On 12/29/2011 11:27 PM, Bora Sahin wrote: > Hi all, > > We have some problems regarding the use of Freesco Logic FL1009 xhci > controller in our embedded board. > > xHCI is connected to the system through PCIe. Our kernel is > linux-2.6.30 so did not have xHCI support. Fresco Logic provided > patches (in the form of file replacements). We applied those patches > and put it into the test. We are using a USB stick (filesystem is > vfat) for this purpose. We can mount the disk without any problem, ls > seems working, however, when we try to read files (60, 70KB) (by using > cat) and/or creating dummy files on it, it crashes. We opened DEBUG > feature and caught some log while this is happening. > I don't know what patches Fresco provided to you, so it's hard to tell the differences from current kernel and debug. Can you try the latest stable kernel? > > ----------- > I suppose the above is okay, and it continues for a while. > ----------- > ... > > ----------- > I think all of the above log is okay. > > Then things start going bad. Look at the below output: > ----------- > > xhci_hcd 0000:00:00.0: Ring deq = 0x313d460 (DMA) > xhci_hcd 0000:00:00.0: Event ring deq = 0x313c570 (DMA) > xhci_hcd 0000:00:00.0: Giveback URB 834087c0, len = 31, status = 0 > xhci_hcd 0000:00:00.0: xhci_handle_event - returned from handle_tx_event > xhci_hcd 0000:00:00.0: In xhci_handle_event > xhci_hcd 0000:00:00.0: `MEM_WRITE_DWORD(3'b000, 64'h79002038, > 64'h313c578, 4'hf); > xhci_hcd 0000:00:00.0: count sg list trbs: > xhci_hcd 0000:00:00.0: sg #0: dma = 0x13b4000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #1: dma = 0x1384000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #2: dma = 0x1388000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #3: dma = 0x138c000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #4: dma = 0x1390000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #5: dma = 0x1394000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #6: dma = 0x1398000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #7: dma = 0x139c000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #8: dma = 0x600000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #9: dma = 0x604000, len = 0x2000 (8192), num_trbs = 1 > xhci_hcd 0000:00:00.0: sg #10: dma = 0x608000, len = 0xa000 (40960), > num_trbs = 2 > xhci_hcd 0000:00:00.0: > usb 1-1: ep 0x2 - urb len = 122880, sglist used, num_trbs = 12 > xhci_hcd 0000:00:00.0: Endpoint state = 0x1 > xhci_hcd 0000:00:00.0: First length to xfer from 1st sglist entry = 8192 > xhci_hcd 0000:00:00.0: sg entry: dma = 0x13b4000, len = 0x2000 (8192), > 64KB boundary at 0x13c0000, end dma = 0x13b6000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d470 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x1384000, len = 0x2000 (8192), > 64KB boundary at 0x1390000, end dma = 0x1386000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d480 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x1388000, len = 0x2000 (8192), > 64KB boundary at 0x1390000, end dma = 0x138a000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d490 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x138c000, len = 0x2000 (8192), > 64KB boundary at 0x1390000, end dma = 0x138e000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d4a0 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x1390000, len = 0x2000 (8192), > 64KB boundary at 0x13a0000, end dma = 0x1392000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d4b0 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x1394000, len = 0x2000 (8192), > 64KB boundary at 0x13a0000, end dma = 0x1396000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d4c0 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x1398000, len = 0x2000 (8192), > 64KB boundary at 0x13a0000, end dma = 0x139a000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d4d0 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x139c000, len = 0x2000 (8192), > 64KB boundary at 0x13a0000, end dma = 0x139e000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d4e0 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x600000, len = 0x2000 (8192), > 64KB boundary at 0x610000, end dma = 0x602000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d4f0 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x604000, len = 0x2000 (8192), > 64KB boundary at 0x610000, end dma = 0x606000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d500 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x608000, len = 0x8000 (32768), > 64KB boundary at 0x610000, end dma = 0x610000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d510 (DMA) > xhci_hcd 0000:00:00.0: sg entry: dma = 0x610000, len = 0x2000 (8192), > 64KB boundary at 0x620000, end dma = 0x612000 > xhci_hcd 0000:00:00.0: Ring enq = 0x313d520 (DMA) > xhci_hcd 0000:00:00.0: `MEM_WRITE_DWORD(3'b000, 32'h79003004, 32'h4, 4'hf); > xhci_hcd 0000:00:00.0: Cancel URB 834086c0 > Here xhci_urb_dequeue() is called, means usb core wants to unlink or kill a URB. Seems the driver does not handle that correctly. To dequeue a URB, xHCI driver needs to do many things, including adding the unfinished TDs to a list, issue a stop endpoint command, giveback the URB when the command is executed successfully, and ring the doorbell to continue transfer or issue a set TRDP command if necessary. It's a complex process. Perhaps you can ask Fresco to check that part. Thanks, Andiry > After that it prints event_ring, command_ring, endpoint_ring etc. You > can find the details in the log file. > > Could you please comment on that? > > Regards, > > PS: I am not subscribed to the list. Please cc to me. > > -- > Bora Sahin > -- > 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 > > -- 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
