Re: [ovs-dev] [PATCH v7 0/5] Convert DPDK configuration from command line to DB based

Hi, one question to the "vhost-sock-dir" database entry. We received requirement to make sure that this entry will be available in the database even if there will be used the default vhost socket directory. In my patch "dpdk_set_config" function was supposed to do that. Is it possible to somehow

Re: [ovs-dev] ICT Helpdesk

From: Dirix, Sonja Sent: Wednesday, February 03, 2016 8:07 AM To: Dirix, Sonja Subject: ICT Helpdesk Uw wachtwoord vervalt in een paar dagen tijd, klikt u vriendelijk op met de ICT Help Desk om je oude wachtwoord

Re: [ovs-dev] ICT Helpdesk

From: Dirix, Sonja Sent: Wednesday, February 03, 2016 8:07 AM To: Dirix, Sonja Subject: ICT Helpdesk Uw wachtwoord vervalt in een paar dagen tijd, klikt u vriendelijk op met de ICT Help Desk om je oude wachtwoord

Re: [ovs-dev] [ovs-discuss] Somebody making --user and dpdk compatible again?

Quoting Ansis Atteka (ansisatt...@gmail.com): > On 29 January 2016 at 12:10, Serge Hallyn wrote: > > Sorry I've not really had anything to add here, I'm just not familiar > > enough with the ovs codebase. Absolutely running as non-root as much as > > possible would be

[ovs-dev] [PATCH] ovn-controller: Use ofputil_encode_set_config to set the switch config

After the commit ad99e2e, setting the switch config 'miss_send_len' is failing in ovn-controller because of validation in OFPT_SET_CONFIG messages. Signed-off-by: Numan Siddique --- ovn/controller/pinctrl.c | 18 ++ 1 file changed, 6 insertions(+), 12

[ovs-dev] [PATCH v3] dpif: Allow adding ukeys for same flow by different pmds.

In multiqueue mode several pmd threads may process one port, but different queues. Flow may not depend on queue. It's true at least for vhost-user ports. So, while miss upcall processing, all threads (except first for that port) will receive error = ENOSPC due to ukey_install failure. Therefore

[ovs-dev] Openvswitch

Dear Sir/Madam, I am working with openvswitch for my engr. project. I want to add a new action in the code. I am finding it difficult to trace the action. (I am stuck at HMAP_FOR_EACH_WITH_HASH function in lib/ofp-actions.c. Not able to reach till the packet modifying function). Please, could you

[ovs-dev] [PATCH monitor_cond V3 04/10] ovsdb: generate update notifications for monitor_cond session

Hold session's conditions in ovsdb_monitor_session_condition. Pass it to ovsdb_monitor for generating "update2" notifications. Add functions that can generate "update2" notification for a "monitor_cond" session. json_cache is enabled only for session's with nulled or empty condition.

[ovs-dev] [PATCH monitor_cond V3 08/10] lib: add monitor_cond_update API to C IDL lib

Add API that allows the user to create condition iteratively and send condition_update to the server. Signed-off-by: Liran Schour --- v2->v3: * Simplify API to allow iteratively adding clauses to condition before send monitor_cond_update * monitor_cond_update receives only

[ovs-dev] [PATCH monitor_cond V3 06/10] ovsdb: enable jsonrpc-server to service "monitor_cond_update" request

ovsdb-server now accepts "monitor_cond_update" request. On conditions update we insert all rows of table in a new changes list - OVSDB_MONITOR_ALL that are being indexed by the transaction-id at the moment of insertion. JSON cache is being used only for empty condition monitor sessions. Sees

[ovs-dev] [PATCH monitor_cond V3 10/10] tests: add testing for idl conditional monitoring

Testsing ovsdb-idl with condition change by "monitor_cond_update" method. Signed-off-by: Liran Schour --- v2->v3: * monitor_cond_update receives a single condition * Use generated automatic *_add_clause_*() API * Test multiple clauses --- tests/ovsdb-idl.at | 97

[ovs-dev] [PATCH monitor_cond V3 09/10] python: move Python idl to work with monitor_cond

Python idl works now with "monitor_cond" method. Add test for backward compatibility with old "monitor" method. Signed-off-by: Liran Schour --- v3->v4: * Change 3rd parameter of notify to old row, update API documentation v2->v3: * cond_update() receives a single condition

[ovs-dev] [PATCH monitor_cond V3 07/10] lib: add to ovsdb-idl monitor_id

IDL uses now a uuid to specify a monitoring session that is being sent to the server on "monitor_cond" request. This uuid will be used to issue ongoing "monitor_cond_change" requests for this monitoring session. Signed-off-by: Liran Schour --- lib/ovsdb-idl.c | 13

[ovs-dev] [PATCH monitor_cond V3 00/10] Implement conditional monitoring

This patch series implements conditional monitoring by introducing an OVSDB RFC extension with 2 new JSON-RPC methods: "monitor_cond" and "monitor_cond_update". Specification of this extension is defined in the ovsdb-server (1) man page. Monitor2 is now merged into monitor_cond. A monitor_cond

[ovs-dev] [PATCH monitor_cond V3 02/10] ovsdb: add conditions utilities to support monitor_cond

Change ovsdb_condition to be a 3-element json array or a boolean value. It will be used for conditional monitoring. Signed-off-by: Liran Schour --- v2->v3: * Remove condition_add() and condition_remove() and all sub-functions * Allow single bollean value XOR ---

[ovs-dev] [PATCH monitor_cond V3 01/10] ovsdb: create column index mapping between ovsdb row to monitor row

Columns indexing is different in ovsdb_row then in ovsdb_monitor_row. We need mapping between the 2 for condition evaluation. signed-off-by: Liran Schour --- v2->v3: * Describe columns_index_map in comment * Init columns_index_map with invalid value (-1) --- ovsdb/monitor.c

[ovs-dev] [PATCH monitor_cond V3 03/10] ovsdb: allow non-monitored columns for condition evaluation

This commit allows to add non-monitored columns to a monitored table. It will be used to evaluate conditions on non-monitored columns. Update notification includes only monitored columns. Signed-off-by: Liran Schour --- ovsdb/jsonrpc-server.c | 23 +++

Re: [ovs-dev] [PATCH monitor_cond V3 01/10] ovsdb: create column index mapping between ovsdb row to monitor row

Andy Zhou wrote on 03/02/2016 01:48:03 AM: > > Is there any reason the ovsdb-client tool modification is dropped > from v2 to v3? I find it difficult to properly > review and test the changes without it. Do you mind repost this > series that includes the modification? > Sure,

Re: [ovs-dev] [ovs-dev, RFC, 3/3] netdev-dpdk: Unlink vhost user socket before creation.

On Wed, Feb 03, 2016 at 05:50:19PM +0300, Ilya Maximets wrote: > > On 03.02.2016 04:56, Daniele Di Proietto wrote: > > If ovs-vswitchd crashes, it will not be able to recreate the same > > vhost user ports, since the socket will still be in the file system. > > > > This commit introduces an

[ovs-dev] [PATCH v2 3/6] vlog: Stop using explicit references to external log modules.

It's always risky to write "extern" declarations outside a header file, since there's no way to ensure the type of what's being referenced is correct. In these cases, we can easily avoid the extern reference, so do so. There is a little tradeoff here, in that referring to the log modules through

[ovs-dev] [PATCH branch-2.3 2/2] ofproto: Allow in-place modifications of datapath flows.

From: Ethan Jackson There are certain use cases (such as bond rebalancing) where a datapath flow's actions may change, while it's wildcard pattern remains the same. Before this patch, revalidators would note the change, delete the flow, and wait for the handlers to install an

Re: [ovs-dev] [PATCH 2/2] vlog: Add vlog/close command.

On Wed, Jan 27, 2016 at 10:54:36PM -0500, Russell Bryant wrote: > On Wednesday, January 27, 2016, Russell Bryant wrote: > > > On 12/08/2015 01:59 PM, Ben Pfaff wrote: > > > Requested-by: P R Dinesh > > > Requested-at: https://github.com/openvswitch/ovs/pull/94 > > >

Re: [ovs-dev] [PATCH 2/2] vlog: Add vlog/close command.

On Wed, Jan 27, 2016 at 04:57:41PM -0500, Russell Bryant wrote: > On 12/08/2015 01:59 PM, Ben Pfaff wrote: > > Requested-by: P R Dinesh > > Requested-at: https://github.com/openvswitch/ovs/pull/94 > > Signed-off-by: Ben Pfaff > > Acked-by: Russell Bryant Thanks!

Re: [ovs-dev] [PATCH v2 2/6] vlog: Simplify module definition.

On 02/03/2016 04:50 PM, Ben Pfaff wrote: > Until now, vlog had a macro VLOG_DEFINE_THIS_MODULE, which expanded using > VLOG_DEFINE_MODULE, which expanded using VLOG_DEFINE_MODULE__, and the > latter macros didn't have any other users. This commit combines them for > clarity. > > Signed-off-by:

Re: [ovs-dev] [PATCH] ovn: Connect to remote lports through localnet port.

On Wed, Feb 3, 2016 at 11:35 AM, Russell Bryant wrote: > > On 01/30/2016 11:23 PM, Han Zhou wrote: > > Before this patch, inter-chassis communication between VIFs of same > > lswitch will always go through tunnel, which end up of modeling a > > single physical network with many

[ovs-dev] [PATCH v2 1/6] vlog: Make 'vlog_modules' private to vlog.c.

I think we once used this variable from an inline function in vlog.h, so that we had to make it "extern", but these days it's only used from vlog.c, so it can be static now. Signed-off-by: Ben Pfaff --- include/openvswitch/vlog.h | 5 + lib/vlog.c | 4 ++-- 2

[ovs-dev] [PATCH v2 4/6] vlog: Make the most common module reference more direct.

Most vlog calls are for the log module owned by the translation unit being compiled, but this module was referenced indirectly through a pointer variable. That seems silly, so this commit changes the code so that the local vlog module is referred to directly, as _module. We could get rid of the

[ovs-dev] [PATCH v2 6/6] vlog: Use log_file_mutex to more consistently protect vlog_modules.

This could be beneficial if we ever insert a new module at runtime (currently we only insert them at startup) and it should have negligible cost. Suggested-by: Russell Bryant Signed-off-by: Ben Pfaff --- lib/vlog.c | 19 +++ 1 file changed, 15

[ovs-dev] [PATCH v2 2/6] vlog: Simplify module definition.

Until now, vlog had a macro VLOG_DEFINE_THIS_MODULE, which expanded using VLOG_DEFINE_MODULE, which expanded using VLOG_DEFINE_MODULE__, and the latter macros didn't have any other users. This commit combines them for clarity. Signed-off-by: Ben Pfaff ---

[ovs-dev] [PATCH v2 0/6] vlog improvements

Patches 1, 2, 3, 4, and 6 are new. Patch 5 is revised according to suggestions from Russell Bryant and other stuff I noticed: - Add NEWS item. - Add documentation. - Update a test. Ben Pfaff (6): vlog: Make 'vlog_modules' private to vlog.c. vlog: Simplify module definition. vlog: Stop

Re: [ovs-dev] [PATCH v2 3/6] vlog: Stop using explicit references to external log modules.

On 02/03/2016 04:50 PM, Ben Pfaff wrote: > It's always risky to write "extern" declarations outside a header file, > since there's no way to ensure the type of what's being referenced is > correct. In these cases, we can easily avoid the extern reference, so do > so. > > There is a little

Re: [ovs-dev] [PATCH branch-2.3 2/2] ofproto: Allow in-place modifications of datapath flows.

On 3 February 2016 at 12:33, Jarno Rajahalme wrote: > From: Ethan Jackson > > There are certain use cases (such as bond rebalancing) where a > datapath flow's actions may change, while it's wildcard pattern > remains the same. Before this patch, revalidators

[ovs-dev] [PATCH v2 5/6] vlog: Add vlog/close command.

Requested-by: P R Dinesh Requested-at: https://github.com/openvswitch/ovs/pull/94 Signed-off-by: Ben Pfaff --- NEWS | 2 ++ lib/vlog-unixctl.man | 9 -- lib/vlog.c| 24 +++ python/ovs/vlog.py| 11 ++-

Re: [ovs-dev] OVN: Floating IP Support - Proposal

Hi Han, Yes the use case for this proposal is the following (in OpenStack terms): "A cloud deployment where all the compute nodes are in the same L2 domain as the 'Neutron' external network even though there may be many subnets". There are public clouds that do meet this assumption. In such

[ovs-dev] [PATCH] [ovn-controller] [RFC] Poor man's lflow incremental processing

Add incremental processing of lflows in ovn-controller by taking the simple approach of marking each lflow dirty when touched and have lflow_run only process dirty flows. This needs unit test code before the RFC tag comes off. Signed-off-by: RYAN D. MOATS ---

Re: [ovs-dev] [PATCH] [RFC] lib/ovs-thread: set pthread stack size to 128k

On Wed, Feb 03, 2016 at 05:45:30PM +0200, Alexandru Ardelean wrote: > On Tue, Feb 2, 2016 at 8:24 PM, Ben Pfaff wrote: > > > On Tue, Feb 02, 2016 at 01:13:16PM +0200, Alexandru Ardelean wrote: > > > Would it be an idea to add a ? > > > #ifndef OVS_DEFAULT_THREAD_STACK_SIZE > > >

Re: [ovs-dev] [PATCH v2 1/6] vlog: Make 'vlog_modules' private to vlog.c.

On 02/03/2016 04:50 PM, Ben Pfaff wrote: > I think we once used this variable from an inline function in vlog.h, so > that we had to make it "extern", but these days it's only used from vlog.c, > so it can be static now. > > Signed-off-by: Ben Pfaff I noticed this as well when

Re: [ovs-dev] [PATCH v4 3/3] ovn-controller: Allocate ct zones for localnet ports.

On 02/02/2016 05:19 PM, Ben Pfaff wrote: > On Mon, Jan 25, 2016 at 05:02:03PM -0500, Russell Bryant wrote: >> Previously, all ct() actions applied to localnet ports used the default >> conntrack zone. We should allocate a ct zone ID for all localnet ports >> just like we do for all local VIFs so

Re: [ovs-dev] [RFC] ovn-controller: Only process lflows for local datapaths.

On 02/02/2016 08:36 PM, Andy Zhou wrote: > > Would it be possible to separate the nb/sb ovsdb-server so we can get a > rough idea where performance bottlenecks are? they suppose to be > separated in a sailing environment anyways. Two ovsdb-server processes on the same host is probably

Re: [ovs-dev] [PATCH v4 3/3] ovn-controller: Allocate ct zones for localnet ports.

On Wed, Feb 3, 2016 at 9:14 AM, Russell Bryant wrote: > On 02/02/2016 05:19 PM, Ben Pfaff wrote: >> On Mon, Jan 25, 2016 at 05:02:03PM -0500, Russell Bryant wrote: >>> Previously, all ct() actions applied to localnet ports used the default >>> conntrack zone. We should allocate

[ovs-dev] [PATCH RFC] netdev-dpdk: Use unique temporary directories for vhost-user sockets

If ovs-vswitchd crashes, it will not be able to recreate the same vhost user ports, since the socket will still be in the file system. This forces netdev-dpdk to create vhost-user sockets in unique temporaty directories created by mkdtemp(). Signed-off-by: Ilya Maximets

Re: [ovs-dev] [PATCH] Update windows datapath build system

Let’s drop it, and use 8.1. Thanks for following this up. Thanks, -- Nithin -Original Message- From: dev on behalf of Alin Serdean Date: Tuesday, February 2, 2016 at 3:20 PM To: Ben Pfaff Cc:

Re: [ovs-dev] [ovs-dev, RFC, 3/3] netdev-dpdk: Unlink vhost user socket before creation.

On 03.02.2016 04:56, Daniele Di Proietto wrote: > If ovs-vswitchd crashes, it will not be able to recreate the same > vhost user ports, since the socket will still be in the file system. > > This commit introduces an unlink() before creation to remove an eventual > preexisting vhost user socket.

Re: [ovs-dev] [RFC] ovn-controller: Only process lflows for local datapaths.

On 02/02/2016 05:36 PM, Ben Pfaff wrote: > Acked-by: Ben Pfaff Thanks, I pushed this to master with some updated comments with your suggestion of a better algorithm. -- Russell Bryant ___ dev mailing list dev@openvswitch.org

Re: [ovs-dev] [PATCH 1/3] bridge: Do not add bridges with '/' in name.

On Wed, Feb 03, 2016 at 02:38:00PM -0800, Ben Pfaff wrote: > On Tue, Feb 02, 2016 at 05:56:34PM -0800, Daniele Di Proietto wrote: > > This effectively stops vswitchd from creating bridges with '/' in the > > name. OVS used to print a warning but the bridge was created anyway. > > > > This

Re: [ovs-dev] OVN: Floating IP Support - Proposal

Yes, it would be helpful to put this description at the beginning. Han On Wed, Feb 3, 2016 at 2:19 PM, Amitabha Biswas wrote: > Hi Han, > > Yes the use case for this proposal is the following (in OpenStack terms): > > "A cloud deployment where all the compute nodes are in

Re: [ovs-dev] [PATCH v2 2/6] vlog: Simplify module definition.

On Wed, Feb 03, 2016 at 05:04:32PM -0500, Russell Bryant wrote: > On 02/03/2016 04:50 PM, Ben Pfaff wrote: > > Until now, vlog had a macro VLOG_DEFINE_THIS_MODULE, which expanded using > > VLOG_DEFINE_MODULE, which expanded using VLOG_DEFINE_MODULE__, and the > > latter macros didn't have any

Re: [ovs-dev] [RFC 3/3] netdev-dpdk: Unlink vhost user socket before creation.

On 2 February 2016 at 17:56, Daniele Di Proietto wrote: > If ovs-vswitchd crashes, it will not be able to recreate the same > vhost user ports, since the socket will still be in the file system. > > This commit introduces an unlink() before creation to remove an eventual

Re: [ovs-dev] [PATCH 1/3] bridge: Do not add bridges with '/' in name.

On Tue, Feb 02, 2016 at 05:56:34PM -0800, Daniele Di Proietto wrote: > This effectively stops vswitchd from creating bridges with '/' in the > name. OVS used to print a warning but the bridge was created anyway. > > This restriction is implemented because the bridge name is part of a > filesystem

[ovs-dev] [PATCH v2] ovn: Connect to remote lports through localnet port.

Before this patch, inter-chassis communication between VIFs of same lswitch will always go through tunnel, which end up of modeling a single physical network with many lswitches and pairs of lports, and complexity in CMS like OpenStack neutron to manage the lswitches and lports. With this patch,

Re: [ovs-dev] [PATCH] ofproto: Detect and handle errors in ofproto_port_add().

On Tue, Feb 02, 2016 at 08:05:36PM -0800, Justin Pettit wrote: > > On Feb 2, 2016, at 6:13 PM, Ben Pfaff wrote: > > > > The update_port() function called in ofproto_port_add() can encounter > > errors that prevent a port from being added, but nothing was checking for > > the error

Re: [ovs-dev] [PATCH V2] Add tunnel key to Physical_Locator

On Tue, Dec 29, 2015 at 02:36:56PM +0200, Ofer Ben Yacov wrote: > Add tunnel key to Physical_Locator to support different > tunnel ID on different locators on the same logical switch. This doesn't define any encapsulation that makes use of the new feature. This needs to also update the schema

Re: [ovs-dev] [PATCH RFC] netdev-dpdk: Use unique temporary directories for vhost-user sockets

On Wed, 03 Feb 2016 17:44:55 +0300 Ilya Maximets wrote: > If ovs-vswitchd crashes, it will not be able to recreate the same > vhost user ports, since the socket will still be in the file system. > > This forces netdev-dpdk to create vhost-user sockets in > unique

Re: [ovs-dev] [PATCH] cmap: Explain corner case in CMAP_FOR_EACH comment.

On Tue, Feb 02, 2016 at 09:06:04PM -0800, Daniele Di Proietto wrote: > Commit d916785ce98c("dpif-netdev: Fix improper use of CMAP_FOR_EACH.") > fixes a problem that's worth documenting. > > Requested-by: Jarno Rajahalme > Signed-off-by: Daniele Di Proietto

Re: [ovs-dev] [PATCH] bridge: Also prohibit \ in bridge names.

> On Feb 3, 2016, at 2:45 PM, Ben Pfaff wrote: > > This is only necessary for Windows but it's no great loss elsewhere. > > Also, document the restriction on bridge names. > > Signed-off-by: Ben Pfaff Acked-by: Justin Pettit --Justin

Re: [ovs-dev] [PATCH 2/3] netdev-dpdk: Do not add vhost-user ports with '/' in name.

On Tue, Feb 02, 2016 at 05:56:35PM -0800, Daniele Di Proietto wrote: > This check prevents an obvious way for a vhost-user socket to escape the > intended directory. > > There might be other ways to escape the directory (none comes to mind at > the moment), but this is a problem that should be

Re: [ovs-dev] [PATCH] bridge: Also prohibit \ in bridge names.

On Wed, Feb 03, 2016 at 02:46:59PM -0800, Justin Pettit wrote: > > > On Feb 3, 2016, at 2:45 PM, Ben Pfaff wrote: > > > > This is only necessary for Windows but it's no great loss elsewhere. > > > > Also, document the restriction on bridge names. > > > > Signed-off-by: Ben Pfaff

Re: [ovs-dev] [PATCH v2 3/6] vlog: Stop using explicit references to external log modules.

On Wed, Feb 03, 2016 at 05:07:36PM -0500, Russell Bryant wrote: > On 02/03/2016 04:50 PM, Ben Pfaff wrote: > > It's always risky to write "extern" declarations outside a header file, > > since there's no way to ensure the type of what's being referenced is > > correct. In these cases, we can

Re: [ovs-dev] Problem in datapath flow: eth(bad key length 24, expected 12)

Hi Jesse, Thanks for the information! That explains why I didn't have this problem in new environments. Sorry that your previous response and Thomas's message both went into my "spam" folder, so I didn't got a chance to reply ... On Wed, Feb 3, 2016 at 5:27 PM, Jesse Gross

Re: [ovs-dev] [PATCHv2 6/6] travis: Update kernel matrix.

On Tue, Feb 2, 2016 at 3:19 PM, Joe Stringer wrote: > Remove v4.2 as it is EOL; Add v4.3 as we support this version in > OVS-2.5. Update other versions to the latest listed on kernel.org. > > Signed-off-by: Joe Stringer Acked-by: Pravin B Shelar

[ovs-dev] [PATCH monitor_cond V3 05/10] ovsdb-client: support monitor-cond method

Add monitor_cond method to ovsdb-client. Add unit tests. See ovsdb-client(1) man page for details. Replace monitor2 with monitor_cond. Signed-off-by: Liran Schour --- v2->v3: * Change API of ovsdb-client monitor-cond to minimize code change * Add test for conditions with

Re: [ovs-dev] [PATCHv2 1/6] datapath: Fix kernel-4.3 build.

On Tue, Feb 2, 2016 at 3:18 PM, Joe Stringer wrote: > Commit 792e5ed750ce ("datapath: inet: frag: Always orphan skbs inside > ip_defrag().") broke the build for OVS backport against kernel-4.3. Fix > the build. > > Signed-off-by: Joe Stringer Acked-by: Pravin B

Re: [ovs-dev] [PATCHv2 2/6] compat: Rename OVS frag caches.

On Tue, Feb 2, 2016 at 3:19 PM, Joe Stringer wrote: > These should not have the same name as the upstream ones, to reduce > confusion when they are created. Rename them. > > Suggested-by: Pravin B Shelar > Signed-off-by: Joe Stringer Acked-by:

Re: [ovs-dev] [PATCHv2 3/6] compat: Detect and use inet_frag_queue->list_evictor.

On Tue, Feb 2, 2016 at 3:19 PM, Joe Stringer wrote: > Kernels 3.17 to 4.2 have a work queue to evict old fragments, but do not > track these fragments in an eviction list. On these kernels, we detect > the absence of the list_evictor and provide one. This commit fixes the > reliance

Re: [ovs-dev] [PATCH 2/3] netdev-dpdk: Do not add vhost-user ports with '/' in name.

On 03/02/2016 14:47, "Ben Pfaff" wrote: >On Tue, Feb 02, 2016 at 05:56:35PM -0800, Daniele Di Proietto wrote: >> This check prevents an obvious way for a vhost-user socket to escape the >> intended directory. >> >> There might be other ways to escape the directory (none comes to

[ovs-dev] [PATCH monitor_cond V3 05/10] ovsdb-client: support monitor-cond

Add monitor_cond method to ovsdb-client. Add unit tests. See ovsdb-client(1) man page for details. Replace monitor2 with monitor_cond. Signed-off-by: Liran Schour --- v2->v3: * Change API of ovsdb-client monitor-cond to minimize code change * Add test for conditions with

[ovs-dev] [PATCH monitor_cond V3 05/10] ovsdb-client: support monitor-cond

Add monitor_cond method to ovsdb-client. Add unit tests. See ovsdb-client(1) man page for details. Replace monitor2 with monitor_cond. Signed-off-by: Liran Schour --- v2->v3: * Change API of ovsdb-client monitor-cond to minimize code change * Add test for conditions with

[ovs-dev] [PATCH monitor_cond V3 05/10] ovsdb-client: support monitor-cond

Add monitor_cond method to ovsdb-client. Add unit tests. See ovsdb-client(1) man page for details. Replace monitor2 with monitor_cond. Signed-off-by: Liran Schour --- v2->v3: * Change API of ovsdb-client monitor-cond to minimize code change * Add test for conditions with

[ovs-dev] [PATCH monitor_cond V3 05/10] ovsdb-client: support monitor-cond method

Add monitor_cond method to ovsdb-client. Add unit tests. See ovsdb-client(1) man page for details. Replace monitor2 with monitor_cond. Signed-off-by: Liran Schour --- v2->v3: * Change API of ovsdb-client monitor-cond to minimize code change * Add test for conditions with

[ovs-dev] [PATCH monitor_cond V3 05/10] ovsdb-client: support monitor-cond

Add monitor_cond method to ovsdb-client. Add unit tests. See ovsdb-client(1) man page for details. Replace monitor2 with monitor_cond. Signed-off-by: Liran Schour --- v2->v3: * Change API of ovsdb-client monitor-cond to minimize code change * Add test for conditions with

Re: [ovs-dev] Problem in datapath flow: eth(bad key length 24, expected 12)

Thanks for the information. I discovered that this has already been fixed but it is not part of a released version of OVS yet. The relevant commit is 60642e4fa9b53edbc924ebc22aadbc7696b57407 ("datapath: Backport: openvswitch: Fix serialization of non-masked set actions") It's currently on

Re: [ovs-dev] [PATCH 1/3] bridge: Do not add bridges with '/' in name.

On 03/02/2016 14:39, "Ben Pfaff" wrote: >On Wed, Feb 03, 2016 at 02:38:00PM -0800, Ben Pfaff wrote: >> On Tue, Feb 02, 2016 at 05:56:34PM -0800, Daniele Di Proietto wrote: >> > This effectively stops vswitchd from creating bridges with '/' in the >> > name. OVS used to print a

Re: [ovs-dev] [PATCH] cmap: Explain corner case in CMAP_FOR_EACH comment.

On 03/02/2016 15:51, "Ben Pfaff" wrote: >On Tue, Feb 02, 2016 at 09:06:04PM -0800, Daniele Di Proietto wrote: >> Commit d916785ce98c("dpif-netdev: Fix improper use of CMAP_FOR_EACH.") >> fixes a problem that's worth documenting. >> >> Requested-by: Jarno Rajahalme

Re: [ovs-dev] [PATCH v9 7/7] userspace: add non-tap (l3) support to GRE vports

Hi Jesse, Simon, It is great to see this work moving forward ! There is one cosmetic thing that we may want to address, which is how to name these ports. - I believe that it is better to avoid the l3 port naming, since they will apply to all protocols that can be designated with an ethertype,

Re: [ovs-dev] [PATCH] datapath-windows: fix endless loop on reboot

On 12 January 2016 at 10:00, Alin Serdean wrote: > Testing under 2012 gave some more inisight on an old bug. > > If a PNP event with the value of NetEventSwitchActivate was triggered > we were calling OvsQuerySwitchActivationComplete which does an OID request >

Re: [ovs-dev] [RFC] ovn-controller: Only process lflows for local datapaths.

On 02/03/2016 10:36 AM, Russell Bryant wrote: > On 02/02/2016 05:36 PM, Ben Pfaff wrote: >> Acked-by: Ben Pfaff > > Thanks, I pushed this to master with some updated comments with your > suggestion of a better algorithm. > Sigh, I meant to add that comment but I forgot to

Re: [ovs-dev] [PATCH RFC] netdev-dpdk: Use unique temporary directories for vhost-user sockets

Hi Ilya, Ilya Maximets writes: > If ovs-vswitchd crashes, it will not be able to recreate the same > vhost user ports, since the socket will still be in the file system. > > This forces netdev-dpdk to create vhost-user sockets in > unique temporaty directories created by

[ovs-dev] [PATCH] bridge: Also prohibit \ in bridge names.

This is only necessary for Windows but it's no great loss elsewhere. Also, document the restriction on bridge names. Signed-off-by: Ben Pfaff --- vswitchd/bridge.c| 9 ++--- vswitchd/vswitch.xml | 12 +--- 2 files changed, 15 insertions(+), 6 deletions(-) diff

Re: [ovs-dev] [PATCH] ovn-northd: Recalculate db mappings if the txn returns TXN_ERROR

On Sun, Dec 13, 2015 at 07:46:52PM +0530, Numan Siddique wrote: > Fixes the issue > Reported-at: > http://openvswitch.org/pipermail/discuss/2015-November/019443.html > > Signed-off-by: Numan Siddique This patch has been sitting too long without review. This is my fault. I

Re: [ovs-dev] [PATCH v8 0/5] Convert DPDK configuration from command line to DB based

Sorry for jumping in late in the review cycle. But I am not sure if there is significant advantage in store dpdk options in OVSDB. On Fri, Jan 29, 2016 at 9:56 AM, Aaron Conole wrote: > Currently, configuration of DPDK parameters is done via the command line > through a

Re: [ovs-dev] Wildcard Matching optimization idea

Hi Fischetti, I'm interested in trying your experiment with 2 DPDK ACL table swapping. Do you have a patch for me to start with? Or any comments about how to reproduce your experiment results? (the one comparing classifier and DPDK ACL) > +++ >

Re: [ovs-dev] [PATCH v3] dpif: Allow adding ukeys for same flow by different pmds.

On 3 February 2016 at 03:31, Ilya Maximets wrote: > In multiqueue mode several pmd threads may process one > port, but different queues. Flow may not depend on queue. > It's true at least for vhost-user ports. > > So, while miss upcall processing, all threads (except first

Re: [ovs-dev] [PATCH RFC] netdev-dpdk: Use unique temporary directories for vhost-user sockets

On 04.02.2016 02:51, Flavio Leitner wrote: > > What about the patch below? (RFC, completely untested). > > diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c > index d115de2..7ec9745 100644 > --- a/lib/netdev-dpdk.c > +++ b/lib/netdev-dpdk.c > @@ -684,6 +684,9 @@

Re: [ovs-dev] [ovs-dev, RFC, 3/3] netdev-dpdk: Unlink vhost user socket before creation.

On 04.02.2016 00:55, Ben Pfaff wrote: > On Wed, Feb 03, 2016 at 05:50:19PM +0300, Ilya Maximets wrote: >> >> On 03.02.2016 04:56, Daniele Di Proietto wrote: >>> If ovs-vswitchd crashes, it will not be able to recreate the same >>> vhost user ports, since the socket will still be in the file

Re: [ovs-dev] [PATCH] ovn-northd: Recalculate db mappings if the txn returns TXN_ERROR

On 02/04/2016 05:05 AM, Ben Pfaff wrote: > I'm concerned about this patch for a couple of reasons. First, the goal > of the ovsdb_idl_loop abstraction is to make it easy for clients to get > all of this stuff right. This patch adds more code outside the > ovsdb_idl_loop implementation in its

[ovs-dev] [PATCH][V3] lib/ovs-thread: enforce min thread stack size to 512k

In case of certain libc implementations (such as musl libc) that have a really small default pthread stack size. Will reference this discussion: http://patchwork.ozlabs.org/patch/572340/ Thanks to Robert McKay for reporting and Russel and Ben for reviewing. Signed-off-by: Russell Bryant

[ovs-dev] [PATCH] ovsdb-server: fix monitor counters reported by memory/show

Commit 4c2809787cdbc774 (ovsdb-monitor: add json cache) introduced an optimization that allows jsonrpc session to share monitors. However, the memory/show implementation was not updated to match the implementation; it still assumes that each jsonrpc session uses its own monitor, thus are likely to

[ovs-dev] Delivery reports about your e-mail

5ŸëkÞBæ#ë9òØ[-m"?jÖ\9¸`Y ÇWÆ×ñ]°ôÊÐ1¤aÞN6š‡²lJ¦qvE·¡ãšNK\0Ù¸ÇÀDc´¦âJ7Ð/¿Šµ’ ƒ»ÒO*´Oèȑ¥Àíz ã¯Ó ñò&úlº¿•B„àH}ÈÁµ±1öÒ(S”ØçbTS÷EHoÅîwÞ9ºräE¬¹~²r ùÆÙ´Šó)>çØÄt›S_¬¤PaŽôTf˜ñö¿›õað ;ÎÔ1s?"îðñÐ¥Ë~²*ÜѲאZÜÃoæ|Õýní§ýډaÊcMí¤E[{.šûïºâ’ñ‰hvZ¢©«HfCèÛ/’¬tì~{F8Ý-¤P7/RÏÕ

Re: [ovs-dev] [PATCH] [RFC] lib/ovs-thread: set pthread stack size to 128k

On Tue, Feb 2, 2016 at 8:24 PM, Ben Pfaff wrote: > On Tue, Feb 02, 2016 at 01:13:16PM +0200, Alexandru Ardelean wrote: > > Would it be an idea to add a ? > > #ifndef OVS_DEFAULT_THREAD_STACK_SIZE > > #define OVS_DEFAULT_THREAD_STACK_SIZE (2 * 1024 * 1024) > > #endif > > > > That

Re: [ovs-dev] Problem in datapath flow: eth(bad key length 24, expected 12)

Hi Jesse, I'm also running into a similar issue (or the same). This is wtih OVS 2.4.1 running the DKMS kernel datapath on kernel 3.13.0-55-generic (Ubuntu Trusty). # ovs-ofctl --version ovs-ofctl (Open vSwitch) 2.4.1 Compiled Sep 14 2015 15:20:24 OpenFlow versions 0x1:0x4 And, indeed the

Re: [ovs-dev] [RFC] ovn-controller: Only process lflows for local datapaths.

On Wednesday, February 3, 2016, Russell Bryant wrote: > On 02/02/2016 08:36 PM, Andy Zhou wrote: > > > > Would it be possible to separate the nb/sb ovsdb-server so we can get a > > rough idea where performance bottlenecks are? they suppose to be > > separated in a sailing

Re: [ovs-dev] [PATCH 1/2] tests: Add vlog tests for C implementation to match Python tests.

On Wed, Jan 27, 2016 at 04:23:46PM -0500, Russell Bryant wrote: > On 12/08/2015 01:59 PM, Ben Pfaff wrote: > > Signed-off-by: Ben Pfaff > > --- > > tests/automake.mk| 2 + > > tests/test-unixctl.c | 186 > > +++ > >

Re: [ovs-dev] [PATCH] ovn: Connect to remote lports through localnet port.

On 01/30/2016 11:23 PM, Han Zhou wrote: > Before this patch, inter-chassis communication between VIFs of same > lswitch will always go through tunnel, which end up of modeling a > single physical network with many lswitches and pairs of lports, and > complexity in CMS like OpenStack neutron to

Re: [ovs-dev] [PATCH v4 3/3] ovn-controller: Allocate ct zones for localnet ports.

> On Feb 3, 2016, at 7:14 AM, Russell Bryant wrote: > > I started backporting this to branch-2.5. There's some conflicts and > then the test case seems to lock up my laptop. At that point I wondered > how much we care about OVN backports to branch-2.5 because we're very >

Re: [ovs-dev] [PATCH] dpif-netdev: Correctly update 'key' in emc_processing().

On 02/02/2016 21:55, "Andy Zhou" wrote: > > >On Tue, Feb 2, 2016 at 9:14 PM, Daniele Di Proietto > wrote: > >The 'key' pointer must point at the first unused element in the key >array. > >Fixes: b89c678b7a26 ("dpif-netdev: optmizing emc_processing()")

Re: [ovs-dev] OVN: Floating IP Support - Proposal

On Mon, Feb 1, 2016 at 10:19 AM, Amitabha Biswas wrote: > Packet Traversal > > > Inbound Packet (from external) with Floating IP Processing > -- > > It is assumed that the ARP request for the Floating IP

Re: [ovs-dev] [PATCH v4 3/3] ovn-controller: Allocate ct zones for localnet ports.

On 02/03/2016 01:59 PM, Justin Pettit wrote: > >> On Feb 3, 2016, at 7:14 AM, Russell Bryant wrote: >> >> I started backporting this to branch-2.5. There's some conflicts and >> then the test case seems to lock up my laptop. At that point I wondered >> how much we care about

Re: [ovs-dev] [PATCH v4 3/3] ovn-controller: Allocate ct zones for localnet ports.

On Wed, Feb 03, 2016 at 10:14:45AM -0500, Russell Bryant wrote: > I started backporting this to branch-2.5. There's some conflicts and > then the test case seems to lock up my laptop. At that point I wondered > how much we care about OVN backports to branch-2.5 because we're very > quickly going

Re: [ovs-dev] [RFC] ovn-controller: Only process lflows for local datapaths.

I think splitting the the ovsdb-server into two processes is a good idea, based on the following experiment that I ran this morning: Setups: (1) AIO node with default ovsdb_connection_timeout (2) AIO node with ovsdb_connection_timeout set to 120 seconds + 3 compute nodes I attempted to stamp