On Thu, 9 Aug 2007, Martin Fahr wrote: > Hi all, > > after I moved my Debian root file system on a 8GB USB memory stick I > observed a very rare disconnect of the stick, leaving me with an totally > unusable system. > dmesg shows something like > |usb 1-5: reset high speed USB device using ehci_hcd and address 5 > |usb 1-5: device descriptor read/64, error -110 > |usb 1-5: device descriptor read/64, error -110 > > However later tests on another computer proved that this is not related > to the root file system on the stick. But as this happens only once > every few hours of usage of the stick, I observe it almost only when the > root file system is there. Just copying files is doesn't take long enough. > > I could reproduce this several times, one time even at a desktop > computer, where the USB stick is connected to the onboard Intel USB > controller (lspci: Intel Corporation 82801G (ICH7 Family) USB2 EHCI > Controller (rev 01)). But maybe this was a totally different failure.. > > I don't know if this is related, but later recovery of the filesystem > (ext3 or reiserfs) showed some serious errors. I lost several files on > my filesystem almost every time when this happened. > At first I thought it's just that the power supply to the USB stick is > cut off. To prove that I've made a new USB cable, where I could switch > off the power lanes. Then I copied some files on the stick and cut the > power lanes. As expected the stick stopped working instantaneously, but > dmesg shows something different: > |usb 1-5: USB disconnect, address 19 > |sd 12:0:0:0: [sdb] Result: hostbyte=0x01 driverbyte=0x00 > |end_request: I/O error, dev sdb, sector 1013880 > and the journaling filesystems never got screwed up that horrible with > this setup. > Thus it's not a simple power-cutoff.. > > My questions: > Anyone experiencing something similar? > Does someone what's happening? > Could my hardware be defective? (stick, controller, ??) > Is a workaround known? > What can I do? What should I do?
Lots of people have experienced this sort of problem (although they generally don't have their root filesystem on the USB device!). Nobody knows what causes it. In general the -71 error indicates some kind of communication failure. But whether it's caused by noise in the cable or crashed firmware in the device, we can't tell. The fact that error recovery fails (the kernel tries to do a USB port reset) indicates that the problem really does lie in the device, at least in part. Maybe it's caused by overheating, maybe it's random, maybe there's a bug in the firmware, maybe the computer tries to transfer data faster than the device can handle... There are lots of guesses and no evidence. It probably is not caused by defective hardware in the computer's USB controller. Unplugging the device and plugging it back in again should fix the fault. Of course, that won't work for you. I don't know of any workaround. My only suggestion is to try using a different storage device. Alan Stern ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Linux-usb-users@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
