OpenStack-Neutron-Fwaas-代码【二】

　　上一节从代码层面来讲解了fwaas的流程，这里通过具体查看iptables规则来说下应用规则的流程：

1.首先通过命令获取当前路由中的规则

#ip netns exec qrouter-[router-uuid] iptables-save -c >> rules

2.查看文件，这里我给定一个实例文件，这里因为我们只关心filter表，其他表去除

# Generated by iptables-save v1.4.21 on Fri May 22 15:59:01 2015
*filter
:INPUT ACCEPT [6425:953808]
:FORWARD ACCEPT [5:408]
:OUTPUT ACCEPT [9:972]  #这里往上是三个内置链

:neutron-filter-top - [0:0] #这是唯一一个非包装链

:neutron-l3-agent-FORWARD - [0:0]
:neutron-l3-agent-INPUT - [0:0]
:neutron-l3-agent-OUTPUT - [0:0]
:neutron-l3-agent-fwaas-defau - [0:0]
:neutron-l3-agent-iv4f7e1ffb5 - [0:0]
:neutron-l3-agent-local - [0:0]
:neutron-l3-agent-ov4f7e1ffb5 - [0:0]
:neutron-openvswi-FORWARD - [0:0]
:neutron-openvswi-INPUT - [0:0]
:neutron-openvswi-OUTPUT - [0:0]
:neutron-openvswi-ia45dd659-2 - [0:0]
:neutron-openvswi-local - [0:0]
:neutron-openvswi-oa45dd659-2 - [0:0]
:neutron-openvswi-sg-chain - [0:0]
:neutron-openvswi-sg-fallback - [0:0]
:neutron-vpn-agen-FORWARD - [0:0]
:neutron-vpn-agen-INPUT - [0:0]
:neutron-vpn-agen-OUTPUT - [0:0]
:neutron-vpn-agen-fwaas-defau - [0:0]
:neutron-vpn-agen-iv4f7e1ffb5 - [0:0]
:neutron-vpn-agen-local - [0:0]
:neutron-vpn-agen-ov4f7e1ffb5 - [0:0]  #这里是包装链，包装名称通过os.path.basename(inspect.stack()[-1][1])[:16]获取
[15606:1868567] -A INPUT -j neutron-vpn-agen-INPUT
[15621:1870099] -A INPUT -j neutron-l3-agent-INPUT
[10406:1260231] -A INPUT -j neutron-openvswi-INPUT #往上三个将input包装链作为跳转规则添加到内置input链中
[5:408] -A FORWARD -j neutron-filter-top
[5:408] -A FORWARD -j neutron-vpn-agen-FORWARD
[5:408] -A FORWARD -j neutron-l3-agent-FORWARD
[0:0] -A FORWARD -j neutron-openvswi-FORWARD #往上三个将forward包装链作为跳转规则添加到内置forward链中
[9:972] -A OUTPUT -j neutron-filter-top
[9:972] -A OUTPUT -j neutron-vpn-agen-OUTPUT
[9:972] -A OUTPUT -j neutron-l3-agent-OUTPUT
[0:0] -A OUTPUT -j neutron-openvswi-OUTPUT #往上三个将output包装链作为跳转规则添加到内置output链中
[14:1380] -A neutron-filter-top -j neutron-vpn-agen-local
[14:1380] -A neutron-filter-top -j neutron-l3-agent-local
[0:0] -A neutron-filter-top -j neutron-openvswi-local
[0:0] -A neutron-l3-agent-FORWARD -o qr-+ -j neutron-l3-agent-iv4f7e1ffb5
[0:0] -A neutron-l3-agent-FORWARD -i qr-+ -j neutron-l3-agent-ov4f7e1ffb5 #将neutron-l3-agent-ov4f7e1ffb5链作为跳转规则添加到neutron-l3-agent-FORWARD
[0:0] -A neutron-l3-agent-FORWARD -o qr-+ -j neutron-l3-agent-fwaas-defau
[0:0] -A neutron-l3-agent-FORWARD -i qr-+ -j neutron-l3-agent-fwaas-defau  #将neutron-l3-agent-fwaas-defau链作为跳转规则添加到neutron-l3-agent-FORWARD
[0:0] -A neutron-l3-agent-INPUT -d 127.0.0.1/32 -p tcp -m tcp --dport 9697 -j ACCEPT #默认规则

[0:0] -A neutron-l3-agent-fwaas-defau -j DROP #neutron-l3-agent-fwaas-defau链中只有一条drop规则
[0:0] -A neutron-l3-agent-iv4f7e1ffb5 -m state --state INVALID -j DROP
[0:0] -A neutron-l3-agent-iv4f7e1ffb5 -m state --state RELATED,ESTABLISHED -j ACCEPT #这两条规则是在初始化neutron-l3-agent-iv4f7e1ffb5链的时候添加的
[0:0] -A neutron-l3-agent-iv4f7e1ffb5 -s 10.10.10.0/24 -d 192.168.1.0/24 -p tcp -j DROP #界面添加的用户规则
[0:0] -A neutron-l3-agent-ov4f7e1ffb5 -m state --state INVALID -j DROP
[0:0] -A neutron-l3-agent-ov4f7e1ffb5 -m state --state RELATED,ESTABLISHED -j ACCEPT #这两条规则是在初始化neutron-l3-agent-ov4f7e1ffb5链的时候添加的
[0:0] -A neutron-l3-agent-ov4f7e1ffb5 -s 10.10.10.0/24 -d 192.168.1.0/24 -p tcp -j DROP #界面添加的用户规则 ,可以看到是同时将界面规则添加到出入链中
[0:0] -A neutron-openvswi-FORWARD -i qg-a45dd659-23 -j neutron-openvswi-sg-chain
[0:0] -A neutron-openvswi-FORWARD -o qg-a45dd659-23 -j neutron-openvswi-sg-chain
[10406:1260231] -A neutron-openvswi-INPUT -j neutron-openvswi-oa45dd659-2
[0:0] -A neutron-openvswi-ia45dd659-2 -m state --state INVALID -j DROP
[0:0] -A neutron-openvswi-ia45dd659-2 -m state --state RELATED,ESTABLISHED -j RETURN
[0:0] -A neutron-openvswi-ia45dd659-2 -s 192.168.1.12/32 -p udp -m udp --sport 67 --dport 68 -j RETURN
[0:0] -A neutron-openvswi-ia45dd659-2 -j neutron-openvswi-sg-fallback
[983:322818] -A neutron-openvswi-oa45dd659-2 -p udp -m udp --sport 68 --dport 67 -j RETURN
[0:0] -A neutron-openvswi-oa45dd659-2 -p tcp -m tcp --dport 22 -j RETURN
[125:41000] -A neutron-openvswi-oa45dd659-2 -p udp -m udp --sport 67 --dport 68 -j DROP
[0:0] -A neutron-openvswi-oa45dd659-2 -m state --state INVALID -j DROP
[0:0] -A neutron-openvswi-oa45dd659-2 -m state --state RELATED,ESTABLISHED -j RETURN
[9298:896413] -A neutron-openvswi-oa45dd659-2 -j neutron-openvswi-sg-fallback
[0:0] -A neutron-openvswi-sg-chain -i qg-a45dd659-23 -j neutron-openvswi-ia45dd659-2
[0:0] -A neutron-openvswi-sg-chain -o qg-a45dd659-23 -j neutron-openvswi-oa45dd659-2
[0:0] -A neutron-openvswi-sg-chain -j ACCEPT
[9298:896413] -A neutron-openvswi-sg-fallback -j DROP
[0:0] -A neutron-vpn-agen-FORWARD -o qr-+ -j neutron-vpn-agen-iv4f7e1ffb5
[0:0] -A neutron-vpn-agen-FORWARD -i qr-+ -j neutron-vpn-agen-ov4f7e1ffb5
[0:0] -A neutron-vpn-agen-FORWARD -o qr-+ -j neutron-vpn-agen-fwaas-defau
[0:0] -A neutron-vpn-agen-FORWARD -i qr-+ -j neutron-vpn-agen-fwaas-defau
[0:0] -A neutron-vpn-agen-INPUT -d 127.0.0.1/32 -p tcp -m tcp --dport 9697 -j ACCEPT
[0:0] -A neutron-vpn-agen-fwaas-defau -j DROP
[0:0] -A neutron-vpn-agen-iv4f7e1ffb5 -m state --state INVALID -j DROP
[0:0] -A neutron-vpn-agen-iv4f7e1ffb5 -m state --state RELATED,ESTABLISHED -j ACCEPT
[0:0] -A neutron-vpn-agen-iv4f7e1ffb5 -s 10.10.10.0/24 -d 192.168.1.0/24 -p tcp -j DROP
[0:0] -A neutron-vpn-agen-ov4f7e1ffb5 -m state --state INVALID -j DROP
[0:0] -A neutron-vpn-agen-ov4f7e1ffb5 -m state --state RELATED,ESTABLISHED -j ACCEPT
[0:0] -A neutron-vpn-agen-ov4f7e1ffb5 -s 10.10.10.0/24 -d 192.168.1.0/24 -p tcp -j DROP
COMMIT
# Completed on Fri May 22 16:48:38 2015

　　

3. 代码中应用规则是这样的流程

代码路径：neutron\agent\linux\iptables_manager.py

def _apply_synchronized(self):
        """Apply the current in-memory set of iptables rules.

        This will blow away any rules left over from previous runs of the
        same component of Nova, and replace them with our current set of
        rules. This happens atomically, thanks to iptables-restore.

        """
        s = [('iptables', self.ipv4)]
        if self.use_ipv6:
            s += [('ip6tables', self.ipv6)]

        for cmd, tables in s:
            args = ['%s-save' % (cmd,), '-c']
            if self.namespace:
                args = ['ip', 'netns', 'exec', self.namespace] + args
            all_tables = self.execute(args, root_helper=self.root_helper)  #将已有的规则保存到all_tables对象中
            all_lines = all_tables.split('\n')
            # Traverse tables in sorted order for predictable dump output
            for table_name in sorted(tables):
                table = tables[table_name]
                start, end = self._find_table(all_lines, table_name) #查表，实际看下保存的结果就可以很容易的看出他的数据结构了
                all_lines[start:end] = self._modify_rules(  # 【3.1】重点看下这个方法
                    all_lines[start:end], table, table_name)

            args = ['%s-restore' % (cmd,), '-c']
            if self.namespace:
                args = ['ip', 'netns', 'exec', self.namespace] + args
            try:
                self.execute(args, process_input='\n'.join(all_lines),
                             root_helper=self.root_helper) #将更新好的规则重新应用
            except RuntimeError as r_error:
                with excutils.save_and_reraise_exception():
                    try:
                        line_no = int(re.search(
                            'iptables-restore: line ([0-9]+?) failed',
                            str(r_error)).group(1))
                        context = IPTABLES_ERROR_LINES_OF_CONTEXT
                        log_start = max(0, line_no - context)
                        log_end = line_no + context
                    except AttributeError:
                        # line error wasn't found, print all lines instead
                        log_start = 0
                        log_end = len(all_lines)
                    log_lines = ('%7d. %s' % (idx, l)
                                 for idx, l in enumerate(
                                     all_lines[log_start:log_end],
                                     log_start + 1)
                                 )
                    LOG.error(_("IPTablesManager.apply failed to apply the "
                                "following set of iptables rules:\n%s"),
                              '\n'.join(log_lines))
        LOG.debug(_("IPTablesManager.apply completed with success"))

【3.1】　过滤规则

def _modify_rules(self, current_lines, table, table_name):
        # Chains are stored as sets to avoid duplicates.
        # Sort the output chains here to make their order predictable.
        unwrapped_chains = sorted(table.unwrapped_chains) #排序非包装链，我们知道就只有一个
        chains = sorted(table.chains) #排序包装链
        remove_chains = table.remove_chains
        rules = table.rules
        remove_rules = table.remove_rules

        if not current_lines:
            fake_table = ['# Generated by iptables_manager',
                          '*' + table_name, 'COMMIT',
                          '# Completed by iptables_manager']
            current_lines = fake_table

        # Fill old_filter with any chains or rules we might have added,
        # they could have a [packet:byte] count we want to preserve.
        # Fill new_filter with any chains or rules without our name in them.
        old_filter, new_filter = [], []
        for line in current_lines:
            (old_filter if self.wrap_name in line else
             new_filter).append(line.strip())  #将原有规则过滤的包数和字节数保存

        rules_index = self._find_rules_index(new_filter)

        all_chains = [':%s' % name for name in unwrapped_chains]  #添加完内置链后第一个添加非包装链
        all_chains += [':%s-%s' % (self.wrap_name, name) for name in chains]  #添加包装链

        # Iterate through all the chains, trying to find an existing
        # match.
        our_chains = []
        for chain in all_chains:
            chain_str = str(chain).strip()

            old = self._find_last_entry(old_filter, chain_str)
            if not old:
                dup = self._find_last_entry(new_filter, chain_str)
            new_filter = [s for s in new_filter if chain_str not in s.strip()]

            # if no old or duplicates, use original chain
            if old or dup:
                chain_str = str(old or dup)
            else:
                # add-on the [packet:bytes]
                chain_str += ' - [0:0]'

            our_chains += [chain_str]  #添加所有的自定义链

        # Iterate through all the rules, trying to find an existing
        # match.
        our_rules = []
        bot_rules = []
        for rule in rules:  #开始添加规则了
            rule_str = str(rule).strip()  #查看下IptablesRule对象后，会看到实现了_str__方法，返回经过处理的字符串
            # Further down, we weed out duplicates from the bottom of the
            # list, so here we remove the dupes ahead of time.

            old = self._find_last_entry(old_filter, rule_str)
            if not old:
                dup = self._find_last_entry(new_filter, rule_str)
            new_filter = [s for s in new_filter if rule_str not in s.strip()]

            # if no old or duplicates, use original rule
            if old or dup:
                rule_str = str(old or dup)
                # backup one index so we write the array correctly
                if not old:
                    rules_index -= 1
            else:
                # add-on the [packet:bytes]
                rule_str = '[0:0] ' + rule_str

            if rule.top:
                # rule.top == True means we want this rule to be at the top.
                our_rules += [rule_str]
            else:
                bot_rules += [rule_str]

        our_rules += bot_rules  #直接添加规则。规则之间的逻辑顺序其实在创建规则的时候已经给定了（因为已经给定了这个规则所在的链）

        new_filter[rules_index:rules_index] = our_rules #替换规则，注意添加完了后文件中的行数变化了
        new_filter[rules_index:rules_index] = our_chains #替换链，位置是在规则前面

        def _strip_packets_bytes(line):
            # strip any [packet:byte] counts at start or end of lines
            if line.startswith(':'):
                # it's a chain, for example, ":neutron-billing - [0:0]"
                line = line.split(':')[1]
                line = line.split(' - [', 1)[0]
            elif line.startswith('['):
                # it's a rule, for example, "[0:0] -A neutron-billing..."
                line = line.split('] ', 1)[1]
            line = line.strip()
            return line

        seen_chains = set()

        def _weed_out_duplicate_chains(line):
            # ignore [packet:byte] counts at end of lines
            if line.startswith(':'):
                line = _strip_packets_bytes(line)
                if line in seen_chains:
                    return False
                else:
                    seen_chains.add(line)

            # Leave it alone
            return True

        seen_rules = set()

        def _weed_out_duplicate_rules(line):
            if line.startswith('['):
                line = _strip_packets_bytes(line)
                if line in seen_rules:
                    return False
                else:
                    seen_rules.add(line)

            # Leave it alone
            return True

        def _weed_out_removes(line):
            # We need to find exact matches here
            if line.startswith(':'):
                line = _strip_packets_bytes(line)
                for chain in remove_chains:
                    if chain == line:
                        remove_chains.remove(chain)
                        return False
            elif line.startswith('['):
                line = _strip_packets_bytes(line)
                for rule in remove_rules:
                    rule_str = _strip_packets_bytes(str(rule))
                    if rule_str == line:
                        remove_rules.remove(rule)
                        return False

            # Leave it alone
            return True

        # We filter duplicates.  Go through the chains and rules, letting
        # the *last* occurrence take precedence since it could have a
        # non-zero [packet:byte] count we want to preserve.  We also filter
        # out anything in the "remove" list.
        new_filter.reverse()
        new_filter = [line for line in new_filter
                      if _weed_out_duplicate_chains(line) and
                      _weed_out_duplicate_rules(line) and
                      _weed_out_removes(line)]
        new_filter.reverse()

        # flush lists, just in case we didn't find something
        remove_chains.clear()
        for rule in remove_rules:
            remove_rules.remove(rule)  #善后工作

        return new_filter #返回过滤完的规则列表