ie53@NTSUVAX.BITNET (HAERIM LEE) (08/26/87)
VAx Ada(tm) LRM 9.5.13 (?) says: If several tasks call the same entry before a corresponding accept statement is reached, the calls are queued. .. The calls are processed in the order of arrival. My question is: Are those entry calls queued in the order of calls? In other words, is the order of call the same as that of arrival? -- Haerim Lee (IE53@NTSUVAX)
tom@telesoft.UUCP (Tom Quiggle @spot) (08/28/87)
> VAx Ada(tm) LRM 9.5.13 (?) says: If several tasks call the same entry ^^^^^^ FYI: 9.5(15) ANSI-STD-1815A-1983 > before a corresponding accept statement is reached, the calls are queued. > .. > The calls are processed in the order of arrival. > > My question is: Are those entry calls queued in the order of calls? > In other words, is the order of call the same as that of arrival? > > -- Haerim Lee (IE53@NTSUVAX) > Not necessarily: In a loosely coupled distributed system, there is a discernable delay between the moment when an entry call is made and the moment the call is placed on the associated entry queue. Depending on the characteristics of the underlying communication system, multiple messages requesting that calls be placed on an entry queue may not reach the destination machine in the order in which the messages were presented to the communication system; therefore the callers may not be queued in the order the calls were generated. On a uniprocessor system there is still a finite amount of processing which must take place in order to execute an entry call. LRM 9.5(10) states: Execution of an entry call statement starts with the evaluation of the entry name; this is followed by any evaluations required for actual parameters in the same manner as for a subprogram call (see 6.4). [Presumably the enqueuing of the call on the associated entry queue occurs immediately after the entry name and actual parameters have been evaluated; although the LRM does not explicitly state this. One could argue that an implementation is free to schedule any task it chooses at this point, and therefore the answer to your question is NO. A reasonable implementation will not take this attitude, so I will further consider the question.] If during the execution of the entry call statement (i.e. the evaluation of the entry name and actual parameters) another task with higher priority becomes eligible for execution (say by completion of a delay statement), the execution of the entry call can be suspended. If this higher priority task proceeded to call the same entry, it will be queued first. As a further aside, the evaluation of the actual parameters to the entry call may result in scheduling actions which alter the order of arrival of entry calls. For instance the evaluation of an actual parameter may include a call to a function which itself calls the same entry. Tom -- Tom Quiggle TeleSoft (619)457-2700x158 ucbvax--| hp-sdd--| telesoft!tom@UCSD.ARPA decvax--+--sdcsvax--+--telesoft ihnp4--| noscvax--|