First pass at watchdog #3
Conversation
|
Some links Looks like 2 different peripherals at least. Very similar, but maybe more in series 1, this implementation is series 0 EFM32 Series 0 consists of: https://www.silabs.com/documents/public/application-notes/an0015.1-efr32-efm32-series-1-watchdog.pdf |
| Self { wdog } | ||
| } | ||
|
|
||
| fn free(self) -> registers::WDOG { |
There was a problem hiding this comment.
Is there precedent in any other HAL for using free for the "undo the construction step" operation?
I'm using .destroy in thunderboard-sltb001a BSC, not sure I had a reference there.
There was a problem hiding this comment.
| #[cfg(feature = "unproven")] | ||
| #[allow(dead_code)] | ||
| impl Watchdog { | ||
| pub fn new(wdog: registers::WDOG) -> Self { |
There was a problem hiding this comment.
To get a usable watchdog, some clock needs to be routed there, and should not be deconfigurable before freeing (destroying, see below) the watchdog first.
In the EFM32GG that clock is WDOGCLK configurable from ULRFCO, LFRCO or LFXO; I think this should be a with_clock (or any better unification of that name) a with other peripherals.
There was a problem hiding this comment.
Then well have to not use the embedded_hal implementation. Not that I care
Doesn't have any new up config (as we dont need it, our bootloader set us up)