On Thu, 23 Jan 2014 16:41:26 -0500 (EST), Alan Stern wrote: > On Thu, 23 Jan 2014, Dennis New wrote: > > > On Wed, 15 Jan 2014 17:04:05 -0500 (EST), Alan Stern wrote: > > > On Wed, 15 Jan 2014, Dennis New wrote: > > > > > > > When listening to my usb-bluetooth audio headset, the usb > > > > subsystem (I think) on my laptop crashes, the bluetooth/audio > > > > is lost, and my mplayer program is left in an "uninterruptible > > > > sleep" (D) state and cannot be killed. The only way to get my > > > > usb-audio working again is to reboot. > > > > > > ... > > > > > > > The crash is difficult to reproduce, sometimes it can occur > > > > after a few minutes of a clean boot, and sometimes after a few > > > > days. After the audio/usb abruptly stops, the syslog simply > > > > says: > > > > > > > > kernel: ALSA sound/usb/endpoint.c:501 timeout: still 3 active > > > > urbs on EP #3 > > > > > > > > And then after a few minutes (after enabling some kernel > > > > debugging options), I get some call traces of the hung task(s), > > > > which have "ohci_urb_dequeue" in common. > > > > > > > > http://dennisn.dyndns.org/guest/pubstuff/debug-usbaudio/crash1.log > > > > http://dennisn.dyndns.org/guest/pubstuff/debug-usbaudio/crash2.log > > > > http://dennisn.dyndns.org/guest/pubstuff/debug-usbaudio/crash3.log > > > > > > The logs show that the khubd thread is hung, waiting for the > > > ohci-hcd driver to finish cancelling a transfer. > > > > > > > The problem is not with my device, since it occurs with another > > > > identical piece of hardware. It maybe be a problem with my > > > > laptop, since my friend does not seem to be experiencing this > > > > problem, although her laptop does not use ohci-pci. > > > > > > > > My usb controllers are: > > > > > > > > 1002:4374 00:13.0 USB controller: Advanced Micro Devices [AMD] > > > > nee ATI IXP SB400 USB Host Controller (prog-if 10 [OHCI]) > > > > 1002:4375 00:13.1 USB controller: Advanced Micro Devices [AMD] > > > > nee ATI IXP SB400 USB Host Controller (prog-if 10 [OHCI]) > > > > 1002:4373 00:13.2 USB controller: Advanced Micro Devices [AMD] > > > > nee ATI IXP SB400 USB2 Host Controller (prog-if 20 [EHCI]) > > > > > > Your first step in attacking this problem should be to use the > > > most up-to-date kernel available, which at the moment is > > > 3.13-rc8. Be sure to enable CONFIG_FRAME_POINTER along with > > > CONFIG_USB_DEBUG. > > > > > > If the problem occurs again, collect a copy of the output from the > > > dmesg command (_not_ a copy of the system log). Then go into > > > /sys/kernel/debug/usb/ohci and collect copies of all the files in > > > the two subdirectories. > > > > > > It may be necessary to apply some debugging patches to get more > > > information. > > > > Okay, so it happened again (after the "possibly usbaudio induced > > crash" none of my usb ports work at all), and I was able to get some > > info from /sys/kernel/debug/usb this time. > > > > http://dennisn.dyndns.org/guest/pubstuff/debug-usbaudio/crash5.log > > Your problem looks a lot like the one in this email thread: > > http://marc.info/?l=linux-usb&m=139041301317106&w=2 > > Check it out. That patches in that thread won't help on your > machine, though, because you're using an ATI motherboard rather than > NVIDIA. > > (Not to mention that it's very hard to test anything in your setup, > because the problem occurs so rarely.) Indeed, "ohci_quirk_zfmicro" (as mentioned in that marc.info link) would crash my kernel/system (I think when some graphics switch would happen) :). So I tried "ohci_quirk_amd700", since there were already some PCI_VENDOR_ID_ATI quirks in ohci-pci.c that were using it, but alas, no luck. After about 6 days, the hang/crash happened again. I wasn't able to get debugging info this time around :s. The mystery continues. -- 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