On Sun, 9 Sep 2018, Klaus Kusche wrote: > Hello, > > On 01/09/2018 17:38, Alan Stern wrote: > > However, the USB layer does set certain quirk bits which can cause > > those other parts to avoid sending certain commands. Perhaps your > > controller needs the BROKEN_FUA flag (see the existing entries in > > drivers/usb/storage/unusual_devs.h with that flag). > > The following seems to fix the problem (I don't know if I did it right): > > --- unusual_uas.h.orig 2018-09-09 10:31:20.440751625 +0200 > +++ unusual_uas.h 2018-09-09 10:32:30.491381466 +0200 > @@ -95,7 +95,7 @@ > "JMicron", > "JMS566", > USB_SC_DEVICE, USB_PR_DEVICE, NULL, > - US_FL_NO_REPORT_OPCODES | US_FL_IGNORE_UAS), > + US_FL_NO_REPORT_OPCODES | US_FL_IGNORE_UAS | US_FL_BROKEN_FUA), > > /* Reported-by: Hans de Goede <hdegoede@xxxxxxxxxx> */ > UNUSUAL_DEV(0x4971, 0x1012, 0x0000, 0x9999, Okay. You should submit this in the proper format for inclusion in the kernel. Follow the instructions in Documentation/process/submitting-patches.rst. > When plugging in the USB enclosure, syslog now says: > > Sep 09 10:36:51 lap kernel: sd 6:0:0:0: [sdc] Disabling FUA > Sep 09 10:36:51 lap kernel: sd 6:0:0:0: [sdc] Write cache: enabled, read cache: > enabled, doesn't support DPO or FUA > > With that, the disk can be mounted, read and written > with no error at the user level. > > There is just one irritating message in the syslog: > > Sep 09 10:42:53 lap kernel: sd 6:0:0:0: [sdc] Synchronizing SCSI cache > Sep 09 10:42:53 lap kernel: sd 6:0:0:0: [sdc] Synchronize Cache(10) failed: > Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK JMicron devices are notorious for not obeying the standard requirements for USB and SCSI. That error is unfortunate, but I don't think much can be done about it. As long as you're careful about not unplugging the drive or turning it off while a filesystem is still mounted, you should be okay. Alan Stern
