On 12/29/2011 02:40 PM, Andrej Podzimek wrote: >>>> Do you connect your USB2.0 devices to USB3 root hub or via a USB3.0 >>>> external hub? >>> >>> I tried both combinations. A direct connection did not work either. >>> Surprisingly, it provoked a complete system freeze (with no reaction to >>> ping) after a couple of attempts. (However, this is not 100% >>> reproducible.) I didn't have netconsole configured, so I don't know what >>> exactly happened. Plugging the USB 2.0 devices to the USB 3.0 hub never >>> caused a freeze. >>> >>>> Please apply the two patches attached and see if they are useful. >>> >>> Unfortunately, the patches did not bring any obvious improvement. My USB >>> 3.0 hard drive worked as usual and legacy devices were ignored as usual. >>> The machine now runs kernel 3.1.1 (which seems to have these patches >>> already included) and there's still the same problem. >>> >> >> Haven't seen such issue on NEC uPD720200 before. Can you enable >> CONFIG_USB_DEBUG and CONFIG_USB_XHCI_HCD_DEBUGGING, send the dmesg >> from bootup, and specify the point when you hot plug a USB2.0 device >> to the root hub? > > Hi, > > I tried this with kernel 3.1.6. Problems with 2.0 devices connected to > 3.0 ports are nearly gone. There is only one issue remaining, probably > related to the active cable. I tested the following connections to the > 3.0 host with a 3.0 hub, 3.0 active cable, 2.0 webcam and 3.0 hard drive: > > What works: > host <-> webcam, drive > host <-> hub <-> webcam, drive > host <-> active cable <-> drive > host <-> active cable <-> hub <-> drive > > What doesn't work: > host <-> active cable <-> webcam > host <-> active cable <-> hub <-> webcam > > It seems that the active cable is "filtering" all the legacy > connections, no matter if connected directly to the active cable or > through a hub. USB 3.0 connections work fine. Is this supposed to > happen? Or is there something wrong with my active cable? > > I'm attaching the kernel.log from bootup. It seems to be truncated, as > the kernel's ringbuffer must have overflown early. (I could of course > enlarge the buffer if that helps...) Furthermore, plugxhci.log contains > the log messages that appeared after plugging in my USB 3.0 hard drive > into the hub connected to the active cable. > > Surprisingly, when I plugged in the webcam through the active cable, > *nothing* new appeared in dmesg. In the "What works" cases, the webcam > caused the usual flood of debug messages and became usable (at least in > cheese). But it was not even detected when connected through the active > cable. The same holds for flashdisks, a usb/serial converter, a DVD+RW > drive and probably all USB 2.0 devices in general. Yet all of them now > work fine when connected to a 3.0 port without the active cable in between. > > I'll probably have to avoid using the active cable. :-( That's it. > OK, I can just assume the active cable is somehow "incompatible" with USB2.0 devices, and when plug in the USB2.0 device via the active cable, no interrupt triggers, no port status change occurs, hence you saw nothing in dmesg and the device is failed to be recognized. But that's only my assumption, perhaps you can try other cables. Thanks, Andiry -- 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
