On Fri, Sep 12, 2014 at 12:36:45PM -0700, Jacob Pan wrote: > On Fri, 12 Sep 2014 17:18:24 +0200 > Maxime Ripard wrote: > > > On Thu, Sep 11, 2014 at 04:15:52PM -0700, Jacob Pan wrote: > > > XPowers AXP288 is a customized PMIC found on some Intel Baytrail-CR > > > platforms. It comes with sub-functions such as USB charging, fuel > > > gauge, ADC, and many LDO and BUCK channels. > > > > > > By extending the existing AXP20x driver, this patchset adds basic > > > support for AXP288 PMIC with GPADC as one MFD cell device driver. > > > It also adds hooks for ACPI opregion handler driver which can be > > > used to handle ACPI requests. > > > > > > Currently, the PMIC driver in this patchset does not support > > > platform data enumeration. But when ACPI _DSD and unified device > > > properties become available, cell devices with platform data will > > > be added. > > > > > > This patch does not use intel_soc_pmic core for i2c and regmap > > > handling in that axp288 shares similar programming interface with > > > other Xpower PMICs supported in axp20x.c. Therefore, extending > > > axp20x.c to include axp288 makes more sense. > > > > > > Changes > > > v3: - put all file rename changes in 1/5 > > > > The variables renaming are still not in 1/5.... > > > 1/5 is for file rename such that the follow up patches are more > readable. Which is exactly my point. So why don't you apply it to the variable renames as well? > There are so many details in variable rename, I think it > belongs to the patch that expands the new device support. This has nothing to do in this patch. Remember that one patch should do one thing. You're obviously doing 2 in the second patch, and just like you pointed out, the renaming just make the whole thing less readable. Maxime -- Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com