linux-kernel - Re: [PATCH] Input: egalax_ts - do not release gpio if probe successful

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: <20170112234057.GA17933@dtor-ws>
Date:   Thu, 12 Jan 2017 15:40:57 -0800
From:   Dmitry Torokhov <dmitry.torokhov@...il.com>
To:     Gary Bisson <gary.bisson@...ndarydevices.com>
Cc:     linux-input@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] Input: egalax_ts - do not release gpio if probe
 successful

On Fri, Jan 13, 2017 at 12:03:43AM +0100, Gary Bisson wrote:
> Hi Dmitri,
> 
> On Thu, Jan 12, 2017 at 10:30:19AM -0800, Dmitry Torokhov wrote:
> > Hi Gary,
> > 
> > On Wed, Jan 11, 2017 at 06:28:41PM +0100, Gary Bisson wrote:
> > > Thus preventing anyone to later modify the interrupt GPIO direction
> > > and/or state without the driver knowing.
> > 
> > I am afraid not releasing gpio after waking up the controller will cause
> > request_irq to fail if we are using the same pin for interrupt and
> > wakeup (as majority of current DTSes do: see
> > arch/arm/boot/dts/imx53-tx53-x13x.dts for example).
> 
> No, keeping the GPIO doesn't prevent from requesting the IRQ.
> 
> However it keeps other drivers/users from changing the GPIO as output
> later on.


Hmm, I think _gpiod_direction_output_raw() will not let them as it
checks FLAG_USED_AS_IRQ, so as long as it's the same line it's OK. But I
guess if they are separate some other component might try to grab it and
mess with it.

OK, in this case please:

- use devm_gpiod_get
- request with GPIOD_OUT_LOW
- do not override the return value with -ENODEV (especially important
  with probe deferrals)
- lose gpiod_direction_output() call
- move everything into egalax_ts_probe() as egalax_wake_up_device() is
  no longer self-contained. 

Thanks!

-- 
Dmitry