linux-kernel - Re: [RFC PATCH 0/3] pwm: add support for duty

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: <fhrderkiwqzc7fw2hptz2la3aco72wlossqbrr4m42qllcpcpw@vhxflah2ex2a>
Date: Fri, 5 Apr 2024 08:23:25 +0200
From: Uwe Kleine-König <u.kleine-koenig@...gutronix.de>
To: Trevor Gamblin <tgamblin@...libre.com>
Cc: linux-pwm@...r.kernel.org, linux-kernel@...r.kernel.org, 
	michael.hennerich@...log.com, nuno.sa@...log.com, dlechner@...libre.com
Subject: Re: [RFC PATCH 0/3] pwm: add support for duty_offset

Hello Trevor,

In general I really like your effort to generalize the pwm framework.
Thanks a lot!

On Thu, Apr 04, 2024 at 08:30:22PM -0400, Trevor Gamblin wrote:
> This series extends the PWM subsystem to support the duty_offset feature
> found on some PWM devices. It includes a patch to enable this feature
> for the axi-pwmgen driver, which can also serve as an example of how to
> implement it for other devices. It also contains a patch adding a new
> pwm_config_full() function mirroring the behavior of pwm_config() but

Please don't. pwm_config() is a function I want to get rid of in the
long term. Consumers that want to make use of it should use
pwm_apply_*().

> with duty_offset included, to help maintain compatibility for drivers
> that don't support the feature.
> 
> The series was tested on actual hardware using a Zedboard. An
> oscilloscope was used to validate that the generated PWM signals matched
> the requested ones. The libpwm [1] tool was also used for testing the
> char device functionality.
> 
> The series is marked RFC as there are some outstanding questions about
> implementation:
> 
> 1. In drivers/pwm/core.c, __pwm_apply() was modified to check that the
> sum of state->duty_offset + state->duty_cycle does not exceed
> state->period, but in the character device section these values are
> being checked separately. Is this intentional? What is the intended
> behavior?

state->duty_offset + state->duty_cycle doesn't necessarily need to be
less or equal to state->period. Consider this waveform, where ^ marks
the start of a period.


        ___       _________       __...
           \_____/         \_____/
        ^               ^

This one has duty_offset = 9 and duty_cycle = 10 while
period = 16 < 10 + 9.

> 2. Should __pwm_apply() explicitly disallow PWM_POLARITY_INVERSED and
> duty_offset together?

While there is no technical need for that, a configuration with both
PWM_POLARITY_INVERSED and duty_offset > 0 is irritating. So I'd say yes,
it should be disallowed. When I created the cdev API I even considered
dropping .polarity for lowlevel drivers and convert them all to
.duty_offset.

Having said that I don't like the addition of .supports_offset to
struct pwm_chip, which only signals a new incomplete evolution of the
pwm framework. Better adapt all drivers and then assume all of them
support it.

> 3. Are there other places that would need duty_offset handling which
> have been missed?

I'm happy you adapted the tracing stuff. I didn't look closely, but I
don't think something important was missed.

> Note that in addition to the other patches in this series, the changes
> to the axi-pwmgen driver rely on [2] and [3], which haven't been picked
> up yet.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)