RE: [PATCH V3 1/7] firmware: imx: Add stubs for !CONFIG_IMX_SCU case

[Anson Huang <anson.huang@nxp.com>]

> Subject: Re: [PATCH V3 1/7] firmware: imx: Add stubs for !CONFIG_IMX_SCU
> case
> 
> On 3/16/20 7:18 PM, Peng Fan wrote:
> >> Subject: RE: [PATCH V3 1/7] firmware: imx: Add stubs for
> >> !CONFIG_IMX_SCU case
> >>
> >> Hi, Alexandre
> >>
> >>> -----Original Message-----
> >>> From: Alexandre Belloni <alexandre.belloni@bootlin.com>
> >>> Sent: Monday, March 16, 2020 5:16 PM
> >>> To: Anson Huang <anson.huang@nxp.com>
> >>> Cc: Shawn Guo <shawnguo@kernel.org>; Peng Fan
> <peng.fan@nxp.com>;
> >>> s.hauer@pengutronix.de; kernel@pengutronix.de; festevam@gmail.com;
> >>> dmitry.torokhov@gmail.com; a.zummo@towertech.it;
> >> rui.zhang@intel.com;
> >>> daniel.lezcano@linaro.org; amit.kucheria@verdurent.com; wim@linux-
> >>> watchdog.org; linux@roeck-us.net; Daniel Baluta
> >>> <daniel.baluta@nxp.com>; gregkh@linuxfoundation.org;
> >>> linux@rempel-privat.de; tglx@linutronix.de; m.felsch@pengutronix.de;
> >>> andriy.shevchenko@linux.intel.com;
> >>> arnd@arndb.de; ronald@innovation.ch; krzk@kernel.org;
> >>> robh@kernel.org; Leonard Crestez <leonard.crestez@nxp.com>; Aisheng
> >>> Dong <aisheng.dong@nxp.com>; linux-arm-kernel@lists.infradead.org;
> >>> linux- kernel@vger.kernel.org; linux-input@vger.kernel.org; linux-
> >>> rtc@vger.kernel.org; linux-pm@vger.kernel.org; linux-
> >>> watchdog@vger.kernel.org; dl-linux-imx <linux-imx@nxp.com>
> >>> Subject: Re: [PATCH V3 1/7] firmware: imx: Add stubs for
> >>> !CONFIG_IMX_SCU case
> >>>
> >>> On 16/03/2020 09:08:53+0000, Anson Huang wrote:
> >>>>> Why do you absolutely need to compile them independently? From a
> >>>>> code coverage point of view, having:
> >>>>>
> >>>>> COMPILE_TEST=y
> >>>>> CONFIG_IMX_SCU=y
> >>>>>
> >>>>> is enough to select and compile the remaining drivers.
> >>>>
> >>>> What I meant is for below case, like using other arch config which
> >>>> does NOT have CONFIG_IMX_SCU selected, ONLY with COMPILE_TEST
> >>>> selected, adding stubs for IMX_SCU APIs can fix such scenario.
> >>>>
> >>>> COMPILE_TEST=y
> >>>> CONFIG_IMX_SCU=n
> >>>>
> >>>
> >>> Why is that an issue? If they don't have IMX_SCU selected, then the
> >>> other SCU driver are not selected either, having stubs doesn't
> >>> change that you will have to select at least one option. Please
> >>> explain what is the issue that is not solved here.
> >>
> >> OK, what I thought is even without IMX_SCU selected, other SCU
> >> drivers still can be selected for build test after adding
> >> "COMPILE_TEST" to the kconfig, like below, if without IMX_SCU API
> >> stubs, the "COMPILE_TEST" can NOT be added to SCU drivers to enable
> >> build test, so I think the IMX_SCU API stubs should be added?
> >
> > Forgot to mention, without stub api, for drivers with " #include
> > <linux/firmware/imx/sci.h> " will met compile error without
> > +#ifdef CONFIG_IMX_SCU
> > +#endif
> >
> > So we have to use ifdef CONFIG_IMX_SCU to guard the include.
> >
> Add "depends on IMX_SCU" to the Kconfig entry for those drivers, and/or
> drop "COMPILE_TEST" from their Kconfig entry.
> 
> Really, COMPILE_TEST is abused here. I start to understand those who
> advocate that it should be removed. This is an excellent case in point.

Yup, COMPILE_TEST should ONLY be added to those independent drivers,
those drivers with dependency on "core" driver should NOT have it added.
SCU drivers are similar.

Anson