Hi,
Sorry for huge delay in answering. Unfortunately we don't have enough 
resources now to prepare clean enough patch to be accepted by community.
Please find the latest version attached.  Driver has passed stress 
tests, but looks like it need seriuos refactoring (it is unnecessarily 
complicated).
We still have polling in slave code. Since imx doesn't generate 
interrupt on "bus busy"/"bus free" events we have to test IBB bit in 
polling loop.
Comparing to v2 version several race conditions were fixed (bus locking 
by slave, breaking slave transaction by starting master xfer). v2 is 
working pretty good in slave-only or master-only mode. It is reasonable 
to add  slave locking test  - sometimes imx6 hw doesn't release data 
line. As workaround we use dummy reading of IMX_I2C_I2DR if driver is 
in  slave mode and IBB bit is set for a long time.

Thanks,
Maxim Syrchin


27.10.2016 10:31, Frkuska, Joshua пишет:
> Hi Maxim, Dmitriy, Wolfram,
>
> If there is no immediate plan for a third release of the below patch 
> set, would it be possible to continue with merging v2 after addressing 
> the remaining concerns?
>
>
> Thank you and regards,
>
> Joshua
>> Hi Maxim,
>>
>> On 2016-03-04 11:06:10 in the thread "Re: [PATCH] i2c: imx: add slave 
>> support. v2"
>> referenced here:   https://patchwork.ozlabs.org/patch/573353/ you said:
>>> Hi Wolfram,
>>> I'm now working on creating new driver version. I think I'll be able to
>>> sent it soon.
>> Do you still plan to release a driver update for an i2c imx driver 
>> slave support?
>>
>> Best regards,
>> Jim Baxter
>>