TAYBENGH@NUSDISCS.BITNET (02/28/91)
Hi netlanders, After reading thru the implementation of socket and TCP/IP from the book The Design and Implementation of the 4.3BSD OS (published by Addison-Wesley), I am puzzled of implementing OOB in the socket. Inside the book (pg 335) the authors describe that multiple OOB messages can be hold in a protocol if the user specify the SO_OOBINLINE option to force the OOB data to be placed in the normal receive queue. However, according to the data structure of socket, so_oobmark is used to mark the offset of the LAST OOB message from the beginning of the receive queue. My question is: How can a so_oobmark to mark different OOB messages? Or does it means that although many OOB messages can be hold in the receive queue, only ONE of them can be received using so_oobmark? If so, what is the purpose of holding multiple messages if we can not differentiate in-band and the rest OOB data? Thanks in advance. - Beng Hang Dept of Information Systems and Computer Science National University of Singapore