Ignacio | 28 Mar 07:36 2015

Avoid fakes, Get Australian Brand Meds

Amigo! Your girlfriends will be ready for anything! Believe me experience! 
Among 10 men suffering from impotence 7 find the answer! Are you in? 
http://x.co/8JBt1

_______________________________________________
dev mailing list
dev <at> openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Alex Wang | 28 Mar 07:44 2015

[PATCH] ofproto-dpif: Set need_revalidate when removing cfm from ofport.

When cfm is deleted from a port, all modules should release their
reference so that the cfm struct can be removed from the global hmap
and freed.  Therein, the reference held by xlate module can only be
released when the need_revalidate flag is set (e.g set to
REV_RECONFIGURE).  And this flag should be set while removing cfm
from ofport.  Unfortunately, this has never been done before and the
bug was hidden by another bug fixed in recent commit a190839
(netdev-vport: Do not update netdev when there is no config change.)

To fix this issue, this commit makes the code set need_revalidate
when removing cfm from ofport.

Signed-off-by: Alex Wang <alexw <at> nicira.com>
---
 ofproto/ofproto-dpif.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index 6402404..42ff747 100644
--- a/ofproto/ofproto-dpif.c
+++ b/ofproto/ofproto-dpif.c
 <at>  <at>  -1915,13 +1915,11  <at>  <at>  static int
 set_cfm(struct ofport *ofport_, const struct cfm_settings *s)
 {
     struct ofport_dpif *ofport = ofport_dpif_cast(ofport_);
+    struct ofproto_dpif *ofproto = ofproto_dpif_cast(ofport->up.ofproto);
     int error = 0;

     if (s) {
         if (!ofport->cfm) {
(Continue reading)

MAILER-DAEMON | 28 Mar 02:27 2015

Message could not be delivered

The original message was received at Sat, 28 Mar 2015 08:27:12 +0700
from openvswitch.org [145.215.108.170]

----- The following addresses had permanent fatal errors -----
dev <at> openvswitch.org

_______________________________________________
dev mailing list
dev <at> openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Russell Bryant | 27 Mar 22:37 2015
Picon

ovn: scope for first iteration

We've discussed wanting to have something to demo by the OpenStack
Summit in Vancouver (May 18-22), which I think is a fantastic goal.  I
wanted to clarify the scope of this first target.

So far in the work I've been doing, I've only been looking at the L2
parts (logical switches and logical ports).  What would you all think of
sticking with that for the first iteration?  We can use the DHCP and L3
agents from Neutron for now.

I've done some looking at the L3 parts of the current design, and I
think there's some more needed.  It would be nice to just make L2 work
and come back to it.

I haven't done an exhaustive review, but here are some of the things
that may need some discussion / changes for L3:

1) As discussed once before, DHCP support would be great to have in OVN.
 We can use the DHCP agent in Neutron in the meantime, though.

2) We've got no way to provide floating IP support.  Also related to
floating IP support, I came across this older message on the discuss
list that hasn't been followed up on (I missed it since it wasn't on dev):

http://openvswitch.org/pipermail/discuss/2015-March/016732.html

3) I don't think there's enough information in the ovn-nb schema to let
ovn-nbd figure out how to configure gateways in the ovn db.  Neutron has
an API concept called "provider networks" which seems to map well to
this (defining how to map Neutron networks to existing physical
networks).  However, I don't see a way to get that information down.
(Continue reading)

Alex Wang | 27 Mar 19:48 2015

[PATCH] netdev-linux: Make htb quantum always no less than mtu.

Currently, ovs uses hardcoded rate2quantum = 10 for each htb qdisc.
When qdisc class's rate is small, the resulting quantum (calculated
by min_rate / rate2quantum) will be smaller than MTU.  This is not
recommended and tc will keep complaining the following in syslog.

localhost kernel: HTB: quantum of class 10003 is small. Consider r2q change.
localhost kernel: HTB: quantum of class 10004 is small. Consider r2q change.
localhost kernel: HTB: quantum of class 10005 is small. Consider r2q change.
localhost kernel: HTB: quantum of class 10006 is small. Consider r2q change.

To fix the issue, this commit makes ovs always use htb quantum no less
than the MTU.

Signed-off-by: Alex Wang <alexw <at> nicira.com>
---
 lib/netdev-linux.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/lib/netdev-linux.c b/lib/netdev-linux.c
index 5e6f01e..5c0fdff 100644
--- a/lib/netdev-linux.c
+++ b/lib/netdev-linux.c
 <at>  <at>  -3497,6 +3497,7  <at>  <at>  static const struct tc_ops tc_ops_sfq = {
 /* HTB traffic control class. */

 #define HTB_N_QUEUES 0xf000
+#define HTB_RATE2QUANTUM 10

 struct htb {
     struct tc tc;
(Continue reading)

Kevin Traynor | 27 Mar 19:06 2015
Picon

[PATCH v2] netdev-dpdk: Put cuse thread into quiescent state.

ovsrcu_synchronize() is used when setting virtio_dev to NULL.
This results in an ovsrcu_quiesce_end() call which means the
cuse thread may not go into quiescent state again for an
indefinite time. Add an ovsrcu_quiesce_start() call to prevent
this.

Signed-off-by: Kevin Traynor <kevin.traynor <at> intel.com>
---
 lib/netdev-dpdk.c |   13 +++++++++----
 1 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index 4e16f39..f69154b 100644
--- a/lib/netdev-dpdk.c
+++ b/lib/netdev-dpdk.c
 <at>  <at>  -1586,6 +1586,11  <at>  <at>  destroy_device(volatile struct virtio_net *dev)
              * setting the virtio_dev to NULL.
              */
             ovsrcu_synchronize();
+            /*
+             * As call to ovsrcu_synchronize() will end the quiescent state,
+             * put thread back into quiescent state before returning.
+             */
+            ovsrcu_quiesce_start();
         }
     }
     ovs_mutex_unlock(&dpdk_mutex);
 <at>  <at>  -1614,6 +1619,8  <at>  <at>  static void *
 start_cuse_session_loop(void *dummy OVS_UNUSED)
 {
(Continue reading)

Daniele Di Proietto | 27 Mar 17:29 2015

[PATCHv4 0/9] Add pmd thread statistics

The goal of this series is to add an appctl command to retrieve detailed
performance statistics from the pmd thread execution. These statistics
include exact match cache and masked classifier hits and rough cycles counters.

This series also introduces a simple lock (taken from the cmap) which is used
to protect (already existing and newly introduced) statistics reads/write.

v3 -> v4:
* Avoid creating an empty structure on MSVC 64-bit builds (thanks Ben and Guru)

v2 -> v3:
* Rename cntlock to seqlock (suggested by Ben)
* Initialize flow stats seqlock in dpif-netdev
* Change OVS_UNLIKELY to OVS_LIKELY in seqlock.h
 (v2 erroneously changed the semantics of this)
* Minor style and typos fixes
* Rebase

v1 -> v2:
* Group writes to stats variables
* Introduce 64-bit stats lock, based on new cntlock (thanks to Jarno)
* Fix a typo (thanks to Jarno)
* Protect existing dpif-netdev stats updates with new locks (thanks to Jarno)
* Protect new stats updates with new locks (thanks to Jarno)
* Change rte_get_timer_cycles() to rte_get_tsc_cycles(). This prevents an OVS
 DPDK build from crashing when DPDK has not been enabled at runtime

Daniele Di Proietto (9):
  dpif-netdev: Account for and free lost packets.
  dpif-netdev: Group statistics updates in the slow path.
(Continue reading)

Daniele Di Proietto | 27 Mar 14:26 2015

[PATCH v3 0/9] Add pmd thread statistics

The goal of this series is to add an appctl command to retrieve detailed
performance statistics from the pmd thread execution. These statistics
include exact match cache and masked classifier hits and rough cycles counters.

This series also introduces a simple lock (taken from the cmap) which is used
to protect (already existing and newly introduced) statistics reads/write.

v2 -> v3:
* Rename cntlock to seqlock (suggested by Ben)
* Initialize flow stats seqlock in dpif-netdev
* Change OVS_UNLIKELY to OVS_LIKELY in seqlock.h
  (v2 erroneously changed the semantics of this)
* Minor style and typos fixes
* Rebase

v1 -> v2:
* Group writes to stats variables
* Introduce 64-bit stats lock, based on new cntlock (thanks to Jarno)
* Fix a typo (thanks to Jarno)
* Protect existing dpif-netdev stats updates with new locks (thanks to Jarno)
* Protect new stats updates with new locks (thanks to Jarno)
* Change rte_get_timer_cycles() to rte_get_tsc_cycles(). This prevents an OVS
  DPDK build from crashing when DPDK has not been enabled at runtime

Daniele Di Proietto (9):
  dpif-netdev: Account for and free lost packets.
  dpif-netdev: Group statistics updates in the slow path.
  lib: Introduce seqlock mechanism.
  lib: Introduce u64-stats-sync.
  dpif-netdev: Use u64_stats_sync when reading/writing stats.
(Continue reading)

daf | 27 Mar 10:47 2015
Picon

Returned mail: Data format error

K"(wQ֪Fj}{,:#;&o1x;rnFQ6Z?|Hƕ/A"IжUWEAuJv
U,ssPB
62a5*!m
͜ޱs-!G8V1:_C~7&Fp6κ6J3؍>ݮsGѹChBܲ۝'Q1ard4wŏxi}ztf֙Nqąt窟P>R
o!lGLHdRݚo?ܷpaewd9(?yZyHozH
ծsϾy1N5f{˪Iu2hv<ae,
~L7Ɗnz.Ycԝ<vK;q*cuxW1,ɣ[KCNX]DgZnxLlY
J?Pў
]6ղQ(K$"ŏ8
v )uN GD$`5żl,cEE.J^q؞6;$"82HLl&
WUl\[goZNL;5d.d[v2Y~ԤA_Yۦf]MCA\8
3O,Șn񔃻D
&ŲcJrդTŗ5yv2]3]gq,^N>H)W|6`%{  T;?KmpNWыQUs~C"ż1~3RYiܐH!A9jcmD6(AXq``Z&B~y߷qͿPIbl꩐r$d}CGFsalҁ!1rN`AuIy8ݵK9ߔX4F?Q4pwJέ0?V{TnKc-G\Gt"~FGx(BU͇|Z:&gi
[S
܈JJ)IK#ӹ"H!Ehw{֜/S5]C-
ݍ5SK~."^4
i)Ƽ!rxµ
'86cOFP{Ö͓"lmpkh64kW7ӄ9g#z0<#:w

\w%"KlZEN^.u
OԠ{?ſ%qxW:z]hcő?Α3ȕFH\!EΊ0H!OSppEn_8\z50fi
sؔMe$68OϓRH'ƸHmǺe2
W!'׾RTwaGbD!..n
B9\K>;¸o
<:N)?3yZZ
irx(wBX/_%B0tяdQt,0t."g].meyCe,6R1c,m~w$k$fI,}agWRC,PR9y8<ەxK
Mi_-`e,(yKbƝx͡`AI%mX2a)9Yהw-lpd&D뜕o"7M^l\[m:Ǚ!Wazrv0.(K
-u
nvӐƎCر"6uQĠQ*E)RC-E`og(wS{WGοLlm̄D3Jh4,FVDpZox'~z7\֭tږR"8s܁Vx2ekW܈Z6O86۬hq
.RZF8W䉯,'z9c3ED"ՉM$!u1|AChxaY,0`-uyʧ;*ԘdC*9N9AH|p'jpԅs,Hb{fϤI1kg h(xO
(Continue reading)

Alex Wang | 27 Mar 02:00 2015

[PATCH] netdev-vport: Do not update netdev when there is no config change.

When there is any update from ovsdb, ovs will call netdev_set_config()
for every vport.  Even though the change is not related to vport, the
current implementation will always increment the per-netdev sequence
number.  Subsequently this could cause even more unwanted effects,
e.g. the recreation of 'struct tnl_port' in ofproto level.

This commit fixes the issue by only updating the netdev when there
is actual configuration change.

Signed-off-by: Alex Wang <alexw <at> nicira.com>
---
 lib/netdev-vport.c |   16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/lib/netdev-vport.c b/lib/netdev-vport.c
index 954ab9b..28db568 100644
--- a/lib/netdev-vport.c
+++ b/lib/netdev-vport.c
 <at>  <at>  -612,9 +612,11  <at>  <at>  set_tunnel_config(struct netdev *dev_, const struct smap *args)
                                &tnl_cfg.out_key_flow);

     ovs_mutex_lock(&dev->mutex);
-    dev->tnl_cfg = tnl_cfg;
-    tunnel_check_status_change__(dev);
-    netdev_change_seq_changed(dev_);
+    if (memcmp(&dev->tnl_cfg, &tnl_cfg, sizeof tnl_cfg)) {
+        dev->tnl_cfg = tnl_cfg;
+        tunnel_check_status_change__(dev);
+        netdev_change_seq_changed(dev_);
+    }
(Continue reading)

Ricky Li | 27 Mar 01:55 2015
Picon

[PATCH] Bug fixed: User space tunneling doesn't support setting tunnel key by flow

e.g. Set tunnel id for encapsulated VxLAN packet (out_key=flow):

ovs-vsctl add-port int-br vxlan0 -- set interface vxlan0 \
    type=vxlan options:remote_ip=172.168.1.2 options:out_key=flow

ovs-ofctl add-flow int-br in_port=LOCAL, icmp,\
    actions=set_tunnel:3, output:1 (1 is the port# of vxlan0)

Output tunnel ID should be modified to 3 with this patch.

=========================================================

Add test case for setting tunnel id in userspace DP

This test case is to test setting tunnel id by flow rule,
in userspace datapath.

e.g. actions=set_tunnel:3,output:[tunnel port]

test case: dnl Check VXLAN tunnel push set tunnel id by flow
 <at> tests/tunnel-push-pop.at

=========================================================

Signed-off-by: Ricky Li <ricky.li <at> intel.com>
---
 lib/netdev-provider.h    |  2 +-
 lib/netdev-vport.c       | 10 ++++++----
 lib/netdev.c             |  5 +++--
 lib/netdev.h             |  3 ++-
(Continue reading)


Gmane