Blob Blame History Raw
#!/usr/bin/env python3
#
# Script to check for inconsistencies between documented mount options
# and implemented kernel options.
# Copyright (C) 2018 Aurelien Aptel (aaptel@suse.com)
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.

import os
import sys
import re
import subprocess
import argparse
from pprint import pprint as P

def extract_canonical_opts(s):
    """
    Return list of option names present in s.
    e.g "opt1=a|opt2=d" => ["opt1", "opt2"])
    """
    opts = s.split("|")
    res = []
    for o in opts:
        x = o.split("=")
        res.append(x[0])
    return res

def extract_kernel_opts(fn):
    STATE_BASE = 0
    STATE_DEF = 1
    STATE_USE = 2
    STATE_EXIT = 3

    state = STATE_BASE
    fmt2enum = {}
    enum2code = {}
    code = ''
    current_opt = ''
    rx = RX()

    def code_add(s):
        if current_opt != '':
            if current_opt not in enum2code:
                enum2code[current_opt] = ''
            enum2code[current_opt] += s

    with open(fn) as f:
        for s in f.readlines():
            if state == STATE_EXIT:
                break

            elif state == STATE_BASE:
                if rx.search(r'cifs_mount_option_tokens.*\{', s):
                    state = STATE_DEF
                elif rx.search(r'^cifs_parse_mount_options', s):
                    state = STATE_USE

            elif state == STATE_DEF:
                if rx.search(r'(Opt_[a-zA-Z0-9_]+)\s*,\s*"([^"]+)"', s):
                    fmt = rx.group(2)
                    opts = extract_canonical_opts(fmt)
                    assert(len(opts) == 1)
                    name = opts[0]
                    fmt2enum[name] = {'enum':rx.group(1), 'fmt':fmt}
                elif rx.search(r'^};', s):
                    state = STATE_BASE

            elif state == STATE_USE:
                if rx.search(r'^\s*case (Opt_[a-zA-Z0-9_]+)', s):
                    current_opt = rx.group(1)
                elif current_opt != '' and rx.search(r'^\s*default:', s):
                    state = STATE_EXIT
                else:
                    code_add(s)
    return fmt2enum, enum2code

def chomp(s):
    if s[-1] == '\n':
        return s[:-1]
    return s

def extract_man_opts(fn):
    STATE_EXIT = 0
    STATE_BASE = 1
    STATE_OPT = 2

    state = STATE_BASE
    rx = RX()
    opts = {}
    ln = 0

    with open(fn) as f:
        for s in f.readlines():
            ln += 1

            if state == STATE_EXIT:
                break

            elif state == STATE_BASE:
                if rx.search(r'^OPTION', s):
                    state = STATE_OPT

            elif state == STATE_OPT:
                if rx.search('^[a-z]', s) and len(s) < 50:
                    s = chomp(s)
                    names = extract_canonical_opts(s)
                    for name in names:
                        if name not in opts:
                            opts[name] = []
                        opts[name].append({'ln':ln, 'fmt':s})
                elif rx.search(r'^[A-Z]+', s):
                    state = STATE_EXIT
    return opts

def format_code(s):
    # remove common indent in the block
    min_indent = None
    for ln in s.split("\n"):
        indent = 0
        for c in ln:
            if c == '\t': indent += 1
            else: break
        if min_indent is None:
            min_indent = indent
        elif indent > 0:
            min_indent = min(indent, min_indent)
    out = ''
    lines = s.split("\n")
    if lines[-1].strip() == '':
        lines.pop()
    for ln in lines:
        out += "| %s\n" % ln[min_indent:]
    return out

def sortedset(s):
    return sorted(list(s), key=lambda x: re.sub('^no', '', x))

def opt_neg(opt):
    if opt.startswith("no"):
        return opt[2:]
    else:
        return "no"+opt

def main():
    ap = argparse.ArgumentParser(description="Cross-check mount options from cifs.ko/man page")
    ap.add_argument("cfile", help="path to connect.c")
    ap.add_argument("rstfile", help="path to mount.cifs.rst")
    args = ap.parse_args()

    fmt2enum, enum2code = extract_kernel_opts(args.cfile)
    manopts = extract_man_opts(args.rstfile)

    kernel_opts_set = set(fmt2enum.keys())
    man_opts_set = set(manopts.keys())

    def opt_alias_is_doc(o):
        enum = fmt2enum[o]['enum']
        aliases = []
        for k,v in fmt2enum.items():
            if k != o and v['enum'] == enum:
                if opt_is_doc(k):
                    return k
        return None

    def opt_exists(o):
        return o in fmt2enum

    def opt_is_doc(o):
        return o in manopts

    print('DUPLICATED DOC OPTIONS')
    print('======================')

    for opt in sortedset(man_opts_set):
        if len(manopts[opt]) > 1:
            lines = ", ".join([str(x['ln']) for x in manopts[opt]])
            print("OPTION %-20.20s (lines %s)"%(opt, lines))
    print()

    print('UNDOCUMENTED OPTIONS')
    print('====================')

    undoc_opts = kernel_opts_set - man_opts_set
    # group opts and their negations together
    for opt in sortedset(undoc_opts):
        fmt = fmt2enum[opt]['fmt']
        enum = fmt2enum[opt]['enum']
        code = format_code(enum2code[enum])
        neg = opt_neg(opt)

        if enum == 'Opt_ignore':
            print("# skipping %s (Opt_ignore)\n"%opt)
            continue

        if opt_exists(neg) and opt_is_doc(neg):
            print("# skipping %s (%s is documented)\n"%(opt, neg))
            continue

        alias = opt_alias_is_doc(opt)
        if alias:
            print("# skipping %s (alias %s is documented)\n"%(opt, alias))
            continue

        print('OPTION %s ("%s" -> %s):\n%s'%(opt, fmt, enum, code))

    print('')
    print('DOCUMENTED BUT NON-EXISTING OPTIONS')
    print('===================================')

    unex_opts = man_opts_set - kernel_opts_set
    # group opts and their negations together
    for opt in sortedset(unex_opts):
        man = manopts[opt][0]
        print('OPTION %s ("%s") line %d' % (opt, man['fmt'], man['ln']))


    print('')
    print('NEGATIVE OPTIONS WITHOUT POSITIVE')
    print('=================================')

    for opt in sortedset(kernel_opts_set):
        if not opt.startswith('no'):
            continue

        neg = opt[2:]
        if not opt_exists(neg):
            print("OPTION %s exists but not %s"%(opt,neg))

# little helper to test AND store result at the same time so you can
# do if/elsif easily instead of nesting them when you need to do
# captures
class RX:
    def __init__(self):
        pass
    def search(self, rx, s, flags=0):
        self.r = re.search(rx, s, flags)
        return self.r
    def group(self, n):
        return self.r.group(n)

if __name__ == '__main__':
    main()