Re: [opensuse-factory] Re: usb midi keyboard broken [3.3.3 regression] [SOLUTION]

["Steffen Müller" <steffen.mueller@radio-frei.de>]

Oh man! After look and try many usr relevant settings I found the malfactor: 
tlp usb autosuspend settings.

I use tlp (energy-saving tools for thinkpad laptops) and they autosuspend usb 
devices by default. Now I have blacklisted my midi keyboard an all is good. I 
can reproduce the problem this way. Seems so that the midi keyboard doesn't 
wake up if it once sleeping.

For all with similar problems: 
edit /etc/default/tlp and set USB_BLACKLIST="<device id from usbls>" or 
disable usb autosuspend entirely with USB_AUTOSUSPEND=0

thanks and sorry for the bother.
steffen


Am Freitag, 27. April 2012, 11:26:12 schrieb Alan Stern:
> On Fri, 27 Apr 2012, Jiri Slaby wrote:
> > On 04/27/2012 09:37 AM, Steffen M�ller wrote:
> > >> [  224.418062] usb 1-1.3: USB disconnect, device number 6 [
> > >> 227.662396] usb 1-1.3: new full-speed USB device number 7 using
> > >> ehci_hcd [  227.749541] usb 1-1.3: New USB device found,
> > >> idVendor=0763, idProduct=0192 [  227.749550] usb 1-1.3: New USB
> > >> device strings: Mfr=1, Product=2, SerialNumber=0 [  227.749556]
> > >> usb 1-1.3: Product: USB Keystation 88es [  227.749560] usb 1-1.3:
> > >> Manufacturer: M-Audio [  229.750236] ALSA midi.c:214 urb status
> > >> -32 [  229.750700] ALSA midi.c:214 urb status -32 [  229.750911]
> > >> ALSA midi.c:214 urb status -32 [  229.751175] ALSA midi.c:214 urb
> > >> status -32 [  229.751413] ALSA midi.c:214 urb status -32 [
> > >> 229.751659] ALSA midi.c:214 urb status -32 [  229.751910] ALSA
> > >> midi.c:214 urb status -32 [  229.752210] ALSA midi.c:214 urb
> > >> status -32 [  229.752459] ALSA midi.c:214 urb status -32 [
> > >> 229.752662] ALSA midi.c:214 urb status -32
> > 
> > Hmm... I suspect some USB patches added in 3.3.3... Alan, Greg, any ideas?
> > 
> > 3.3.2 is reported to work.
> 
> No ideas.  It would be good to see a usbmon trace, with as many other
> USB devices unplugged as possible.  Even better would be to see also a
> second usbmon trace from a working kernel.
> 
> Alan Stern