Re: [PATCH 0/5] watchdog: eiois200_wdt: Add EIO-IS200 Watchdog Driver

From: Wenkai
Date: Wed Oct 11 2023 - 00:09:18 EST




Guenter Roeck 於 10/6/2023 10:16 PM 寫道:
On Fri, Oct 06, 2023 at 05:27:48PM +0800, Wenkai wrote:

Guenter Roeck 於 10/6/2023 11:02 AM 寫道:
On Thu, Oct 05, 2023 at 04:51:18PM +0800, advantech.susiteam@xxxxxxxxx wrote:
From: Wenkai <advantech.susiteam@xxxxxxxxx>

This patch series aims to add support for the Advantech EIO-IS200
Embedded Controller's watchdog timer to the Linux kernel. The EIO-IS200
is a widely used embedded controller, and this series introduces a
native driver for its watchdog timer functionality within the Linux
ecosystem.

I am not going to review this patch series. This is just ne watchdog driver.
One patch is sufficient.

Guenter
Hi Guenter,

Advantech's EIO-IS200 watchdog supports 5 output pins: RESET, Power
Button, SCI, IRQ, and GPIO. The most traditional scenario is that the
Pretimeout triggers IRQ, and the timeout triggers RESET.

However, unfortunately, for industrial usages, there are various use
cases, which require certain mechanisms and logic to manage which signal
is output when Pretimeout and timeout expire. I am concerned that
consolidating all these features into a single patch for upstream may
lead to confusion and make the source code less readable and
understandable.

The 1st patch in your series doesn't even compile. I don't call that
understandable.

Oh, it fails to compile because you include a non-existing file from
../mfd directly and because you select a non-existing configuration option
instead of depending on it.

None of those is even remotely acceptable. Are you seriously sending me
a series of patches that don't even build to review ?

I understand that the patches don't meet the expected quality standards.
The compile issue is due to my MFD core driver, which is currently under
review and has not been merged yet.

I would also like to seek your advice on how to best proceed with the
sub-drivers like the watchdog driver. Should I wait for my core MFD
driver to be successfully merged before submitting the sub-drivers, or
let Jones Lee review my core MFD driver and all its sub-drivers, or is
there another approach that you recommend?
Therefore, I have divided the implementation into 5 separate patches,
aiming to make the code more comprehensible and acceptable. If it's
acceptable to you, I am more than willing to provide a single patch as
per your preference.

Frankly, your series is one more nail in the coffin. I am now seriously
considering to resign as co-maintainer of the watchdog subsystem.

Guenter

I also appreciate your patience, dedication, and valuable contributions
to the Linux community. Your longstanding efforts and expertise are
commendable and have been instrumental in advancing the Linux ecosystem.
I understand that upstream review can be a meticulous and vital, albeit
thankless, task. I don't want my actions to cause any inconvenience or
distress, especially to someone as esteemed as you are in the Linux
community. Your insights and guidance are incredibly valuable to all of
us.

Once again, thank you for your understanding, and I am committed to
delivering high-quality code for your review.

Best regards,
Wenkai