watchdog_enable() confuses milliseconds and microseconds #1238

New issue

Closed

opened 2023-02-10 03:43:25 -06:00 by awabom · 5 comments

awabom commented

2023-02-10 03:43:25 -06:00

(Migrated from github.com)

In the file:

https://github.com/raspberrypi/pico-sdk/blob/master/src/rp2_common/hardware_watchdog/include/hardware/watchdog.h

delay_ms is described as:

"delay_ms Number of milliseconds before watchdog will reboot without watchdog_update being called. Maximum of 0x7fffff, which is approximately 8.3 seconds"

But 0x7fffff is 8388607 which is 8.3 seconds only if the unit is microseconds. The text also mentions "then the delay_ms parameter will not be in microseconds".

So I guess it is microseconds? Then I also think the parameter should be named delay_us, not delay_ms.

In the file: https://github.com/raspberrypi/pico-sdk/blob/master/src/rp2_common/hardware_watchdog/include/hardware/watchdog.h delay_ms is described as: "delay_ms Number of milliseconds before watchdog will reboot without watchdog_update being called. Maximum of 0x7fffff, which is approximately 8.3 seconds" But 0x7fffff is 8388607 which is 8.3 seconds only if the unit is microseconds. The text also mentions "then the delay_ms parameter will not be in microseconds". So I guess it is microseconds? Then I also think the parameter should be named delay_us, not delay_ms.

lurch commented

2023-02-10 04:28:29 -06:00

(Migrated from github.com)

Hmmm, https://github.com/raspberrypi/pico-sdk/blob/master/src/rp2_common/hardware_watchdog/watchdog.c#L55 does

        // Note, we have x2 here as the watchdog HW currently decrements twice per tick
        load_value = delay_ms * 1000 * 2;

        if (load_value > 0xffffffu)
            load_value = 0xffffffu;

so it looks like the documentation is correct in describing the parameter as milliseconds (with the hardware internally using microseconds), but it's the maximum-value which is wrong? As the parameter is in milliseconds, I guess the actual max-value is 8388. (which gives a timeout only 1215 microseconds less (i.e. 1.2ms) than the watchdog's maximum timeout)

Hmmm, https://github.com/raspberrypi/pico-sdk/blob/master/src/rp2_common/hardware_watchdog/watchdog.c#L55 does ``` // Note, we have x2 here as the watchdog HW currently decrements twice per tick load_value = delay_ms * 1000 * 2; if (load_value > 0xffffffu) load_value = 0xffffffu; ``` so it looks like the documentation is correct in describing the parameter as milliseconds (with the hardware internally using microseconds), but it's the maximum-value which is wrong? As the parameter **is** in milliseconds, I guess the actual max-value is 8388. (which gives a timeout only 1215 microseconds less (i.e. 1.2ms) than the watchdog's maximum timeout)

awabom commented

2023-02-10 19:04:07 -06:00

(Migrated from github.com)

(which gives a timeout only 1215 microseconds less (i.e. 0.1ms) than the watchdog's maximum timeout)

Sorry, but 1215 microseconds is 1.2 milliseconds... :)

I haven't tested the actual delay, but indeed the parameter looks to be milliseconds, and the maximum value was wrong.

(which gives a timeout only 1215 microseconds less (i.e. 0.1ms) than the watchdog's maximum timeout) Sorry, but 1215 microseconds is 1.2 milliseconds... :) I haven't tested the actual delay, but indeed the parameter looks to be milliseconds, and the maximum value was wrong.

lurch commented

2023-02-11 03:47:01 -06:00

(Migrated from github.com)

Sorry, but 1215 microseconds is 1.2 milliseconds... :)

LOL of course - must have been end-of-the-week brain 😆 Have edited my previous comment, thanks.

> Sorry, but 1215 microseconds is 1.2 milliseconds... :) LOL of course - must have been end-of-the-week brain :laughing: Have edited my previous comment, thanks.

lurch commented

2023-06-01 08:50:47 -05:00

(Migrated from github.com)