add power_table.py start of creating markdown Appendix D tables

diff --git a/src/openpower/decoder/power_table.py b/src/openpower/decoder/power_table.py
new file mode 100644 (file)
index 0000000..da35cdb
--- /dev/null
+++ b/src/openpower/decoder/power_table.py
@@ -0,0 +1,58 @@
+import pathlib
+from openpower.decoder.power_enums import find_wiki_dir 
+from openpower.decoder.power_insn import (Database, MarkdownDatabase,
+                                          FieldsDatabase, PPCDatabase,
+                                          IntegerOpcode, PatternOpcode,
+                                          parse, Section)
+root = find_wiki_dir()
+root = pathlib.Path(root)
+mdwndb = MarkdownDatabase()
+fieldsdb = FieldsDatabase()
+
+# create by-sections first. these will be the markdown tables
+sections = {}
+insns = {}
+path = (root / "insndb.csv")
+with open(path, "r", encoding="UTF-8") as stream:
+    for section in parse(stream, Section.CSV):
+        sections[str(section.path)] = section
+        insns[str(section.path)] = []
+for (name, section) in sections.items():
+        print (name, section)
+
+# enumerate all instructions and drop them into sections
+db = Database(root)
+for insn in db:
+    insns[str(insn.section.path)].append(insn)
+    print (insn)
+
+def do_table(insns, section):
+    start, end = section.bitsel.start, section.bitsel.end
+    print ("start-end", start, end)
+    bitlen = end-start+1
+    half = bitlen // 2
+    xomask = ((1<<bitlen)-1) << start
+    lowermask = (1<<half)-1
+    uppermask = (1<<(bitlen-half))-1
+    table_entries = {}
+    for i in range(1<<bitlen):
+        # calculate row, column
+        lower = i & lowermask
+        upper = (i>>half) & uppermask
+        print (i, bin(lower), bin(upper))
+        if lower not in table_entries:
+            table_entries[lower] = {}
+        table_entries[lower][upper] = i
+        # create an XO
+        key = i << start
+        print ("search", i, hex(key))
+        # start hunting
+        for insn in insns:
+            opcode = insn.opcode
+            if not isinstance(opcode, list):
+                opcode = [opcode]
+            for op in opcode:
+                if ((op.value & op.mask & xomask) == (key & op.mask & xomask)):
+                    print ("match", i, hex(key), insn.name)
+
+do_table(insns['minor_30.csv'], sections['minor_30.csv'])