Thanks for your advice.I have done this job .
Best Regards,
Mark
> Subject: Re: [Ryu-devel] how to increase openflow buffer size
> To: [email protected]
> CC: [email protected]
> From: [email protected]
> Date: Fri, 15 Jan 2016 16:28:09 +0900
>
> Hi,
>
> On 2016年01月14
Currently, some actions or match fields are missing from Ryu-Documentation.
These patches update the list and add the documents for Nicira Extension
Structures.
IWASE Yusuke (2):
doc: ofproto_ref: Update supported actions/matches list
doc: Add Nicira Extension Structures references
doc/sourc
Signed-off-by: IWASE Yusuke
---
ryu/ofproto/ofproto_v1_2_parser.py | 16
ryu/ofproto/ofproto_v1_3_parser.py | 16
ryu/ofproto/ofproto_v1_4_parser.py | 14 ++
ryu/ofproto/ofproto_v1_5_parser.py | 13 +
4 files changed, 59 insertions(+)
dif
Signed-off-by: IWASE Yusuke
---
doc/source/developing.rst | 1 +
doc/source/nicira_ext_ref.rst | 31 +++
ryu/ofproto/nx_match.py | 28
3 files changed, 60 insertions(+)
create mode 100644 doc/source/nicira_ext_ref.rst
diff --gi
Sigh. Needed to check that the datapath being unregistered is the same as the
one returned from _get_switch(), especially since a switch re-connection could
cause an OFPStateChange event for a DEAD_DISPATCHER transition to be queued
after the new datapath is added to self.dps.
Without this chec
On Jan 17, 2016, at 6:00 PM, Victor Orlikowski
mailto:[email protected]>> wrote:
I will fix this shortly, and send another iteration of the topology patch.
OK. A little later than I’d hoped, but let’s try this again.
Both patches attached, only the topology patch has changed.
Apply both to a clean
Signed-off-by: IWASE Yusuke
---
.../json/of13/libofproto-OFP13-features_reply.packet.json| 9 +
.../json/of13/libofproto-OFP13-get_config_reply.packet.json | 6 ++
.../ofproto/json/of13/libofproto-OFP13-set_config.packet.json| 6 ++
.../json/of15/libofproto-OFP15-
Add the following OFP13/15 Messages.
- OFP13/15 Handshake
Implemented: OFPT_FEATURES_REPLY
TODO:OFPT_FEATURES_REQUEST
- OFP13/15 Switch Configuration Messages
Implemented: OFPT_SET_CONFIG, OFPT_GET_CONFIG_REPLY
TODO:OFPT_GET_CONFIG_REQUEST
Signed-off-by: IWAS
Signed-off-by: IWASE Yusuke
---
.../of13/libofproto-OFP13-features_reply.packet| Bin 0 -> 32 bytes
.../of13/libofproto-OFP13-get_config_reply.packet | Bin 0 -> 12 bytes
.../packet_data/of13/libofproto-OFP13-set_config.packet| Bin 0 -> 12 bytes
.../of15/libofproto-OFP15
Currently, test_ofctl compares the sent messages in binary, so it
is difficult to figure out the differences of the sent messages
and the expected messages.
This patch fixes to compare the sent messages in json representation
in order to make easy to debug.
Signed-off-by: IWASE Yusuke
---
ryu/t
The followings fix test_ofctl to compare the sent messages in
json representation in order to make easy to debug and include
some improvements.
IWASE Yusuke (2):
test_ofctl: Compare sent messages in json representation
lib/ofctl_v1_*: Confirm hw_addr is str in mod_port_behavior
ryu/lib/ofctl
Signed-off-by: IWASE Yusuke
---
ryu/lib/ofctl_v1_0.py | 2 +-
ryu/lib/ofctl_v1_2.py | 2 +-
ryu/lib/ofctl_v1_3.py | 2 +-
3 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/ryu/lib/ofctl_v1_0.py b/ryu/lib/ofctl_v1_0.py
index cb1e84e..430d63f 100644
--- a/ryu/lib/ofctl_v1_0.py
+++ b/ry
Actually, I just thought of another bug in the topology module.
If the switch reconnects, a DEAD_DISPATCHER OFPStateChange event may be queued
for after the completion of the processing of the OFPStateChange event for the
MAIN_DISPATCHER.
In this case, _get_switch() will return a value for the
Thought of a corner case in _unregister() to handle.
Signed-off-by: Victor J. Orlikowski
diff --git a/ryu/topology/event.py b/ryu/topology/event.py
index c54152b..e7b682c 100644
--- a/ryu/topology/event.py
+++ b/ryu/topology/event.py
@@ -41,6 +41,11 @@ class EventSwitchLeave(EventSwitchBase):
On Jan 17, 2016, at 1:36 PM, Victor Orlikowski
mailto:[email protected]>> wrote:
One of them is the revised changes to topology.
The other is the stability patch.
You should apply these, one after the other, to a clean Ryu 3.29.
I thought of another corner case in topology, that is already handled i
New revision, based on comments from Zhao Jun.
Signed-off-by: Victor J. Orlikowski
diff --git a/ryu/topology/event.py b/ryu/topology/event.py
index c54152b..e7b682c 100644
--- a/ryu/topology/event.py
+++ b/ryu/topology/event.py
@@ -41,6 +41,11 @@ class EventSwitchLeave(EventSwitchBase):
On Jan 17, 2016, at 9:29 AM, 俊 赵
mailto:[email protected]>> wrote:
OK, I'd like to try that patch. Thank you.
OK.
Two patches attached this time.
One of them is the revised changes to topology.
The other is the stability patch.
You should apply these, one after the other, to a clean Ryu 3.2
OK, I'd like to try that patch. Thank you.
Zhao Jun
From: [email protected]
To: [email protected]
Date: Sun, 17 Jan 2016 14:04:46 +
CC: [email protected]
Subject: Re: [Ryu-devel] [maybe a bug] dpset : multiple connections
Ok; I will revise the patch to add it to _EVENTS.
Ok; I will revise the patch to add it to _EVENTS.
The failure to close connections is the reason for the patch I am in the middle
of testing.
We encountered the failure to close connections several months ago, but it took
us a while to discover the causes, and to sort out what was going on.
I p
I tried the patch.
If we add event.EventSwitchReconnected in _EVENTS ryu.topology.switches's
Switches class, the patch will work fine!
But it seems that it does not correctly close the older connection when facing
multiple connections as shown in picture in attachment.
Thanks.
Zhao Jun
From: v.
20 matches
Mail list logo
