[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y3Q/k0hoPcnQhS8L@google.com>
Date: Tue, 15 Nov 2022 17:40:35 -0800
From: Dmitry Torokhov <dmitry.torokhov@...il.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: Gregory Clement <gregory.clement@...tlin.com>,
Sebastian Hesselbarth <sebastian.hesselbarth@...il.com>,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ARM: mvebu: switch to using gpiod API in pm-board code
On Wed, Nov 16, 2022 at 02:10:29AM +0100, Andrew Lunn wrote:
> > - ret = gpio_direction_output(pic_gpios[i], 0);
> > - if (ret < 0) {
> > - gpio_free(pic_gpios[i]);
> > + pic_gpios[i] = fwnode_gpiod_get_index(of_fwnode_handle(np),
> > + "ctrl", i, GPIOD_OUT_HIGH,
> > + name);
>
> The old code passes value=0 to gpio_direction_output(). For
> fwnode_gpiod_get_index() you pass GPIOD_OUT_HIGH. Is this correct?
Yes, gpiod API works on logical states, whereas old gpio API used signal
levels. In arch/arm/boot/dts/armada-xp-gp.dts ctrl-gpios are described
as "active low":
cpus {
pm_pic {
ctrl-gpios = <&gpio0 16 GPIO_ACTIVE_LOW>,
<&gpio0 17 GPIO_ACTIVE_LOW>,
<&gpio0 18 GPIO_ACTIVE_LOW>;
};
};
so gpiolib will translate GPIOD_OUT_HIGH to 0 when setting final state
of the pin.
There are discussions to rename GPIOD_OUT_HIGH and friends to something
like active/inactive for better clarity, but that has not happened yet.
Thanks.
--
Dmitry
Powered by blists - more mailing lists