On 08.02.2013 11:58, Greg KH wrote: > On Fri, Feb 08, 2013 at 08:36:24PM +0100, Matthias Schniedermeyer wrote: > > On 08.02.2013 08:35, Greg KH wrote: > > > On Fri, Feb 08, 2013 at 09:22:09AM +0100, Matthias Schniedermeyer wrote: > > > > On 07.02.2013 07:05, Greg KH wrote: > > > > > On Thu, Feb 07, 2013 at 02:42:05PM +0100, Matthias Schniedermeyer wrote: > > > > > > Hi > > > > > > > > > > > > > > > > > > After updating from 3.7.2 to 3.7.6 disconnecting a USB3 device from a > > > > > > xhci-port isn't detected properly anymore. After removing a 32GB stick > > > > > > the only line in syslog i can see is: > > > > > > sdc: detected capacity change from 31625052160 to 0 > > > > > > /dev/sdc is still there. > > > > > > Reconnecting the device results in "usb 3-1: USB disconnect, device number 2" > > > > > > immediatly followed by the normal slew of "new device found" messages. > > > > > > > > > > > > I tested the usb-stick on an ehci-port and there is worked normaly. > > > > > > > > > > > > I'm quite sure i read a similar bug-report a little while ago on LKML, > > > > > > is this bug already known and on the way to be fixed (in 3.7.7)? > > > > > > > > > > Possibly, we have a bunch of USB 3 patches queued up for 3.7.7, give me > > > > > a day or so to get 3.7.7-rc1 out for review, and it would be great if > > > > > you could test it then to see if it solves the issue for you or not. > > > > > > > > And the short answert to that is: No > > > > > > Ugh. > > > > > > Can you run 'git bisect' to track down the offending commit? > > > > The change is in v3.7.2..v3.7.3 > > > > Git bisect points to this commit: > > f7965c0846d74b270e246c1470ca955d5078eb07 > > > > After i "patch -R"ed that in 3.7.6 detection of disconnect worked again. > > > > And in anticipation of the next question: > > Yes that bug is in 3.8-rc6 (& latest of a few minutes ago) too. > > "patch -R"ing the diff worked (only tested with latest). > > Thanks for tracking it down. > > > It appears that there are either not that many people that disconnect > > USB3 devices and/or people that reguarly update to bleeding edge > > (stable) kernels. ;-) > > Heh, maybe, but disconnecting USB 3 devices seems to work for me here, > so don't be so sure about that :) Maybe There is a CONFIG options that has an influence. In regards to USB i only have the bare-minimum of options set to support the devices i use. grep USB .config | grep ^CONFIG CONFIG_BT_HCIBTUSB=m CONFIG_SND_USB=y CONFIG_SND_USB_AUDIO=m CONFIG_USB_HID=y CONFIG_USB_ARCH_HAS_OHCI=y CONFIG_USB_ARCH_HAS_EHCI=y CONFIG_USB_ARCH_HAS_XHCI=y CONFIG_USB_SUPPORT=y CONFIG_USB_COMMON=y CONFIG_USB_ARCH_HAS_HCD=y CONFIG_USB=y CONFIG_USB_ANNOUNCE_NEW_DEVICES=y CONFIG_USB_XHCI_HCD=m CONFIG_USB_EHCI_HCD=m CONFIG_USB_UHCI_HCD=m CONFIG_USB_STORAGE=m CONFIG_USB_SERIAL=m CONFIG_USB_SERIAL_FTDI_SIO=m I have this problem on 3 different computers each of a different (Intel) CPU generation and corrosponding chipset (Westmere/H55/NEC, Sandy Bridge/H67/NEC, Ivy Bridge/Z77/Intel) running identical kernels. There are at least 2 different XHCI chips (Intel & NEC) and at least 3 different mass-storage device chipsets. (1 USB-stick, several HDDs in USB3 enclosures with 2 different chipsets) -- Matthias -- 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
