On Tue, 22 Jan 2008 01:31:45 -0500, Robert A. Rosenberg wrote: >At 1:38 PM -0600 on 1/21/08, Paul Gilmartin wrote about Re: How does >ATTACH pass address of ECB to child?: > >>When the mother needs the daughter's assistance, it WAITs on the >>second ECB, clears it, and POSTS the first. > >ONLY IF the mother task has nothing else it can do while it is >waiting for the daughter to be ready to do something for it. >Otherwise the mother task just periodically tests ECB2 and does its >POST once ECB1 has been WAIT'ed on by the daughter. > By the protocol I described, the daughter POSTs ECB2 only after it has cleared ECB1 and is ready to WAIT on ECB1. There is no need for the mother task then to make any decision based on the state of ECB1 -- it may POST either before or after the daughter WAITs.
And I deliberately omitted any mention of WAIT on multiple ECBs and EVENTS, only partly because I'm unfamiliar with them. Interestingly, UNIX has a very similar wait() function which waits only on completion of daughter processes. And I had to submit a PMR on incorrect operation of Unix System Services waitpid(), which operated correctly if issued before the daughter process terminated, but gave incorrrect results if the daughter terminated before the mother issued the waitpid(). -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
