This patch set implements two ideas improve ovsdb connection scaling:
* Allow multiple jsonrpc connection to share a single ovsdb monitor, thus
maintaining a single set of changes.
* With in the same monitor, cache the json object generated for an update.
If the same update is requested by a
jsonrpc-server.c has two main functions. One deals with handling the
jsonrpc connections, the other deals with monitoring the database.
Currently, each jsonrpc connections has its own set of DB monitors.
This can be wasteful if a number of connections shares the same
monitors.
This patch, and a f
To hide ovsdb_monitor_table object from ovsdb_jsonrpc serve.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 22 ++
1 file changed, 14 insertions(+), 8 deletions(-)
diff --git a/ovsdb/jsonrpc-server.c b/ovsdb/jsonrpc-server.c
index 1241537..973cc46 100644
--- a/ovsdb/j
Change ovsdb_jsonrpc_parse_monitor_request() to make
ovsdb_monitor_table an opaque object.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 73 +-
1 file changed, 54 insertions(+), 19 deletions(-)
diff --git a/ovsdb/jsonrpc-server.c b/ovsdb/j
To make ovsdb_monitor an opaque to ovsdb_jsonrpc server object.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/ovsdb/jsonrpc-server.c b/ovsdb/jsonrpc-server.c
index 5fdf08b..ba0e28c 100644
--- a/ovsdb/jsonrpc-server
Now it simply calls ovsdb_monitor_compose_table_update(), which
is actually creates the json object.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 15 +++
1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/ovsdb/jsonrpc-server.c b/ovsdb/jsonrpc-server.c
index 973c
split out per monitoring needs_flush() into
ovsdb_monitor_needs_flush().
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 25 +
1 file changed, 17 insertions(+), 8 deletions(-)
diff --git a/ovsdb/jsonrpc-server.c b/ovsdb/jsonrpc-server.c
index a25a6f1..0630664 10064
rename ovsdb_jsonrpc_monitor_get_initial() to
ovsdb_monitor_get_initial()
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/ovsdb/jsonrpc-server.c b/ovsdb/jsonrpc-server.c
index 0630664..886062c 100644
--- a/ovs
Add ovsdb_minitor_destory() function to properly cleanup ovsdb_monitor.
It is also responsible for unhook from the replica chain.
The replica destroy callback is now called
ovsdb_monitor_destroy_callback()
Minor variable renaming in ovsdb_monitor_create() to make it
more consistent.
Signed-off-b
Added new files ovsdb-monitor.[ch] for monitor backend functions.
There is no functional changes.
Signed-off-by: Andy Zhou
---
ovsdb/automake.mk | 2 +
ovsdb/jsonrpc-server.c | 577 +---
ovsdb/jsonrpc-server.h | 3 +
ovsdb/ovsdb-monitor.c | 5
Add ovsdb_minitor_destory() function to properly cleanup ovsdb_monitor.
It is also responsible for unhook from the replica chain.
The replica destroy callback is now called
ovsdb_monitor_destroy_callback()
Minor variable renaming in ovsdb_monitor_create() to make it
more consistent.
Signed-off-b
Refactoring ovsdb_monitor_get_initial() to not generate JSON object.
It only collect changes within the ovsdb_monitor().
ovsdb_jsonrpc_monitor_compose_table_update() is then used to generate
JSON object.
This change will also make future patch easier.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-
Currently, each ovsdb-monitor points to a single jsonrpc_monitor object.
This means there is 1:1 relationship between them.
In case multiple jsonrpc-monitors needs to monitor the same tables and
the columns within them, then can share a single ovsdb-monitor, so the
updates only needs to be maintai
Added new files ovsdb-monitor.[ch] for monitor backend functions.
There is no functional changes.
Signed-off-by: Andy Zhou
---
ovsdb/automake.mk | 2 +
ovsdb/jsonrpc-server.c | 577 +---
ovsdb/jsonrpc-server.h | 3 +
ovsdb/ovsdb-monitor.c | 5
Refactoring ovsdb_monitor_get_initial() to not generate JSON object.
It only collect changes within the ovsdb_monitor().
ovsdb_jsonrpc_monitor_compose_table_update() is then used to generate
JSON object.
This change will also make future patch easier.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-
With N:1 mappings, multiple jsonrpc server may be servicing the rpc
connection at a different pace. ovsdb-monitor thus needs to maintain
different change sets, depends on connection speed of each rpc
connections. Connections servicing at the same speed can share the
same change set.
Transaction ID
Currently, each ovsdb-monitor points to a single jsonrpc_monitor object.
This means there is 1:1 relationship between them.
In case multiple jsonrpc-monitors needs to monitor the same tables and
the columns within them, then can share a single ovsdb-monitor, so the
updates only needs to be maintai
Currently, each monitor table contains a single hmap 'changes' to
track updates. This patch introduces a new data structure
'ovsdb_monitor_changes' that stores the updates 'rows' tagged by
its first commit transaction id. Each 'ovsdb_monitor_changes' is
refenece counted allowing multiple jsonrpc_mo
Although multiple jsonrpc monitor can share the same ovsdb monitor, each
change still needs to transalted into json object from scratch. This
can be wastful is mutiple jsonrpc monitors are interested in the
same changes.
Json chche impoves this by keeping an copy of json object generated
for trans
Signed-off-by: Andy Zhou
---
ovsdb/ovsdb-monitor.c | 124 ++
1 file changed, 85 insertions(+), 39 deletions(-)
diff --git a/ovsdb/ovsdb-monitor.c b/ovsdb/ovsdb-monitor.c
index a1efa66..3259f3c 100644
--- a/ovsdb/ovsdb-monitor.c
+++ b/ovsdb/ovsdb-mo
jsonrpc_monitor_compse_update() seems fit better than
jsonrpc_monitor_compose_table_update(), since it composes changes
from all tables.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 12 ++--
ovsdb/ovsdb-monitor.c | 4 ++--
ovsdb/ovsdb-monitor.h | 4 ++--
3 files changed, 10
Currently, each monitor table contains a single hmap 'changes' to
track updates. This patch introduces a new data structure
'ovsdb_monitor_changes' that stores the updates 'rows' tagged by
its first commit transaction id. Each 'ovsdb_monitor_changes' is
refenece counted allowing multiple jsonrpc_mo
Add ovsdb_monitor_add_jsonrpc_monitor(). This change will allow
ovsdb_monitor to be reference counted.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 2 ++
ovsdb/ovsdb-monitor.c | 16 +++-
ovsdb/ovsdb-monitor.h | 3 +++
3 files changed, 16 insertions(+), 5 deletions(-)
di
Signed-off-by: Andy Zhou
---
ovsdb/ovsdb-monitor.c | 124 ++
1 file changed, 85 insertions(+), 39 deletions(-)
diff --git a/ovsdb/ovsdb-monitor.c b/ovsdb/ovsdb-monitor.c
index a1efa66..3259f3c 100644
--- a/ovsdb/ovsdb-monitor.c
+++ b/ovsdb/ovsdb-mo
Add ovsdb_monitor_add_jsonrpc_monitor(). This change will allow
ovsdb_monitor to be reference counted.
Signed-off-by: Andy Zhou
---
ovsdb/jsonrpc-server.c | 2 ++
ovsdb/ovsdb-monitor.c | 16 +++-
ovsdb/ovsdb-monitor.h | 3 +++
3 files changed, 16 insertions(+), 5 deletions(-)
di
Store ovsdb monitor in global hmap. A newly created ovsdb monitor
object will first search the global hmap for a possible match. If
one is found, the existing ovsdb monitor is used instead.
With this patch, jsonrpc monitor and ovsdb monitor now have N:1 mapping.
Signed-off-by: Andy Zhou
---
ovs
Although multiple jsonrpc monitor can share the same ovsdb monitor, each
change still needs to transalted into json object from scratch. This
can be wastful is mutiple jsonrpc monitors are interested in the
same changes.
Json chche impoves this by keeping an copy of json object generated
for trans
With N:1 mappings, multiple jsonrpc server may be servicing the rpc
connection at a different pace. ovsdb-monitor thus needs to maintain
different change sets, depends on connection speed of each rpc
connections. Connections servicing at the same speed can share the
same change set.
Transaction ID
It appears that I sent out some old patch files in error. Sorry.
The set of 18 is the set for review. The set of 20 is an older set,
please disregard them.
On Thu, Mar 19, 2015 at 12:08 AM, Andy Zhou wrote:
> This patch set implements two ideas improve ovsdb connection scaling:
>
> * Allow mul
The original message was received at Thu, 19 Mar 2015 16:59:42 +0800 from
openvswitch.org [142.149.211.165]
- The following addresses had permanent fatal errors -
dev@openvswitch.org
- Transcript of session follows -
... while talking to 18.174.192.14:
554 5.0.0 Service unavailab
Hello again,
At lib/dpif.c in function "dpif_operate" there is a line that is being
called.
dpif->dpif_class->operate(dpif, ops, chunk);
This is possibly the actions are performed, but I have no clue how to know
what function this line leads to.
In lib/dpif-provider.h I can find the definition f
Hi Ben,
This was originally authored by Mark as part of an earlier patch set -
http://openvswitch.org/pipermail/dev/2015-March/052123.html
I am handling the rework of same hence the dual sign-off.
Is there a preferred way to handled situations like this?
Cheers,
Billy
-Original Message--
To indicate that the author is different from the submitter, the first
line in the body of the email should be:
From: Author Name
"git send-email" does that automatically if the author and submitter are
different.
On Thu, Mar 19, 2015 at 11:41:19AM +, O Mahony, Billy wrote:
> Hi Ben,
I change the author to Mark and applied this to master. Thank you!
On Thu, Mar 19, 2015 at 06:34:09AM -0700, Ben Pfaff wrote:
> To indicate that the author is different from the submitter, the first
> line in the body of the email should be:
> From: Author Name
> "git send-email" does th
Thanks for the hint! I was starting to get confused.
On Thu, Mar 19, 2015 at 01:50:31AM -0700, Andy Zhou wrote:
> It appears that I sent out some old patch files in error. Sorry.
>
> The set of 18 is the set for review. The set of 20 is an older set,
> please disregard them.
>
>
>
> On Thu, M
Thanks, Ben.
I'll follow the method you outlined for subsequent patches.
Billy.
-Original Message-
From: Ben Pfaff [mailto:b...@nicira.com]
Sent: Thursday, March 19, 2015 1:39 PM
To: O Mahony, Billy
Cc: dev@openvswitch.org
Subject: Re: [ovs-dev] [PATCH v2] automake.mk: Improve schema ch
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
View IBSP/UNESCO Attached Letter
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
The design was come up after inputs and discussions with multiple
people, including (in alphabetical order) Aaron Rosen, Ben Pfaff,
Ganesan Chandrashekhar, Justin Pettit, Russell Bryant and Somik Behera.
Signed-off-by: Gurucharan Shetty
---
ovn/CONTAINERS.OpenStack.md | 114
It is ovn-nbd and not ovs-nbd.
Signed-off-by: Gurucharan Shetty
---
ovn/ovn-architecture.7.xml |8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/ovn/ovn-architecture.7.xml b/ovn/ovn-architecture.7.xml
index a5b0807..035527f 100644
--- a/ovn/ovn-architecture.7.xml
+++
Looks good. It looks like there's another one in "ovn-nb.xml".
Acked-by: Justin Pettit
--Justin
> On Mar 19, 2015, at 7:32 AM, Gurucharan Shetty wrote:
>
> It is ovn-nbd and not ovs-nbd.
>
> Signed-off-by: Gurucharan Shetty
> ---
> ovn/ovn-architecture.7.xml |8
> 1 file chang
On Thu, Mar 19, 2015 at 07:32:28AM -0700, Gurucharan Shetty wrote:
> It is ovn-nbd and not ovs-nbd.
>
> Signed-off-by: Gurucharan Shetty
Acked-by: Ben Pfaff
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
port_unref() shouldn't immediately free the port and close the netdev,
because otherwise threads that still have a pointer to the port would
crash.
This commit fixes the problem by introducing an ovsrcu_synchronize()
call in port_unref().
I chose not to use ovsrcu_postpone(), because postponing t
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 03/19/2015 10:31 AM, Gurucharan Shetty wrote:
> The design was come up after inputs and discussions with multiple
> people, including (in alphabetical order) Aaron Rosen, Ben Pfaff,
> Ganesan Chandrashekhar, Justin Pettit, Russell Bryant and Somik Behera.
>
> Signed-off-by: Gurucharan Shetty
>
>
> Could you summarize the difference from v2? I don't recall any real
> discussion about v2 on the list other than our brief exchange about the
> 4096 CIFs limitation.
Sorry about that. There are only 2 differences between the [RFC v2]
and this more official patch.
1. Expanded introduction to CI
On Thu, Mar 19, 2015 at 10:07:40AM -0700, Gurucharan Shetty wrote:
> >
> > Could you summarize the difference from v2? I don't recall any real
> > discussion about v2 on the list other than our brief exchange about the
> > 4096 CIFs limitation.
> Sorry about that. There are only 2 differences betw
> From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Daniele Di
> Proietto
> Sent: Thursday, March 12, 2015 6:05 PM
>
> If rte_mempool_create() fails with ENOMEM, try asking for a smaller
> mempools. This patch enables OVS DPDK to run on systems without 1GB
> hugepages
>
> Signed-off-by
> -Original Message-
> From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Traynor,
> Kevin
> Sent: Monday, March 16, 2015 6:46 PM
[snip]
> > >
> > > Hi, I've reviewed this patchset - few comments/questions on it…
> >
> > Hi Kevin, thanks
> >
> > > I haven't tested yet - but I'm
> -Original Message-
> From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Daniele Di
> Proietto
> Sent: Thursday, March 12, 2015 6:05 PM
> This series improves OVS configuration with DPDK in three ways:
>
> * netdev-dpdk is patched to work on smaller systems (without 1GB
> hugep
Hi,
Would you be interested in Acquiring 3Com Users contacts list of 2015?
Information Fields - Name, Title, Email, Phone, Company Name, Physical
Address, City, State, Zip Code, Country, Web Address, Employee Size,
Revenue Size and Industry.
We also have other technology users like:
> On 19 Mar 2015, at 18:05, Gray, Mark D wrote:
>
>> From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Daniele Di
>> Proietto
>> Sent: Thursday, March 12, 2015 6:05 PM
>
>>
>> If rte_mempool_create() fails with ENOMEM, try asking for a smaller
>> mempools. This patch enables OVS DPDK
> From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Gray, Mark
> D
> Sent: Thursday, March 19, 2015 6:19 PM
> To: Traynor, Kevin; Daniele Di Proietto
> Subject: Re: [ovs-dev] [PATCH 0/6] DPDK: simplify configuration
>
> [snip]
> > > >
> > > > Hi, I've reviewed this patchset - few comment
> It might be worth giving the back-story on rationale for change #2.
Here is the backstory:
Before the introduction to Container (spawned inside VMs) networking
in OVN, OVN's architecture was designed with VMs in mind. In this
case, ovn-controller could create a record in Bindings table as soon
as
This RFC patch makes use of the vhost-user implementation that will
be available in DPDK 2.0. Submitting as RFC as this implementation
depends on a stable version of DPDK 2.0 and following that, OVS
support for DPDK 2.0. As such, this patch can be considered experimental.
This patch builds on top
This patch adds support for a new port type to the userspace datapath
called dpdkvhostuser. It adds to the existing infrastructure of
vhost-cuse, however disables vhost-cuse ports in favour of vhost-user
ports.
A new dpdkvhostuser port will create a unix domain socket which when
provided to QEMU i
>
> I wasn’t aware of this restriction. It’s (very) briefly stated in
> INSTALL.DPDK.md under ‘Restrictions:’, though.
>
> Do you maybe want to send a patch to clarify this? You clearly have a better
> understanding of the issue :-)
>
Sure thing, I will do this.
___
Raúl,
The dip class functions are implemented by the different dip classes. Linux and
Windows data paths are interfaced by lib/dpif-netlink.c and the userspace
datapath is implemented in lib/dpif-netlink.c. You’ll find a class struct with
the local function pointers in those files, just search
ovn-nbctl is intended to be a utility for manually interacting with
the OVN northbound db. A real consumer of OVN, such as OpenStack,
should be using ovsdb directly. However, a utility like this is
useful during development and testing of both ovn itself as well as
the OpenStack Neutron plugin.
I suppose my only question with this patch is why we can't just
allocate MIN_NB_MBUF sized pools in the first place.
Acked-by: Ethan Jackson
On Thu, Mar 12, 2015 at 11:04 AM, Daniele Di Proietto
wrote:
> If rte_mempool_create() fails with ENOMEM, try asking for a smaller
> mempools. This patch
Acked-by: Jarno Rajahalme
Ben may also want to have a look, as commits 98de6bebb (dpif-netdev: Fix
another use-after-free in port_unref().) and 87400a3d4cc4a (dpif-netdev: Fix
use-after-free in port_unref().) by him have fixed earlier problems related to
this. I did not check the changes in th
Thanks Jarno for the hints!
I found this function looking for ", struct dpif_op " term in every file.
static void dpif_linux_operate(struct dpif *dpif_, struct dpif_op **ops,
size_t n_ops)
The main lines in that function of my interest are:
dpif_linux_encode_execute(dpif->dp_ifindex, execute, &a
Any custom actions need to be supported by the datapaths, otherwise they fail
flow setup validation. Linux datapath (kernel module) is part of the linux
kernel sources, but if you need to try out something you can test your feature
with the userspace datapath first, then maybe with the ova-tree
> On Mar 19, 2015, at 4:02 PM, Russell Bryant wrote:
>
> ovn-nbctl is intended to be a utility for manually interacting with
> the OVN northbound db. A real consumer of OVN, such as OpenStack,
> should be using ovsdb directly. However, a utility like this is
> useful during development and te
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
i_дH
]Ï\¡r©¢ÞUEü#3wmeAgëòdä¢Ù&µÑs,Pæ²mÏSëbJ5÷ÆgÕ
îù¤óaó¾z0O&#MÁälrôNi]öctÖÇ]8ì¬êÆEþ
UBk¦vøÚðx®VàüÐEÃ6Vw1ølQ#pónKwt¬ m Lk'øÅ_
äýÍ°ÑÚûËg,;;Ó#!1f KgÇV
\NÁÔô¨§U)wêöHù(ÎS¦ùs5¾w¬.8N(ë?éÏì×ÏkÙ· ÏYêþeÛÊ¥?EÅd¬ºB|ZÔ¾O·
J3ËîHwó8dÊhÆÚ.ùQú]M6j¦`Y¿PG®&v'gn6·àÈVêå7æýÁ
On Wed, 18 Mar 2015 14:47:25 -0700: Alex Wang wrote:
> This commit makes the bash completion scripts be installed to
> /etc/bash_completion.d through 'make install' or package installation.
> This will make the scripts available for each bash session.
> An alternative is to put scripts
On Thu, Mar 19, 2015 at 5:08 PM, Peter Amidon wrote:
> On Wed, 18 Mar 2015 14:47:25 -0700: Alex Wang wrote:
>
>
> > This commit makes the bash completion scripts be installed to
> > /etc/bash_completion.d through 'make install' or package installation.
> > This will make the scripts availa
Signed-off-by: Jarno Rajahalme
---
ofproto/ofproto-dpif-xlate.c | 66 +-
ofproto/ofproto-dpif-xlate.h |6
2 files changed, 39 insertions(+), 33 deletions(-)
diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
index 0e28c77..
The flow key should be the same regardless of whether a live bucket is
found or not, as it would be confusing that the flow key would be
different (different mask bits) after the last group bucket goes dead.
In general, the megaflow algorithm expects the mask bits be set as
soon as we read the hea
We used to roll back group bucket changes only for 'all' and
'indirect' group types, but the expected semantics of all group types
is that any changes by the group bucket are not visible after the
group has been executed.
Signed-off-by: Jarno Rajahalme
---
ofproto/ofproto-dpif-xlate.c | 37 +++
Prior to this patch MPLS recirculation was not performed on a table
lookup following an MPLS_POP action. This patch refactors MPLS
recirculation triggering so that a table action can be re-done after
recirculation if that table action follows an MPLS_POP action.
Recirculation for a patch port tra
On Thu, Mar 5, 2015 at 1:42 PM, Kevin Traynor wrote:
> This patch adds support for a new port type to userspace datapath
> called dpdkvhost. This allows KVM (QEMU) to offload the servicing
> of virtio-net devices to its associated dpdkvhost port. Instructions
> for use are in INSTALL.DPDK.
>
> Thi
On Wed, 18 Mar 2015 14:47:24 -0700: Alex Wang wrote:
> Modifying $COMP_WORDBREAKS in completion script is not the recommended
> as it is a global variable and the modification could affect the behavior
> of other completion scripts. As a workaround, this commit uses the
> _get_comp_word
Hello everyone,
why are we using a in_port action, when we have output action ?
xlate_output_action has a default case where we are checking the port
should not be same as input port .
Why this check is required ?
Thanks and regards ,
Deepak
___
dev mai
On Fri, Mar 20, 2015 at 09:25:37AM +0530, deepaksingh wrote:
> why are we using a in_port action, when we have output action ?
> xlate_output_action has a default case where we are checking the port
> should not be same as input port .
> Why this check is required ?
Because OpenFlow says so.
_
Please don't drop the mailing list.
You will break OpenFlow compatibility.
What are you trying to accomplish?
On Fri, Mar 20, 2015 at 09:30:26AM +0530, deepaksingh wrote:
> Thanks Ben .
>
> If i will remove this check, will it create any problem ?
>
> On Fri, Mar 20, 2015 at 9:28 AM, Ben Pfaff
Please don't drop the list!
On Fri, Mar 20, 2015 at 09:40:23AM +0530, deepaksingh wrote:
> Hi Ben ,
>
> the L3 flows for wireless-client communicate with wireless-client in
> another VLAN will look like below.
>
> In_port=ats,vlan=10,src_mac=client1's
> MAC,dst_mac=vRouter'sMAC,eth_type=0x800,ds
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Hey Ben,
Do you have more comments? This series should be ready for commit~
Thanks,
Alex Wang,
On Thu, Mar 19, 2015 at 8:44 PM, Peter Amidon wrote:
> On Wed, 18 Mar 2015 14:47:24 -0700: Alex Wang wrote:
>
>
> > Modifying $COMP_WORDBREAKS in completion script is not the recommended
> > as
Hi,
this patch set implements the group select selection method extension that
I circulated some months ago.
The implementation makes use of a group experimenter property and
thus depends on Open Flow 1.5 groups (ONF EXT-250).
The last patch of the series adds an implementation of a hash selecti
NTR selection method
Signed-off-by: Simon Horman
---
This patch does not define any selection method names, however,
a subsequent patch proposes a "hash" selection method.
v4
* No change
v3
* Use NTR instead of NMX as Netronome extension prefix
v2
* Use array of TLVs rather than OF1.1 match f
NTR selection method
Signed-off-by: Simon Horman
Acked-by: Ben Pfaff
---
v4
* Included Ben Pfaff's Ack
v3
* Use NTR instead of NMX as Netronome extension prefix
v2
* Initial post
---
Documentation/automake.mk | 2 +
Documentation/group-selection-method-property.txt
This is in preparation for supporting group mod and desc reply
messages with an NTR selection method group experimenter property.
Currently decoding always fails as it only allows properties for known
selection methods and no selection methods are known yet. A subsequent
patch will propose a hash
Include NTR selection method experimenter group property in
in group mod request and group desc reply.
NTR selection method
Signed-off-by: Simon Horman
---
v4
* Simplify append_group_desc() by copying properties from struct ofpgroup
which now has a properties field.
* Do not add fields parame
Only the default existing behaviour is translated.
All other methods are rejected for now.
NTR selection method
Signed-off-by: Simon Horman
---
v4
* Update group_dpif_get_selection_method() to reflect
that struct ofgroup now has a props field.
* Do not needlessly include openflow/netronome-ex
NTR selection method
Signed-off-by: Simon Horman
---
v4
* Use mf_parse_value() instead of mf_parse() in parse_select_group_field()
as there is no mask to parse.
* Document that new features are only supported in OVS2.4+ with *OF1.5+
* '\0' is the null character. '0' is not!
* Use bitmap of fie
This is intended as a usable demonstration of how
the NTR selection method extension might may be used.
NTR selection method
Signed-off-by: Simon Horman
---
v4
* Update group_dpif_get_fields and group_dpif_get_selection_method_param to
reflect that struct ofgroup now has a props field.
* Use
The AutoAttach table is new in OVS 2.3.90, but ovs-vsctl was creating a
record in the table unconditionally whenever it created a new bridge.
This caused a gratuitous incompatibility with older databases, which can
be a problem during upgrade. This commit switches to creating the
AutoAttaach recor
The first two commits fix fairly serious upgrade problems specifically
related to the new auto-attach feature. The third commit is less critical;
it fixes a similar upgrade bug in ovs-vswitchd specifically. The fourth
fixes a long-standing bug in ovsdb-server that I've known about for a few
weeks
Otherwise this creates a gratuitous incompatibility with older databases,
which can be a problem in upgrade situations.
VMware-BZ: #1413562
Reported-by: Alex Wang
Signed-off-by: Ben Pfaff
---
utilities/ovs-vsctl.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/utilitie
Until now, if ovs-vsctl (or another client of the C ovsdb-idl library) was
compiled against a schema that had a column or table that was not in the
database actually being used (e.g. during an upgrade), and the column or
table was selected for monitoring, then ovsdb-idl would fail to get any
data a
I realized that this bug existed only a few weeks ago, but it has been
there since the earliest ovsdb-server code.
Signed-off-by: Ben Pfaff
---
ovsdb/jsonrpc-server.c | 26 ++
ovsdb/ovsdb-server.1.in | 29 -
tests/ovsdb-server.at | 2 +-
3
93 matches
Mail list logo
