On Mon, Feb 13, 2012 at 4:46 AM, Patil, Rachna rachna@ti.com wrote:
Hi Govindraj,
On Mon, Feb 13, 2012 at 14:29:56, Govindraj wrote:
Hi Peter,
On Mon, Feb 13, 2012 at 1:52 PM, Peter Meerwald pmeerw@pmeerw.net wrote:
Hello,
the recent OMAP USB changes (43b62393da6dfdd7f503d1b37344463a05ea47b5, 29321c0518d99494ab2a250e5b4f75b3add83b11) fail on beagleboard-xm rev. C, see below
The errors have nothing to do with ehci patch merges, (I had tested it on beagle rev c board before posting it out)
Looking into it closely the below two patches are causing this issue:
[...] commit 498cbdfe62a8330f6c89765bdd15e60328a26511 Author: Patil, Rachna rachna@ti.com Date: Sun Jan 22 23:46:23 2012 +0000
ARM: AM33XX: Add AM33XX I2C driver support
commit 2faa76196af4b3e93bcb9e38ed9090cbd3b06db3 Author: Patil, Rachna rachna@ti.com Date: Sun Jan 22 23:44:12 2012 +0000
ARM: I2C: I2C Multi byte address support [...]
Reverting these two patches I don't see these errors on my beagle rev c board logs as in here [1]
I have tested these patches on AM335x EVM, AM335x Beaglebone and OMAP35xx. I did not face any issues w.r.t these EVM's. I tried i2c probe, read/write and all of it worked as expected. Also from the logs in [1] I still see an I2C error " wait_for_bb: I2C_STAT=1000", which is bus busy indication.
To be clear, I assume you're looking for one of the omap37xx platforms to see why the busy changes have broken it. Thanks.