On Mon, 8 Sep 2014, Mark wrote: > Thanks very much for that! I booted with usb-storage.quirks=059b:0040:s to > enable the single-LUN quirk. No change in dmesg output on connecting the > drive/cable. I also compiled the kernel to add both US_FL_SINGLE_LUN and > US_FL_BULK_IGNORE_TAG quirks. [For that I was using a VirtualBox VM and > had booted the host with the single-LUN quirk.] I can post usbmon logs for > both cases if that would be of interest. (I didn't paste them here to > avoid bloating this message too much; there are other usbmon logs below.) > > I noticed a couple of lines in the JAZ USB Adapter manual: > "When using the Jaz USB Adapter, you cannot connect other devices or > cables to your Jaz drive." > "NOTE: If you are connecting two Jaz drives via two Jaz USB Adapters to > your computer, set each drive to a different SCSI ID number." > > The first point could be a limitation of either the cable firmware or > Iomega's driver software. [There's also a physical limitation; with an > external Jaz drive, you can't plug a cable into the other connector when > the SCSI-USB cable is connected, because the connectors are too close.] > Presumably the second point relates to Iomega's Windows driver. > > I managed to find an old Jaz drive and tested it with my cable. It seems > to work fine, so the cable firmware does restrict itself to working with > Jaz drives. I had the drive SCSI ID set to 4, so it seems the firmware > scans the bus and talks to the first(?) Jaz drive it sees. > > However there were still four messages saying "usb 7-1: reset full speed > USB device number 2 using uhci_hcd" in dmesg output. But the drive could > be accessed, INQUIRY returned sensible data etc. Those reset messages > didn't appear on booting with the single-LUN quirk. Take a look at the > usbmon logs below. So it seems the Iomega SCSI-USB cable needs > US_FL_SINGLE_LUN. Indeed, all the messages attempting to probe LUN 1 failed. That's what caused the resets. On the other hand, even with the SINGLE_LUN flag and a Jaz drive attached, it didn't really work. Every TEST UNIT READY command got a Not Ready; No Medium Present error response. It's not clear what messages the adapter expects to receive. If you can sniff the USB data on a system where it works (Windows most likely) then perhaps it could be made to work on Linux. But how much time do you really want to spend on this? Unless you've got some highly important data stored on those old Jaz drives, it doesn't seem worthwhile. The adapter connects at a measly 12 Mb/s (which means transfer rates well under 1 MB/s in practice) and the drives probably don't have a large storage capacity. A small USB flash drive would hold more data and communicate much faster. 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
