On Sat, Oct 27, 2012 at 07:37:50PM -0700, Matthew Hall wrote: > Sarah, > > I finally had occasion to experiment with this further, as more USB 3.0 > devices have become available at decent prices. > > I discovered that, while the card reader would fail with strange errors, a 32 > GB USB 3.0 flash drive from Patriot was able to write out a 4GB file of > /dev/zero at 39.5 MiB / second without experiencing a single error. > > This makes me wonder if there could be a problem with the peripheral itself > worth trying to identify. Upon connection and mount attempt you see this: > > [76008.931517] usb 4-1: reset SuperSpeed USB device number 3 using xhci_hcd > [76008.946972] usb 4-1: Parent hub missing LPM exit latency info. Power management will be impacted. > [76008.948441] xhci_hcd 0000:03:00.0: xHCI xhci_drop_endpoint called with disabled ep ffff8803e7b87500 > [76008.948447] xhci_hcd 0000:03:00.0: xHCI xhci_drop_endpoint called with disabled ep ffff8803e7b87540 > [76009.427691] sd 6:0:0:2: [sdj] Media Changed > [76009.427696] sd 6:0:0:2: [sdj] > [76009.427698] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE > [76009.427700] sd 6:0:0:2: [sdj] > [76009.427702] Sense Key : Unit Attention [current] > [76009.427707] sd 6:0:0:2: [sdj] > [76009.427710] Add. Sense: Not ready to ready change, medium may have changed > [76009.427712] sd 6:0:0:2: [sdj] CDB: > [76009.427714] Write(10): 2a 00 00 00 80 00 00 00 08 00 > [76009.427723] end_request: I/O error, dev sdj, sector 32768 > [76009.427727] Buffer I/O error on device sdj1, logical block 0 > [76009.427729] lost page write due to I/O error on sdj1 > [76009.434558] sdj: sdj1 Hmm, I'm not a storage expert, but maybe the USB storage mailing list could help. It's possible this device needs a particular storage quirk. How big is the device? I'm wondering if sector 32768 is at the end of the drive. It's interesting, since that's exactly 2^(15). Is the failure always for that sector, or does it fail on different sectors? It might help the storage experts if you could take a USB mon trace when the device is plugged in. Directions on how to do that are here: http://lxr.linux.no/#linux+v3.6.5/Documentation/usb/usbmon.txt > It's a bit different from what used to show up on the old kernel. No such > errors appear with the Patriot drive either. 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
