Documentation/connector/connector.txt

   1 /*****************************************/
   2 Kernel Connector.
   3 /*****************************************/
   4
   5 Kernel connector - new netlink based userspace <-> kernel space easy
   6 to use communication module.
   7
   8 The Connector driver makes it easy to connect various agents using a
   9 netlink based network.  One must register a callback and an identifier.
  10 When the driver receives a special netlink message with the appropriate
  11 identifier, the appropriate callback will be called.
  12
  13 From the userspace point of view it's quite straightforward:
  14
  15         socket();
  16         bind();
  17         send();
  18         recv();
  19
  20 But if kernelspace wants to use the full power of such connections, the
  21 driver writer must create special sockets, must know about struct sk_buff
  22 handling, etc...  The Connector driver allows any kernelspace agents to use
  23 netlink based networking for inter-process communication in a significantly
  24 easier way:
  25
  26 int cn_add_callback(struct cb_id *id, char *name, void (*callback) (struct cn_msg *, struct netlink_skb_parms *));
  27 void cn_netlink_send(struct cn_msg *msg, u32 __group, int gfp_mask);
  28
  29 struct cb_id
  30 {
  31         __u32                   idx;
  32         __u32                   val;
  33 };
  34
  35 idx and val are unique identifiers which must be registered in the
  36 connector.h header for in-kernel usage.  void (*callback) (void *) is a
  37 callback function which will be called when a message with above idx.val
  38 is received by the connector core.  The argument for that function must
  39 be dereferenced to struct cn_msg *.
  40
  41 struct cn_msg
  42 {
  43         struct cb_id            id;
  44
  45         __u32                   seq;
  46         __u32                   ack;
  47
  48         __u32                   len;            /* Length of the following data */
  49         __u8                    data[0];
  50 };
  51
  52 /*****************************************/
  53 Connector interfaces.
  54 /*****************************************/
  55
  56 int cn_add_callback(struct cb_id *id, char *name, void (*callback) (struct cn_msg *, struct netlink_skb_parms *));
  57
  58  Registers new callback with connector core.
  59
  60  struct cb_id *id               - unique connector's user identifier.
  61                                   It must be registered in connector.h for legal in-kernel users.
  62  char *name                     - connector's callback symbolic name.
  63  void (*callback) (struct cn..) - connector's callback.
  64                                   cn_msg and the sender's credentials
  65
  66
  67 void cn_del_callback(struct cb_id *id);
  68
  69  Unregisters new callback with connector core.
  70
  71  struct cb_id *id               - unique connector's user identifier.
  72
  73
  74 int cn_netlink_send(struct cn_msg *msg, u32 __groups, int gfp_mask);
  75
  76  Sends message to the specified groups.  It can be safely called from
  77  softirq context, but may silently fail under strong memory pressure.
  78  If there are no listeners for given group -ESRCH can be returned.
  79
  80  struct cn_msg *                - message header(with attached data).
  81  u32 __group                    - destination group.
  82                                   If __group is zero, then appropriate group will
  83                                   be searched through all registered connector users,
  84                                   and message will be delivered to the group which was
  85                                   created for user with the same ID as in msg.
  86                                   If __group is not zero, then message will be delivered
  87                                   to the specified group.
  88  int gfp_mask                   - GFP mask.
  89
  90  Note: When registering new callback user, connector core assigns
  91  netlink group to the user which is equal to its id.idx.
  92
  93 /*****************************************/
  94 Protocol description.
  95 /*****************************************/
  96
  97 The current framework offers a transport layer with fixed headers.  The
  98 recommended protocol which uses such a header is as following:
  99
 100 msg->seq and msg->ack are used to determine message genealogy.  When
 101 someone sends a message, they use a locally unique sequence and random
 102 acknowledge number.  The sequence number may be copied into
 103 nlmsghdr->nlmsg_seq too.
 104
 105 The sequence number is incremented with each message sent.
 106
 107 If you expect a reply to the message, then the sequence number in the
 108 received message MUST be the same as in the original message, and the
 109 acknowledge number MUST be the same + 1.
 110
 111 If we receive a message and its sequence number is not equal to one we
 112 are expecting, then it is a new message.  If we receive a message and
 113 its sequence number is the same as one we are expecting, but its
 114 acknowledge is not equal to the acknowledge number in the original
 115 message + 1, then it is a new message.
 116
 117 Obviously, the protocol header contains the above id.
 118
 119 The connector allows event notification in the following form: kernel
 120 driver or userspace process can ask connector to notify it when
 121 selected ids will be turned on or off (registered or unregistered its
 122 callback).  It is done by sending a special command to the connector
 123 driver (it also registers itself with id={-1, -1}).
 124
 125 As example of this usage can be found in the cn_test.c module which
 126 uses the connector to request notification and to send messages.
 127
 128 /*****************************************/
 129 Reliability.
 130 /*****************************************/
 131
 132 Netlink itself is not a reliable protocol.  That means that messages can
 133 be lost due to memory pressure or process' receiving queue overflowed,
 134 so caller is warned that it must be prepared.  That is why the struct
 135 cn_msg [main connector's message header] contains u32 seq and u32 ack
 136 fields.
 137
 138 /*****************************************/
 139 Userspace usage.
 140 /*****************************************/
 141
 142 2.6.14 has a new netlink socket implementation, which by default does not
 143 allow people to send data to netlink groups other than 1.
 144 So, if you wish to use a netlink socket (for example using connector)
 145 with a different group number, the userspace application must subscribe to
 146 that group first.  It can be achieved by the following pseudocode:
 147
 148 s = socket(PF_NETLINK, SOCK_DGRAM, NETLINK_CONNECTOR);
 149
 150 l_local.nl_family = AF_NETLINK;
 151 l_local.nl_groups = 12345;
 152 l_local.nl_pid = 0;
 153
 154 if (bind(s, (struct sockaddr *)&l_local, sizeof(struct sockaddr_nl)) == -1) {
 155         perror("bind");
 156         close(s);
 157         return -1;
 158 }
 159
 160 {
 161         int on = l_local.nl_groups;
 162         setsockopt(s, 270, 1, &on, sizeof(on));
 163 }
 164
 165 Where 270 above is SOL_NETLINK, and 1 is a NETLINK_ADD_MEMBERSHIP socket
 166 option.  To drop a multicast subscription, one should call the above socket
 167 option with the NETLINK_DROP_MEMBERSHIP parameter which is defined as 0.
 168
 169 2.6.14 netlink code only allows to select a group which is less or equal to
 170 the maximum group number, which is used at netlink_kernel_create() time.
 171 In case of connector it is CN_NETLINK_USERS + 0xf, so if you want to use
 172 group number 12345, you must increment CN_NETLINK_USERS to that number.
 173 Additional 0xf numbers are allocated to be used by non-in-kernel users.
 174
 175 Due to this limitation, group 0xffffffff does not work now, so one can
 176 not use add/remove connector's group notifications, but as far as I know,
 177 only cn_test.c test module used it.
 178
 179 Some work in netlink area is still being done, so things can be changed in
 180 2.6.15 timeframe, if it will happen, documentation will be updated for that
 181 kernel.