This is a follow-up to commit fe0e2dbc60.
If `nrf_clock_control_request_sync()` ends up with a timeout, before
returning it must cancel the request that was not fulfilled on time.
Otherwise, the request may actually finish successfully a bit later,
but the caller will not be aware that the clock needs to be released
(since the call resulted in an error). And actually even more serious
problem is that because the `req` structure is placed on stack, after
the function returns, the contents of this structure will be probably
overwritten with some other data, so if the request finishes at that
point, an attempt to execute the callback function pointed by this
structure will most likely cause a crash.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Due to the possibility of simultaneous accesess to LRCCONF registers,
additional management is required.
Signed-off-by: Adam Kondraciuk <adam.kondraciuk@nordicsemi.no>
Add custom clock_control API for nRF platforms that allows requesting
clocks with specified minimal required attributes (accuracy, precision,
and frequency). Provide an implementation of this API for FLL16M, HFXO,
HSFLL, and LFCLK controllers in the nRF54H20 SoC.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
