As I thought about your observation, the only case which might
need CONNECTION_STATE is when the connection is opened
unidirecitonally. I suppose unidirectional communication could
be determined from the MAX_I2T_DATA_SIZE and MAX_T2I_DATA_SIZE
parameters. We need one or the other for this case.
If this is the only case, then I'll remove CONNECTION_STATE
and keep both MAX_I2T_DATA_SIZE and MAX_T2I_DATA_SIZE.
Can anyone else think of a case where the message or stream
transport needs to go from bi- to uni-directional before closing?
If so, respond before 8/3/98 or else I'll remove it from the
command set proposal.
- Greg Shue, HP
Akihiro Shimura wrote:
>> > E) CONNECTION_STATE ( 2 bits unsigned int
>> I do not think it is necessary for the initiator to control connection
> of each direction separately at transport level.
>> I suppose the initiator can do it by not issuing either
> TRANSPORT_I2T_DATA or TRANSPORT_T2I_DATA command, while
> uni-directional target may want to declare its capability.
>> Because we are providing bi-directional transport, I think, whether
> clients use it as uni-directional or bi-directional is up to the
> clients.
>> Any suggestions?
--
Greg Shue
Hewlett-Packard Company
All-in-One Division gregs at sdd.hp.com
----------------------------------------------------------------