CFP MSA Management Protocol Implementation

All about Sub-20 Multi Interface USB Adapter USB to I2C, SPI, GPIO, RS232, RS485, Ir, LCD

Moderator: serg

Post Reply
xol
Site Admin
Posts: 241
Joined: Sat Aug 29, 2009 8:04 am

CFP MSA Management Protocol Implementation

Post by xol »

HI,

SUB-20 currently supports IEEE 802.3 Clause 22 and Caluse 45 Management Interface Implementation via MDC MDIO
http://www.xdimax.com/sub20/doc/HTML/mdio_functions.htm

We are currently working on extending it to support CFP MSA Rev 1.4 Specification by http://www.cfp-msa.org/
Actually it is pretty the same MDC MDIO but at hight speed of up to 4MHz

We have some arguing about right interpretation of the specification. The question is WHEN Host (in our case it is SUB-20) should scan MDIO line referenced to MDC rising edge. Here is some background.

This is taken from CFP MSA Rev 1.4
CFP-MDIO.jpg
CFP-MDIO.jpg (53.66 KiB) Viewed 9076 times
and
CFP-DOC.jpg
CFP-DOC.jpg (172.87 KiB) Viewed 9076 times
As you can see CFP MSA reference to IEEE 802.3 Clause 45.
IEEE-MDIO.jpg
IEEE-MDIO.jpg (43.55 KiB) Viewed 9076 times
1. t_delay of 0..175ns is a delay between rising edge of MDC and stable MDIO. Is it correct?
2. To support all kinds of PHY's Host should scan MDIO 175ns after rising edge om MDC right?
3. We have some customer with t_delay = 0ns. But MDIO on their implementation is stable for only 65ns after MDC rising edge and if we scan MDIO after 175ns we get wrong data. Do you think such implementation is compliant with a spec?

This is the illustration. SUB-20 reads the last bit as "1" but customer says it should be "0". Who is right?
Yellow signal is MDC
Green signal is MDIO
LastBit.jpg
LastBit.jpg (27.15 KiB) Viewed 9073 times
Thanks Everybody.

Post Reply