From: Steve Twiss <stwiss.opensource@diasemi.com>
To: Lee Jones <lee.jones@linaro.org>
CC: LINUX-KERNEL <linux-kernel@vger.kernel.org>,
        DEVICETREE <devicetree@vger.kernel.org>,
        Dmitry Torokhov <dmitry.torokhov@gmail.com>,
        Eduardo Valentin <edubezval@gmail.com>,
        Guenter Roeck <linux@roeck-us.net>,
        LINUX-INPUT <linux-input@vger.kernel.org>,
        LINUX-PM <linux-pm@vger.kernel.org>,
        LINUX-WATCHDOG <linux-watchdog@vger.kernel.org>,
        Liam Girdwood <lgirdwood@gmail.com>, Mark Brown <broonie@kernel.org>,
        "Mark Rutland" <mark.rutland@arm.com>,
        Rob Herring <robh+dt@kernel.org>,
        "Support Opensource" <Support.Opensource@diasemi.com>,
        Wim Van Sebroeck <wim@iguana.be>, Zhang Rui <rui.zhang@intel.com>
Subject: RE: [PATCH V3 5/9] mfd: da9061: MFD core support
Thread-Topic: [PATCH V3 5/9] mfd: da9061: MFD core support
Thread-Index: AQHSM5IFFMHFxbniuEmfO9QmgeGvmaDFw7wAgAGADMA=
Date: Mon, 7 Nov 2016 15:25:00 +0000
Message-ID: <6ED8E3B22081A4459DAC7699F3695FB7018CCE7CB6@SW-EX-MBX02.diasemi.com>
References: <cover.1477929725.git.stwiss.opensource@diasemi.com>
 <72145ea5055e95d7b6f05168c641927e4f3830d9.1477929725.git.stwiss.opensource@diasemi.com>
 <20161102142854.GS13127@dell>
In-Reply-To: <20161102142854.GS13127@dell>
Accept-Language: en-GB, en-US
Content-Language: en-US
x-kse-attachmentfiltering-interceptor-info: protection disabled
x-kse-serverinfo: sw-ex-cashub01.diasemi.com, 9
x-kse-antivirus-interceptor-info: scan successful
x-kse-antivirus-info: Clean, bases: 07/11/2016 13:28:00
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from base64 to 8bit by mail.home.local id uA7FPEnM032587
Content-Length: 2283
Lines: 68

On 02 November 2016 14:29, Lee Jones wrote:

> On Mon, 31 Oct 2016, Steve Twiss wrote:
> > From: Steve Twiss <stwiss.opensource@diasemi.com>
> > 
> > @@ -475,7 +855,25 @@ static int da9062_i2c_probe(struct i2c_client *i2c,
> >  		return -EINVAL;
> >  	}
> >
> > -	chip->regmap = devm_regmap_init_i2c(i2c, &da9062_regmap_config);
> > +	switch (chip->chip_type) {
> > +	case(COMPAT_TYPE_DA9061):
> > +		cell = da9061_devs;
> > +		cell_num = ARRAY_SIZE(da9061_devs);
> > +		irq_chip = &da9061_irq_chip;
> > +		config = &da9061_regmap_config;
> > +		break;
> > +	case(COMPAT_TYPE_DA9062):
> > +		cell = da9062_devs;
> > +		cell_num = ARRAY_SIZE(da9062_devs);
> > +		irq_chip = &da9062_irq_chip;
> > +		config = &da9062_regmap_config;
> > +		break;
> > +	default:
> > +		dev_err(chip->dev, "Unrecognised chip type\n");
> > +		return -ENODEV;
> > +	}
> 
> I very much dislike when MFD and OF functionality is mixed.
> 
> In your case you can use da9062_get_device_type() to dynamically
> interrogate the device and register using the correct MFD cells that
> way.

Hi Lee,

It's the device tree that decides what the chip type is. It's not chip
interrogation in this case. The ordering dictates this I think: to access the
hardware ID register, a regmap definition is needed first. But because the
correct I2C register map requires a knowledge of what chip is being used,
it becomes a circular dependency.

To solve this dependency, I define the chip type (DA9061 or DA9062) in the
device tree and assign the correct regmap first before accessing any registers.

> > +	chip->regmap = devm_regmap_init_i2c(i2c, config);
> >  	if (IS_ERR(chip->regmap)) {
> >  		ret = PTR_ERR(chip->regmap);
> >  		dev_err(chip->dev, "Failed to allocate register map: %d\n",
> > @@ -493,7 +891,7 @@ static int da9062_i2c_probe(struct i2c_client *i2c,
> >
> >  	ret = regmap_add_irq_chip(chip->regmap, i2c->irq,
> >  			IRQF_TRIGGER_LOW | IRQF_ONESHOT |IRQF_SHARED,
> > -			-1, &da9062_irq_chip,
> > +			-1, irq_chip,
> 
> What is -1?

.. it's a request for an irq_base.
http://lxr.free-electrons.com/source/kernel/irq/irqdesc.c#L477

Is there a reason I shouldn't be doing that?
There doesn't seem to be a #define anywhere, and using -1 seems
to be the standard in the kernel at the moment.

Regards,
Steve