linux-kernel - Re: [PATCH v3 1/2] gpio: sch: Add edge event support

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] [day] [month] [year] [list]

Date:   Mon, 9 Dec 2019 18:36:58 +0200
From:   Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To:     Jan Kiszka <jan.kiszka@...mens.com>
Cc:     Linus Walleij <linus.walleij@...aro.org>,
        Bartosz Golaszewski <bgolaszewski@...libre.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-gpio@...r.kernel.org,
        Mika Westerberg <mika.westerberg@...ux.intel.com>,
        ACPI Devel Maling List <linux-acpi@...r.kernel.org>,
        "Rafael J . Wysocki" <rafael.j.wysocki@...el.com>
Subject: Re: [PATCH v3 1/2] gpio: sch: Add edge event support

On Fri, Nov 22, 2019 at 04:33:05PM +0100, Jan Kiszka wrote:
> On 22.11.19 12:12, Andy Shevchenko wrote:
> > On Wed, Nov 20, 2019 at 08:20:13PM +0100, Jan Kiszka wrote:

> > > +	switch (type & IRQ_TYPE_SENSE_MASK) {
> > > +	case IRQ_TYPE_EDGE_RISING:
> > > +		rising = 1;
> > > +		break;
> > > +	case IRQ_TYPE_EDGE_FALLING:
> > > +		falling = 1;
> > > +		break;
> > > +	case IRQ_TYPE_EDGE_BOTH:
> > > +		rising = 1;
> > > +		falling = 1;
> > > +		break;
> > > +	default:
> > > +		return -EINVAL;
> > > +	}

> > Won't we need to set up IRQ handler here and use handle_bad_irq() during
> > initialization phase?
> 
> Why? This is just defining the edge type, not whether an interrupt could be
> generated or not. Also, we only have edge events here, so no reason to
> switch types.

OK.

> > > +	irq_base = devm_irq_alloc_descs(&pdev->dev, -1, 0, sch->chip.ngpio,
> > > +					NUMA_NO_NODE);
> > > +	if (irq_base < 0)
> > > +		return irq_base;
> > > +	sch->irq_base = irq_base;
> > > +
> > > +	gc = devm_irq_alloc_generic_chip(&pdev->dev, "sch_gpio", 1, irq_base,
> > > +					 NULL, handle_simple_irq);
> > > +	if (!gc)
> > > +		return -ENOMEM;
> > > +
> > > +	gc->private = sch;
> > > +	ct = gc->chip_types;
> > > +
> > > +	ct->chip.irq_mask = sch_irq_mask;
> > > +	ct->chip.irq_unmask = sch_irq_unmask;
> > > +	ct->chip.irq_set_type = sch_irq_type;
> > > +
> > > +	ret = devm_irq_setup_generic_chip(&pdev->dev, gc,
> > > +					  IRQ_MSK(sch->chip.ngpio),
> > > +					  0, IRQ_NOREQUEST | IRQ_NOPROBE, 0);
> > > +	if (ret)
> > > +		return ret;
> > 
> > Shan't we do this in the (similar) way how it's done in pinctrl-cherryview.c
> > driver? (Keep in mind later patches which are going to be v5.5)
> > 
> 
> Can you be a bit more specific for me? Do you mean the pattern
> gpiochip_irqchip_add / gpiochip_set_chained_irqchip? What would be the
> difference / benefit? And how would I link sch_sci_handler to that pattern?

Now we have struct irq_chip is part of GPIO chip, so, we may use it and supply
needed callbacks and settings before calling gpiochip_add_data().

Will it work in this case?

-- 
With Best Regards,
Andy Shevchenko