Re: MUSB driver on AM3352 dropping USB packets

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi,

On Thu, May 05, 2016 at 12:22:33PM +0000, Andrew Goodbody wrote:
> > From: Bin Liu [mailto:b-liu@xxxxxx]
> > Hi,
> > 
> > On Wed, May 04, 2016 at 03:48:50PM +0000, Andrew Goodbody wrote:
> > > Hi,
> > >
> > > I have been investigating communication issues with iPads. When the
> > > system is busy it seems that the musb driver is silently dropping
> > > occasional packets. I have a usbmon trace that does not show the
> > > packet and I have a trace from a hardware USB analyser that does show
> > > the packet. So the device is correctly sending the packet, it is even
> > > being ACKed, but it is not passed up to the application. The packet is
> > > a bulk transfer packet of 20 bytes. Can anyone please give me pointers
> > > to where to go looking for the problem? The syslog shows nothing
> > > relevant.
> > 
> > What is the part number on the am3352 chip?
> 
> AM3352BZCZ100
> 
> > What kernel version do you use?
> 
> 4.5.0
> 
> > Is musb cppi dma enabled? If so, does the problem still happen when CPPI
> > disabled?
> 
> Yes. Yes. When testing with PIO I did get the message "Rx interrupt with no errors or packet!".
> 
> > First try to turn on dynamic debug log in musb_host.c to check if musb
> > receives the packet or not.
> > 
> > Regards,
> > -Bin.
> 
> I am having problems doing this. If I enable the whole file then I get
> lots of messages on the console about /dev/kmsg buffer overrun. There
> are more then 26 million packets in the hardware trace and I have not
> worked out how to correlate any of the possible message from dynamic
> debug with those packets even when I enable some of the dynamic debug
> lines.  I can see a few messages about "DMA complete but packet still
> in FIFO, CSR 2103" and just the occasional "extra TX2 ready, csr 2100"
> when I enable some of the lines for dynamic debug.

Well, this issue would not be easy to debug. Is this with your custom
board? If so, have you run EyeDiagram test to rule out signal integrity
problem? Are you able to reproduce it with any TI EVM, such as
Beaglebone Black? If so, please explain the detail of the test case, I
could try to reproduce it on my side.

> 
> Andrew

Regards,
-Bin.
--
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



[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux