search

saeaadl / arinc653

0 stars 0 forks source link

Clarify use of Scheduling_Protocol #3

Open stevevestal opened 5 years ago

stevevestal commented 5 years ago

The annex states that the Scheduling_Protocol is required on virtual processors (partitions). The text of the standard says, "Scheduling_Protocol defines the scheduling policy used inside each ARINC 653 partition." Only in an example does it show that this is RMS for a partition.

The text does not state that Scheduling_Protocol is required on processors (A653 modules). It is listed in the table of processor properties, but that could be interpreted as meaning it is allowed but optional.

The recommendation is to clarify in the text that the standard Scheduling_Protocol for partitions is RMS; and that the Scheduling_Protocol is required for processors with value ARINC653 (so that system models can unambiguously mix ARINC 653 processors with processors that use other scheduling protocols).

jjhugues commented 4 years ago

Pierre recalled that this was suggested in one of the draft to have such enumerator, got dropped at some point.

Having an annex-specific enumerator would reduce the complexity to analyze the context and select the right tool configuration

Have the scheduling_protocol mandatory for processors is a better approach and clarifies the design intent

jjhugues commented 3 years ago

For the records, OSATE already has an ARINC653 scheduling protocol. If memory serves me, Ellidiss tools also have it. I added to the document that this enumerator must be used