On Wed, 29 May 2013, Joe Julian wrote: > On 05/22/2013 12:27 PM, Alan Stern wrote: > > On Wed, 22 May 2013, Joe Julian wrote: > > > >> On 05/21/2013 03:20 PM, Joe Julian wrote: > >>> I have about 100 of these creditcard/check scanners that are dropping > >>> events. I was able to find some overflows that I assume are probably > >>> related to the problem, "usb 4-1: ctrl urb status -75 received". > >>> > >>> Short of asking the vendor to fix their product or compiling a custom > >>> kernel, are there any other options? Can the MaxPacketSize or the > >>> buffer size be overridden somehow (/sys/bus/usb/devices... maybe)? > >> I have started a dialog with UIC (Uniform Industrial Corp) about a fix, > >> but I'm not very hopeful. Is there anything I can do to work around the > >> overflow? > > Without knowing more (like where the overflows occur and what data was > > expected), it is impossible to answer this question. A usbmon trace > > would help. > I spent the day at one of our stores capturing data from a small handful > of customer loyalty cards I picked up for this purpose. I was able to > capture 2 bad scans. > > If I'm reading this correctly, there's nothing wrong that could be > corrected from the linux side, could you confirm? I agree. In fact, the usbmon trace doesn't show any errors at all. > I think that overflow was a red herring as none occurred during the bad > scans. Could be. All the transfers to endpoint 0 in the trace are to turn some output indicator on or off. Maybe an LED or something of the sort. > The best example is the last scan of the Qdoba card that ends in "I" > (7th from the bottom). What is that an example of? > The usbmon capture is at > http://joejulian.name/media/uploads/usbcapture.usbmon.gz > > The expected scan data was successive scans of the following cards: > %B6277200522629830^^391200077861?[ > %B6277200522629848^^391200017860?S > %B6277200522629855^^391200027724?S > %B6277200522629863^^391200055339?[ > %B6277200522629871^^391200079146?\ > %B6010565032591577^QDOBA/VALUELINK^25010004000060072408 ?@ > %B6010565032591494^QDOBA/VALUELINK^25010004000060073301 ?C > %B6010565032591700^QDOBA/VALUELINK^25010004000060076767 ?L > %B6010565032591551^QDOBA/VALUELINK^25010004000060073264 ?I > %B6010565032591536^QDOBA/VALUELINK^25010004000060075630 ?K > %B6010565032591528^QDOBA/VALUELINK^25010004000060072347 ?F > %B6010565032591569^QDOBA/VALUELINK^25010004000060075724 ?E > %B6010565032591544^QDOBA/VALUELINK^25010004000060075630 ?N > %B6010565032591502^QDOBA/VALUELINK^25010004000060074630 ?M > %B6010565032591510^QDOBA/VALUELINK^25010004000060074982 ?H I don't have any way to match up these strings against the data in the trace. For example, it looks like the scan of the first card got something like 69 keystrokes. They don't seem to bear any relation to the expected scan data above; for example, the fifth and sixth keystrokes aren't the same. 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
