On Fri, 27 Apr 2012 12:52:19 +0200, Linus Walleij <[email protected]> wrote:
> On Fri, Apr 27, 2012 at 12:08 PM, Dmitry Eremin-Solenikov
> <[email protected]> wrote:
>
> > Add a driver to use GPIO pins available on several AMD south bridges
> > (currently only AMD 8111 is supported).
> >
> > Signed-off-by: Dmitry Eremin-Solenikov <[email protected]>
>
> Mainly nitpicks, it's looking good overall...
Hi Dimtry,
Do you have an updated version of this patch to address Linus' comments?
g.
>
> > + spin_lock_irqsave(&agp->lock, flags);
> > + temp = ioread8(agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
> > + temp = (temp & 0x1c) | (!!value);
>
> This is a bit idomatic but I think I understand what's going on. However this
> 0x1C magic could be better of with some #define AMD_8111_GPIO_FOO don't
> you think?
>
> > +static int amd_gpio_get(struct gpio_chip *chip, unsigned offset)
> > +{
> > + struct amd_gpio *agp = to_agp(chip);
> > + u8 temp;
> > +
> > + temp = ioread8(agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
> > +
> > + dev_dbg(&agp->pdev->dev, "Getting gpio %d, reg=%02x\n", offset, temp);
> > +
> > + return !!(temp & 0x20);
>
> And this could be #define AMD_8111_GPIO_INVAL or so (I bet this bit has a
> name in the datasheet, right?)
>
> > +static int amd_gpio_dirout(struct gpio_chip *chip, unsigned offset, int value)
> > +{
> > + struct amd_gpio *agp = to_agp(chip);
> > + u8 temp;
> > + unsigned long flags;
> > +
> > + spin_lock_irqsave(&agp->lock, flags);
> > + temp = ioread8(agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
> > + temp = (temp & 0x10) | 0x4 | (!!value);
>
> And so on...
>
> > + iowrite8(temp, agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
> > + spin_unlock_irqrestore(&agp->lock, flags);
> > +
> > + dev_dbg(&agp->pdev->dev, "Dirout gpio %d, value %d, reg=%02x\n", offset, !!value, temp);
> > +
> > + return 0;
> > +}
>
> > +static int __init mod_init(void)
>
> amd_gpio_init() maybe? These names are nice for e.g. ftrace.
>
> > + printk(KERN_INFO "AMD-8111 GPIO detected\n");
>
> pr_info() & friends are nice shorthands for this.
>
> > +static void __exit mod_exit(void)
>
> amd_gpio_mod_exit()?
>
> > +MODULE_AUTHOR("The Linux Kernel team");
>
> Don't be shy, put in your own name :-)
>
> Yours,
> Linus Walleij
--
Grant Likely, B.Sc, P.Eng.
Secret Lab Technologies, Ltd.
On Sat, May 19, 2012 at 2:35 AM, Grant Likely <[email protected]> wrote:
> On Fri, 27 Apr 2012 12:52:19 +0200, Linus Walleij <[email protected]> wrote:
>> On Fri, Apr 27, 2012 at 12:08 PM, Dmitry Eremin-Solenikov
>> <[email protected]> wrote:
>>
>> > Add a driver to use GPIO pins available on several AMD south bridges
>> > (currently only AMD 8111 is supported).
>> >
>> > Signed-off-by: Dmitry Eremin-Solenikov <[email protected]>
>>
>> Mainly nitpicks, it's looking good overall...
>
> Hi Dimtry,
>
> Do you have an updated version of this patch to address Linus' comments?
Colleagues,
I was a little bit busy with job & private topics. I plan to post
updated version
either during the weekend, or on Mon-Wed.
>
> g.
>
>>
>> > + ? ? ? spin_lock_irqsave(&agp->lock, flags);
>> > + ? ? ? temp = ioread8(agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
>> > + ? ? ? temp = (temp & 0x1c) | (!!value);
>>
>> This is a bit idomatic but I think I understand what's going on. However this
>> 0x1C magic could be better of with some #define AMD_8111_GPIO_FOO don't
>> you think?
>>
>> > +static int amd_gpio_get(struct gpio_chip *chip, unsigned offset)
>> > +{
>> > + ? ? ? struct amd_gpio *agp = to_agp(chip);
>> > + ? ? ? u8 temp;
>> > +
>> > + ? ? ? temp = ioread8(agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
>> > +
>> > + ? ? ? dev_dbg(&agp->pdev->dev, "Getting gpio %d, reg=%02x\n", offset, temp);
>> > +
>> > + ? ? ? return !!(temp & 0x20);
>>
>> And this could be #define AMD_8111_GPIO_INVAL or so (I bet this bit has a
>> name in the datasheet, right?)
>>
>> > +static int amd_gpio_dirout(struct gpio_chip *chip, unsigned offset, int value)
>> > +{
>> > + ? ? ? struct amd_gpio *agp = to_agp(chip);
>> > + ? ? ? u8 temp;
>> > + ? ? ? unsigned long flags;
>> > +
>> > + ? ? ? spin_lock_irqsave(&agp->lock, flags);
>> > + ? ? ? temp = ioread8(agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
>> > + ? ? ? temp = (temp & 0x10) | 0x4 | (!!value);
>>
>> And so on...
>>
>> > + ? ? ? iowrite8(temp, agp->pm - PMBASE_OFFSET + PMBASE_GPIO(offset));
>> > + ? ? ? spin_unlock_irqrestore(&agp->lock, flags);
>> > +
>> > + ? ? ? dev_dbg(&agp->pdev->dev, "Dirout gpio %d, value %d, reg=%02x\n", offset, !!value, temp);
>> > +
>> > + ? ? ? return 0;
>> > +}
>>
>> > +static int __init mod_init(void)
>>
>> amd_gpio_init() maybe? These names are nice for e.g. ftrace.
>>
>> > + ? ? ? printk(KERN_INFO "AMD-8111 GPIO detected\n");
>>
>> pr_info() & friends are nice shorthands for this.
>>
>> > +static void __exit mod_exit(void)
>>
>> amd_gpio_mod_exit()?
>>
>> > +MODULE_AUTHOR("The Linux Kernel team");
>>
>> Don't be shy, put in your own name :-)
>>
>> Yours,
>> Linus Walleij
>
> --
> Grant Likely, B.Sc, P.Eng.
> Secret Lab Technologies, Ltd.
--
With best wishes
Dmitry