On 19-Jun-17 17:05, Andy Shevchenko wrote:

On Mon, Jun 19, 2017 at 6:58 PM, Wolfram Sang [off-list ref] wrote:

quoted

quoted

Believe me I've looked for mistakes like this but I've read the code so many
times I don't think I can see the obvious anymore.

Yes, I know this. And you worked hard on this slave feature,
acknowledged.

Patches 1-4 look good to me from what I glimpsed. I largely trust here
the *much* appreciated review from Jarkko and Andy. Thanks a lot, guys!

I wonder if we haven't reached a state where Luis just could fix the
buildbot error (missing 'select I2C_SLAVE') and the thing pointed out by
Andy and we handle further small fixes incrementally during the v4.13
cycle? AFAICS there is no major show-stopper, or am I wrong?

I would go with the following plan:
1. Push 1-4
2. Resend 5-6 with addressed pointed issues for one more (fast) round

For the Kbuild error I fixed it by changing

ifneq ($(CONFIG_I2C_DESIGNWARE_SLAVE),n)
i2c-designware-core-objs += i2c-designware-slave.o

to:

ifeq ($(CONFIG_I2C_DESIGNWARE_SLAVE),y)
i2c-designware-core-objs += i2c-designware-slave.o

I have test it with the .config provided and it works now.