Makefiles: Add $(AM_V_GEN) annotations to clean up "make" output.

The Open vSwitch "make" output was still pretty verbose even when
configured with --enable-silent-rules.  This cleans it up.

Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Joe Stringer <joestringer@nicira.com>

FAQ: Replace reference to OVS 2.2 with 2.3.

We never released 2.2.0, so references to it are confusing.

Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Gurucharan Shetty <gshetty@nicira.com>

bridge: Fix bug where IDL wakeup causes 100% CPU.

Commit 9c537baf613a16e (bridge: Refactor the stats and status update.)
inadvertently changed the way that the status transaction is destroyed,
which could cause the main thread to constantly wake up.

The bug occurs when ovsdb_idl_txn_commit() returns TXN_INCOMPLETE and
there are no further changes to connectivity or 'status_txn_try_again'.

- ovsdb_idl_run() receives the transaction reply and updates the
  transaction status to TXN_SUCCESS.
- status_update_wait() detects that the transaction is in progress, and
  the status is TXN_SUCCESS so wakes up the main thread immediately.
- run_status_update() is meant to destroy the transaction now that it is
  finished, however the logic is never run because there were no changes.
- Repeat the wakeup every time the main loop runs.

This patch fixes the behaviour by ensuring that ovsdb_idl_txn_commit()
gets a chance to run whenever there is an ongoing status transaction.

Bug was found by unloading and reloading the kernel module, then
immediately restarting ovs-vswitchd.

Signed-off-by: Joe Stringer <joestringer@nicira.com>
Acked-by: Alex Wang <alexw@nicira.com>

datapath-windows: Build fix.

Recently we changed the signature of NlAttrParse.
Function OvsGetVport was not updated accordingly.

Fixed the same.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>

AUTHORS: Add Sorin Vinturis.

Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Missed packets are not queued.

Currently, whenever there's a missed packet, the ovs driver allocates
memory and copies the packet even if there's no packet queue setup from
userspace. Then if there's no queue created, the packet is released and
dropped.

The solution was to check for the existence of the userspace queue before
trying to allocate and add a new missed packet to the queue. If there is
no userspace queue created, the original packet is dropped without creating
a new missed packet.

Signed-off-by: Sorin Vinturis <svinturis@cloudbasesolutions.com>
Reported-by: Nithin Raju <nithin@vmware.com>
Reported-at: https://github.com/openvswitch/ovs-issues/issues/32
Tested-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

flow.c: Correct the comment for miniflow_equal().

It's for miniflow equalness check, not flow.

Signed-off-by: Wang Sheng-Hui <shhuiw@gmail.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

AUTHORS: Add Samuel Ghinet.

Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Rename OvsFlowNlNewCmdHandler to OvsFlowNlCmdHandler

OvsFlowNlNewCmdHandler NEW/DEL/SET/FLUSH commands hence renaming it to
OvsFlowNlCmdHandler.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Flow.c: DEL_FLOWS command handler.

Added changes to handle DEL_FLOWS (FLUSH) scenario.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Flow.c: FLOW_DEL command handler.

Registered FLOW_DEL command handler. The same command
handler as FLOW_ADD is good enough to handle FLOW_DEL
case as well with minor changes for checking to action
attribute.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Flow.c: FLOW_SET command handler.

Registered FLOW_SET command handler. The same command
handler as FLOW_ADD is good enough to handle FLOW_SET
case as well.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Flow.c: FLOW_NEW command handler.

This patch covers the changes needed to support FLOW_NEW command.
API _OvsFlowMapNlToFlowPutFlags has a bug, which will be fixed
with the patches for FLOW_DEL.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Flow.c : Basic support for add-flow.

This patch covers basic changes in registering add flow handler.
And declaring FLOW related attribute parsing policies.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Netlink: Fixed NlAttrParseNested

NlAttrParseNested was using the whole netlink payload for iteration.
This is not correct, as it would lead to exceeding the
nested attribute boundries. Fixed the same in this patch.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Netlink: Allow support for NESTED Attributes in NlAttrValidate

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Netlink: Add NlAttrLen API

Added an API to retrieve the attribute length.
Added 2 more API for BE16 and BE8 attribute parsing.
Fixed a trailing whitespace issue.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Netlink: Add NlFillOvsMsg API for creating Netlink message headers.

Added NlFillOvsMsg API in Netlink.c This API will be used to populate
netlink message headers.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: move OVS_MESSAGE to Netlink.h

Moved the structure OVS_MESSAGE to Netlink.h.
This change is done for following reasons.

a. Patch 2 in this series provides a generic API in Netlink.c
for creating netlink message. That API needs OVS_MESSAGE.
Including Datapath.h in Netlink.c/h gives compilation error.

b. OVS_MESSAGE defines netlink messages hence moving it to
Netlink.h looked fine to me.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Tested-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Fix broken build by adding new file to distribution.

Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Add WMI Script that updates Hyper-V friendly port names.

The following script leverage's the advantages of WMI infrastructure
offered in Hyper-V.

This scripts allows the user to change the
Msvm_EthernetPortAllocationSettingData property of a VM network adapter
connected to a Hyper-V Virtual Switch.

Usage:
import-module .\OVS.psm1
$vnic = Get-VMNetworkAdapter VM1
Connect-VMNetworkAdapter -VMNetworkAdapter $vnic -SwitchName external
$vnic | Set-VMNetworkAdapterOVSPort -OVSPortName ovs-port-1

VM1 - is a VM on top of a Hyper-V
external - is a Hyper-V Virtual Switch

Signed-off-by: Alessandro Pilotti <apilotti@cloudbasesolutions.com>
Signed-off-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Tested-by: Nithin Raju <nithin@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Implement event read handler.

The Read event handler is executed when user mode issues a socket
receive on an MC socket associated with the event queue. A new IOCTL
READ command is used to differentiate between transaction based and packet
miss sockets.
An entry for the handler will be added once the Control family (logically
should have been added to the Vport family)
implementation checked in.
User mode code for setting the socket type will follow

Signed-off-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: NUL character should be left out during VPORT hash lookup

While calculating the hash on a VPORT name, we don't include the NUL character.
We should be doing the same while doing lookup as well.

We set the required minimum length of the name to be 2 so that the string has
at least one valid character.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

ovs-ofctl: Correct help text for add-groups

It is add-groups rather than add-group that takes FILE as an argument.

Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

AUTHORS: Add Simon Horman <simon.horman@netronome.com>

Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Add Netlink vport command get.

The transactional get vport command.
This command uses the netlink transactional errors.

Signed-off-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Add file NetlinkError.h.

Contains error codes for netlink transactional errors.
These errors are passed to the "error" field (INT) of the NL_MSG_ERR struct.
The userspace requires them to be negative values: the nl_msg_nlmsgerr userspace
function transforms them from negative to positive values.

These error codes correspond to the userspace error codes defined in:
"C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\include\errno.h"

Signed-off-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Implement vport dump Netlink command.

Functionality for vport dump.
Later, when we will add more netlink dump commands, some common code will need
to be split to functions.

Notes:
a) the current implementation of vport assumes the datapath feature
"multiple upcall pids" is not used. A single upcall pid is used now.
c) the vxlan destination udp port is currently a constant. When it will become
configurable, the vport options netlink attribute will become relevant.

Signed-off-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Acked-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: fix OVS_VPORT_TYPE

The windows ovs kernel uses an OVS_VPORT_TYPE enum that is incompatible with
the userspace counterpart (enum ovs_vport_type from openvswitch.h). We must use
the same enum type for the netlink communication to work properly.

This patch makes the fix: "typedef enum ovs_vport_type OVS_VPORT_TYPE" and
changes the afferent kernel driver code:
o) vport types synthetic and emulated turn to: netdev
o) vport type internal turns to: internal
o) vport type external truns to: netdev (plus, we hold a field in vport,
"isExternal"

Signed-off-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

ovs-vtep: Use shlex module to split args.

string.split() function splits a quoted string if there is a whitespace
inside the quote.
ex: The following code snippet will output ['printing', '"No', 'Diagnostic"']
args = 'printing "No Diagnostic"'
print args.split()

The above is a problem if we run the following command through vtep_ctl().
vtep-ctl set tunnel $uuid bfd_status:diagnostic="No Diagnostic"

The workaround is to use the split() function from shlex module.

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ariel Tubaltsev <atubaltsev@vmware.com>
Acked-by: Justin Pettit <jpettit@nicira.com>

ovs-vtep: Store physical switch name globally.

ovs-vtep is an emulator and it works only on one
physical switch. This switch name is stored in the variable
'ps_name' and then passed around. An upcoming commit requires
access to this variable at more places and it is easier if this
variable is global.

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ariel Tubaltsev <atubaltsev@vmware.com>
Acked-by: Justin Pettit <jpettit@nicira.com>

ovs-vtep: Clear left-over local mac information.

Before destroying a logical switch, cleanup any left over local
mac information in Ucast_Macs_Local or Mcast_Macs_Local table.
We need to do this to atleast cleanup the 'unknown-dst' information
added in the Mcast_Macs_Local table while creating the Logical_Switch
class in setup_ls().

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ariel Tubaltsev <atubaltsev@vmware.com>
Acked-by: Justin Pettit <jpettit@nicira.com>

vtep-ctl: Add Tunnel table to vtep_ctl_table_class.

This is needed to create, get, set records in the Tunnel table.

(We need to add the Tunnel table's 'local' and 'remote' columns
that point to the Physical_Locator record to cache because vtep-ctl
commands like 'add-ucast-local' will try to add an entry in
Physical_Locator table based on the contents of the cache.)

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ariel Tubaltsev <atubaltsev@vmware.com>
Acked-by: Justin Pettit <jpettit@nicira.com>

README.ovs-vtep: Remotes can be connected through VTEP's manager table.

Reported-by: Ziyou Wang <ziyouw@vmware.com>
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Justin Pettit <jpettit@nicira.com>

lib/ofproto: Sync RSTP operational state after configuration changes.

Otherwise the RSTP port operational state could be out of sync until
the next time the port's carrier status changes.

Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Daniele Venturino <daniele.venturino@m3s.it>

netdev-dpdk: Fix crash when there is no pci numa info.

When kernel cannot obtain the pci numa info, the numa_node file
in corresponding pci directory in sysfs will show -1.  Then the
rte_eth_dev_socket_id() function will return it to ovs.  On
current master, ovs assumes rte_eth_dev_socket_id() always
returns non-negative value.  So using this -1 in pmd thread
creation will cause ovs crash.

To fix the above issue, this commit makes ovs always check the
return value of rte_eth_dev_socket_id() and use numa node 0 if
the return value is negative.

Reported-by: Daniel Badea <daniel.badea@windriver.com>
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

netdev: Fix error check.

Reported-by: Daniel Badea <daniel.badea@windriver.com>
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

ofproto-dpif-rid: remove unused return value of rid_pool_add()

The return value of rid_pool_add() is never used
so the code may be slightly simplified by removing it.

Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>

ofproto-dpif-rid: remove struct rid_map

struct rid_map only has one member which is a struct hmap.
This allows for a slight simplification of the code by removing
struct rid_map and using a struct hmap directly instead.

Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>

rstp.at: Fix intermittent test failure.

Sub-test "RSTP - dummy interface" checks the ovs-vswitchd.log
output immediately after command execution.  The check may
fail if the write of new log is delayed by the IO thread.

This commit fixes the above issue by waiting for the
ovs-vswitchd.log output.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Gurucharan Shetty <gshetty@nicira.com>

ofproto-dpif-rid: correct logic error in rid_pool_alloc_id()

When searching through the valid ids an id should
be used if is not found rather than if it is found.

It appears to me that without this change duplicate recirculation
ids may used in cases where the last recirculation id has
been allocated; selection loops back to the beginning of the pool and;
reaches a recirculation id that is still in use.

As the number of recirculation ids is currently RECIRC_ID_N_IDS = 1024 this
does not seem beyond the bounds of possibility.

I have not verified that such a scenario can actually occur.  But it seems
that a likely consequence would be that some packets may be forwarded
incorrectly.

Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>

ovs-atomic-msvc: Fix 64 bit atomic read/writes.

MSVC converts 64 bit read/writes into two instructions (uses 'mov' as
seen through cl //FAs). So there is a possibility that an interrupt can
make a 64 bit read/write non-atomic even when 8 byte aligned. So we cannot
use a simple assignment. Use a full memory barrier function instead.

Reported-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>

AUTHORS: Add Selvamuthukumar <smkumar@merunetworks.com>.

Signed-off-by: Ben Pfaff <blp@nicira.com>

ofp-actions: Fix error code for invalid table id.

Send OFPET_BAD_INSTRUCTION/OFPBIC_BAD_TABLE_ID if table is invalid
in goto table instruction.

Signed-off-by: Selvamuthukumar <smkumar@merunetworks.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

jsonrpc: Notify excessive sending backlog.

This commit adds a log message to notify the excessive backlog
for jsonrpc.  Expectedly, this message should never be printed.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

bridge: Refactor the stats and status update.

This commit refactors the stats and status update in bridge_run()
by moving the corresponding code to separate functions.  This
makes the code more organized.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

bridge: Rate limit the statistics update.

When ovs is running with large topology (e.g. large number of
interfaces), the stats update to ovsdb becomes huge and normally
requires multiple run of ovsdb jsonrpc message processing loop to
consume.

To prevent the periodic stats update from backlogging in the
jsonrpc sending queue, this commit adds rate limiting logic
which only allows new update if the previous one is done.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Acked-by: Flavio Leitner <fbl@redhat.com>

datapath: Constify various function arguments

Help produce better optimized code.

Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>

datapath: Restore OVS_CB after skb_segment.

OVS needs to segments large skb before sending it for miss
packet handling to userspace. but skb_gso_segment uses
skb->cb. This corrupted OVS_CB which result in following panic.

[  735.419921] BUG: unable to handle kernel paging request at 00000014000001b2
[  735.423168] Oops: 0000 [#1] SMP DEBUG_PAGEALLOC
[  735.445097] RIP: 0010:[<ffffffffa05df0d7>]  [<ffffffffa05df0d7>] ovs_nla_put_flow+0x37/0x7c0 [openvswitch]
[  735.468858] Call Trace:
[  735.470384]  [<ffffffffa05d7ec2>] queue_userspace_packet+0x332/0x4d0 [openvswitch]
[  735.471741]  [<ffffffffa05d8155>] queue_gso_packets+0xf5/0x180 [openvswitch]
[  735.481862]  [<ffffffffa05da9f5>] ovs_dp_upcall+0x65/0x70 [openvswitch]
[  735.483031]  [<ffffffffa05dab81>] ovs_dp_process_packet+0x181/0x1b0 [openvswitch]
[  735.484391]  [<ffffffffa05e2f55>] ovs_vport_receive+0x65/0x90 [openvswitch]
[  735.492638]  [<ffffffffa05e5738>] internal_dev_xmit+0x68/0x110 [openvswitch]
[  735.495334]  [<ffffffff81588eb6>] dev_hard_start_xmit+0x2e6/0x8b0
[  735.496503]  [<ffffffff81589847>] __dev_queue_xmit+0x3c7/0x920
[  735.499827]  [<ffffffff81589db0>] dev_queue_xmit+0x10/0x20
[  735.500798]  [<ffffffff815d3b60>] ip_finish_output+0x6a0/0x950
[  735.502818]  [<ffffffff815d55f8>] ip_output+0x68/0x110
[  735.503835]  [<ffffffff815d4979>] ip_local_out+0x29/0x90
[  735.504801]  [<ffffffff815d4e46>] ip_queue_xmit+0x1d6/0x640
[  735.507015]  [<ffffffff815ee0d7>] tcp_transmit_skb+0x477/0xac0
[  735.508260]  [<ffffffff815ee856>] tcp_write_xmit+0x136/0xba0
[  735.510829]  [<ffffffff815ef56e>] __tcp_push_pending_frames+0x2e/0xc0
[  735.512296]  [<ffffffff815e0593>] tcp_sendmsg+0xa63/0xd50
[  735.513526]  [<ffffffff81612c2c>] inet_sendmsg+0x10c/0x220
[  735.516025]  [<ffffffff81566b8c>] sock_sendmsg+0x9c/0xe0
[  735.518066]  [<ffffffff81566d41>] SYSC_sendto+0x121/0x1c0
[  735.521398]  [<ffffffff8156801e>] SyS_sendto+0xe/0x10
[  735.522473]  [<ffffffff816df5e9>] system_call_fastpath+0x16/0x1b

Reported-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Andy Zhou <azhou@nicira.com>

datapath: Fix double free when ovs_nla_copy_actions() fails

ovs_nla_copy_actions() already frees the allocated actions buffers,
ovs_flow_cmd_new() will free it a second time when jumping to
err_kfree_acts.

Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

ovs-pki: Use SHA-1 instead of SHA-512 as message digest.

Commit 9ff33ca7 (ovs-pki: Use SHA-512 instead of MD5 as message
digest.) changes the message digest algorithm to SHA-512.  This
seems to break the unit tests on some xenserver 5.6/6.0 builds
causing the error: "SSL_connect: error:0D0C50A1:asn1 encoding
routines:ASN1_item_verify:unknown message digest algorithm".

As a solution, this commit changes the message digest algorithm
to SHA-1 which works for both the above xenserver builds and
centos 7.

VMware-BZ: #1319116

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

datapath: compat: Fix compilation for 2.6.32 kernel

Define alloc_netdev() using alloc_netdev_mq() which is available on all
kernel supported by OVS.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>

datapath: Remove pkt_key from OVS_CB.

OVS keeps pointer to packet key in skb->cb, but the packet key is
store on stack. This could make code bit tricky. So it is better to
get rid of the pointer.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Andy Zhou <azhou@nicira.com>

datapath: restore OVS_FLOW_CMD_NEW notifications

Since commit fb5d1e9e127a ("openvswitch: Build flow cmd netlink reply only if needed."),
the new flows are not notified to the listeners of OVS_FLOW_MCGROUP.

This commit fixes the problem by using the genl function, ie
genl_has_listerners() instead of netlink_has_listeners().

Signed-off-by: Samuel Gauthier <samuel.gauthier@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

datapath: Provide compatibility for kernels up to 3.17

Port datapath to work with kernrels up to 3.17 and use 3.16.2 as
the new kernel for CI testing.

Tested with 3.14, 3.16.2, and net-next (3.17).

Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Co-authored-by: Madhu Challa <challa@noironetworks.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

datapath: fix sparse warning.

datapath/linux/datapath.c:1418:28: warning: symbol
'i' shadows an earlier one
datapath/linux/datapath.c:1396:18: originally declared here

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>

datapath: change the data type of error status to atomic_long_t

Change the date type of error status from u64 to atomic_long_t, and use atomic
operation, then remove the lock which is used to protect the error status.

The operation of atomic maybe faster than spin lock.

Cc: Pravin Shelar <pshelar@nicira.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

datapath: Remove support to set vport stats.

This was required for old compatibility code which update stats
on fake bond interface. Now vswitchd has dropped it. This
support was always deprecated, so finally removing it.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Andy Zhou <azhou@nicira.com>

dpif-netdev: Fix (packet) memory leaks in the slow path.

If a packet didn't match a rule in the fast path classifier its memory was
never freed. The issue was particularly clear with DPDK devices because it was
not possible to process more than ~250000 DPDK mbufs in the slow path.

This commit fixes the problem by:
* calling dpif_packet_delete() if the upcalls are disabled
* passing may_steal==true to dp_netdev_execute_actions() during normal upcall
  processing

Signed-off-by: Daniele Di Proietto <ddiproietto@vmware.com>
Acked-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

ovs-pki: Use SHA-512 instead of MD5 as message digest.

This fixes numerous testsuite failures of the form "SSL_connect:
error:0D0C50A1:asn1 encoding routines:ASN1_item_verify:unknown message
digest algorithm" on systems that disable MD5 in OpenSSL.  Centos 7 is one
example.  Presumably it increase security as well for anyone who generates
certificates based on a new configuration created by the new ovs-pki.

Reported-by: Robert Strickler <anomalyst@gmail.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

dpif-netdev: Allow multi-rx-queue, multi-pmd-thread configuration.

This commits adds the multithreading functionality to OVS dpdk
module.  Users are able to create multiple pmd threads and set
their cpu affinity via specifying the cpu mask string similar
to the EAL '-c COREMASK' option.

Also, the number of rx queues for each dpdk interface is made
configurable to help distribution of rx packets among multiple
pmd threads.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

ovs-numa: Add support for cpu-mask configuration.

This commit adds support in ovs-numa module for reading a user
specified cpu mask, which configures the availability of the cores.

The cpu mask has the format of a hex string similar to the EAL '-c
COREMASK' option input or the 'taskset' mask input.  The lowest order
bit corresponds to the first CPU core.  Bit value '1' means the
corresponding core is available.

An upcoming patch will allow user to configure the mask via OVSDB.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

dpif-netlink: rename linux_flow variable to datapath_flow

In the flow related functions, there's a stack variable called
'linux_flow'. Since this code is not specific to Linux anymore,
in this patch, we rename the variable to 'datpath_flow'.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: add OVS_DP_CMD_SET and OVS_DP_CMD_GET transaction support

In this patch, we add support for two commands, both of them are issued
as part of transactions semantics from userspace:
1. OVS_DP_CMD_SET is used to get the properties of a DP as well as set
some properties. The set operations does not seem to make much sense for
the Windows datpath right now.
2. There's already support for OVS_DP_CMD_GET command issued via the
dump semantics from userspace. Turns out that userspace can issue
OVS_DP_CMD_GET as a transaction.

There's lot of common code between these two commands. Hence combining
the implementation and the review.

Also refactories some of the code in the implementation of dump-based
OVS_DP_CMD_GET, and updated some of the comments.

Validation:
- With these series of patches, I was able to run the following command:
> .\utilities\ovs-dpctl.exe show
system@ovs-system:
    lookups: hit:0 missed:22 lost:0
    flows: 0
- I got so far as to hit the PORT_DUMP command which is currently not
implemented.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Tested-by: Nithin Raju <nithin@vmware.com>
Reported-at: https://github.com/openvswitch/ovs-issues/issues/38
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

extract-odp-netlink-windows-dp-h: add definition of IFNAMSIZ

The Windows kernel datapath needs the definition of 'IFNAMSIZ' for
specifying attribute sizes in netlink policies. Adding the definition
of 'IFNAMSIZ' to be part of OvsDpInterface.h similar to ETH_ADDR_LEN.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

lib/netlink-socket.c: add support for nl_transact() on Windows

In this patch, we add support for nl_transact() on Windows using
the OVS_IOCTL_TRANSACT ioctl that sends down the request and gets
the reply in the same call to the kernel.

This is obviously a digression from the way it is implemented in
Linux where all the sends are done at once using sendmsg() and
replies are received one at a time.

Initial implementation was in the Linux way using multiple writes
followed by reads, but decided against it since it is not efficient
and also it complicates the state machine in the kernel.

The Windows implementation has equivalent code for handling corner
cases and error coditions similar to Linux. Some of it is not
applicable yet. Eg. the Windows kernel does not embed an error
in the netlink message itself. There's userspace code nevertheless
for this.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: add OvsCompareString() to compare strings

In this patch we implement a utility function to compare ANSI
strings using the Rtl* functions. As much as possible, in an
NDIS driver, we stick to Rtl* functions for memory/string
manipulation.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: return TRUE on success in NlAttrValidate

Signed-off-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

travis: Allow testsuite to run with GCC or Clang.

I don't see why the testsuite is supported only with GCC.

Acked-by: Thomas Graf <tgraf@noironetworks.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

travis: Include testsuite.log on failure.

Acked-by: Thomas Graf <tgraf@noironetworks.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

netdev-dpdk: Fix a bug in netdev_dpdk_set_multiq().

Commit 5a0340 (dpif-netdev: Create multiple tx/rx queues when
adding dpdk interface.) introduced a bug which causes the function
netdev_dpdk_set_multiq() never resetting the tx queues.  This bug
could cause pmd thread accessing unassigned memory, resulting in
segfault.

This commit fixes the bug.

Reported-by: Ethan Jackson <ethan@nicira.com>
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

netdev-dpdk: Fix a typo.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

coverage: Remove unused macro.

Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Alex Wang <alexw@nicira.com>

FAQ: Fix a newline problem

Fix a newline problem in commit dd63a57e55daddccbbd4f0bedfdc86b6827b6b1a
("FAQ: Add an entry about reconfiguration")

Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>

FAQ: Add an entry about reconfiguration

It seems that the behaviour is not so intuitive.
cf. https://bugs.launchpad.net/neutron/+bug/1346861

Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Acked-by: Ben Pfaff <blp@nicira.com>

netdev-dpdk: Pass queue id to dpdk_do_tx_copy().

Since dpdk_do_tx_copy() will be called by both pmd and
non-pmd thread, it should take the queue id as input.
The current ovs always uses NON_PMD_THREAD_TX_QUEUE
as queue id, which causes unprotected multi-access
to the same queue.

This commit fixes the issue by passing the queue id
to dpdk_do_tx_copy().

Reported-by: Ethan Jackson <ethan@nicira.com>
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

datapath-windows: NetLink kernel side, Event subscription and notification

This code handles an event notification subscription for a user mode thread
which joins an MC group. The event wait handler queues an IRP which is
completed upon change in a port state.

Signed-off-by: Eitan Eliahu <eliahue@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Acked-by: Samuel Ghinet <sghinet@cloudbasesolutions.com>

dpif-linux: Rename dpif-netlink; change to compile with MSVC.

The patch contains the necessary modifications to compile and also to run
under MSVC.

Added the files to the build system and also changed dpif_linux to be under
a more generic name dpif_windows.

Added a TODO under the windows part in case we want to implement another
counterpart for epoll functions.

Signed-off-by: Alin Gabriel Serdean <aserdean@cloudbasesolutions.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

ofproto: Warn about excessive rule counts in OpenFlow tables.

Frequently we've run into controller bugs which result in hundreds of
thousands, or even millions of rules being installed in an OpenFlow
table.  This isn't something trouble-shooters naturally think of to
check for, so it's nice to have a low rate warning message to hint at
the potential problem.

Signed-off-by: Ethan Jackson <ethan@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

dpif-netdev: Store miniflow length in exact match cache

This optimization is done to avoid calling count_1bits(), which, if
the popcnt istruction is not available might is slow. popcnt may not
be available because:

- We are running on old hardware
- (more likely) We're using a generic build (i.e. packaged OVS from a
  distro), not tuned for the specific CPU

Signed-off-by: Daniele Di Proietto <ddiproietto@vmware.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>

dpif-netdev: Introduce netdev_flow_key_* functions

netdev_flow_key is a miniflow with the following constraints:

1) It is used only inside dpif-netdev.c.
2) It always has inline values.
3) It contains only miniflows created by miniflow_extract().

Therefore, by using these new functions instead of the miniflow_*
ones, we get the following (performance related) benefits:

- Because of (1) the functions can be inlined.
- Because of (2) and (3) the netdev_flow_key can be treated as POD.
  Specifically, because of (3), we can do comparisons with memcmp,
  since if the map is different the miniflow must be different.

Signed-off-by: Daniele Di Proietto <ddiproietto@vmware.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>

ofproto-dpif-xlate: Wildcard skb_priority if QoS is disabled

This optimization should give a small performance benefit to the userspace
datapath.

Signed-off-by: Daniele Di Proietto <ddiproietto@vmware.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>

datapath-windows/Netlink: Add optional flag in policy.

Added the optional flag in policy structure. This would allow
caller to avoid checks for mandatory attributes if parsing
succeeds.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: Add support for getting 64 and 16 bit attributes.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows/Netlink: Added support for variable length attributes in validation.

Added minor fix for allowing support for variable lenghth attributes in
parsing policy.

Signed-off-by: Ankur Sharma <ankursharma@vmware.com>
Acked-by: Nithin Raju <nithin@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

ovs-dev.py: Support running the clang binaries.

They have slightly different support characteristics, so it's nice to
easily switch between them for testing.

Signed-off-by: Ethan Jackson <ethan@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

ovs-dev.py: Support additional optimization flags.

They may or may not make a difference, but there's no reason not to
support passing them.

Signed-off-by: Ethan Jackson <ethan@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

datapath-windows: use the Netlink set API and need new APIs

In this change:
1. we refactor the code that fills up information about the DP into
a seprate function.
2. use the netlink set APIs to fill up the netlink attributes.
3. we define a OVS_DP_STATS to be a typedef of 'struct ovs_dp_stats'
in keeping with the Windows kernel naming conventions.
4. In the absence of netlink set API, I had put in an ASSERT earlier
that the output buffer should be limited to 512 bytes. This is not
true anymore. The netlink set API checks for bounds of the buffer.
Hence removed the ASSERT.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Acked-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

datapath-windows: fix bug in NlBufCopyAtTailUninit

We should be returning value of tail before the increment
and not after.

Signed-off-by: Nithin Raju <nithin@vmware.com>
Acked-by: Ankur Sharma <ankursharma@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>

ofproto-dpif-xlate: Suppress some warnings on non-Linux OSes

These warnings were introduced by
commit 7d031d7e511aeea8dd45348922fe8e3bbdd2956e
("ofproto-dpif-xlate: Work around Linux netdev_max_backlog limit.")
and found by --enable-Werror build on NetBSD.

Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Signed-off-by: Ben Pfaff <blp@nicira.com>

compiler: Define NO_RETURN for MSVC.

To prevent warnings such as "Not all control paths return a value",
we should define NO_RETURN for MSVC.

Currently for gcc, we add NO_RETURN at the end of function declaration.
But for MSVC, "__declspec(noreturn)" is needed at the beginning of function
declaration. So this commit moves NO_RETURN to the beginning of the function
declaration as it works with gcc and clang too.

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

Fix remaining "uninitialized local variable" used warning by MSVC.

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

Fix remaining "void function returning a value" warning by MSVC.

MSVC complains about a void function returning a value if there is a
statement of the form - 'return foo()' even if foo() has a void return
type.

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>

ovs-atomic-msvc: Disable a compiler warning.

MSVC does not support c11 style atomics for the C compiler.
Windows has different InterLocked* functions for different data
sizes.  ovs-atomic-msvc.h maps the api in ovs-atomic.h (which is similar
to c11 atomics) to the available atomic functions in Windows. In some
cases, this causes compiler warnings about mismatched data sizes because
the generated code has 'if else' conditions on different data sizes and
proper casting is not possible.

In current OVS code base, we get one compiler warning through ovs-rcu.h
which says "‘void *’ differs in levels of indirection from LONGLONG."
This comes from the following in ovs-atomic-msvc.h for atomic_read64():
*(DST) = InterlockedOr64((int64_t volatile *) (SRC), 0);
when *DST is a void pointer (because InterLockedOr64 returns LONGLONG).
But this code path is only every hit for 64 bit data. So it should be safe to
disable the warning. (Any real bugs in api calls would hopefully be caught
while compiling on Linux using gcc/clang).

Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Eitan Eliahu <eliahue@vmware.com>

netdev-dpdk: Fix thread-safety breach.

dpdk_eth_dev_init() must be called with dpdk_mutex.  However,
netdev_dpdk_set_multiq() fails to follow this rule.  This commit
fixes this breach.

Found by clang.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

netdev-dpdk: Make get_config() report correct queue info.

With the separation of tx queue and rx queue configuration
in netdev-dpdk module, the netdev_dpdk_get_config() can no
longer report 'n_rxq' as tx queue configuration.

This commit fixes the above issue.

Reported-by: Daniele Di Proietto <ddiproietto@vmware.com>
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Daniele Di Proietto <ddiproietto@vmware.com>

dpif-netdev: Create multiple pmd threads by default.

With this commit, ovs by default will create one pmd thread
for each numa node and pin the pmd thread to available cpu
core on the numa node.

NON_PMD_CORE_ID (currently 0) is used to reserve a particular
cpu core for the I/O of all non-pmd threads.  No pmd thread
can be pinned to this reserved core.

As side-effects of this commit:

-  pmd thread will not be created, if there is no dpdk interface
   from the corresponding numa node added to ovs.

- the exact-match cache for non-pmd threads is removed from
  'struct dp_netdev'.  Instead, all non-pmd threads will use
  the exact-match cache defined in the 'struct dp_netdev_pmd_thread'
  for NON_PMD_CORE_ID.

- the rx packet processing functions are refactored to use
  'struct dp_netdev_pmd_thread' as input.

- the 'netdev_send()' function will be called with the proper
  queue id.

- both pmd and non-pmd threads can call the dpif_netdev_execute().
  so, use a per-thread key to help recognize the calling thread.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

netdev-dpdk: Remove the tx queue spinlock.

The previous commit makes OVS create one tx queue for each
cpu core, each pmd thread will use a separate tx queue.
Also, tx of non-pmd threads on dpdk interface is all through
'NON_PMD_THREAD_TX_QUEUE', protected by the 'nonpmd_mempool_mutex'.
Therefore, the spinlock is no longer needed.  And this commit
removes it from 'struct dpdk_tx_queue'.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

netdev-dpdk: Add indicator for flushing tx queue.

Previous commit makes OVS create one tx queue for each cpu
core.  An upcoming patch will allow multiple pmd threads be
created and pinned to cpu cores.  So each pmd thread will use
the tx queue corresponding to its core id.

Moreover, the pmd threads running on different numa node than
the dpdk interface (called non-local pmd thread) will not
handle the rx of the interface.  Consequently, there need to
be a way to flush the tx queues of the non-local pmd threads.

To address the queue flushing issue, this commit introduces a
new flag 'flush_tx' in the 'struct dpdk_tx_queue' which is
set if the queue is to be used by a non-local pmd thread.
Then, when enqueueing the tx pkts, if the flag is set, the tx
queue will always be flushed immediately after the enqueue.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>

dpif-netdev: Create multiple tx/rx queues when adding dpdk interface.

Before this commit, ovs creates one tx and one rx queue for
each dpdk interface and uses only one poll thread for handling
I/O of all dpdk interfaces.  An upcoming patch will allow multiple
poll threads be created.  As a preparation, this commit changes
the dpif-netdev to create multiple tx/rx queues when the dpdk
interface is added.

Specifically, the number of rx queues will still be one per-dpdk
interface for this commit.  But upcoming work will allow user
create multiple rx queues.  The number of tx queues will be the
number of cpu cores on the machine.  Although not all the tx queues
will be used, each poll thread will have its own queue for
transmission on the dpdk interface.

Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>