On Fri, 2 Dec 2011, Michael Below wrote: > Hi, > > yesterday I decided to get a webcam and got the above mentioned error > with two different uvc-supported models (Logitech C270, MS HD 5000). > > I have a HP dc5850 microtower, it has an AMD Phenom II with AMD780 > chipset (Radeon 3100 integrated graphics). I am running Debian testing, > kernel 3.1.4. Each time I plugged the Logitech C270 into the system, two > other USB devices (Samsung Printer, DLink wireless) stopped working. I > only got the other devices working again through a reboot, simple > unloading/reloading of the modules wouldn't help. Most of the times also > the USB mouse (MS Wireless 5000) would quit (the keyboard is PS/2). I > installed the webcam successfully on Windows 7, but when booting into > Linux the USB devices stopped working. > > I returned the Logitech and got a Microsoft HD-5000 webcam instead. I > installed it in Windows 7 and rebooted to Linux. I was able to make a > test call with Skype on Linux and see my test image there, but when I > quit Skype the USB devices stopped working, and rebooting with the > webcam attached didn't help, i had to reboot again without the webcam. > > There is a similar problem described here, but without a solution: > http://lkml.org/lkml/2011/4/25/22 > > I asked for help on the linux-uvc-devel list: Alexey Fisher > advised me to try older kernel versions and certain USB options. I > tried Debian kernels 2.6.32 and 2.6.38, both would boot with USB > up, but as soon as I accessed the webcam, USB would go down with > similar error messages as current kernels. The problem is that the EHCI controller hardware does not turn off the periodic schedule quickly enough when the computer tells it to. It's a hardware fault, not a driver problem. Other people have encountered the same problem. For example, you can read through these two threads (actually only one thread but the Subject: got changed in midstream): http://marc.info/?l=linux-usb&m=131974283732241&w=2 http://marc.info/?l=linux-usb&m=132009844030500&w=2 > Setting "echo -1 > /sys/module/usbcore/parameters/autosuspend" or "echo > Y > /sys/module/usbcore/parameters/old_scheme_first" didn't help > either. No, because this has nothing to do with suspending or device initialization. > Today I was able to work around this problem by plugging the > camera into the frontside USB ports, before I had been using the > backside ports. Seems to be a different controller, in the log > files it is now usb 2-4 instead of 1-4 or the like. That's odd. I wonder why the two controllers should behave differently. Maybe there's another device attached to bus 2 which keeps the periodic schedule active all the time. > I reported this as a Debian bug, there is additional information in the > bug report: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=650736 > > I am attaching parts of my kern.log. If you need more info, I am glad to > give it... Let's see the output from lsusb and lspci. Alan Stern -- 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
