Re: What is expected behavior of watchdog with WDT_FLAG_RESET_SOC?

Carles Cufi

Hi Katsuhiro,

Thank you for reply. I understand situation.
I will try to add myself as collaborator of watchdog.
Thank you, that will certainly help.

But it seems that Zephyr project rules need two or more reviewers to
proceed pull requests. And of course, I cannot review my patch.
Maybe it does not help to merge or review my patch even if I will be as
collaborator. RISC-V area is also in same situation...
Of course, but there are things to consider here:

- If you indeed end up becoming a maintainer, you will be able to take certain decisions over questions or issues that show up in Pull Requests, effectively helping to move progress forward.
- As a consequence of that, if a Pull Request comes from the maintainer, then the requirement for two approvals stands, but the approvals are obviously automatically assumed to come from non-maintainers, so that means you don't need to wait for a maintainer that is non-existent if there are disagreements

In general, it is my opinion that having a maintainer or at least a collaborator will help in getting things merged, regardless of whether the patches come from that maintainer or anybody else.


Best Regards,
Katsuhiro Suzuki

On 2020/12/01 4:50, Carles Cufi wrote:
Hi Katsuhiro,

The whole watchdog subsystem is orphaned in fact.
We need maintainers for it, so if anyone wants to volunteer that would
be much appreciated.



-----Original Message-----
From: <>
On Behalf Of Katsuhiro Suzuki via
Sent: 30 November 2020 18:14
To: zephyr-devel <>
Subject: Re: [Zephyr-devel] What is expected behavior of watchdog


It seems that watchdog driver is orphaned.
I would be appreciate it if anyone inform about that...

Best Regards,
Katsuhiro Suzuki

On 2020/11/28 20:35, Katsuhiro Suzuki wrote:
Hello All,

I'm implementing Watchdog driver for HiFive1 Rev.b.
This watchdog can reset SoC immediately when counter is reaching
timeout. It's suitable feature of WDT_FLAG_RESET_SOC.

But tests/drivers/watchdog/wdt_basic_api/ expected to be calling
back the function 'wdt_int_cb0' even if with WDT_FLAG_RESET_SOC.
I think we cannot do that because SoC reset from watchdog is
immediate and not raise interrupt.

Do other boards pass this watchdog test? If so, I have to add
emulated SoC reset codes in interrupt handler of watchdog.

Best Regards,
Katsuhiro Suzuki

Join to automatically receive all group messages.