On Sat, Jun 3, 2017 at 8:53 PM, Sathyanarayanan Kuppuswamy Natarajan
[off-list ref] wrote:

On Sat, Jun 3, 2017 at 10:32 AM, Andy Shevchenko
[off-list ref] wrote:

quoted

On Sat, Jun 3, 2017 at 8:28 PM, Sathyanarayanan Kuppuswamy Natarajan
[off-list ref] wrote:

quoted

Let's assume we applied patches 1 and 2, does it mean the regression
to the existing behaviour?

Yes, Just applying patch 1 & 2 and not apply patch 3 will create
regression in thermal driver.

Since there is no compilation dependency between patch 2 & 3 and they
are from two different domains, I submitted them separately to make it
easier for review.

And this is exactly bisectability (level 2 :-) ) issue.
Basically you confirmed you need to squash them like I said.

quoted

(For example, device worked improperly in some cases -> stopped
working completely)

-- 
With Best Regards,
Andy Shevchenko