[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ed99a966-6c10-4156-846a-4ae02621dcf3@kernel.org>
Date: Wed, 14 Jan 2026 16:58:58 +0100
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Nandor Han <nandor.han@...ealthcare.com>, wim@...ux-watchdog.org,
linux@...ck-us.net, robh@...nel.org, krzk+dt@...nel.org,
conor+dt@...nel.org, shawnguo@...nel.org, s.hauer@...gutronix.de
Cc: kernel@...gutronix.de, festevam@...il.com,
linux-watchdog@...r.kernel.org, devicetree@...r.kernel.org,
imx@...ts.linux.dev, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/2] dt-bindings: watchdog: fsl-imx: document continue
in low power mode
On 14/01/2026 14:22, Nandor Han wrote:
>
> On 1/12/26 15:55, Krzysztof Kozlowski wrote:
>> CAUTION: This email originated from outside of GE HealthCare. Only open links or attachments if you trust the sender. Report suspicious emails using Outlook’s “Report” button.
>>
>> On 12/01/2026 14:08, Nandor Han wrote:
>>> Property "fsl,wdt-continue-in-low-power" allows the watchdog to continue
>>> running in low power modes (STOP and DOZE). By default, the watchdog is
>>> suspended in these modes. This property provides the option to keep the
>>> watchdog active during low power states when needed.
>>>
>>> Signed-off-by: Nandor Han <nandor.han@...ealthcare.com>
>> Do not attach (thread) your patchsets to some other threads (unrelated
>> or older versions). This buries them deep in the mailbox and might
>> interfere with applying entire sets. See also:
>> https://elixir.bootlin.com/linux/v6.16-rc2/source/Documentation/process/submitting-patches.rst#L830
>
> Apologies, will not do that on the next version.
>
>
> <snip>
>
>>> + fsl,wdt-continue-in-low-power:
>>> + $ref: /schemas/types.yaml#/definitions/flag
>>> + description: |
>> Do not need '|' unless you need to preserve formatting.
>
> Thanks, will update.
>
>>> + If present, the watchdog device continues to run in low power modes
>>> + (STOP and DOZE).
>> Not much improved. My questions how this petting is suppose to work
>> remain unanswered.
>>
>> Again, system load is really not relevant whether watchdog works or not
>> in low power mode.
>
> Our use case is that we do not allow the CPU (i.MX7) to enter low-power modes.
That's very close to OS policy... or some sort of runtime choice.
>
> To enforce this, we explicitly enable the watchdog during low-power states so that if the system enters such a mode for any reason, the watchdog will reset the CPU.
So finally this is explained. You actually want to trigger the bite by
lack of petting.
You can as well set the bit moment you put the system to suspend. You do
not need DT flag for this.
I unfortunately do not see how this could be anything else than OS
policy, so it is not suitable for DT. Maybe some rephrasing could help
and that's what I would expect usually, but I already raised objections
and then you sent v2 with the same ignoring these objections (really,
never do that), so I doubt that discussing will change anything and you
need very specific answer: no.
> Since in STOP and DOZE modes the CPU is halted and systemd can no longer service the watchdog, the watchdog will eventually trigger a reset.
Best regards,
Krzysztof
Powered by blists - more mailing lists