Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[802.3ae] RE: Clause 49, Receive State Machine

Hi Pat,
   There is another thing that looks wierd to me is that
0x33 1e 1e 1e 1e D0 D1 D1 D2 (four errors with S4)
   are valid S frame,
but when an error frame is followed by a S frame, S frame is
still treated as error frame.

-----Original Message-----
From: Tuan Hoang [mailto:tuanh@xxxxxxxxxxxx]
Sent: Tuesday, January 15, 2002 4:41 PM
To: stds-802-3-hssg@xxxxxxxx; 'THALER,PAT (A-Roseville,ex1)'
Cc: Sheila Qiu (E-mail); Hongtao Jiang (E-mail)
Subject: Clause 49, Receive State Machine

Hi Pat,

Definition of R_BLOCK_TYPE in page 366, clause 49 as:

Values: C; The vector contains s sync header of 10 and one of the following:
           a) A block type field of 0x1e and eight valid control characters
other than /E/;
           b) A block type field of 0x2d or 0x4b, a valid O code, and four
valid control characters;

        T; The vector contains a sync header of 10, a block type field of
0x87, 0x99, 0xaa, 0xb4, 0xcc, 0xd2, 0xe1 or 0xff and all control characters
are valid.

 When a 64b/66b data stream contains the End Of Packet frame followed by
Error frames as:

                           0xaa D0 D1 "00000"  0x00  0x1e  0x1e  0x1e  0x1e
(T frame with four /E/s)
                           0x1e 0x1e 0x1e 0x1e 0x1e 0x1e 0x1e 0x1e 0x1e
(Error frame)
                           0x1e 0x1e 0x1e 0x1e 0x1e 0x1e 0x1e 0x1e 0x1e
(Error frame)

Based on the C definition, and on the Receive state machine flow, Receive
state machine should stuck in the RX_T until a C block frame, or a S block
frame is received.  However, from the flow presentation of the Receive state
machine, it looks like the intention of the RX_T state is only transitional,
meaning that RX_T duration is only one clock at a time.  Can you clarify