Report
7 months, 3 weeks ago.

Does read_high_resolution_us perform properly when low 32 bits of the master ticker wrap to zero

I'm concerned that read_high_resolution_us, called from background, fetches the low and high parts of the low level microsecond timer with separate instructions. If the two fetches were done exactly when the low word wrapped to zero, the two fetched values might not be in sync.

read_high_resolution_us uses the core_util_critical_section routines but does that stop the incrementing of the low level microsecond timer?

Comment on this question
Be the first to answer this question.

You need to log in to post a question