|
| 1 | +# Retrieves the sequence of PCode ops for each function, and organizes these lists |
| 2 | +# into a JSON file |
| 3 | +# |
| 4 | +# @category colemankane |
| 5 | +# |
| 6 | +# Use the Python json library |
| 7 | +import json |
| 8 | + |
| 9 | +# Add the Python argument parser |
| 10 | +from argparse import ArgumentParser |
| 11 | + |
| 12 | +# Import some of the Ghidra classes we will be using |
| 13 | +from ghidra.util.task import ConsoleTaskMonitor |
| 14 | + |
| 15 | +# Map each Ghidra PCode op to a unique byte value |
| 16 | +pcode_maps = { |
| 17 | + 'BOOL_AND': 'A', |
| 18 | + 'BOOL_NEGATE': 'B', |
| 19 | + 'BOOL_OR': 'C', |
| 20 | + 'BOOL_XOR': 'D', |
| 21 | + 'BRANCH': 'E', |
| 22 | + 'BRANCHIND': 'F', |
| 23 | + 'CALL': 'G', |
| 24 | + 'CALLIND': 'H', |
| 25 | + 'CALLOTHER': 'I', |
| 26 | + 'CAST': 'J', |
| 27 | + 'CBRANCH': 'K', |
| 28 | + 'COPY': 'L', |
| 29 | + 'CPOOLREF': 'M', |
| 30 | + 'EXTRACT': 'N', |
| 31 | + 'FLOAT_ABS': 'O', |
| 32 | + 'FLOAT_ADD': 'P', |
| 33 | + 'FLOAT_CEIL': 'Q', |
| 34 | + 'FLOAT_DIV': 'R', |
| 35 | + 'FLOAT_EQUAL': 'S', |
| 36 | + 'FLOAT_FLOAT2FLOAT': 'T', |
| 37 | + 'FLOAT_FLOOR': 'U', |
| 38 | + 'FLOAT_INT2FLOAT': 'V', |
| 39 | + 'FLOAT_LESS': 'W', |
| 40 | + 'FLOAT_LESSEQUAL': 'X', |
| 41 | + 'FLOAT_MULT': 'Y', |
| 42 | + 'FLOAT_NAN': 'Z', |
| 43 | + 'FLOAT_NEG': 'a', |
| 44 | + 'FLOAT_NOTEQUAL': 'b', |
| 45 | + 'FLOAT_ROUND': 'c', |
| 46 | + 'FLOAT_SQRT': 'd', |
| 47 | + 'FLOAT_SUB': 'e', |
| 48 | + 'FLOAT_TRUNC': 'f', |
| 49 | + 'INDIRECT': 'g', |
| 50 | + 'INSERT': 'h', |
| 51 | + 'INT_2COMP': 'i', |
| 52 | + 'INT_ADD': 'j', |
| 53 | + 'INT_AND': 'k', |
| 54 | + 'INT_CARRY': 'l', |
| 55 | + 'INT_DIV': 'm', |
| 56 | + 'INT_EQUAL': 'n', |
| 57 | + 'INT_LEFT': 'o', |
| 58 | + 'INT_LESS': 'p', |
| 59 | + 'INT_LESSEQUAL': 'q', |
| 60 | + 'INT_MULT': 'r', |
| 61 | + 'INT_NEGATE': 's', |
| 62 | + 'INT_NOTEQUAL': 't', |
| 63 | + 'INT_OR': 'u', |
| 64 | + 'INT_REM': 'v', |
| 65 | + 'INT_RIGHT': 'w', |
| 66 | + 'INT_SBORROW': 'x', |
| 67 | + 'INT_SCARRY': 'y', |
| 68 | + 'INT_SDIV': 'z', |
| 69 | + 'INT_SEXT': '0', |
| 70 | + 'INT_SLESS': '1', |
| 71 | + 'INT_SLESSEQUAL': '2', |
| 72 | + 'INT_SREM': '3', |
| 73 | + 'INT_SRIGHT': '4', |
| 74 | + 'INT_SUB': '5', |
| 75 | + 'INT_XOR': '6', |
| 76 | + 'INT_ZEXT': '7', |
| 77 | + 'LOAD': '8', |
| 78 | + 'MULTIEQUAL': '9', |
| 79 | + 'NEW': '0', |
| 80 | + 'PCODE_MAX': '~', |
| 81 | + 'PIECE': '!', |
| 82 | + 'POPCOUNT': '@', |
| 83 | + 'PTRADD': '#', |
| 84 | + 'PTRSUB': '$', |
| 85 | + 'RETURN': '%', |
| 86 | + 'SEGMENTOP': '^', |
| 87 | + 'STORE': '&', |
| 88 | + 'SUBPIECE': '*', |
| 89 | + 'UNIMPLEMENTED': '-' |
| 90 | +} |
| 91 | + |
| 92 | +# Initialize an empty dict for the "all functions" report |
| 93 | +fn_report = {} |
| 94 | + |
| 95 | +# Set up parser for the script arguments |
| 96 | +arg_parser = ArgumentParser(description="P-Code statistical analysis", prog='script', prefix_chars='+') |
| 97 | + |
| 98 | +# The "raw" JSON output file containing all the details |
| 99 | +arg_parser.add_argument('+o', '++output', required=True, help='Output file for JSON') |
| 100 | + |
| 101 | +# Optionally, a text file for the "signature strings" |
| 102 | +arg_parser.add_argument('+s', '++strings', required=False, default=None, help='Output file for strings') |
| 103 | + |
| 104 | +# Parse the arguments like a normal Python program |
| 105 | +args = arg_parser.parse_args(args=getScriptArgs()) |
| 106 | + |
| 107 | +# the Program.getFunctionManager() provides an interface to navigate the functions |
| 108 | +# that Ghidra has found within the program. The getFunctions() method will provide |
| 109 | +# an iterator that allows you to walk through the list forward (True) or |
| 110 | +# backward (False). |
| 111 | +for fn in getCurrentProgram().getFunctionManager().getFunctions(True): |
| 112 | + |
| 113 | + # Get the earliest instruction defined within the function, to start our exploration |
| 114 | + instr = getInstructionAt(fn.getBody().getMinAddress()) |
| 115 | + |
| 116 | + # Walk through each instruction that's determined to be part of this function |
| 117 | + while instr and instr.getMinAddress() <= fn.getBody().getMaxAddress(): |
| 118 | + if fn.getBody().contains(instr.getMinAddress()): |
| 119 | + # Iterate across the list of P-Code operations that are expanded from |
| 120 | + # the parsed machine instruction |
| 121 | + for pcode_op in instr.getPcode(): |
| 122 | + |
| 123 | + # Get the string name of the PCode operation |
| 124 | + pcode_name = pcode_op.getMnemonic() |
| 125 | + |
| 126 | + # Create a new empty list for this function the first time we get a valid instruction |
| 127 | + # This way we can easily assume to use .append() below |
| 128 | + if fn.getName() not in fn_report: |
| 129 | + fn_report[fn.getName()] = [] |
| 130 | + |
| 131 | + # Push the PCode op name on the end of the list |
| 132 | + fn_report[fn.getName()].append(pcode_name) |
| 133 | + |
| 134 | + # Advance to the next instruction |
| 135 | + instr = instr.getNext() |
| 136 | + |
| 137 | +# Now, open the file provided by the user, and write the JSON into it |
| 138 | +with open(args.output, 'w') as outfile: |
| 139 | + outfile.write(json.dumps(fn_report)) |
| 140 | + |
| 141 | +if args.strings: |
| 142 | + # Finally, if provided, condense the PCode lists into a string where each byte represents a single |
| 143 | + # PCode opcode, as mapped in the pcode_maps data structure |
| 144 | + with open(args.strings, 'w') as stringsfile: |
| 145 | + for fn in fn_report.keys(): |
| 146 | + outstr = ''.join(pcode_maps[x] for x in fn_report[fn]) |
| 147 | + stringsfile.write('{fn}:{s}\n'.format(fn=fn, s=outstr)) # Write one signature per line |
0 commit comments