Alan Stern <stern@...> writes: > > On Mon, 25 Apr 2011, Thomas Lindroth wrote: > > > On 04/23/2011 10:17 PM, Alan Stern wrote: > > > On Sat, 23 Apr 2011, Thomas Lindroth wrote: > > >> On 04/23/2011 05:31 PM, Alan Stern wrote: > > >>> Your separate post didn't include the dmesg showing what happened when > > >>> you plugged the mouse into an EHCI controller. In fact, a usbmon trace > > >>> of those events would be more useful than a dmesg log. Can you post > > >>> one of those? It doesn't need to start until just before you plug in > > >>> the mouse. > > >> > > >> I made some usbmon dumps of the mouse getting plugged in. I ran > > >> usbmon > file, plugged the mouse in and ctrl + c after about 20 sec. > > > > > > This explains the problem with the EHCI controller. The mouse can't be > > > used because the kernel thinks it would require more periodic bandwidth > > > than is available on the bus. Now, I have no idea why 2.6.39 should > > > behave differently from 2.6.38 in this respect. > > Connecting the mouse to a EHCI port fails on both 2.6.28 and .29 but it > > used to be possible to connect the mouse to a XHCI port prior to .29. > > (Presumably you mean .38 and .39 rather than .28 and .29.) > > Oh, sorry, I must have misunderstood your earlier message. Now it > makes more sense. > > > > You have a bunch of devices all connected to the same EHCI controller. > > > If you can attach the mouse to a different port, one that goes to the > > > other EHCI controller, it might work better. > > I looked into the periodic file and noticed that my Microsoft keyboard > > 4000 used a whole lot of periodic bandwidth for some reason. I assumed > > that a keyboard is a low bandwidth device and put it in the same > > controller as the mouse. Moving the keyboard to the front usb panel > > makes it possible to connect the mouse at the back panel :) > > Heh. A keyboard is low _speed_ ... but that very fact means it uses > more bandwidth than you might think. Low speed signalling is 8 times > slower than full speed, therefore it requires 8 times more bandwidth. > > > Thanks for you help. > > You're welcome. > > Alan Stern > > -- > To unsubscribe from this list: send the line "unsubscribe linux-usb" in > the body of a message to majordomo@... > More majordomo info at http://vger.kernel.org/majordomo-info.html > > I see these posts are from 2011, but I have the exact same dmesg output after attaching a USB2.0 sound card to my 3.0 ports on kernel 4.2.0. It shows over and over again the error: [ 38.659093] usb 1-1: Not enough bandwidth for altsetting 2 [ 38.659095] usb 1-1: 1:2: usb_set_interface failed (-22) [ 38.661184] xhci_hcd 0000:04:00.0: ERROR: unexpected command completion code 0x11. [ 38.661192] usb 1-1: Not enough bandwidth for altsetting 2 [ 38.661195] usb 1-1: 1:2: usb_set_interface failed (-22) [ 38.662383] xhci_hcd 0000:04:00.0: ERROR: unexpected command completion code 0x11. [ 38.662392] usb 1-1: Not enough bandwidth for altsetting 2 [ 38.662396] usb 1-1: 1:2: usb_set_interface failed (-22) Am I experiencing the same problem as the original poster or something different? Is there a way to force a 3.0 port to use the ehci driver? I can unbind the port from the xhci driver but then it doesn't allow me to bind to the ehci driver (I do have the module installed and even in use, as confirmed by lsmod). Any help is much appreciated, I have been dealing with this for over a month now. -- 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
