[Intel-gfx] [PATCH 6/8] drivers/pwm: Add Crystalcove (CRC) PWM driver

[Paul Bolle]

[Added Paul Gortmaker.]

Hi Shobhit,

On Fri, 2015-06-19 at 12:16 +0530, Shobhit Kumar wrote:
> So what is the exact big problem with this ?

The main problem I have is that it's hard to read a submitter's mind.
And, I think, in cases like this we need to know if the submitter just
made some silly mistake or that the mismatch (between Kconfig type and
code) was intentional. So each time such a mismatch is spotted the
submitter ought to be asked about it.

(I'd guess that one or two new drivers are submitted _each_ day. And
these mismatches are quite common. I'd say I receive answers like:
- "Oops, yes bool should have been tristate"; or
- "Oops, forgot to clean up after noticing tristate didn't work"; or
- "I just copy-and-pasted a similar driver, the module stuff isn't
  actually needed"
at least once a week. Not sure, I don't keep track of this stuff.)

Furthermore, it appears that Paul Gortmaker is on a mission to, badly
summarized, untangle the module and init code. See for instance
https://lkml.org/lkml/2015/5/28/809 and
https://lkml.org/lkml/2015/5/31/205 .

Now, I don't know whether (other) Paul is bothered by these MODULE_*
macros. But Paul did submit a series that adds
builtin_platform_driver(), see https://lkml.org/lkml/2015/5/10/131 .
That new macro ensures built-in only code doesn't have to use
module_platform_driver(), which your patch also uses. So perhaps Paul
can explain some of the non-obvious issues caused by built-in only code
using module specific constructs.

> I can anyway shove out these macros to end the discussion.

I'd rather convince you than annoy you into doing as I suggested.

> BTW whether you  buy the argument or not, please do treat yourself
> with ice cream for being able to make such a comment.

Will do.

Thanks,


Paul Bolle