Message-ID: <4114007.EkZL1biBx8@debian64>
Date:	Thu, 12 May 2016 14:07:24 +0200
From:	Christian Lamparter <chunkeey@...glemail.com>
To:	Alexandre Courbot <gnurou@...il.com>
Cc:	linux-gpio@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org,
	Álvaro Fernández Rojas <noltari@...il.com>,
	Alexander Shiyan <shc_work@...l.ru>,
	Linus Walleij <linus.walleij@...aro.org>,
	Andy Shevchenko <andy.shevchenko@...il.com>,
	Joachim Eastwood <manabian@...il.com>
Subject: Re: [PATCH v9 1/2] gpio: mmio: add DT support for memory-mapped GPIOs

On Thursday, May 12, 2016 07:26:32 PM Alexandre Courbot wrote:
> On Wednesday, May 11, 2016 6:34:34 PM JST, Christian Lamparter wrote:
> > From: Álvaro Fernández Rojas <noltari@...il.com>
> >
> > This patch adds support for defining memory-mapped GPIOs which
> > are compatible with the existing gpio-mmio interface. The generic
> > library provides support for many memory-mapped GPIO controllers
> > that are found in various on-board FPGA and ASIC solutions that
> > are used to control board's switches, LEDs, chip-selects,
> > Ethernet/USB PHY power, etc.
> >
> > For setting GPIO's there are three configurations:
> 
> s/GPIO's/GPIOs
OK

> > 	1. single input/output register resource (named "dat"),
> > 	2. set/clear pair (named "set" and "clr"),
> > 	3. single output register resource and single input resource
> > 	   ("set" and dat").
> >
> > The configuration is detected by which resources are present.
> > For the single output register, this drives a 1 by setting a bit
> > and a zero by clearing a bit.  For the set clr pair, this drives
> > a 1 by setting a bit in the set register and clears it by setting
> > a bit in the clear register. The configuration is detected by
> > which resources are present.
> 
> The last sentence of this paragraph repeats for first one.
Ok

> >
> > For setting the GPIO direction, there are three configurations:
> > 	a. simple bidirectional GPIOs that requires no configuration.
> > 	b. an output direction register (named "dirout")
> > 	   where a 1 bit indicates the GPIO is an output.
> > 	c. an input direction register (named "dirin")
> > 	   where a 1 bit indicates the GPIO is an input.
> >
> > The first user for this binding is "wd,mbl-gpio".
> >
> > Reviewed-by: Andy Shevchenko <andy.shevchenko@...il.com>
> > Signed-off-by: Álvaro Fernández Rojas <noltari@...il.com>
> > Signed-off-by: Christian Lamparter <chunkeey@...glemail.com>
> > ---
> >  static void bgpio_write8(void __iomem *reg, unsigned long data)
> >  {
> > @@ -569,6 +571,58 @@ static void __iomem *bgpio_map(struct 
> > platform_device *pdev,
> >  	return devm_ioremap_resource(&pdev->dev, r);
> >  }
> >  
> > +#ifdef CONFIG_OF
> > +static int bgpio_basic_mmio_parse_dt(struct platform_device *pdev,
> > +				     struct bgpio_pdata *pdata,
> > +				     unsigned long *flags)
> > +{
> > +	struct device *dev = &pdev->dev;
> > +
> > +	pdata->base = -1;
> > +
> > +	if (of_property_read_bool(dev->of_node, "no-output"))
> > +		*flags |= BGPIOF_NO_OUTPUT;
> 
> I don't think it is a good idea to add "generic" properties. Whether a 
> controller is capable of output or not should be determined by its 
> compatible string only, and not a vague property.

Well, meet the gpios on the MBL. If you want to figure out how WD wired
up these two GPIOs (one is input, the other output), I can sent you a
built-yourself MBL kit. It just needs a 3,5" drive and a 12V 2A power
plug with a standard 5.5mm plug.

> > +
> > +	return 0;
> > +}
> > +
> > +static const struct of_device_id bgpio_of_match[] = {
> > +	{ .compatible = "wd,mbl-gpio", .data = bgpio_basic_mmio_parse_dt },
> 
> Mmm cannot you determine whether your controller is capable of output or 
> not just from the compatible property here? If so, the 
> bgpio_basic_mmio_parse_dt seems to be unneeded. If not, then this is 
> dependent on the wd,mbl-gpio binding and should be renamed accordingly.
Sadly I don't know of any method. The device has two GPIOs one at 0x4e0000000.
The other one is at 0x4e0100000. The address tells me that there are two 
external chips connected to the EBC (memory bank - RAM, ROM and DMA chips
go here according to IBM's documentations). Which is not the place you
would expect peripherals. 

> > @@ -646,6 +709,7 @@ MODULE_DEVICE_TABLE(platform, bgpio_id_table);
> >  static struct platform_driver bgpio_driver = {
> >  	.driver = {
> >  		.name = "basic-mmio-gpio",
> > +		.of_match_table = of_match_ptr(bgpio_of_match),
> >  	},
> >  	.id_table = bgpio_id_table,
> >  	.probe = bgpio_pdev_probe,
> 
> It seems to me that this patch does two things:
> 
> 1) Add code to support device tree lookup
> 2) Add support for "wd,mbl-gpio".
> 
> If true, these two things should be in their own patches. You should also 
> have another patch that adds the DT bindings for "wd,mbl-gpio", so I would 
> do things in that order:

The DT bindings have been merged. That's why I dropped it from the rebase.

> 1/3: DT support for basic-mmio-gpio
Sadly, adding the "basic-mmio-gpio" binding is not possible without a ACK from
the device tree maintainers. They have voiced their concerns. I think this was
your post of the discussion on it:
<http://www.spinics.net/lists/devicetree/msg124613.html>

That's why the series was updated around v5 and v6 to use the "wd,mbl-gpio"
binding.

So yes, I wanted to go this route in the beginning as well. But no go.
If we find more devices we could have a "basic-mmio-gpio" class. But
for now, we have to start somewhere.

> 2/3: DT bindings for "wd,mbl-gpio" (and have them validated by the DT 
> people - e.g. do you really need a "reg" property or is it here just to fit 
> with what bgpio_pdev_probe expects? More about this on 2/2)
> 3/3: Support for "wd,mbl-gpio" in basic-mmio-gpio
Yes, that would have been nice. And I agree it was the way to do it. But
without the wd,mbl-gpio mapping I would add the bgpio_parse_dt function
without any caller. (As I can't add the compatible = "basic-mmio-gpio", ...)

Regards,
Christian

Powered by blists - more mailing lists