On 04/11/2016 04:03 AM, George Broz wrote:
On 10 April 2016 at 10:47, Marek Vasut marex@denx.de wrote:
On 04/09/2016 12:40 AM, George Broz wrote:
On 8 April 2016 at 05:36, Marek Vasut marex@denx.de wrote:
On 04/08/2016 07:16 AM, Stefan Roese wrote:
On 08.04.2016 01:51, George Broz wrote:
<snip>
>>> Try with the attached patch (and probably with dcache off) >> >> The patch applied cleanly. The behavior is unchanged with both >> dcache on and off. The "good" sticks still work, and "bad" sticks still don't. > > OK. Then I should probably go hunting for Kingston DataTraveler SE9, > right ? Can you give me a link to the stick you have, so I know what > crappy device to look for ? Thanks!
Here it is [1] - I have the 8GB version.
I think there will always be crappy sticks that don't work... but do you have any advice as to what properties will/might generally cause a problem?
[1] http://www.amazon.com/Kingston-Digital-DataTraveler-DTSE9H-16GBZET/dp/B00DYQ...
I have exactly this stick here (16GiB) version. And it is detected just fine in both, current mainline Armada XP (theadorable) and x86 boards (conga-qeval20-qa3-e3845). Here my lsusb output:
I bought the kingston stick and it's not detected on SoCFPGA SoCkit at all. Ouch :-(
[...] Best regards, Marek Vasut
For what it's worth - here is the marking on the OTG chip on the de0_nano_soc:
SMSC 3300-EZK A1515AC13 515AR3A ASETV
OK, that's the standard/recommended USB3300 PHY. I will keep fiddling with the Kingston SE9 USB stick to see what's going on, that's probably some other issue than the cache issue though.
On my third order for an OTG USB mini cable from Amazon, I finally got an actual OTG cable!
On the SoCKit, using this cable the "dwc_otg_core_host_init: Timeout!" messages no longer appear.
A few of the USB sticks I have here are immediately recognized and function normally - a first for me with the latest version of u-boot. For others, like the Kingston SE9 stick, I get the same result as you - it's not detected at all.
I have yet to find one of the non-working USB sticks on the SoCKit fail with the same sort of long timeout followed by the zombie behavior exhibited by the DE0/Kingston combination.
Thanks for checking. I also have the SE9 here and it fails indeed, I will have to look into it later.