linux-kernel - Re: [RFC PATCH RESEND 3/3] pinctrl: upboard: Add UP2 pinctrl and gpio driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 26 Apr 2018 03:38:36 +0100
From:   Javier Arteaga <javier@...tex.com>
To:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc:     Linus Walleij <linus.walleij@...aro.org>,
        Dan O'Donovan <dan@...tex.com>,
        Mika Westerberg <mika.westerberg@...ux.intel.com>,
        Heikki Krogerus <heikki.krogerus@...ux.intel.com>,
        Lee Jones <lee.jones@...aro.org>,
        Jacek Anaszewski <jacek.anaszewski@...il.com>,
        Pavel Machek <pavel@....cz>, linux-gpio@...r.kernel.org,
        linux-leds@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH RESEND 3/3] pinctrl: upboard: Add UP2 pinctrl and
 gpio driver

On Wed, Apr 25, 2018 at 07:49:12PM +0300, Andy Shevchenko wrote:
> > For reference, here's the relevant ASL from the UP2 platform
> > controller.
> 
> It should be in Documentation file or in commit message.

Perfect, I just wasn't sure whether dumping ASL this way would be
acceptable in commit history. In that case, I think I prefer the commit
body over keeping a Documentation/ file elsewhere that's easy to miss.

> >  static const struct mfd_cell upboard_up2_mfd_cells[] = {
> > +	{ .name = "upboard-pinctrl" },
> 
> I guess it should be 3 lines.

> 
> >  	UPBOARD_LED_CELL(upboard_up2_led_data, 0),
> >  	UPBOARD_LED_CELL(upboard_up2_led_data, 1),
> >  	UPBOARD_LED_CELL(upboard_up2_led_data, 2),
> 
> ...and honestly I would not use this macro and put 4 cells explicitly
> here.

The idea was to avoid repeating .platform_data + .pdata_size over and
over as there's a few LEDs per board, but explicit is good too.

May just go with your suggestion as that seems to be more popular than
macros in other mfd drivers.

> > +static int upboard_gpio_request_enable(struct pinctrl_dev *pctldev,
> > +				       struct pinctrl_gpio_range
> > *range,
> > +				       unsigned int pin)
> > +{
> > +	const struct pin_desc * const pd = pin_desc_get(pctldev,
> > pin);
> > +	const struct upboard_pin *p;
> > +	int ret;
> > +
> 
> > +	if (!pd)
> > +		return -EINVAL;
> 
> When it's possible?

Just reread pin_request() from pinctrl core. You're right, it isn't.
I'll take out the check.

> > +	if (!pd)
> > +		return -EINVAL;
> 
> Ditto.

As above.

> 
> > +	struct upboard_pinctrl *pctrl =
> > +		container_of(gc, struct upboard_pinctrl, chip);
> 
> Do define and use to_upboard_pinctrl().

Will do.

> > +	if (offset + 1 > pctrl->nsoc_gpios || !pctrl-
> > >soc_gpios[offset])
> > +		return ERR_PTR(-ENODEV);
> 
> When this is a case?

Another unnecessary check if gpiolib already guarantees valid inputs.

> > +static int upboard_gpio_get_direction(struct gpio_chip *gc, unsigned
> > int offset)
> > +{
> > +	struct gpio_desc *desc = upboard_offset_to_soc_gpio(gc,
> > offset);
> > +
> 
> Split above to definition and assignment pieces. Put assignment
> immediately before condition.
> 
> > +	if (IS_ERR(desc))
> > +		return PTR_ERR(desc);

I'll do that.

> > +static struct regmap_field * __init upboard_field_alloc(struct device
> > *dev,
> > +							struct regmap
> > *regmap,
> > +							unsigned int
> > base,
> > +							unsigned int
> > number)
> 
> You should really understand what __init means and when it's appropriate
> to use it.

As in the other email: the intention was to drop probe() and funcs only
used on module init, but I appear to have misunderstood something here.

> > +static int __init upboard_pinctrl_probe(struct platform_device *pdev)
> > +{
> > +	struct acpi_device * const adev = ACPI_COMPANION(&pdev->dev);
> 
> Huh, const in that place? Why?

You're right, it isn't a usual pattern. Dropped.

> > +	if (!pdev->dev.parent)
> > +		return -EINVAL;
> > +
> > +	upboard = dev_get_drvdata(pdev->dev.parent);
> > +	if (!upboard)
> > +		return -EINVAL;
> 
> Same comment as per LED driver.

I'll address that too.

> > +	if (strcmp(acpi_device_hid(adev), "AANT0F01"))
> > +		return -ENODEV;
> 
> Huh?

Ugh, this is left over from a bit of code that selected the right
pinctrl_desc* for each UP HID. Of course, it doesn't make sense now.

I'll take that out.

> > +		((struct pinctrl_pin_desc *)pd)->drv_data = pin;
> 
> What is that?! I mean ugly casting.

I agree, it's an eyesore. The intention was to drop const from
pinctrl_desc->pins as we're replacing the pins' drv_data on init.

It does look like I'm going at this the wrong way though. I'll take
another stab at it (pointers welcome of course).

> > +MODULE_LICENSE("GPL");
> 
> License mismatch.

Will fix.

Thanks again for your time Andy! I really appreciate your help :)