[U-Boot] ARM: omap3: Implement dpll5 (HSUSB clk) workaround for OMAP36xx/AM/DM37xx according to errata sprz318e.

[Robert Nelson]

On Tue, Aug 20, 2013 at 5:15 AM, Andreas Naumann <dev at andin.de> wrote:
> Hi Roger,
>
>
>>
>> What are the symptoms you see when this issue triggers?
>
>
> The symptoms are erroneous USB transaction, seen with a USB port analyzer.
> These only sometimes (not always) stall the USB communication, e.g. a USB
> mass storage device cant be read any longer.
>
>
>>
>> What is the test case to trigger the error? Is it just running any USB I/O
>> for
>> long enough time?
>
>
> Our scenario to reproduce was rebooting a warmed up board (either let it run
> for >5min or heat up in climate chamber).
>
> However, the beagle probably uses a 26 MHz crystal oscillator (our board
> uses 19.2MHz), so the PLL5 dividers may be set in a way that the problem
> never occurs.

The xM uses a 26Mhz, but the errata still applies, as a number of
customer boards do show the issue..

http://www.ti.com/lit/er/sprz319e/sprz319e.pdf
page 113

>> I have a beagle-xm (DM3730 ES1.2) with me and would like to reproduce the
>> error.

Roger, this only seems to effect a small number of xM's, as it seems
to vary on pll drift.  So if your xM is fine, i do have a spare xM C,
that pretty reliably shows the issue after transferring a large amount
of data over the usb port...  I had traded a good xM with a customer
such that i could keep re-basing our out of tree dpll5 tweak..

Regards,

-- 
Robert Nelson
http://www.rcn-ee.com/