|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] RE: iSCSI: Use of the A bit>>>>> "Kevin" == A-Roseville,ex1 <LEMAY> writes: Kevin> I also do not like the comment about having to wait until Kevin> maxBurstsize byte to set the A-bit. Kevin> I wanted to use it as Eddie suggested. To force the initiator Kevin> to acknowledge the IO so that I can release resources on the Kevin> target. Kevin> This will be especially useful for hardware accelerated IOs Kevin> where the chip cannot declare the IO complete until it has Kevin> been Ack'd. Kevin> Could we just change the wording to: ---- For sessions with Kevin> ErrorRecoveryLevel 1 or higher the target sets this bit to 1 Kevin> to indicate that it requests from the initiator a positive Kevin> acknowledgement for the data received. The target should use Kevin> the A bit moderately; the A-bit MAY only be set to 1 at the Kevin> end of a data sequence. Kevin> On receiving a Data-In PDU with the A bit set to 1 the Kevin> initiator MUST issued a SNACK of type DataACK. If the Kevin> initiator has detected holes in the input sequence, it MUST Kevin> postpone issuing the SNACK of type ACKN until the holes are Kevin> filled. --- Sounds right to me. paul
Home Last updated: Thu Mar 14 17:18:08 2002 9126 messages in chronological order |