Changes between Initial Version and Version 1 of Architecture

Timestamp:

May 30, 2006, 4:54:12 PM (19 years ago)

Author:

soujak

Comment:

--

Architecture

@@ +1 @@
+= Project architecture =
+
+== 1. File system organization ==
+Kaya Phase 1 distribution is subdivided in several folders in order to 
+efficiently organize files and logically separate them into different sections.
+Such division is thought for being ready to receive Phase 2 implementation 
+providing a well structured environment.
+ * `src/`: sources folder containing Phase 1 primitive functions source code,
+   as described below in paragraph 2;
+ * `test/`: testing folder, containing a debugging program (`p1test.c`), related
+   environment generated by the Makefile "`test`" target; this rule compiles,
+   links, creates uMPS executable and virtual devices and finally run it on the
+   uMPS virtual machine (see also README);
+ * `doc/`: documentation root containing various developers informations about 
+   project internals;
+ * `include/`: library headers
+   * `*.e`: Phase 1 primitive functions interfaces: `pcb.e` and `st.e` 
+     respectively relative to `pcb.c` and `st.c`;
+   * `*.h`: system constants and macro definitions (`const.h`), types and data 
+     structures definitions (`types.h`);
+ * `lib/`: Phase 1 libraries objects files, ready for being linked.
+
+== 2. Source modules organization ==
+Following the evident logical division between two entities we have to 
+instantiate - processes and semaphores -, we opted for a two-modules subdivision:
+ * `pcb.c`: collects primitive functions for PCBs initialization/ allocation/
+   deallocation, process queue/ tree maintenance;
+ * `st.c`: collects primitive functions for SD tree initialization and PCB
+   enqueueing to/dequeueing from semaphores.
+
+{{{
+ * `check.c` contains an error check routine which is able to perform a simple type
+   check on the own data structures of Phase 1.
+}}}