Re: [PATCH v2] pinctrl: baytrail: Clear direct_irq_en flag on broken configs

[Andy Shevchenko <andriy.shevchenko@xxxxxxxxx>]

On Fri, Jan 07, 2022 at 06:41:30PM +0200, Andy Shevchenko wrote:
> On Fri, Jan 07, 2022 at 03:23:43PM +0100, Hans de Goede wrote:

...

> > +static bool byt_direct_irq_sanity_check(struct intel_pinctrl *vg, int pin, u32 value)
> > +{
> > +	void __iomem *reg;
> > +	int i, j;
> > +
> > +	if (!(value & (BYT_TRIG_POS | BYT_TRIG_NEG))) {
> > +		dev_warn(vg->dev,
> > +			 FW_BUG "pin %i: direct_irq_en set without trigger, clearing\n", pin);
> > +		return false;
> > +	}
> > +
> > +	reg = vg->communities->pad_regs + BYT_DIRECT_IRQ_REG;
> 
> > +	for (i = 0; i < 16; i += 4) {
> > +		value = readl(reg + i);
> > +		for (j = 0; j < 4; j++) {
> > +			if (((value >> j * 8) & 0xff) == pin) {
> 
> Can it be like
> 
> 	u32 direct_irq[16];
> 	void __iomem *reg;
> 	void *match;
> 
> 
> 	memcpy_fromio(...);
> 	match = memchr(...);
> 	if (match)
> 		dev_dbg();
> 	else
> 		dev_warn();
> 
> 	return !!match;
> 
> ?
> 
> > +				dev_dbg(vg->dev, "Pin %i: uses direct IRQ %d (APIC %d)\n",

> > +					pin, i + j, 0x43 + i + j);

Why 0x43 is hard coded?

> > +				return true;
> > +			}
> > +		}
> > +	}
> > +
> > +	dev_warn(vg->dev,
> > +		 FW_BUG "pin %i: direct_irq_en set but no IRQ assigned, clearing\n", pin);
> > +	return false;
> > +}

-- 
With Best Regards,
Andy Shevchenko