On Thu, 30 Oct 2008, Karl Kiniger wrote: > OK, did some experiments with two drives - 1TB Samsung and ICY box. > > Test setup: ICY box mounted r/o and running tar -cvf /dev/null -C /mnt/icybox > > While tar is running: power off/on or pull/plug USB cable on the > other drive. > > Result: pulling USB from Samsung or powering off does not make a > difference. Pulling cable or turning off produces log entries like > > Oct 30 20:14:32 kipc2 kernel: usb 1-3: USB disconnect, address 6 <== Samsung > and one or two lines > Oct 30 20:14:32 kipc2 kernel: usb 1-8: reset high speed USB device using ehci_hcd and address 3 > from the ICY box. > > Power on or connect of USB cable had no crosstalk so far. > > Two times out of about 15 tries the ICY box was hung completely, in the > remaining cases it recovered by itself. However it was enough to power > cycle the ICY - it was not necessary to reload the ehci driver. > > Without activity (tar) I could only see the usb 1-3 disconnects. At the > moment I dont want to experiment with writing to the drives while > playing USB games. (applies also to the OneTouch below) > > The same happens when unplugging the OneTouch. There are not > always disconnects on the ICY but mostly. This message turned up as I was going through old email. I thought you'd like to know that a patch to address this matter has been accepted; it should appear in 2.6.30. If you would like to try it out, the patch is: http://www.kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/gregkh-04-usb/usb-ehci-add-software-retry-for-transaction-errors.patch 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
