Specialized data structure designed specifically for DAQ.

Guiding design requirements:

• Single producer (CAN_IRQs w same priority)
• Multiple Consumer (SD, ETH)
• the SD tail must never miss a message (or else mark overflow)
• the ETH tail does "best effort streaming"
• high throughput, lock free

Assumptions:

• Even though there are actually two producing ISRs (CAN1 and CAN2), they have the same priority and cannot preempt each other, so we can treat them as a single producer for the purposes of this data structure.
• We do not care if the follower tail "drops" data, as long as the master tail is guaranteed to consume all messages in the queue.

Note

todo more detailed docs

timing diagram

sd fsm