[openib-general] Send Side RMPP and OpenSM GetTableResp

Hal Rosenstock halr at voltaire.com
Mon May 30 12:09:44 PDT 2005


Hi Sean,

When running a Solaris 10 client against OpenSM with RMPP supported
turned on, I see the following sequence:

SA Client					SA
SA GetTable ---------------------------->
   TID 6
SA GetTable ---------------------------->
   TID 7
SA GetTable ---------------------------->
   TID 8
            <----------------------------    SA GetTableResp
					RMPP flags 0x07 (Data, First, Last)
					SegmentNumber 1
					PayloadLength 0x8C
					TID 6
SA GetTable ---------------------------->
RMPP flags 0x02 (ACK)
SegmentNumber 1
NewWindowLast 6
TID 6
            <----------------------------    SA GetTableResp
					RMPP flags 0x07 (Data, First, Last)
					SegmentNumber 1
					PayloadLength 0x8C
					TID 7
SA GetTable ---------------------------->
RMPP flags 0x02 (ACK)
SegmentNumber 1
NewWindowLast 6
TID 7
            <----------------------------    SA GetTableResp
					RMPP flags 0x03 (Data, First)
					SegmentNumber 1
					PayloadLength 0x370
					TID 8
SA GetTable ---------------------------->
RMPP flags 0x02 (ACK)
SegmentNumber 1
NewWindowLast 6
TID 8
            <----------------------------    SA GetTableResp				
					RMPP flags 0x01 (Data)
					SegmentNumber 2
					PayloadLength 0x370
					TID 8
            <----------------------------    SA GetTableResp				
					RMPP flags 0x01 (Data)
					SegmentNumber 3
					PayloadLength 0x370
					TID 8
            <----------------------------    SA GetTableResp				
					RMPP flags 0x05 (Data, Last)
					SegmentNumber 4
					PayloadLength 0x34
					TID 8
SA GetTable ---------------------------->
RMPP flags 0x02 (ACK)
SegmentNumber 1
NewWindowLast 6
TID 8
            <----------------------------    SA GetTableResp
					RMPP flags 0x01 (Data)
					SegmentNumber 5
					PayloadLength 0x34
					TID 8


I presume the reACKing is used as a keep alive so a response timeout
(Resp) does not occur. The SA client is using RRespTime of 0xE. The
OpenIB side sets this field to 0 (not sure if this affects the SA client
side).

Some questions on the RMPP sender side (SA):

1. I wouldn't think that reACKing the same segment (1) by the receiver
(SA client end) would cause the sender side to send segment 5. 

2. In the resend, the header (everything up to SA data) appears to be
good but the data appears to be garbage.

Have you tested RMPP retransmission ? Have you tested simultaneous
transactions in progress ?

Thanks.

-- Hal




More information about the general mailing list