On Thu, Oct 17, 2013 at 03:45:24AM +0200, bjoern.glaessner@xxxxxxxxxxxxxx wrote: > Hi Sarah... > > Sorry, that I contact you directly via email address. > > But our pain with this bug is so annoying, that - after a bunch > of web searches without a solution - I came to the conclusion to ask you > in personal. > > You are the developer and no one knows more about USB3.0 under Linux. > > > First, I wanna recall a bug report from an other user: > > http://www.spinics.net/lists/linux-usb/msg93401.html > > This is pretty the same issue and I only want to mention, that there are > a lot of other users, who suffer this in our Linux community. No, that's not the same issue. That's the same error message, but their report does not match yours at all. > So, now to the hardware setup: > > - LogiLink USB3.0 Dual SATA HDD Docking Station > http://i01.twenga.com/computer/externe-festplattengehaeuse/logilink-quickport-usb-3.0-p_33593001vb.jpg > > - 2 x 3TB TOSHIBA DT01ABA300 - Retail Kit (PA4293E-1HN0) hard disk > drives, formatted with EXT4 > > Linux: Ubuntu 13.04 with kernel 3.8.0-31-generic > > - Mainboard: ASRock Z86M/USB3 > http://www.asrock.com/mb/Intel/Z68MUSB3/index.de.asp What kernel are you running? Please run `uname -a`. > Harddisks working like charm, until they were not needed for a time > (aprox. 15 minutes). > > After this, the harddisks BOTH umount COMPLETELY from system and are NOT > recoverable. Only possibility is to unplug the USB3.0 cable and re-plug > it to mainboard USB3.0 port or shutting down power and repower the > harddisk docking station. Yeah, it looks like from the log that they disconnect from the bus. > "lsusb" or "fdisk" doesn't show any device after umount. > > > > "dmesg" looks like this: > > ================================= I really need the parts of the log before this. > [45446.007668] xhci_hcd 0000:03:00.0: Timeout while waiting for address > device command > [45446.007720] xhci_hcd 0000:03:00.0: WARN Set TR deq ptr command for > freed stream ID 0 It looks like you're using the uas driver. Can you verify that you are by running `lsmod | grep uas`? That driver has been marked as BROKEN for a long time. What happens if you blacklist it in /etc/modprobe.d/ and use the usb-storage driver instead? Can you send me the output of `sudo lsusb -v` when the devices are plugged in (and hopefully before they freeze?). I want to see if the device supports UAS. > [45451.211534] xhci_hcd 0000:03:00.0: Timeout while waiting for address > device command > [45451.415500] usb 4-1: device not accepting address 2, error -62 > [45451.503598] usb 4-1: USB disconnect, device number 2 > [45451.503612] sd 5:0:0:1: Device offlined - not ready after error > recovery > [45451.815345] Buffer I/O error on device sdc1, logical block 365985792 > [45451.815348] lost page write due to I/O error on sdc1 > [45451.815349] JBD2: Error -5 detected when updating journal superblock > for sdc1-8. > [45452.062727] xhci_hcd 0000:03:00.0: xHCI xhci_drop_endpoint called > with disabled ep ffff880086843b00 > [45452.062730] xhci_hcd 0000:03:00.0: xHCI xhci_drop_endpoint called > with disabled ep ffff880086843b40 > [45452.064017] Buffer I/O error on device sdd1, logical block 121667584 > [45452.064019] lost page write due to I/O error on sdd1 > [45452.064021] JBD2: Error -5 detected when updating journal superblock > for sdd1-8. > [45452.187552] xhci_hcd 0000:03:00.0: Bad Slot ID 1 > [45452.187554] xhci_hcd 0000:03:00.0: Could not allocate xHCI USB device > data structures > [45452.187559] hub 4-0:1.0: couldn't allocate port 1 usb_device > > ================================= > > In my opinion the APM was shutting down the hard disks after a few > minutes. But why the umount completely ? I have no idea what you mean by APM. > Wouldn't it be better to stay connected to the mount points and > re-activate them once the were used ?? > > > Is there a way to tell the system NOT to shut down or to deactivate > power management for this specal USB3.0 device ??? The only power management I know of for USB devices is auto-suspend, which is controlled through /sys/bus/usb/devices/../power/control. It's set to 'disabled' by default, and even if it was enabled, the USB device doesn't go into suspend while the hard drive is mounted. > Best regards and have a nice weekend ! > > Please let me know if you need further or detailed information. Can you compile a kernel from my fun-streams-branch? Just follow these directions http://kernelnewbies.org/KernelBuild, and instead of cloning Linus' tree, run git clone git://git.kernel.org/pub/scm/linux/kernel/git/sarah/xhci.git Once that finishes: cd xhci git checkout -b fun-streams-fixes origin/fun-streams-fixes And compile and install that kernel. That branch contains a lot of fixes for the UAS driver. To test them, make sure you're not blacklisting the uas driver, and reboot into your new kernel. Sarah Sharp -- 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
