2010/10/11 Grazvydas Ignotas <nota...@gmail.com>:
> On Mon, Oct 11, 2010 at 1:53 PM, Ming Lei <tom.leim...@gmail.com> wrote:
>> 2010/10/11 Grazvydas Ignotas <nota...@gmail.com>:
>>>>> which effectively disables DMA due to a bug in it, so it's probably
>>>>> DMA problems on my board. Here is my log:
>>>>
>>>> Yes, your issue may be related with ep1out Rx DMA, but I am not sure
>>>> since the very important message between timestamp 39.256011 and
>>>> 85.210693 is not provided by you, why not post all messages?
>>>
>>> I did not remove any messages (only added annotation), there was
>>> simply no output. I suppose something wrong happens after
>>
>> If so, no any rx interrupt for ep1out comes after request is queued since
>> musb_g_rx is not called from your log.
>>
>> There is two possibilities:
>>
>> - usb host does not send any packets to ep1out
>>
>> - packets has been sent from usb host to musb gadget, but omap3
>> is not notified by musb rx interrupt, or there is no musb rx
>> interrupt.
>>
>> You can trace usb traffic in usb host to see if there are packets sent
>> to ep1out, usbmon can do it for you, please see Documentation/usb/usbmon.txt
>> for reference. If no packets are traced for ep1out, your problem is nothing
>> to do with musb, and may should be related with usb host driver(usbnet?).
>
> ok here are usbmon logs, taken using same scenario:
> http://notaz.gp2x.de/misc/pnd/linux_next_20101008_usbmon
>
> and here is usbmon log when musb runs in PIO mode and ping works:
> http://notaz.gp2x.de/misc/pnd/linux_next_20101008_usbmon_nodma
>
> So maybe it's ep2in issue after all?
>
Yes, it is really with ep2in, and musb ep2in always return more data to
host so cause overflow in host.
I see what triggered your issue, please try the patch in the link below:
http://marc.info/?l=linux-usb&m=128576496614332&w=2
--
Lei Ming
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
