On Mon, Nov 14, 2011 at 04:51:43PM -0500, starlight@xxxxxxxxxxx wrote: > Hello, > > I'm experiencing a write-test failure with > a WDC MyBook 3TB. Hoping someone may > recognize the issue and perhaps have > some advice. > > Briefly > > # dd bs=1024k if=/dev/zero of=test oflag=direct > dd: writing `test': Read-only file system > 10017+0 records in > 10016+0 records out > 10502537216 bytes (11 GB) copied, 132.675 s, 79.2 MB/s > > verbosity follows, attached below. > > Running with CentOS 6 kernel 2.6.32-71.29.1.el6.x86_64 > but I've found someone else experiencing the same > problem with latest kernel: > > http://forums.gentoo.org/viewtopic.php?p=6871586#6871586 > > Willing to run the test with current kernel.org > if desired. Same drive works perfectly when > attached to a CentOS 5 system via USB2. Have > trouble tickets open with both WD and Highpoint. > > hardware: > > USB3 controller: ASMedia ASM1042 USB controller > (Highpoint RocketU 1144A card) > > drive: WDBACW0030HBK-00 > (WDC MyBook Essential 3TB) > (latest firmware 1016 flashed) A patch was sent out last week for the 3.2 kernel that fixes an issue with the AsMedia host controller: d31c285 xhci: Set slot and ep0 flags for address command. It's marked for the stable kernel, so it should make it into 3.1 soon. It's possible that the problem lies with the combination of your AsMedia host controller and the device, not with the device alone. The patch that went in fixes a case where the xHCI driver feels the host controller bad data, but only if the device needs a reset (which some storage devices do when they decide they can't handle a particular SCSI command). So the MyBook might trigger the xHCI driver bug, but your new drive appears not to. Any chance you can try the latest kernel from Linus' tree and see if it fixes your problem? > Experienced a generally similar failure when > trying it out with a VIA VL800 USB3 controller > (SYBA card). Messages in 'dmesg' were different. > > I find it interesting that the other person > seeing this issue has an Intel USB3 controller > (i.e. reference-implementation hardware). Seems > to reduce the possibility of an issue at the > USB3 layer and point to a higher-layer firmware > or driver issue, perhaps related to timing or > latency. If other host controllers are strict about checking input like the AsMedia host is, then they would also run into this bug. I know it doesn't show up on the NEC xHCI host controller. 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
