linux-kernel - Re: [PATCH v2 1/4] gpiolib: Add "unknown" direction support

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 linux-cve-announce 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:	Thu, 17 Feb 2011 14:06:06 +0100
From:	Lars-Peter Clausen <lars@...afoo.de>
To:	Alexander Stein <alexander.stein@...tec-electronic.com>
CC:	Eric Miao <eric.y.miao@...il.com>,
	Peter Tyser <ptyser@...-inc.com>, linux-kernel@...r.kernel.org,
	Alek Du <alek.du@...el.com>,
	Samuel Ortiz <sameo@...ux.intel.com>,
	David Brownell <dbrownell@...rs.sourceforge.net>,
	Uwe Kleine-K?nig <u.kleine-koenig@...gutronix.de>,
	Mark Brown <broonie@...nsource.wolfsonmicro.com>,
	Joe Perches <joe@...ches.com>,
	Alan Cox <alan@...rguk.ukuu.org.uk>,
	Grant Likely <grant.likely@...retlab.ca>
Subject: Re: [PATCH v2 1/4] gpiolib: Add "unknown" direction support

On 02/17/2011 01:21 PM, Alexander Stein wrote:
> On Thursday 17 February 2011, 13:03:54 Lars-Peter Clausen wrote:
>> On 02/17/2011 08:33 AM, Eric Miao wrote:
>>> On Thu, Feb 17, 2011 at 8:56 AM, Peter Tyser <ptyser@...-inc.com> wrote:
>>>> Previously, gpiolib would unconditionally flag all GPIO pins as inputs,
>>>> regardless of their true state.  This resulted in all GPIO output pins
>>>> initially being incorrectly identified as "input" in the GPIO sysfs.
>>>>
>>>> Since the direction of GPIOs is not known prior to having their
>>>> direction set, instead set the default direction to "unknown" to prevent
>>>> user confusion.  A pin with an "unknown" direction can not be written or
>>>> read via sysfs; it must first be configured as an input or output before
>>>> it can be used.
>>>
>>> Hrm... that's why I don't like the original definition of gpio_request()
>>> which is vague on the pin configurations.
>>
>> Actually it doesn't say anything at all about the current configuration at
>> all. Requesting a pin grants you exclusive access to that pin, if it
>> succeeds. So it is solely about ownership and not about configuration.
> 
> Well, ownership is a bit misleading here. You must request a GPIO to change 
> its direction. But to set or get a value this isn't required.

Well, it is not enforced in the actual code, but the GPIO API is based on
convention and I would consider it a misusage of the API to call
gpio_{set,get}_value without requesting the pin and having configured its
direction prior to it.

> In general one could expect if you requested a GPIO you are the only one to 
> call any function on it. On the other hand, this may be bad in some situations 
> where you want to read a GPIO value from different places.

Sharing GPIOs in read-only mode, is indeed something that is not covered by the
GPIO API. It might be worth adding a gpio_request_shared, which would only
permit setting the direction to input. Futher gpio_request_shared calls would
be allowed but gpio_request calls would fail.

- Lars
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/