linux.git/net/netlink, branch v3.10-rc2

netlink: kconfig: move mmap i/o into netlink kconfig

2013-05-01T19:02:42+00:00

Currently, in menuconfig, Netlink's new mmaped IO is the very first
entry under the ``Networking support'' item and comes even before
``Networking options'':

  [ ]   Netlink: mmaped IO
  Networking options  --->
  ...

Lets move this into ``Networking options'' under netlink's Kconfig,
since this might be more appropriate. Introduced by commit ccdfcc398
(``netlink: mmaped netlink: ring setup'').

Signed-off-by: Daniel Borkmann 
Signed-off-by: David S. Miller

netlink: Fix skb ref counting.

2013-05-01T18:57:03+00:00

Commit f9c2288837ba072b21dba955f04a4c97eaa77b1e (netlink:
implement memory mapped recvmsg) increamented skb->users
ref count twice for a dump op which does not look right.

Following patch fixes that.

CC: Patrick McHardy 
Signed-off-by: Pravin B Shelar 
Signed-off-by: David S. Miller

genetlink: fix possible memory leak in genl_family_rcv_msg()

2013-04-27T03:25:39+00:00

'attrbuf' is malloced in genl_family_rcv_msg() when family->maxattr &&
family->parallel_ops, thus should be freed before leaving from the error
handling cases, otherwise it will cause memory leak.

Introduced by commit def3117493eafd9dfa1f809d861e0031b2cc8a07
(genl: Allow concurrent genl callbacks.)

Signed-off-by: Wei Yongjun 
Signed-off-by: David S. Miller

genl: Allow concurrent genl callbacks.

2013-04-25T05:43:15+00:00

All genl callbacks are serialized by genl-mutex. This can become
bottleneck in multi threaded case.
Following patch adds an parameter to genl_family so that a
particular family can get concurrent netlink callback without
genl_lock held.
New rw-sem is used to protect genl callback from genl family unregister.
in case of parallel_ops genl-family read-lock is taken for callbacks and
write lock is taken for register or unregistration for any family.
In case of locked genl family semaphore and gel-mutex is locked for
any openration.

Signed-off-by: Pravin B Shelar 
Signed-off-by: David S. Miller

netlink: fix compilation after memory mapped patches

2013-04-24T18:26:55+00:00

Depending of the kernel configuration (CONFIG_UIDGID_STRICT_TYPE_CHECKS), we can
get the following errors:

net/netlink/af_netlink.c: In function ‘netlink_queue_mmaped_skb’:
net/netlink/af_netlink.c:663:14: error: incompatible types when assigning to type ‘__u32’ from type ‘kuid_t’
net/netlink/af_netlink.c:664:14: error: incompatible types when assigning to type ‘__u32’ from type ‘kgid_t’
net/netlink/af_netlink.c: In function ‘netlink_ring_set_copied’:
net/netlink/af_netlink.c:693:14: error: incompatible types when assigning to type ‘__u32’ from type ‘kuid_t’
net/netlink/af_netlink.c:694:14: error: incompatible types when assigning to type ‘__u32’ from type ‘kgid_t’

We must use the helpers to get the uid and gid, and also take care of user_ns.

Fix suggested by Eric W. Biederman .

Signed-off-by: Nicolas Dichtel 
Signed-off-by: David S. Miller

netlink: Fix build with mmap disabled.

2013-04-23T19:39:03+00:00

net/netlink/diag.c: In function 'sk_diag_put_rings_cfg':
net/netlink/diag.c:28:17: error: 'struct netlink_sock' has no member named 'pg_vec_lock'
net/netlink/diag.c:29:29: error: 'struct netlink_sock' has no member named 'rx_ring'
net/netlink/diag.c:31:30: error: 'struct netlink_sock' has no member named 'tx_ring'
net/netlink/diag.c:33:19: error: 'struct netlink_sock' has no member named 'pg_vec_lock'

Reported-by: Randy Dunlap 
Signed-off-by: David S. Miller

netlink: fix typo in net/netlink/af_netlink.c

2013-04-23T17:09:33+00:00

Signed-off-by: Stephen Rothwell 
Signed-off-by: David S. Miller

netlink: add RX/TX-ring support to netlink diag

2013-04-19T18:57:58+00:00

Based on AF_PACKET.

Signed-off-by: Patrick McHardy 
Signed-off-by: David S. Miller

netlink: add flow control for memory mapped I/O

2013-04-19T18:57:58+00:00

Add flow control for memory mapped RX. Since user-space usually doesn't
invoke recvmsg() when using memory mapped I/O, flow control is performed
in netlink_poll(). Dumps are allowed to continue if at least half of the
ring frames are unused.

Signed-off-by: Patrick McHardy 
Signed-off-by: David S. Miller

netlink: implement memory mapped recvmsg()

2013-04-19T18:57:58+00:00

Add support for mmap'ed recvmsg(). To allow the kernel to construct messages
into the mapped area, a dataless skb is allocated and the data pointer is
set to point into the ring frame. This means frames will be delivered to
userspace in order of allocation instead of order of transmission. This
usually doesn't matter since the order is either not determinable by
userspace or message creation/transmission is serialized. The only case
where this can have a visible difference is nfnetlink_queue. Userspace
can't assume mmap'ed messages have ordered IDs anymore and needs to check
this if using batched verdicts.

For non-mapped sockets, nothing changes.

Signed-off-by: Patrick McHardy 
Signed-off-by: David S. Miller