> >> +     /*
> >> +      * If we did not get one ACK from slave when writing data, we should
> >> +      * dump all registers to check I2C status.
> >
> > Why? I would say no. NACK from a slave can always happen, e.g. when an
> > EEPROM is busy erasing a page.
> 
> For our I2C controller databook, if the master did not get one ACK
> from slave when writing data to salve, we should send one STOP signal
> to abort this data transfer or generate one repeated START signal to
> start one new data transfer cycle. Considering our I2C usage

Yes, so far so good.

> scenarios, we should dump registers to analyze I2C status and notify
> to user to re-start new data transfer.

I disagree here. You notify the users by returning -EIO. The upper layer
(e.g. the i2c client driver) will handle it, like the EEPROM driver
might retry after a while. This all is expected behaviour, so no need to
print the registers to the logfile.

If you really, really want to keep it, make it debug output. But I think
the sentence "we should dump all registers" needs to be rephrased.

> As I explained before, in our Spreadtrum platform, our regulator
> driver will depend on I2C driver and the regulator driver uses
> subsys_initcall() level to initialize. Moreover some other drivers
> like GPU, they will depend on regulator to set voltage and they also
> need initialization much earlier.
> 
> Since it is arch_initcall() level, Andy suggested I should get rid of
> tristate (use bool) and drop module.h here and all leftovers like
> MODULE_*() calls including module_exit().

I see. So the driver is really so essential for proper bootup that it is
not even allowed to be unloaded. I might make an exception here and
allow arch_initcall() then. But I do wonder: did you try deferred
probing all over the place?