[ofw] WinOF rc5 uDAPL working?

[Rolf Welde Skeie]

I'm trying to connect two machines back to back using WinOF rc5.
The machines are running Windows Server 2003 R2 x64.

Got my OpenSM up and running and IPoIB and WSD works just fine.

However, when I try the dapltest program I'm out of luck:
The client says:
Warning: conn_event_wait DAT_CONNECTION_EVENT_BROKEN

On the server side I get:
Warning: conn_event_wait DAT_CONNECTION_EVENT_NON_PEER_REJECTED

I also wonder why I have to use ibnic0 when IbalHca0 is the one
configured in dat.conf?

(I also tried switching WSD off and on to no effect)

Some more details below:

C:\Documents and Settings\Administrator>dapltest -T T -D ibnic0 -s
10.1.1.10 server SR 4096 2 client SR 4096 2
DAT Registry: Started (dat_init)
DAT Registry: SR Load return 0
DT_cs_Client: Starting Test ...
Server Name: 10.1.1.10
Server Net Address: 10.1.1.10
Warning: conn_event_wait DAT_CONNECTION_EVENT_BROKEN
DT_cs_Client: bad connection event


C:\Program Files\WinOF>dapltest -T S -D ibnic0 -d -d -d -d -d
DAT Registry: Started (dat_init)
DAT Registry: SR Load return 0
Server_Cmd.debug:       1
Server_Cmd.dapl_name: ibnic0
DT_cs_Server: IA ibnic0 opened
DT_cs_Server: PZ created
DT_cs_Server: EP created
DT_cs_Server: PSP created
lmr_create    [000000000034B400, 6000]
lmr_create OK [0x34b400, 6000, lctx=28002f00]
Recv 0  000000000034B400
Recv 1  000000000034D400
SrvInfo 2  000000000034F400
*****  DAPL  Characteristics  *****
Provider: ibnic0  Version 1.0  DAPL 1.1
Adapter: InfiniHost (Tavor) by Mellanox Technolgy Inc. Version 25218.160
Supporting:
         64512 EPs with 16384 DTOs and 4 RDMA/RDs each
         65408 EVDs of up to 131071 entries  (default S/R size is 16/16)
         IOVs of up to 30 elements
         131056 LMRs (and 0 RMRs) of up to 0xffffffffffffffff bytes
         Maximum MTU 0x80000000 bytes, RDMA 0x80000000 bytes
         Maximum Private data size 92 bytes
         Local IP address 10.1.1.10
***** ***** ***** ***** ***** *****
DT_cs_Server: Posting 2 recvs
Post-Recv #0 [000000000034B400, 2000]
Post-Recv #1 [000000000034D400, 2000]
Dapltest: Service Point Ready - ibnic0
DT_cs_Server: Waiting for Connection Request
DT_cs_Server: Accepting Connection Request
DT_cs_Server: Awaiting connection ...
Warning: conn_event_wait DAT_CONNECTION_EVENT_NON_PEER_REJECTED
DT_cs_Server: error awaiting conn-established event
DT_cs_Server: Waiting for clients to all go away...
DT_cs_Server: Cleaning up ...


C:\Program Files\WinOF>more \dat\dat.conf

IbalHca0 u1.1 nonthreadsafe default C:\Windows\dapl.dll mv_dapl.1.1
"IbalHca0 1"
  ""
# Intel MPI
#ibnic0 u1.1 threadsafe default C:\Windows\dapl.dll ri.1.1 "IbalHca0 1" " "

Regards
-- 
Rolf Welde Skeie

Scali - www.scali.com
Scaling the Linux Datacenter