meaning of Pi value in AIF2

Hi Lee,

The AIF2 design spec doesn't mention Pi Check with regard to CPRI, only OBSAI. With that, here is relevant information concerning Pi and Rx Mac:

Pi Check & Measure:

The RM is detecting the presence of the K-characters and the result of this comparison is transferred (with optimal latency) to the AT. Later, the AT determines if the last K-character was a frame boundary demarcation and sends this information to the AT. The reason for the two signals is in regards to timing accuracy. The first signal is extremely accurate while the second potentially has a degree of timing error or latency associated with it.

The AT uses the strobes from RM to determine if the frame boundary arrived at the expected time, and to capture the PhyT timer value at the time of frame boundary. This check and capture are then further qualified with the indication of second strobe indicating the frame boundary.

Within a link, the K28.7 marks the 10ms frame boundary. The arrival of the link frame is permitted to arrive between +/- 5ms relative to the AT (RP1) PHY timer frame boundary. The Link timing is permitted to drift to some limited degree. A total of 64 bytes of drift can be tolerated by AIF2.

The RM (receiver) has a 64 byte buffer which can be used to accommodate timing drift. The user effectively selects the buffering by programming the RM to fill the buffer to the {0, 2, 4, 8, 16} water mark (at the beginning of frame). The allowable drift is the total of 64 bytes, centered on the chosen, initial water mark.

AT: Pi measurement:

There is a need to check for incoming data alignment to a boundary within a frame. A signal is received from the AIF2 RX MAC which is used to check and capture the PHYT timer value. This value is then compared in hardware to the allowable window. If this time is outside the allowable window, an error interrupt is generated.