Re: [PATCHv3 3/4] watchdog: rti-wdt: attach to running watchdog during probe

[Guenter Roeck]

On 7/14/20 12:13 AM, Tero Kristo wrote:
[ ... ]

>>> Well, you can re-program it but... It does not take effect until next window starts, so basically we need to take care of the currently running window anyways after which re-programming it doesn't make much sense anymore. And handling the switch from one setup to next would add extra complexity.
>>>
>>
>> Seems to me that hardware team really made an effort to make the
>> watchdog as difficult to program as possible :-(.
> 
> Yea, it is surprisingly difficult to program... when watchdogs in principle are extremely simple pieces of HW. This claims to be some automotive grade watchdog, which means it has the window in place.
> 

We are getting a bit off topic, but that almost sounds like
"automotive" translates to "it must be hard to program".
I can not think of a valid reason for such a window, no matter
the circumstances. Either case, dealing with the window in the
kernel in that situation (ie if some specification states that
the window must exist) would be the wrong solution and circumvent
the reason for having it in the first place.

Not that this matters here, of course. I am just wondering what the impact
would be if the driver is indeed used in such an application. For example,
some compliance test code might try to ping the watchdog outside the window
to test its reaction. Such test code would fail if the ping is delayed
by the kernel.

Guenter