linux-kernel - Re: [PATCH v2] meson saradc: add iio device attrib to switch channel 7 mux

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] [day] [month] [year] [list]

Message-ID: <20230604135228.5a5d743d@jic23-huawei>
Date:   Sun, 4 Jun 2023 13:52:28 +0100
From:   Jonathan Cameron <jic23@...nel.org>
To:     Martin Blumenstingl <martin.blumenstingl@...glemail.com>
Cc:     George Stark <GNStark@...rdevices.ru>,
        "lars@...afoo.de" <lars@...afoo.de>,
        "neil.armstrong@...aro.org" <neil.armstrong@...aro.org>,
        "khilman@...libre.com" <khilman@...libre.com>,
        "jbrunet@...libre.com" <jbrunet@...libre.com>,
        "andriy.shevchenko@...ux.intel.com" 
        <andriy.shevchenko@...ux.intel.com>,
        "nuno.sa@...log.com" <nuno.sa@...log.com>,
        "linux-iio@...r.kernel.org" <linux-iio@...r.kernel.org>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-amlogic@...ts.infradead.org" 
        <linux-amlogic@...ts.infradead.org>, kernel <kernel@...rdevices.ru>
Subject: Re: [PATCH v2] meson saradc: add iio device attrib to switch
 channel 7 mux

On Mon, 29 May 2023 22:04:18 +0200
Martin Blumenstingl <martin.blumenstingl@...glemail.com> wrote:

> Hi George,
> 
> On Sun, May 28, 2023 at 11:52 PM George Stark <GNStark@...rdevices.ru> wrote:
> [...]
> > > Based on the limited description we have here, I'm not understanding why
> > > you don't just express this as a set of channels. One channel per mux
> > > setting, with the in_voltageX_label providing the information on what the
> > > channel is connected to.
> > >
> > > This is an interesting facility, so good to enable for high precision calibration
> > > but we still want to map it to standards signals.  Userspace doesn't
> > > care that these are all being measured via the same input 7 - which
> > > is itself probably an input to a MUX.
> > >
> > > Jonathan  
> >
> > Hello Jonathan
> >
> > Thanks for the review.
> >
> > Your idea of exposing the mux setting as iio channels is very
> > interesting and at least worth trying.
> > The sysfs approach was chosen because of the code changes are simple and
> > neat (compare to channels approach).
> > Also calibration by using those mux inputs are already supported in the
> > driver (performed at probe stage) so I expect very special usecases for
> > those mux settings like debug or device production stage tests. In those
> > usescases hardware specific knowledge is required anyway.  

Agreed that using channels for this is more complex code wise, but the
avoidance of custom ABI is usually worthwhile.  However I get your point
that this is weird debug / corner case paths anyway.
However...


> Another downside to the debugfs approach is user support:
> If someone reports odd values on ADC channel 7 then we need to make
> sure to double check if the mux has been altered from userspace (the
> calibration during initialization ensures to leave channel 7 in a
> consistent state, while a user may change the mux, forget about it and
> then complain that values are wrong).

This problem raises red flags for me.  If the channel wasn't otherwise
useful then treating it as a weird debug thing is fine, but if a normal
read can end up returning weird answers because of leftover configuration
that is custom ABI, so standard code isn't even aware of it, that's bad.

Hence I would not want to see any path under which that can happen.
A read of in_voltage7_raw should continue to return the normal value
whatever this patch adds.

Jonathan

> 
> 
> Best regards,
> Martin