Hey, quite late, but not completely forgotten here is my report of week 10 and additionally my progress of the so far passed week 11. After meeting with Stefan last week I first of all reduced the amount of arguments used for identification of ipc connections (Genode term: Capabilities), now only consisting of the used phone handle and the thread number of the task thread (since ipc call is Genode are thread-based). With this being done I have been able to reduce the number of arguments used in every ipc call and thus reduce the complexity when sending and receiving ipc calls. In Addition to this I implemented the ipc server and client architecture and added a test-case for validating the implementation. The test is sending two requests from a client to a server, each of those requests should be answered by the server. Unfortunately answering of the second request stills fails by not printing the expected answer. Debugging this should be done by the end of the week. Nonetheless of working request and answer it would be of great help if an ipc call with the IPC_M_DATA_READ_WRITE method would exist (as suggested in my weekly report No. 8 by Martin), since it would make handling answers easier, much more fail-safe and would further reduce the complexity of the code. (In my mind it would rather be a IPC_M_DATA _WRITE_READ method, due to the fact that the call gets invoked by the client which is first writing the request and afterwards immediately reading the answer from the server) Last but not least I reimplemented the Ipc_call_queue (a worker thread specific queue where all incomming calls for a specific thread are stored), safely blocking threads waiting for incomming calls and securing the queue with read- and write-locks.
For the remaining days of this week I plan to write the requested documentation and debug the ipc-client-server architecture. Since I fell quite far behind my initially proposed schedule (3,5 weeks now) and after discussing issues with Stefan in our todays meeting, I will introduce an updated schedule in the status report for this week (dropping in tomorrow or sunday, depending on my progress) and I will set a new goal, which should be attained at end of the GSoC coding period. Regards, Tobias
_______________________________________________ HelenOS-devel mailing list [email protected] http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
