On Wed, Aug 4, 2010 at 4:31 AM, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: > On Tue, 3 Aug 2010, Sarah Sharp wrote: > What was the exact failure case for fdisk? > > Would it hang or would it quit with some sort of error message? # fdisk /dev/sdb Unable to open /dev/sdb > Neither of these will help. The usbmon log shows that the drive is > claiming no medium is present. usb-storage debugging will just show > the same thing, and fdisk won't work. Right. That was what I noticed before. linux/drivers/scsi/sd.c sdkp->media_present = 0. > > One thing I noticed is that your device is running in high speed mode. > > It should have switched over to SuperSpeed mode after it was reset by > > the USB core. Can you get it to enumerate as SuperSpeed by plugging the > > device in after an unload/reload of the xHCI driver? > > That may be related to the problem. Perhaps the drive gets confused > when trying to connect at SuperSpeed, and then falls back to high speed > and stops working. I've tried all sorts of things and combinations with unloading/loading usb_storage and xhci_hcd. Is this what you're looking for? It seems to have problem with the "short transfer" ... hub 3-0:1.0: unable to enumerate USB device on port 3 usb 3-1: new SuperSpeed USB device using xhci_hcd and address 2 xhci_hcd 0000:05:00.0: WARN: short transfer on control ep xhci_hcd 0000:05:00.0: WARN: short transfer on control ep xhci_hcd 0000:05:00.0: WARN: short transfer on control ep xhci_hcd 0000:05:00.0: WARN: short transfer on control ep I've trimmed the detailed dmesg here if it helps ... xhci_hcd 0000:05:00.0: WARN: short transfer on control ep xhci_hcd 0000:05:00.0: Waiting for status stage event xhci_hcd 0000:05:00.0: Event ring deq = 0xbac064d0 (DMA) xhci_hcd 0000:05:00.0: // Write event ring dequeue pointer, preserving EHB bit xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 64'hffffc90017d40638, 64'hbac064d0, 4'hf); xhci_hcd 0000:05:00.0: xhci_handle_event - returned from handle_tx_event xhci_hcd 0000:05:00.0: In xhci_handle_event xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 64'hffffc90017d40638, 64'hbac064d8, 4'hf); xhci_hcd 0000:05:00.0: op reg status = 00000008 xhci_hcd 0000:05:00.0: ir set irq_pending = 00000003 xhci_hcd 0000:05:00.0: Event ring dequeue ptr: xhci_hcd 0000:05:00.0: @bac064d0 bac06940 00000000 01000000 01018001 xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 32'hffffc90017d40024, 32'h8, 4'hf); xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 32'hffffc90017d40620, 32'h3, 4'hf); xhci_hcd 0000:05:00.0: In xhci_handle_event xhci_hcd 0000:05:00.0: xhci_handle_event - OS owns TRB xhci_hcd 0000:05:00.0: xhci_handle_event - calling handle_tx_event xhci_hcd 0000:05:00.0: In handle_tx_event xhci_hcd 0000:05:00.0: handle_tx_event - ep index = 0 xhci_hcd 0000:05:00.0: handle_tx_event - checking for list empty xhci_hcd 0000:05:00.0: handle_tx_event - getting list entry xhci_hcd 0000:05:00.0: handle_tx_event - looking for TD xhci_hcd 0000:05:00.0: handle_tx_event - found event_seg = ffff88023268b840 xhci_hcd 0000:05:00.0: Event TRB with TRB type ID 32 xhci_hcd 0000:05:00.0: Offset 0x00 (buffer lo) = 0xbac06940 xhci_hcd 0000:05:00.0: Offset 0x04 (buffer hi) = 0x0 xhci_hcd 0000:05:00.0: Offset 0x08 (transfer length) = 0x1000000 xhci_hcd 0000:05:00.0: Offset 0x0C (flags) = 0x1018001 xhci_hcd 0000:05:00.0: DMA address or buffer contents= 3133172032 xhci_hcd 0000:05:00.0: Successful control transfer! xhci_hcd 0000:05:00.0: Ring deq = 0xbac06930 (DMA) xhci_hcd 0000:05:00.0: Ring deq = 0xbac06940 (DMA) xhci_hcd 0000:05:00.0: Ring deq = 0xbac06950 (DMA) xhci_hcd 0000:05:00.0: Event ring deq = 0xbac064e0 (DMA) xhci_hcd 0000:05:00.0: // Write event ring dequeue pointer, preserving EHB bit xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 64'hffffc90017d40638, 64'hbac064e0, 4'hf); xhci_hcd 0000:05:00.0: Giveback URB ffff88023241c0c0, len = 16, status = 0 xhci_hcd 0000:05:00.0: xhci_handle_event - returned from handle_tx_event xhci_hcd 0000:05:00.0: In xhci_handle_event xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 64'hffffc90017d40638, 64'hbac064e8, 4'hf); xhci_hcd 0000:05:00.0: Queueing ctrl tx for slot id 1, ep 0 xhci_hcd 0000:05:00.0: Endpoint state = 0x1 xhci_hcd 0000:05:00.0: Ring enq = 0xbac06960 (DMA) xhci_hcd 0000:05:00.0: Ring enq = 0xbac06970 (DMA) xhci_hcd 0000:05:00.0: Ring enq = 0xbac06980 (DMA) xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 32'hffffc90017d40804, 32'h1, 4'hf); xhci_hcd 0000:05:00.0: op reg status = 00000008 xhci_hcd 0000:05:00.0: ir set irq_pending = 00000003 xhci_hcd 0000:05:00.0: Event ring dequeue ptr: xhci_hcd 0000:05:00.0: @bac064e0 bac06960 00000000 0d0000d5 01018001 xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 32'hffffc90017d40024, 32'h8, 4'hf); xhci_hcd 0000:05:00.0: `MEM_WRITE_DWORD(3'b000, 32'hffffc90017d40620, 32'h3, 4'hf); xhci_hcd 0000:05:00.0: In xhci_handle_event xhci_hcd 0000:05:00.0: xhci_handle_event - OS owns TRB xhci_hcd 0000:05:00.0: xhci_handle_event - calling handle_tx_event xhci_hcd 0000:05:00.0: In handle_tx_event xhci_hcd 0000:05:00.0: handle_tx_event - ep index = 0 xhci_hcd 0000:05:00.0: handle_tx_event - checking for list empty xhci_hcd 0000:05:00.0: handle_tx_event - getting list entry xhci_hcd 0000:05:00.0: handle_tx_event - looking for TD xhci_hcd 0000:05:00.0: Offset 0x00 (buffer lo) = 0xbac06960 xhci_hcd 0000:05:00.0: Offset 0x04 (buffer hi) = 0x0 xhci_hcd 0000:05:00.0: Offset 0x08 (transfer length) = 0xd0000d5 xhci_hcd 0000:05:00.0: Offset 0x0C (flags) = 0x1018001 xhci_hcd 0000:05:00.0: DMA address or buffer contents= 3133172064 xhci_hcd 0000:05:00.0: WARN: short transfer on control ep Thanks, Jeff -- 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
