Re: [PATCH 20/51] Input: atmel_mxt_ts - Set default irqflags when there is no pdata

From: Yufeng Shen <miletus@chromium.org>
To: Nick Dyer <nick.dyer@itdev.co.uk>
Cc: Benson Leung <bleung@chromium.org>,
	Dmitry Torokhov <dmitry.torokhov@gmail.com>,
	Henrik Rydberg <rydberg@euromail.se>,
	Daniel Kurtz <djkurtz@chromium.org>,
	Joonyoung Shim <jy0922.shim@samsung.com>,
	Alan Bowens <Alan.Bowens@atmel.com>,
	linux-input <linux-input@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Peter Meerwald <pmeerw@pmeerw.net>,
	Olof Johansson <olofj@chromium.org>
Subject: Re: [PATCH 20/51] Input: atmel_mxt_ts - Set default irqflags when there is no pdata
Date: Mon, 26 May 2014 14:17:39 -0400	[thread overview]
Message-ID: <CAPDwgkMyiYwvh0nzrpJ2E8zOYmF2pf7B95O+MR0=CtgszPtWhQ@mail.gmail.com> (raw)
In-Reply-To: <5380938E.5010408@itdev.co.uk>

On Sat, May 24, 2014 at 8:41 AM, Nick Dyer <nick.dyer@itdev.co.uk> wrote:
> Yufeng Shen wrote:
>> On Thu, May 22, 2014 at 10:29 AM, Nick Dyer <nick.dyer@itdev.co.uk> wrote:
>>> Dmitry Torokhov wrote:
>>>>>> Make the irqflags default to be IRQF_TRIGGER_FALLING if no platform data is
>>>>>> provided.
>>>>
>>>> I think if there is no platform data we should use 0 as IRQ falgs and
>>>> assume that IRQ line is properly configured by the board code or via
>>>> device tree.
>>>
>>> Benson/Yufeng - do you still have a requirement to probe without platform
>>> data or device tree? I'm just merging in some changes to add device tree
>>> support, and it would simplify things a bit if I can drop this patch.
>>
>> It has been working for quite a while for boards/devices that don't
>> provide platform data. If we drop the default IRQ flags, sure we can add
>> code for each board to configure the IRQ separately, but that's just
>> adding extra work. Is there strong reason why we should not do the
>> default setting in the driver if it is not already configured in
>> platform data?
>
> OK, I will keep it in my tree for the moment, since you are using it.
>
> The reason I checked is that in general, I would like to be conservative
> about what is pushed upstream, because it will need maintaining for a long
> time.
>
> The other reason is that in fact Atmel recommend IRQF_TRIGGER_LOW for these
> chips, not IRQF_TRIGGER_FALLING, so there is a bit of an inconsistency here.

I think I chose IRQF_TRIGGER_FALLING is because when I do a search in
the upstream
code where platform is configured, the irq is always set to be
IRQF_TRIGGER_FALLING
so I was assuming  IRQF_TRIGGER_FALLING is a safe bet.

But you would definitely know better than me on this since the atmel
chips that I have
access to are quite limited.