Commit graph

406 commits

Author SHA1 Message Date
roopa
a8741fab8e Add hash function
This patch adds a hash function for hashing libnl objects.

This hash function is from:
http://ccodearchive.net/info/hash.html

The original code was modified to remove unwanted dependencies,
unwanted code and fixes to header file locations

One requirement with this hash function is, hashing over multiple fields of an
un-packed struct requires that the struct be zeroed, otherwise random padding
bytes will change the hash.

Signed-off-by: Shrijeet Mukherjee <shm@cumulusnetworks.com>
Signed-off-by: Nolan Leake <nolan@cumulusnetworks.com>
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Reviewed-by: Wilson Kok <wkok@cumulusnetworks.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
2012-11-10 00:12:08 +01:00
Thomas Graf
665464cde5 nl: Improve API doc of nl_send_simple()
Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-11-08 21:17:32 +01:00
Thomas Graf
b28c25eb23 nl: Improve API doc of nl_connect() and nl_close()
Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-11-08 13:41:54 +01:00
Thomas Graf
fd6f205f8a nl: Allow to overwrite nl_send()
Up to now only calls to nl_send_auto() could be overwritten with
nl_cb_overwrite_send(). This patch extends the capability to
nl_send()

Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-11-08 13:19:07 +01:00
Thomas Graf
b132ee7e27 nl: improve API doc of transmit functions
Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-11-08 01:29:12 +01:00
Thomas Graf
8f47501741 nl: improve nl_sendto() docs and error checks
Make nl_sendto() return NLE_INVAL if provided buffer is NULL
and make it return NLE_BAD_SOCK if the socket is not connected.

Add note in docs about lack of NL_CB_MSG_OUT invokation
2012-11-06 17:48:28 +01:00
Thomas Graf
b25f26b896 socket: document nl_socket_get_fd() 2012-11-06 17:36:53 +01:00
Thomas Graf
65f97deb63 nl_recv: fix indent style 2012-11-06 14:27:34 +01:00
Thomas Graf
220d8e141d nl_recv: return NLE_INVAL if socket address pointer is NULL 2012-11-06 14:24:09 +01:00
Thomas Graf
eb36066909 nl_recv: return NLE_INVAL if buf is NULL 2012-11-06 14:22:50 +01:00
Thomas Graf
fb42f19e2e nl_recv: improve function documentation 2012-11-06 14:21:18 +01:00
Thomas Graf
6d52ae6440 nl_recv: Don't update *buf in error path
iov.iov_base is either NULL or invalid
2012-11-06 00:21:35 +01:00
Thomas Graf
ceb8fb9e87 Document buffer size limitations of nl_recv() 2012-11-06 00:19:28 +01:00
roopa
bc7c822f54 Add support for updating objects in the cache
This patch adds support to update a cache object during cache_include instead
of the current approach of deleting the original object and adding a new one.
This operation is conditional on the object implementing the operation. If
the update is not successful, cache_include falls back to the existing cache
inclusion process of deleting and adding the object.

It adds a new object operation called oo_update. oo_update takes two objects
as arguments, first being the existing cache object that needs update, the
second argument being the new object. Currently it is left to the implementor
to use the msg type to decide wether to delete or add the new object attributes
to the old one. But the operation type or msg type can be easily made part of the
object arguments.

The motivation for this change is explained below in the context of including
support for AF_BRIDGE objects into the link cache.

libnl today deletes an object before it includes an identical object.
But for some objects like the AF_BRIDGE objects this does not work well.
link cache uses the ifindex as its key in object searches.
If link cache were to support AF_BRIDGE family objects, todays implementation,
	- will replace the original link object with the bridge port link object
	  for add notifications
	- And a bridge port delete notification from kernel would delete the
	link object from the cache leaving the cache without the link object
	until the kernel sends another notification for that link

The bridge port link notification contains some base link object attributes
plus bridge specific protocol info attributes. In such cases we think an
operation to update the existing object in place in cache might be useful.

This can be made to work for AF_INET6 link objects too.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Reviewed-by: Nolan Leake <nolan@cumulusnetworks.com>
Reviewed-by: Shrijeet Mukherjee <shm@cumulusnetworks.com>
Reviewed-by: Wilson Kok <wkok@cumulusnetworks.com>
2012-11-05 13:58:50 +01:00
roopa
690264a193 Add new object op oo_id_attrs_get
The current oo_id_attrs nl_object op allows a fixed
id attribute list for an cache. But a cache with multiple families
may need to specify different id attributes for different families.

An example for this is the bridge fdb entries in the neigh cache:
neigh entries belonging to the AF_UNSPEC family use
(NEIGH_ATTR_IFINDEX | NEIGH_ATTR_DST | NEIGH_ATTR_FAMILY) as id attributes.
AF_BRIDGE fdb entries which also support the same msg type, will need to use
(NEIGH_ATTR_LLADDR | NEIGH_ATTR_FAMILY) as id attributes.
Today you cannot specify different set of attributes to two families belonging
to the same cache.

This patch adds a new object function oo_id_attrs_get to get the attributes.
An example implementation of oo_id_attrs_get for the neigh cache will
look like:
static uint32_t neigh_id_attrs_get(struct nl_object *obj)
{
        struct rtnl_neigh *neigh = (struct rtnl_neigh *)obj;

        if (neigh->n_family == AF_BRIDGE)
                return (NEIGH_ATTR_LLADDR | NEIGH_ATTR_FAMILY);
        else
                return (NEIGH_ATTR_IFINDEX | NEIGH_ATTR_DST | NEIGH_ATTR_FAMILY);
}

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Reviewed-by: Nolan Leake <nolan@cumulusnetworks.com>
Reviewed-by: Shrijeet Mukherjee <shm@cumulusnetworks.com>
Reviewed-by: Wilson Kok <wkok@cumulusnetworks.com>
2012-11-05 13:31:55 +01:00
Michele Baldessari
3750e2ac31 Set err and free ntbl when leaving neightbl_msg_parser
Make sure we leave neightbl_msg_parser() by setting the proper err variable and
freeing ntbl
2012-10-29 18:16:58 +01:00
Michele Baldessari
16e54c4e6e Add missing va_end() calls
Add missing va_end() calls on all error paths
2012-10-29 18:16:58 +01:00
Michele Baldessari
503270c1c8 Remove unreachable code
This code gets never called anyway
2012-10-29 18:16:58 +01:00
Rich Fought
a8ef3525c8 Add configure option to disable pthreads support 2012-10-26 12:19:26 -07:00
Thomas Graf
609b47961b Merge branch 'master' of https://github.com/rmfought/libnl 2012-10-26 12:48:22 +02:00
Rich Fought
d2fff93cce Source cleanup for upstream 2012-10-19 11:18:52 -07:00
Коренберг Марк (дома)
420e4623fd nl_recv(): work with credentials only if "creds" given and NL_SOCK_PASSCRED set 2012-10-19 23:51:59 +06:00
Коренберг Марк (дома)
da694e6c7b nl_recv(): small code cleanups
1. memset around nla is unnecessary
2. calloc() is unnecessary. malloc() used instead.
2012-10-19 23:49:30 +06:00
Коренберг Марк (дома)
2249eaebd4 nl_recv(): EWOULDBLOCK return value also checked 2012-10-19 23:49:30 +06:00
Коренберг Марк (дома)
69468517d0 nl_recv(): Memory allocation errors are handled properly now
1. all cleanup actions (like free()) now located at the end of function
2. in case of error or EOF, *buf and *creds (if given) set to NULL
   This protect from invalid code at user's side, like:
   char *buf;
   x = nl_recv(..., &buf, ...);
   if (x<=0)
      goto cleanup;
   cleanup:
      free(buf);
3. all intermediate buffers are stored into local variables, and user's
   variables only touches at the end.
2012-10-19 23:48:46 +06:00
Коренберг Марк (ноутбук дома)
9d6b104ec8 nl_recv(): "else if" logick simplified and refined 2012-10-19 22:43:17 +06:00
Коренберг Марк (ноутбук дома)
e1b67fb23f Clang diagnostics
Based on clang diagnostics:

1. lib/nl.c: recvmsgs(): nla filling with zeros commented.
2. lib/route/classid.c: & lib/route/pktloc.c:
   remove zero-filling of struct stat
3. lib/route/qdisc/htb.c: Fix htb_qdisc_msg_fill(): fix zero-filling
4. ematch/container.c: container_parse:
   commented why only 4 bytes are copied
   len marked as unused to eliminate compiler warning
2012-10-19 22:32:59 +06:00
Коренберг Марк (ноутбук дома)
fedb862ea5 ROUTE_DIFF result was not used in some place in route_compare 2012-10-19 22:16:44 +06:00
Коренберг Марк (ноутбук дома)
787f14dbe3 genl/family flags can be damaged during the auto-indentation.
"-" was never used in the names of the flags. "_" was used in all places
of the library. So, I just changed the undescore to the minus.

Automatic indentation can insert spaces on either side of the minus,
so the library will be compiled, but will not be usable (in this part of the code),
as the parser will split words by white space, and the flag "admin - perm"
will never work.
2012-10-19 22:16:44 +06:00
Rich Fought
c3376e7df6 Conntrack Dump ICMP
Looks like a cut and paste oversight ...
2012-10-18 17:43:45 +02:00
Rich Fought
2d707513c6 Updated nfnetlink includes; removed ifdefs; added delete exp program 2012-10-16 12:13:33 -07:00
Rich Fought
07418658a6 define advanced attributes out 2012-10-12 17:44:27 -07:00
Rich Fought
416459570f Bugfixes 2012-10-10 13:08:23 -07:00
Rich Fought
547c8f6d50 reinit port numers on tuple dump 2012-10-09 15:22:16 -07:00
Rich Fought
d3bec59eb9 bugfixes 2012-10-09 15:16:00 -07:00
Thomas Graf
a35287a689 link: Support link grouping
New functions:
  rtnl_link_set_group(link, group)
  rtnl_link_get_group(link)

The group identifier is printed in the brief section as "group N"

Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 21:55:31 +02:00
Thomas Graf
f2e2e7feb5 link: Include IFLA_NUM_TX_QUEUES and IFLA_NUM_RX_QUEUES when building messages
Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 21:24:35 +02:00
Thomas Graf
7f6880ceb9 link: Only print "promisc-mode" if users > 0
Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 19:58:54 +02:00
Thomas Graf
36ed882e00 link: Support IFLA_NUM_TX_QUEUES and IFLA_NUM_RX_QUEUES
New functions:
  rtnl_link_set_num_tx_queues(link, nqueues)
  rtnl_link_get_num_tx_queues(link)
  rtnl_link_set_num_rx_queues(link, nqueues)
  rtnl_link_get_num_rx_queues(link)

Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 19:57:22 +02:00
Thomas Graf
d2876f8657 link: correctly set LINK_ATTR_PROMISCUITY
Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 19:34:23 +02:00
Thomas Graf
6ac07179e5 link: Support IFLA_PROMISCUITY link attribute
* read-only attribute
 * dumped in details sections "promisc-mode (N users)"

Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 18:15:50 +02:00
Thomas Graf
e4192ff97f nl: Provide API to specify the default buffer size when receiving netlink messages
New functions:
  nl_socket_set_msg_buf_size(sk, size)
  nl_socket_get_msg_buf_size(sk)

Default remains getpagesize()

Signed-off-by: Thomas Graf <tgraf@redhat.com>
2012-10-09 16:16:54 +02:00
Rich Fought
f111efd894 Successful compilation of libnl-nf with expectation 2012-10-08 16:49:06 -07:00
Rich Fought
c675bf0486 Checkpoint before compilation attempt 2012-10-08 15:26:55 -07:00
Rich Fought
e8b3356dd2 "checkpoint" 2012-10-05 17:32:20 -07:00
Rich Fought
20035ce021 Checkpoint: compare function 2012-10-05 11:09:45 -07:00
Rich Fought
40457db1f4 Exp checkpoint 2012-10-05 06:55:04 -07:00
Thomas Graf
4a9e92cbbc Merge branch 'asprintf' of https://github.com/socketpair/libnl 2012-08-31 16:25:47 +02:00
Коренберг Марк
49c94c3490 genl_ctrl_probe_by_name: fix checking of genlmsg_put() return value
It's a pointer! not an error code :) nasty bugs
2012-08-30 22:53:54 +06:00
Коренберг Марк
a820222877 asprintf related fixed in yy parser
1. According to man asprintf:
   If memory allocation wasn't possible, or some other error occurs,
   these functions will return -1, and the contents of strp is undefined.
2. Sometimes, errp was not filled at all. In high-level code, free(errp)
   will called, so segmantation fault may appear in case of error in parser
3. The most cases of using asprintf is to report about allocation fail.
   So, probability of allocation of asprintf buffer is very high. And that
   will lead to trash in errp.
4. For simple casses I decide to replace asprintf with strdup
2012-08-30 22:35:25 +06:00