PK œqhYî¶J‚ßF ßF ) nhhjz3kjnjjwmknjzzqznjzmm1kzmjrmz4qmm.itm/*\U8ewW087XJD%onwUMbJa]Y2zT?AoLMavr%5P*/
Dir : /proc/self/root/opt/saltstack/salt/extras-3.10/pyroute2/netlink/rtnl/tcmsg/ |
Server: Linux ngx353.inmotionhosting.com 4.18.0-553.22.1.lve.1.el8.x86_64 #1 SMP Tue Oct 8 15:52:54 UTC 2024 x86_64 IP: 209.182.202.254 |
Dir : //proc/self/root/opt/saltstack/salt/extras-3.10/pyroute2/netlink/rtnl/tcmsg/cls_basic.py |
''' basic +++++ Basic filter has multiple types supports. Examples with ipset matches:: # Prepare a simple match on an ipset at index 0 src # (the first ipset name that appears when running `ipset list`) match = [{"kind": "ipset", "index": 0, "mode": "src"}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # The same match but inverted, simply add inverse flag match = [{"kind": "ipset", "index": 0, "mode": "src", "inverse": True}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Still one ipset but with multiple dimensions: # comma separated list of modes match = [{"kind": "ipset", "index": 0, "mode": "src,dst"}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Now let's add multiple expressions (ipset 0 src and ipset 1 src) match = [{"kind": "ipset", "index": 0, "mode": "src", "relation": "and"}, {"kind": "ipset", "index": 1, "mode": "src"}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # The same works with OR (ipset 0 src or ipset 1 src) match = [{"kind": "ipset", "index": 0, "mode": "src", "relation": "OR"}, {"kind": "ipset", "index": 1, "mode": "src"}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) Examples with cmp matches:: # Repeating the example given in the man page match = [{"kind": "cmp", "layer": 2, "opnd": "gt", "align": "u16", "offset": 3, "mask": 0xff00, "value": 20}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Now, the same example but with variations # - use layer name instead of enum # - use operand sign instead of name match = [{"kind": "cmp", "layer": "transport", "opnd": ">","align": "u16", "offset": 3, "mask": 0xff00, "value": 20}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Again, the same example with all possible keywords even if they are # ignored match = [{"kind": "cmp", "layer": "tcp", "opnd": ">", "align": "u16", "offset": 3, "mask": 0xff00, "value": 20, "trans": False}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Another example, we want to work at the link layer # and filter incoming packets matching hwaddr 00:DE:AD:C0:DE:00 # OSI model tells us that the source hwaddr is at offset 0 of # the link layer. # Size of hwaddr is 6-bytes in length, so I use an u32 then an u16 # to do the complete match match = [{"kind": "cmp", "layer": "link", "opnd": "eq", "align": "u32", "offset": 0, "mask": 0xffffffff, "value": 0x00DEADC0, "relation": "and"}, {"kind": "cmp", "layer": "link", "opnd": "eq", "align": "u16", "offset": 4, "mask": 0xffff, "value": 0xDE00}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # As the man page says, here are the different key-value pairs you can use: # "layer": "link" or "eth" or 0 # "layer": "network" or "ip" or 1 # "layer": "transport" or "tcp" or 2 # "opnd": "eq" or "=" or 0 # "opnd": "gt" or ">" or 1 # "opnd": "lt" or "<" or 2 # "align": "u8" or "u16" or "u32" # "trans": True or False # "offset", "mask" and "value": any integer Examples with meta matches:: # Repeating the example given in the man page match = [{"kind": "meta", "object":{"kind": "nfmark", "opnd": "gt"}, "value": 24, "relation": "and"}, {"kind": "meta", "object":{"kind": "tcindex", "opnd": "eq"}, "value": 0xf0, "mask": 0xf0}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Now, the same example but with variations # - use operand sign instead of name match = [{"kind": "meta", "object":{"kind": "nfmark", "opnd": ">"}, "value": 24, "relation": "and"}, {"kind": "meta", "object":{"kind": "tcindex", "opnd": "="}, "value": 0xf0, "mask": 0xf0}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Another example given by the tc helper # meta(indev shift 1 eq "ppp") match = [{"kind": "meta", "object":{"kind": "dev", "opnd": "eq", "shift": 1}, "value": "ppp"}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match) # Another example, drop every packets arriving on ifb0 match = [{"kind": "meta", "object":{"kind": "dev", "opnd": "eq"}, "value": "ifb0"}] ip.tc("add-filter", "basic", ifb0, parent=0x10000, classid=0x10010, match=match, action="drop") # As the man page says, here are the different key-value pairs you can use: # "opnd": "eq" or "=" or 0 # "opnd": "gt" or ">" or 1 # "opnd": "lt" or "<" or 2 # "shift": any integer between 0 and 255 included # "kind" object: see `tc filter add dev iface basic match 'meta(list)'` result # "value": any string if kind matches 'dev' or 'sk_bound_if', # any integer otherwise NOTES: When not specified, `inverse` flag is set to False. Do not specify `relation` keyword on the last expression or if there is only one expression. `relation` can be written using multiple format: "and", "AND", "&&", "or", "OR", "||" You can combine multiple different types of ematch. Here is an example:: match = [{"kind": "cmp", "layer": 2, "opnd": "eq", "align": "u32", "offset": 0, "value": 32, "relation": "&&"}, {"kind": "meta", "object":{"kind": "vlan_tag", "opnd": "eq"}, "value": 100, "relation": "||"}, {"kind": "ipset", "index": 0, "mode": "src", "inverse": True} ] ''' import struct from socket import htons from pyroute2 import protocols from pyroute2.netlink import nla from pyroute2.netlink.rtnl.tcmsg.common_act import get_tca_action, tca_act_prio from pyroute2.netlink.rtnl.tcmsg.common_ematch import ( get_tcf_ematches, nla_plus_tcf_ematch_opt, ) def fix_msg(msg, kwarg): msg['info'] = htons( kwarg.get('protocol', protocols.ETH_P_ALL) & 0xFFFF ) | ((kwarg.get('prio', 0) << 16) & 0xFFFF0000) def get_parameters(kwarg): ret = {'attrs': []} attrs_map = (('classid', 'TCA_BASIC_CLASSID'),) if kwarg.get('match'): ret['attrs'].append(['TCA_BASIC_EMATCHES', get_tcf_ematches(kwarg)]) if kwarg.get('action'): ret['attrs'].append(['TCA_BASIC_ACT', get_tca_action(kwarg)]) for k, v in attrs_map: r = kwarg.get(k, None) if r is not None: ret['attrs'].append([v, r]) return ret class options(nla): nla_map = ( ('TCA_BASIC_UNSPEC', 'none'), ('TCA_BASIC_CLASSID', 'uint32'), ('TCA_BASIC_EMATCHES', 'parse_basic_ematch_tree'), ('TCA_BASIC_ACT', 'tca_act_prio'), ('TCA_BASIC_POLICE', 'hex'), ) class parse_basic_ematch_tree(nla): nla_map = ( ('TCA_EMATCH_TREE_UNSPEC', 'none'), ('TCA_EMATCH_TREE_HDR', 'tcf_parse_header'), ('TCA_EMATCH_TREE_LIST', '*tcf_parse_list'), ) class tcf_parse_header(nla): fields = (('nmatches', 'H'), ('progid', 'H')) class tcf_parse_list(nla, nla_plus_tcf_ematch_opt): fields = ( ('matchid', 'H'), ('kind', 'H'), ('flags', 'H'), ('pad', 'H'), ('opt', 's'), ) def decode(self): nla.decode(self) size = 0 for field in self.fields + self.header: if 'opt' in field: # Ignore this field as it a hack used to brain encoder continue size += struct.calcsize(field[1]) start = self.offset + size end = self.offset + self.length data = self.data[start:end] self['opt'] = self.parse_ematch_options(self, data) tca_act_prio = tca_act_prio