linux-kernel - Re: RFC: gpio: mmio: add support for 3 direction regs

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]

Message-ID: <6f609cb3-b294-0e66-a68f-73ad7ee4ac9a@mentor.com>
Date:   Thu, 3 Jan 2019 10:59:09 +0200
From:   Vladimir Zapolskiy <vladimir_zapolskiy@...tor.com>
To:     "Fried, Ramon" <ramon.fried@...ux.intel.com>,
        <linus.walleij@...aro.org>, <bgolaszewski@...libre.com>,
        <linux-gpio@...r.kernel.org>
CC:     <linux-kernel@...r.kernel.org>,
        <vladimir.kondratiev@...ux.intel.com>
Subject: Re: RFC: gpio: mmio: add support for 3 direction regs

On 01/03/2019 10:51 AM, Fried, Ramon wrote:
> 
> On 1/3/2019 10:07, Vladimir Zapolskiy wrote:
>> Hi Ramon,
>>
>> On 01/03/2019 09:36 AM, Fried, Ramon wrote:
>>> Hi.
>>>
>>> I'm working on a driver for STA2X11 GPIO controller who seems to fit
>>> best to the generic mmio driver,
>> I hope you have seen the existing driver drivers/gpio/gpio-sta2x11.c
> I surely did. we have the same IP in our soc but it was changed a lot
> internally, don't want to litter the original.
>>
>>> the only problem I have is with the dir register case. The STA2X11
>>> has 3 registers for dir, one for data, one for set and one for
>>> clear. The generic-mmio driver has support for this fashion for the
>>> dat & set & clear registers but not for dirout/dirin registers.
>>>
>>> I wonder if support for this is generic enough to deserve a patch, if
>>> so I'm willing to quickly add this support, if not, adding a flag
>>> such as below, will allow partly using the generic mmio driver only
>>> for set/get and the direction can be handled outside the driver.
>>>
>> If gpio-mmio fits well, then it might be simpler to set a flag
>> BGPIOF_UNREADABLE_REG_DIR, then call bgpio_init() and then overwrite
>> .direction_input, .direction_output and .get_direction callbacks,
>> as a reference you can take a look at gpio-74xx-mmio.c
> 
> Nice.
> That's an option I didn't think of, better than setting the flag.
> what about adding the generic support ?

Setting a GPIO direction over three different registers doesn't sound
as a suitable candidate for generalization, also your particular
implementation is partial, relies on external platform/driver code
to preset the directions and it does not allow to change direction
in runtime, so just renounce the idea.

--
Best wishes,
Vladimir