Re: [RFC][PATCH 2/5] hi655x-pmic: Make hi655x pmic logic probe child nodes in the dt

From: Lee Jones <lee.jones@linaro.org>
To: John Stultz <john.stultz@linaro.org>
Cc: lkml <linux-kernel@vger.kernel.org>,
	Dmitry Torokhov <dmitry.torokhov@gmail.com>,
	Rob Herring <robh+dt@kernel.org>, Pawel Moll <pawel.moll@arm.com>,
	Mark Rutland <mark.rutland@arm.com>,
	Ian Campbell <ijc+devicetree@hellion.org.uk>,
	Kumar Gala <galak@codeaurora.org>,
	Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org>,
	Wei Xu <xuwei5@hisilicon.com>, Guodong Xu <guodong.xu@linaro.org>,
	Chen Feng <puck.chen@hisilicon.com>
Subject: Re: [RFC][PATCH 2/5] hi655x-pmic: Make hi655x pmic logic probe child nodes in the dt
Date: Thu, 9 Jun 2016 15:43:45 +0100	[thread overview]
Message-ID: <20160609144345.GF2385@dell> (raw)
In-Reply-To: <CALAqxLXiyn1KOvE4C6acU8n-0ANT=WWf0sNxP3-CnuAPybnSEg@mail.gmail.com>

On Wed, 08 Jun 2016, John Stultz wrote:

> On Wed, Jun 8, 2016 at 7:31 AM, Lee Jones <lee.jones@linaro.org> wrote:
> >
> > Please use full buffer width.  Wrapping at 50 chars is not necessary.
> 
> But then how would you be able to review patches on your phone?! :)

Sideways! ;)

> >> --- a/drivers/mfd/hi655x-pmic.c
> >> +++ b/drivers/mfd/hi655x-pmic.c
> >> @@ -39,6 +39,11 @@ static const struct regmap_irq hi655x_irqs[] = {
> >>       { .reg_offset = 0, .mask = RESERVE_INT },
> >>  };
> >>
> >> +static const struct of_device_id of_hi655x_pmic_child_match_tbl[] = {
> >> +     { .compatible = "hisilicon,hi6552-powerkey", },
> >> +     { /* end */ }
> >> +};
> >> +
> >>  static const struct regmap_irq_chip hi655x_irq_chip = {
> >>       .name = "hi655x-pmic",
> >>       .irqs = hi655x_irqs,
> >> @@ -122,6 +127,9 @@ static int hi655x_pmic_probe(struct platform_device *pdev)
> >>
> >>       platform_set_drvdata(pdev, pmic);
> >>
> >> +     /* populate sub nodes */
> >> +     of_platform_populate(np, of_hi655x_pmic_child_match_tbl, NULL, dev);
> >> +
> >
> > Oh, holey poop, no!
> >
> > Please don't mix DT and MFD registration like this.
> >
> > Drivers should contain either of_platform_populate() or
> > mfd_add_devices(), but never both.
> >
> >>       ret = mfd_add_devices(dev, PLATFORM_DEVID_AUTO, hi655x_pmic_devs,
> >>                             ARRAY_SIZE(hi655x_pmic_devs), NULL, 0, NULL);
> >>       if (ret) {
> 
> 
> So apologies for my ignorance here. I'm not quite sure I see what the
> right thing to do. I tried adding a "hi65xx-powerkey" entry in the
> hi655x_pmic_devs, but then mfd_add_devices doesn't seem to provide any
> of the irq info from the DT.
> 
> The only way I can think to make it work w/o calling to populate
> sub-nodes is to pull the powerkey subnode out of the pmic node. It
> seems to work, but I'm not sure if it properly describes the hardware
> then.  Sorry again for being a bit clueless here.

Since we use mfd_add_devices() already, you need to create a cell in
this file for the power-key driver and omit it from DT altogether.  If
your priority is to have each sub-device represented in DT, then you
need to stop using mfd_add_devices() and have of_platform_populate()
do your bidding.

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog