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  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sun, 19 Apr 2020 09:54:49 -0300
From:   Ezequiel Garcia <ezequiel@...guardiasur.com.ar>
To:     Andy Shevchenko <andy.shevchenko@...il.com>
Cc:     Paul Cercueil <paul@...pouillou.net>,
        Artur Rojek <contact@...ur-rojek.eu>,
        Dmitry Torokhov <dmitry.torokhov@...il.com>,
        Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Jonathan Cameron <jic23@...nel.org>,
        Heiko Stuebner <heiko@...ech.de>,
        linux-input <linux-input@...r.kernel.org>,
        devicetree <devicetree@...r.kernel.org>,
        linux-iio <linux-iio@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [RESEND PATCH v5 3/5] IIO: Ingenic JZ47xx: Add touchscreen mode.

On Fri, 17 Apr 2020 at 18:54, Andy Shevchenko <andy.shevchenko@...il.com> wrote:
>
> On Sat, Apr 18, 2020 at 12:45 AM Paul Cercueil <paul@...pouillou.net> wrote:
> > Le sam. 18 avril 2020 à 0:42, Andy Shevchenko
> > <andy.shevchenko@...il.com> a écrit :
> > > On Sat, Apr 18, 2020 at 12:18 AM Paul Cercueil <paul@...pouillou.net>
> > > wrote:
> > >>  Le sam. 18 avril 2020 à 0:13, Andy Shevchenko
> > >>  <andy.shevchenko@...il.com> a écrit :
> > >>  > On Sat, Apr 18, 2020 at 12:05 AM Paul Cercueil
> > >> <paul@...pouillou.net>
> > >>  > wrote:
> > >>  >>  Le ven. 17 avril 2020 à 23:59, Andy Shevchenko
> > >>  >>  <andy.shevchenko@...il.com> a écrit :
> > >>  >>  > On Fri, Apr 17, 2020 at 11:21 PM Artur Rojek
> > >>  >> <contact@...ur-rojek.eu>
> > >>  >>  > wrote:
> > >>  >
> > >>  > ...
> > >>  >
> > >>  >>  >>  +       irq = platform_get_irq(pdev, 0);
> > >>  >>  >
> > >>  >>  > Before it worked w/o IRQ, here is a regression you introduced.
> > >>  >>
> > >>  >>  Before it simply did not need the IRQ, which is provided by the
> > >>  >>  devicetree anyway. No regression here.
> > >>  >
> > >>  > Does it work without IRQ? Or it was a dead code till now?
> > >>  > For me it's clear regression. Otherwise something is really wrong
> > >> in a
> > >>  > process of development of this driver.
> > >>
> > >>  Nothing wrong here. The IRQ was not used by the driver for the
> > >>  functionality it provided before. It is required now to support the
> > >>  touchscreen channels.
> > >
> > > This is exactly what's wrong.
> > > Previous DTS for my (hypothetical) case has no IRQ defined. Everything
> > > works, right?
> > > Now, due to this change it breaks my setup. Don't you see the problem?
> >
> > The IRQ has been provided by every concerned DTS file since the
> > introduction of this driver and the related bindings, even though it
> > was not used by the driver.
>
> Can you speak for all possible DTSs/DTBs in the wild?
> Okay, in any case it will be problem of maintainers and yours if
> somebody complains.
> I'm not going to push this anyway -- your choice.
>
> But I see a (potential) regression.
>

So, there are a few things to keep in mind here.

Let's abstract ourselves from this specific driver
for a minute.

First, and just as Andy pointed out, we can never be fully
sure about DTBs out there. These could be out of tree,
so out of our control. By introducing a new requirement
we break them, which may be seen as a regression.

Second, the interrupt is not required as per
current mainline bindings/iio/adc/ingenic,adc.txt,
so it is perfectly legal for users to not have an interrupt
specified.

Now, back to this case, I think we can get away with this
change, provided this hardware is not that widespread
among developers/users that follow upstream closely.

I suspect anyone developing a serious platform
with this SoC is most likely using some vendor kernel.

If that is not the case, i.e. if you have users _actually_
using this upstream driver, then we should consider
making the interrupt optional instead of required.

Or we can also just break it and hope nobody
complaints.

BTW, this series looks great and I'm happy
to see JZ47xx activity :-)

Arthur: perhaps you can consider converting the txt dt binding
to yaml?

Cheers,
Ezequiel

Powered by blists - more mailing lists