Re: [PATCH] iio: dummy: evgen: use irq_sim

From: Bartosz Golaszewski <brgl@bgdev.pl>
To: Lars-Peter Clausen <lars@metafoo.de>
Cc: Jonathan Cameron <jic23@kernel.org>,
	Hartmut Knaack <knaack.h@gmx.de>,
	Peter Meerwald-Stadler <pmeerw@pmeerw.net>,
	linux-iio@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] iio: dummy: evgen: use irq_sim
Date: Thu, 28 Sep 2017 12:55:06 +0200	[thread overview]
Message-ID: <CAMRc=MdR+osCmb0mG5a53Ggk-s8HQoJ4x7VCvC8cfwqNC1EfvQ@mail.gmail.com> (raw)
In-Reply-To: <d8279093-c98a-20d9-ada4-8995243f4480@metafoo.de>

2017-09-28 10:04 GMT+02:00 Lars-Peter Clausen <lars@metafoo.de>:
> On 09/27/2017 09:23 PM, Bartosz Golaszewski wrote:
>>>>               return ret;
>>>> -     }
>>>> -     iio_evgen->chip.name = iio_evgen_name;
>>>> -     iio_evgen->chip.irq_mask = &iio_dummy_event_irqmask;
>>>> -     iio_evgen->chip.irq_unmask = &iio_dummy_event_irqunmask;
>>>> -     for (i = 0; i < IIO_EVENTGEN_NO; i++) {
>>>> -             irq_set_chip(iio_evgen->base + i, &iio_evgen->chip);
>>>> -             irq_set_handler(iio_evgen->base + i, &handle_simple_irq);
>>>> -             irq_modify_status(iio_evgen->base + i,
>>>> -                               IRQ_NOREQUEST | IRQ_NOAUTOEN,
>>>> -                               IRQ_NOPROBE);
>>>> -     }
>>>> -     init_irq_work(&iio_evgen->handler.work, iio_dummy_work_handler);
>>>> +
>>>> +     iio_evgen->base = irq_sim_irqnum(&iio_evgen->irq_sim, 0);
>>> I saw you introduced irq_sim_baseirq(), to get rid of ->base. But as far as
>>> I can see the only remaining places where we need the base is to do the
>>> reverse lookup from IRQ to index. It would be nice if the irq_sim had a
>>> function for that, then we wouldn't have to know about the base at all.
>>>
>>
>> I'm not sure I understand. Irq sim doesn't know anything about iio
>> data structures, so how would such a reverse lookup work in this case?
>
> Reverse lookup in the sense of translating IRQ number to offset. All we ever
> do with the base in the IIO code is `irq - base` to get the offset. I'd hide
> that calculation in a helper in the irq_sim code. This nicely splits
> functionality and implementation, the IIO code doesn't have to know how to
> get offset from the IRQ.

I see. Sounds like a good improvement for 4.16. In the meantime, I'll
send v2 with the missing kfree() added.

Thanks,
Bartosz