From mboxrd@z Thu Jan  1 00:00:00 1970
From: Mika Westerberg <mika.westerberg@linux.intel.com>
Subject: Re: Re: [PATCH] pinctrl: intel: Implements gpio free function
Date: Thu, 21 Mar 2019 14:36:37 +0200
Message-ID: <20190321123637.GJ3622@lahna.fi.intel.com>
References: <1553135724-38331-1-git-send-email-zhuchangchun@cvte.com>
 <20190321084420.GG3622@lahna.fi.intel.com>
 <2019032119195575582546@cvte.com>
 <20190321120324.GI3622@lahna.fi.intel.com>
 <2019032120213955866649@cvte.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <2019032120213955866649@cvte.com>
Sender: linux-kernel-owner@vger.kernel.org
To: "zhuchangchun@cvte.com" <zhuchangchun@cvte.com>
Cc: "andriy.shevchenko" <andriy.shevchenko@linux.intel.com>, "linus.walleij" <linus.walleij@linaro.org>, linux-gpio <linux-gpio@vger.kernel.org>, linux-kernel <linux-kernel@vger.kernel.org>, hendychu <hendychu@aliyun.com>
List-Id: linux-gpio@vger.kernel.org

On Thu, Mar 21, 2019 at 08:22:40PM +0800, zhuchangchun@cvte.com wrote:
>    On Thu, Mar 21, 2019 at 07:20:56PM +0800, zhuchangchun@cvte.com wrote:
>    >    After you unexport GPIO it can go back to any previous mode it
>    was. If
>    >    you need to use it as GPIO then why unexport it in the first
>    place?
>    >    --> because we need use the GPIO for device reset, we want to
>    recover
>    >    the GPIO value after reset done.If we don't unexport it, then
>    reboot
>    >    the
>    >    intel SOC(not power off), the GPIO's value will stay until the
>    GPIO be
>    >    reinit.
>    >    I tested the GPIO signal with oscilloscope, if  power off the SOC,
>    >    then
>    >    power on, the GPIO value can recover its original value. This will
>    >    infruence
>    >    the peripheral device working condition.
> 
>    If I understand correctly you have some peripheral connected to Intel
>    based board and one GPIO is used to reset the peripheral.
> 
>    You say it works fine if you power cycle the board but it does not when
>    you issue 'reboot' command. Did I understood the problem correctly?
> 
>    --> Yes ,  so I add the -> free callback, it fixes the problem.

OK, so you then write 1 to the sysfs to assert reset, right? And in your
patch you change the mode to input effectively deasserting the reset.

Why not simply do following instead of unexport?

  # echo 0 > /sys/class/gpio/gpioXX/value

or even

  # echo 0 > /sys/class/gpio/gpioXX/value
  # echo in > /sys/class/gpio/gpioXX/direction

Am I missing something?