linux-kernel - Re: [PATCH] arm64: tegra: Enable PWM fan on the Jetson TX1 Devkit

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: <CALHNRZ8GCBwjVOFoeGrP3Rbcy2NVzkahi_i0gR8qAYr2tA0SOA@mail.gmail.com>
Date: Thu, 1 May 2025 17:31:27 -0500
From: Aaron Kling <webgeek1234@...il.com>
To: Tomasz Maciej Nowak <tmn505@...il.com>
Cc: Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>, 
	Conor Dooley <conor+dt@...nel.org>, Thierry Reding <thierry.reding@...il.com>, 
	Jonathan Hunter <jonathanh@...dia.com>, devicetree@...r.kernel.org, 
	linux-tegra@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] arm64: tegra: Enable PWM fan on the Jetson TX1 Devkit

On Wed, Apr 23, 2025 at 8:21 AM Tomasz Maciej Nowak <tmn505@...il.com> wrote:
>
> W dniu 23.04.2025 o 06:31, Aaron Kling pisze:
> > On Tue, Apr 22, 2025 at 11:19 AM Tomasz Maciej Nowak <tmn505@...il.com> wrote:
> >>
> >> W dniu 22.04.2025 o 17:58, Aaron Kling pisze:
> >>> On Tue, Apr 22, 2025 at 9:52 AM Tomasz Maciej Nowak <tmn505@...il.com> wrote:
> >>>>
> >>>> Hi.
> >>>>
> >>>> W dniu 21.04.2025 o 00:42, Aaron Kling via B4 Relay pisze:
> >>>>> From: Aaron Kling <webgeek1234@...il.com>
> >>>>>
> >>>>> This is based on 6f78a94, which enabled added the fan and thermal zones
> >>>>> for the Jetson Nano Devkit. The fan and thermal characteristics of the
> >>>>> two devkits are similar, so usng the same configuration.
> >>>>
> >>>> Does this work on Your DevKit? Doesn't on mine, the fan won't budge. Maybe the
> >>>> revision difference? What I'm using ATM is [1] and [2]. Because inverted polarity
> >>>> of PWM, not submitted since that'll need the driver changes [3],[4].
> >>>
> >>> I would have sworn I verified this before sending it in. I've had the
> >>> patches for some time. But you are correct, this does not work as-is.
> >>> Maybe I lost something cleaning up for submission or just plain
> >>> misremembered the verification. I will send a v2 once I've fixed and
> >>> verified. Apologies to the list for the bad submission.
> >>>
> >>> For inverted polarity, listing them backwards already has precedence
> >>> in mainline, see the Banana Pi R3 dt. This makes me want to double
> >>> check the existing Nano pwm-fan entry in mainline, though. Cause I
> >>> thought all the t210 devices were the same in regards to pwm fan
> >>> inversion. And it doesn't have reversed entries.
> >>
> >> That Banana Pi R3 reverse levels look ugly, but if it's permitted I'm not against.
> >>
> >> I would assume they fixed that in Nano, since PWM controller doesn't implement
> >> inverted polarity in hardware. Looking at Switch [5] it seems the TX1 DevKit was
> >> used for developing, since they replicated the issue.
> >>
> >> 5. https://github.com/fail0verflow/switch-linux/commit/b23e8b89081415f2a63bc625db041c8092e2a8a2
> >>   >
> >>> Sincerely,
> >>> Aaron Kling
> >>>
> >>>>
> >>>> 1. https://github.com/tmn505/linux/commit/a78c520ec94aeab2c9dc8e1f46597c4174ff957d
> >>>> 2. https://github.com/tmn505/linux/commit/99beee4f0cd5d3a6f30e1829d823c11cb8b54bac
> >>>> 3. https://libera.irclog.whitequark.org/tegra/2024-07-19#36707118;
> >>>> 4. https://libera.irclog.whitequark.org/tegra/2024-10-14#37145211;
> >>>>
> >>>> Regards
> >>>>
>
> [snip]
>
> > Mmm, so this is strange. I am currently unable to get the fan to move
> > on any t210 device. But it works just fine on t186, such as the tx2 nx
> > setup supported by mainline. Tomasz, does your change work on 6.12 or
> > current mainline? Even if I match your changes on the tx1 devkit, I
> > get nothing. The pwm duty cycle is changing as expected, per the
> > debugfs pwm listing. Gpio state for pin 4 of the tca9539@74 matches
> > the 4.9 kernel when the fan is running. Best I can tell, it should be
> > working. But it's not.

For documentation purposes, I figured out why I couldn't get the fan
to work on any of my test devices. I'm using the Android boot stack,
which apparently doesn't do the pinmuxing that the L4T boot stack
does. The downstream Android setup does pinmuxing in the kernel. And
mainline support has been primarily designed against the L4T boot
stack, doing pinmuxing in the bootloader and not in kernel. I've
worked around this locally by setting the pinctrl name on the p2597
pinmux node to 'default', so the kernel does pinmuxing. Which matches
things back up between the two bootloaders. I'll be sending a working
v2 shortly.

Sincerely,
Aaron