When homing involves multiple micro-controllers the Klipper code arranges for all the micro-controllers to generate "keep-alive" messages. This is designed to avoid an endstop signal not propagating due to a communication outage.
However, the existing code scheduled all of the micro-controllers to send their keep-alive message at the same time. It's a little more optimal to stagger the updates so as to reduce instantaneous bandwidth and processing load. This PR also slightly increases the rate of keep-alive messages to better handle some "corner cases" with lost messages.
When homing involves multiple micro-controllers the Klipper code arranges for all the micro-controllers to generate "keep-alive" messages. This is designed to avoid an endstop signal not propagating due to a communication outage.
However, the existing code scheduled all of the micro-controllers to send their keep-alive message at the same time. It's a little more optimal to stagger the updates so as to reduce instantaneous bandwidth and processing load. This PR also slightly increases the rate of keep-alive messages to better handle some "corner cases" with lost messages.
This was also discussed at https://klipper.discourse.group/t/configurable-timeout/12032
-Kevin