Forum Discussion
1 Reply
- Altera_Forum
Honored Contributor
The MAC takes care of managing packets to and from the phy.
TSE MAC and other MAC cores, usually exploit DMA to transfer incoming data to memory, so you don't necessarily need to pool at high rate, provided you have enough descriptor memory. The maximum delay you can accepts will then depend from your upper layer protocol requirements on latency. I myself have a design where I use a highest priority task which periodically blocks network task up to 0.3ms every 1ms and the lower priority network task is not affected by this (apart a slight decrease in speed, I mean kB/sec transferred)