Hi, (it's always a good idea to Cc maintainers of drivers you're having trouble with :) Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> writes: > On Sun, 15 Jan 2017, Vincent Pelletier wrote: > >> > An endpoint can not work in both directions, that's not how USB >> > endpoints work at all. I think you are confusing the hardware with your >> > configuration :) >> >> I do agree that I have never noticed in the wild a device using a >> non-zero endpoint (4 LSb) in both directions (1 MSb) - and maybe there >> is no host OS support anywhere. > > I have seen such devices mentioned in the mailing list. They work > fine. > >> On the spec level though, I am less sure. Quoting the USB (v2) spec: > ... > > The USB spec allows such things. > >> So there is either a bug in the HCI side, or in the UDC side: >> - either the HCI should support bidirectional-ish endpoints (same >> 4 LSb, different 1 MSb). Would this limitation be a hardware one ? > > I don't know about DWC. The host controllers drivers I maintain work > okay when two endpoints have the same number but different directions. > > What sort of host controller do you test with? > > What does usbmon show? also, which kernel are you running on dwc3 side? Which platform is it? In order to have any idea of what's going on, I need tracepoints from dwc3. # mkdir -p /t # mount -t tracefs none /t # cd /t # echo 10240 > buffer_size_kb # echo 1 > events/dwc3/enable # echo 0 > events/dwc3/dwc3_readl/enable # echo 0 > events/dwc3/dwc3_writel/enable (run tests, cause it to fail) # cp /t/trace /path/to/non-volatile/media/dwc3-trace.txt compress and send me traces > What happens when you run your test with dummy-hcd? this is a good idea too. >> - or the UDC endpoint auto-allocation mechanism must skip an endpoint >> when it is already used in the other direction > > This could be a bug in the UDC driver. We don't even know which kernel he's running on dwc3 side :-) It could also be a bug on his python functionfs lib. At this point, it could be many things. Vincent, can you compile my xhci-cleanup branch from my usb.org tree (see MAINTAINERS file for the URL. Hint, it's on git.kernel.org) and run it on your xHCI side and *also* capture tracepoints from xHCI? # mkdir -p /t # mount -t tracefs none /t # cd /t # echo 10240 > buffer_size_kb # echo 1 > events/xhci-hcd/enable (run tests, cause it to fail) # cp /t/trace /path/to/non-volatile/media/xhci-trace.txt again, compress and send me traces. -- balbi -- 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
