they can be DMA'ed and there'll be issues with cache consistency on > > non-x86 arches. You should allocate the data with kmalloc(). > > Although, on the second thought, maybe I'm wrong in this case... not really > > sure about sending -- receiving (to the .data section) could certainly be harmful... > > Hmm, do we actually send anything here? The "size" passed to > usb_control_msg() is 0 so I don't think we use that data at all... Good point. Perhaps the 0 is a typo, in which case data does get sent and the buffer must be kmalloc'ed. If the 0 is correct then the buffer should be NULL, not "\x0A\x01A" (and what's the purpose of the leading '0' in the second byte?). In addition, although the bRequestType specifies USB_DIR_OUT, the pipe value is usb_rcvctrlpipe. Is this transfer meant to be IN or OUT? Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
