Re: [PATCH 2/2] HID: mcp2221: configure GP pins for GPIO function

From: rishi gupta <gupt21@gmail.com>
To: Tobias Junghans <tobias.junghans@inhub.de>
Cc: "open list:HID CORE LAYER" <linux-input@vger.kernel.org>,
	Jiri Kosina <jikos@kernel.org>,
	Benjamin Tissoires <benjamin.tissoires@redhat.com>,
	Linus Walleij <linus.walleij@linaro.org>
Subject: Re: [PATCH 2/2] HID: mcp2221: configure GP pins for GPIO function
Date: Mon, 30 Aug 2021 18:47:34 +0530	[thread overview]
Message-ID: <CALUj-gvwPJW6qND7_8eBPLjdXe9OienMeSNu7zAgq6Q-YEitiQ@mail.gmail.com> (raw)
In-Reply-To: <3382390.V25eIC5XRa@306e6e011c15>

By mistake during development it may happen or a rogue application can
knowingly play with our hardware (commercial product may be
vulnerable). What are your thoughts?

-Rishi

On Mon, Aug 30, 2021 at 6:39 PM Tobias Junghans
<tobias.junghans@inhub.de> wrote:
>
> Hi Rishi,
>
> thank you for your questions. I agree with you that one usually would
> reprogram the flash when manufacturing commercial products. However there's not
> always the need to do so if things can be done in software as well. The code
> will do no harm since a GPIO line initially is configured as input
> (MCP2221_GP_GPIO_DIR_IN) when being requested. Like with any other GPIO
> (driver), it's up to the user to take care of not configuring both ends as
> outputs with conflicting pull downs/ups or logic levels. Also the driver's
> default behaviour remains unchanged, i.e. it will not change the GP pin config
> unless requested explicitly.
>
> So all the proposed patch does it is to make the GPIO functions work as
> expected OOTB when explicitly controlling them with the appropriate tools or
> interfaces (libgpiod/sysfs).
>
> Best regards
>
> Tobias
>
>
> > Hi Tobias,
> >
> > To me it sounds like we are discussing about commercial product
> > (predefined internal flash fw) v/s prototype (we want to play with
> > settings at runtime with ease).
> >
> > Let us assume a GPx pin is configured as input and pulled up in
> > hardware board originally. A microcontroller's GPIO is configured as
> > output and connected to this GPx on MCP2221.
> > MCP2221 (GPx, input, pulled up) <----------- (output, no pull up/down)
> > STM32 Microcontroller
> >
> > 1. The STM32 Microcontroller drives this pin and set it to logic low
> > 2. Driver using this patch configure this GPIO on mcp2221 end as
> > output and drives it to logic high
> > It is like two devices trying to drive same (physical wire) GPIO pin
> > at the same time. How we plan to handle this.
> >
> > Will the GPx side will fuse or malfunction because of infinite current flow
> > ?
> >
> > Regards,
> > Rishi
>
>