74cacf6067
Get rid of OFP_MAXLEN.
...
Statistics or vendor extensions can fill up the whole 64 kB.
2008-09-10 20:05:49 -07:00
d0fb7347b9
Delete OFPST_SWITCH and make it available as a Nicira vendor extension.
2008-09-10 15:00:47 -07:00
39136b0fb9
Make switches send error messages when they receive a bad request.
2008-09-10 13:22:25 -07:00
696db2c3b6
Add vendor-specific messages and statistics to OpenFlow protocol.
2008-09-10 13:21:34 -07:00
a70ca44e6d
Fix uninitialized data read in secchan.
...
Found by valgrind.
2008-09-10 13:21:34 -07:00
176265ed07
Use new method to describe table entries in OpenFlow wire protocol.
...
Older versions of OpenFlow pre-defined a few different types of tables and the
Features Reply message would indicate how many of those entries existed. This
was very inflexible, so now the Features Reply just indicates how many tables
there are. The controller now sends a Table Stats request to find out how
many entries are supported and the fields that can be wildcarded.
This change also removes the "buffer_mb" field from the Features Reply, since
it was not being used and seemed to be of limited value.
2008-09-10 13:03:25 -07:00
b5e9402df5
Implement 802.1D Spanning Tree Protocol.
2008-09-09 14:15:57 -07:00
9bf3d9301e
Zero out otherwise uninitialized padding in make_unbuffered_packet_out().
...
Found with valgrind.
2008-09-09 14:12:57 -07:00
8f56e1fd10
Fix instructions for running ofp-pki: "new-pki" is now "init".
...
Thanks to Dan for pointing out the problem.
2008-09-09 14:12:57 -07:00
71f13ed0b6
Send of0 packets from workqueue, to avoid recursive locking of ofN device.
...
lockdep reported the following warning, showing that dev_queue_xmit()
for of0 and for another ethernet device (eth0?) were taking locks of
the same class. It was not a problem in this particular case, because
of0 != eth0, but if a flow were to send a packet back to of0 then we
would have a deadlock.
Solution: use a workqueue to send packets to avoid recursive locking.
We will still waste a lot of CPU time if we get a packet that loops back
to of0. Solution for that still needed.
=============================================
[ INFO: possible recursive locking detected ]
2.6.26.5 #1
---------------------------------------------
memcheck/1258 is trying to acquire lock:
(_xmit_ETHER){-+..}, at: [<c025b126>] __qdisc_run+0xa0/0x18a
but task is already holding lock:
(_xmit_ETHER){-+..}, at: [<c024f3cd>] dev_queue_xmit+0x23c/0x334
other info that might help us debug this:
9 locks held by memcheck/1258:
#0 : (genl_mutex){--..}, at: [<c0262037>] genl_rcv+0x12/0x2b
#1 : (dp_mutex){--..}, at: [<c88e9a13>] dp_genl_openflow+0x58/0x91 [openflow_mo
d]
#2 : (rcu_read_lock){..--}, at: [<c024e749>] netif_receive_skb+0x9e/0x328
#3 : (rcu_read_lock){..--}, at: [<c0267f77>] ip_local_deliver_finish+0x2a/0x1d7
#4 : (slock-AF_INET/1){-+..}, at: [<c027f1b8>] tcp_v4_rcv+0x29f/0x5ef
#5 : (rcu_read_lock){..--}, at: [<c024f308>] dev_queue_xmit+0x177/0x334
#6 : (_xmit_ETHER){-+..}, at: [<c024f3cd>] dev_queue_xmit+0x23c/0x334
#7 : (rcu_read_lock){..--}, at: [<c88ea214>] dp_dev_xmit+0x0/0x6c [openflow_mod
]
#8 : (rcu_read_lock){..--}, at: [<c024f308>] dev_queue_xmit+0x177/0x334
stack backtrace:
Pid: 1258, comm: memcheck Not tainted 2.6.26.5 #1
[<c02c11cc>] ? printk+0xf/0x13
[<c01368c5>] __lock_acquire+0x8c2/0xbe4
[<c0134d35>] ? add_lock_to_list+0x64/0x8a
[<c0107a78>] ? native_sched_clock+0x82/0x94
[<c0136c3e>] lock_acquire+0x57/0x73
[<c025b126>] ? __qdisc_run+0xa0/0x18a
[<c02c3919>] _spin_lock+0x1c/0x49
[<c025b126>] ? __qdisc_run+0xa0/0x18a
[<c025b126>] __qdisc_run+0xa0/0x18a
[<c024f382>] dev_queue_xmit+0x1f1/0x334
[<c88e9cde>] xmit_skb+0x5b/0x65 [openflow_mod]
[<c88e9e60>] dp_output_port+0x178/0x1ae [openflow_mod]
[<c88eaf91>] do_output+0x2a/0x4c [openflow_mod]
[<c88eb11f>] execute_actions+0x16c/0x198 [openflow_mod]
[<c88eb622>] run_flow_through_tables+0xe9/0xf6 [openflow_mod]
[<c88eb63e>] fwd_port_input+0xf/0x3d [openflow_mod]
[<c88ea260>] dp_dev_xmit+0x4c/0x6c [openflow_mod]
[<c88ea214>] ? dp_dev_xmit+0x0/0x6c [openflow_mod]
[<c024f042>] dev_hard_start_xmit+0x20f/0x276
[<c024f3e2>] dev_queue_xmit+0x251/0x334
[<c026bfcb>] ip_finish_output+0x1ea/0x222
[<c026c081>] ip_output+0x7e/0x83
[<c026b37a>] ip_local_out+0x18/0x1b
[<c026baa2>] ip_queue_xmit+0x288/0x2c9
[<c0136a0c>] ? __lock_acquire+0xa09/0xbe4
[<c0111264>] ? kernel_map_pages+0xfc/0x113
[<c027ddad>] ? tcp_v4_send_check+0x80/0xba
[<c027a281>] tcp_transmit_skb+0x695/0x6cd
[<c016060a>] ? __kmalloc_track_caller+0xee/0x12a
[<c024aacc>] ? __alloc_skb+0x51/0xff
[<c027a43c>] tcp_send_ack+0xdf/0xe7
[<c02781e4>] tcp_rcv_state_process+0x389/0xc33
[<c027eebd>] tcp_v4_do_rcv+0x3bd/0x419
[<c027f2c7>] tcp_v4_rcv+0x3ae/0x5ef
[<c026805f>] ip_local_deliver_finish+0x112/0x1d7
[<c0268185>] ip_local_deliver+0x61/0x6a
[<c0267d14>] ip_rcv_finish+0x2a4/0x2c3
[<c0267f23>] ip_rcv+0x1f0/0x21a
[<c024e990>] netif_receive_skb+0x2e5/0x328
[<c024ea52>] process_backlog+0x7f/0xca
[<c024d867>] net_rx_action+0x72/0x127
[<c0120ea1>] __do_softirq+0x7b/0xf2
[<c0105a5d>] do_softirq+0x66/0xb3
[<c024ec69>] netif_rx_ni+0x29/0x2e
[<c88ea2cd>] dp_dev_recv+0x4d/0x6c [openflow_mod]
[<c88e9e06>] dp_output_port+0x11e/0x1ae [openflow_mod]
[<c88eaf91>] do_output+0x2a/0x4c [openflow_mod]
[<c88eb11f>] execute_actions+0x16c/0x198 [openflow_mod]
[<c88eb312>] recv_flow+0x1c7/0x2a5 [openflow_mod]
[<c88eb14b>] ? recv_flow+0x0/0x2a5 [openflow_mod]
[<c88eab37>] fwd_control_input+0x53/0x60 [openflow_mod]
[<c88e9a27>] dp_genl_openflow+0x6c/0x91 [openflow_mod]
[<c02621ce>] genl_rcv_msg+0x17e/0x198
[<c0262050>] ? genl_rcv_msg+0x0/0x198
[<c02614c6>] netlink_rcv_skb+0x30/0x76
[<c0262043>] genl_rcv+0x1e/0x2b
[<c0261053>] netlink_unicast+0x1a9/0x20f
[<c02612dc>] netlink_sendmsg+0x223/0x230
[<c0245384>] sock_sendmsg+0xca/0xe1
[<c012c501>] ? autoremove_wake_function+0x0/0x33
[<c015f6ce>] ? cache_free_debugcheck+0x2a3/0x2be
[<c0107a78>] ? native_sched_clock+0x82/0x94
[<c0134382>] ? lock_release_holdtime+0x1a/0x115
[<c01f52b7>] ? copy_from_user+0x34/0x11b
[<c024b6a5>] ? verify_iovec+0x40/0x6f
[<c02454da>] sys_sendmsg+0x13f/0x192
[<c0168d51>] ? pipe_write+0x434/0x43f
[<c0107a78>] ? native_sched_clock+0x82/0x94
[<c0107a78>] ? native_sched_clock+0x82/0x94
[<c0134382>] ? lock_release_holdtime+0x1a/0x115
[<c0107a78>] ? native_sched_clock+0x82/0x94
[<c0134382>] ? lock_release_holdtime+0x1a/0x115
[<c0246189>] sys_socketcall+0x14e/0x169
[<c0102d43>] ? restore_nocheck+0x12/0x15
[<c0102ce2>] syscall_call+0x7/0xb
=======================
2008-09-09 12:53:47 -07:00
b5bcab1b47
Add ability to monitor both ends of secchan connections from dpctl.
2008-09-08 10:05:13 -07:00
9e7df18a42
vconn: New function vconn_get_name().
2008-09-08 10:05:13 -07:00
7949a13230
Re-add extensions submodule that was accidentally dropped.
2008-09-08 09:25:25 -07:00
7fe9460749
Compatibility fixes for 2.6.18 kernel
...
Functions added in the 2.6.22 kernel in tcp.h and skbuff.h
have been added to the compatibility headers. Additionally
the data argument was dropped in 2.6.22 for the work queue
macros, requiring them to be redfined for previous kernels.
Based on changes from David Erickson.
2008-09-05 21:46:20 -07:00
e71fcda320
Do not send in-band control traffic to the controller.
...
This commit prevents the secure channel from sending any traffic to
or from the controller up to the controller over its OpenFlow connection.
Instead, such traffic is always handled completely locally. This is
what the changes to in-band control up until now have been converging on:
each time, we have reduced the category of traffic that can be sent up
to the controller because we have found that doing so can cause a control
loop. This change goes all the way.
2008-09-05 17:09:53 -07:00
c907aae9f1
Use make_openflow() function to simplify send_features_request().
2008-09-04 14:24:45 -07:00
c572b9a572
Fix typo in comment.
2008-09-04 14:24:45 -07:00
9b13a5792f
Revert "Do not send in-band control traffic to the controller."
...
This reverts commit 123301f8b5
2008-09-04 13:53:26 -07:00
4870c3cc79
Simplify code by using rconn_send_with_limit() helper function.
2008-09-04 09:48:27 -07:00
2137c4a80e
Initialize otherwise uninitialized variable.
...
Found by valgrind.
2008-09-04 09:47:33 -07:00
b5e16b0723
Respin "Make vconns keep track of their names and include them in log messages."
...
This time, leave out the segfaults.
2008-09-04 09:47:05 -07:00
7d7d973d8b
Add serial number to description status message.
2008-09-03 17:51:42 -07:00
af04b055a7
Revert "Make vconns keep track of their names and include them in log messages."
...
This reverts commit a39a2431c1
2008-09-03 17:43:30 -07:00
bb93ad418c
Also disable atexit hooks in fatal_signal_fork().
...
This fixes a problem in ofp-discover, where it would daemonize itself
after setting an IP address on one of its interfaces. The daemonize
function would call fatal_signal_fork() then exit(0), which would in turn
cause the netdev code to disable the interface and thereby remove the IP
address.
The bug that this fixes was introduced in commit 3cc1ae6a3
2008-09-03 10:10:39 -07:00
123301f8b5
Do not send in-band control traffic to the controller.
...
This commit prevents the secure channel from sending any traffic to
or from the controller up to the controller over its OpenFlow connection.
Instead, such traffic is always handled completely locally. This is
what the changes to in-band control up until now have been converging on:
each time, we have reduced the category of traffic that can be sent up
to the controller because we have found that doing so can cause a control
loop. This change goes all the way.
2008-09-02 15:44:58 -07:00
cfc54b8793
Minimal fix for in-band control loop with hub.
...
When a hub has two switches and the controller on three of its ports,
control traffic from the controller to switch 1 is seen by switch 2 as
data traffic, and switch 2 sends it up over its OpenFlow connection. That
control traffic is in turn seen by the switch 1 as data traffic, and so
switch 1 sends it up over its OpenFlow connection. And so on: this is a
loop.
This commit fixes the situation by not sending up over the OpenFlow
connection traffic to the controller that arrived on the controller's
port, as well as traffic from the controller.
2008-09-02 15:42:34 -07:00
a39a2431c1
Make vconns keep track of their names and include them in log messages.
2008-09-02 14:28:22 -07:00
b67a54201f
Update ext.
2008-09-02 13:44:49 -07:00
13e9c27b20
Make "make dist", "make clean" work without extensions submodule present.
2008-08-29 10:05:14 -07:00
27dadb6bc6
Fix detection of OpenFlow extensions submodule.
2008-08-29 09:56:01 -07:00
7b8b6bac10
Fix "make distcheck" error due to failure to clean files added by Automake.
2008-08-28 17:47:01 -07:00
263aa64c49
Add support for extensions submodule.
2008-08-28 17:36:33 -07:00
80e0e1e029
Remove duplicate prefixes on switch-status items.
2008-08-28 14:25:05 -07:00
686e34c8b7
Make the secchan report a few more switch-status items.
2008-08-28 14:25:05 -07:00
3cc1ae6a32
Add ability to run fatal signal hooks upon normal termination too.
...
This is convenient since in many cases we want these hooks to run in
both cases.
2008-08-28 14:24:39 -07:00
7129a314b2
Add license notice to acinclude.m4, configure.ac.
2008-08-28 14:24:39 -07:00
e9ab3dbf6d
Make -P or --pidfile keep programs from running if already running.
2008-08-28 13:17:16 -07:00
3059085830
Fix test for fcntl return value.
...
POSIX only says that -1 is an error; other return values have unspecified
meanings.
2008-08-28 13:15:32 -07:00
637bae3222
Make OpenFlow kernel module print protocol version.
2008-08-28 09:54:50 -07:00
7518061f7d
Allow multiple -l or --listen options on secchan command line.
...
Bug #149 .
2008-08-27 11:50:11 -07:00
ab43666b20
vlog: Apply rate-limiting everywhere it seems warranted.
2008-08-26 13:25:21 -07:00
1a41889356
vlog: Implement rate-limiting.
2008-08-26 13:25:19 -07:00
c1d17df989
vlog: Simplify vlog() by using dynamic-string functions.
2008-08-26 13:25:17 -07:00
2667dfda01
New functions ds_put_strftime(), ds_chomp().
2008-08-26 13:25:17 -07:00
394270a0e5
New header for saturating arithmetic.
2008-08-26 12:04:41 -07:00
5d3042a11f
Eliminate dpctl dependency on /sbin/ifconfig.
2008-08-26 12:03:14 -07:00
23e60ce176
Allow multiple ports on dpctl addif and delif commands, for convenience.
2008-08-26 12:02:59 -07:00
e3d722e73b
Implement new OpenFlow "switch statistics" feature.
...
This should be useful for reporting the status of a switch on its
front-panel display (for those switches that have one) or in the
controller UI.
2008-08-26 11:53:51 -07:00
b057b10a74
rconn: Add functions for reporting connection statistics.
2008-08-26 11:51:17 -07:00
3e464f2c68
Add functions for reporting dhcp client statistics.
2008-08-26 11:51:17 -07:00
Ben Pfaff
Justin Pettit