Thanks for the explanation and the link Russell.
Is there a regular meeting where this topic might be discussed?
On Mar 5, 2016 02:40, "Russell Bryant" wrote:
> There's a lot of work happening to improve ovsdb performance (both on the
> client and server sides). There's testing happening in mul
Dear user of openvswitch.org, Mail system administrator of openvswitch.org
would like to inform you that,
We have received reports that your account was used to send a huge amount of
junk e-mail messages during the recent week.
Obviously, your computer had been infected and now contains a trojan
Hi Joe,
Thanks for your reply. The reason is that line 181:
ofpact_finish(ofpacts, &bundle->ofpact);
https://github.com/openvswitch/ovs/blob/34abaa3deaa430ca0b50453865d2e042a5132165/lib/bundle.c#L181
Could also call into ofpbuf_put_zero, which frees the memory pointed by
bundle.
Regards,
Wil
(Responding to both patches here as I'm having email reception problems)
On 3 March 2016 at 15:12, Jarno Rajahalme wrote:
> Commit 38876d31 fixed setting 'msg' when resizing an ofpbuf, but
> failed to fix the same issue in ofpbuf_clone_with_headroom(). Without
> this fix the newly cloned ofpbuf'
Yes, you are correct. I am misled by the fact that the bit mask starts with two
0s from right to left (has discontiguous 1-bits). This can be
explained by the byte order.
However, there are still some cases that byte order cannot explain. I installed
two rules,
ovs-ofctl add-flow br0
in_p
On 4 March 2016 at 18:00, William Tu wrote:
> Address pointed by bundle could be obsolete/free'd when
> realloc, called from ofpbuf_put_zero(), returns new address.
> Reported by Valgrind 367: ovs-ofctl parse-flows (NXM)
>
> Invalid write of size 4
> bundle_parse__ (bundle.c:200)
> bundle_
Address pointed by bundle could be obsolete/free'd when
realloc, called from ofpbuf_put_zero(), returns new address.
Reported by Valgrind 367: ovs-ofctl parse-flows (NXM)
Invalid write of size 4
bundle_parse__ (bundle.c:200)
bundle_parse_load (bundle.c:272)
parse_bundle_load (ofp-actio
when ctx->ovnsb_idl_txn is null, binding_run exits early
and does not add any local_datapaths, but patch_run
doesnt check this, and ends up deleting localnet ports,
because there are no local datapaths for them,
They get readded in a subsequent run causing unnecessary
deletion and readdition.
Sign
Address pointed by header_ptr might be free'd due to realloc
happened at ofpbuf_put_uninit() and ofpbuf_put_hex(). Reported
by valgrind 379: check TCP flags expression in OXM and NXM.
Invalid write of size 4
nx_match_from_string_raw (nx-match.c:1510)
nx_match_from_string (nx-match.c:1538)
On Fri, Mar 04, 2016 at 02:51:42PM +0800, Peng He wrote:
> Hi,
> I installed some rulesets generated by ClassBench (a ruleset benchmark
> for generating ACL and FW rules) in Open vSwitch, and I find many megaflow's
> mask has
> discontiguous bits. Like mask = 0xfc in network byte order. Aft
It is possible for ofpbuf_put() to realloc a newly allocated address,
casuing the previously referenced pointer, ogds, points to old/free'd
address. The issue is generated by forcing ofpbuf_put() to use newly
allocated buffer and valgrind reports invalid write. The similiar syndrome
is reported at:
On Fri, Mar 4, 2016 at 1:49 PM, Flavio Leitner wrote:
> On Thu, 25 Feb 2016 12:15:10 +0530
> bscha...@redhat.com wrote:
>
> > Perviously it was using the platform's runtime directory which can be
> > different from the runtime directory of ovsdb-server started by the
> > openvswitch service
> >
>
On Thu, 25 Feb 2016 12:15:10 +0530
bscha...@redhat.com wrote:
> Perviously it was using the platform's runtime directory which can be
> different from the runtime directory of ovsdb-server started by the
> openvswitch service
>
> Signed-off-by: Babu Shanmugam
> ---
> rhel/.gitignore
> On Mar 4, 2016, at 9:41 AM, Ryan Moats wrote:
>
> Russell did make a similar comment later on in the patch set and I have to
> admit
> that I did it both ways (on one test, I changed the initialization and on the
> other, I just changed the ovn-[nb|sb]ctl commands) as I wasn't sure which
> w
Justin Pettit wrote on 03/04/2016 01:20:52 AM:
> From: Justin Pettit
> To: Russell Bryant
> Cc: Ryan Moats/Omaha/IBM@IBMUS, ovs dev ,
> arnaldimich...@gmail.com
> Date: 03/04/2016 01:21 AM
> Subject: Re: [ovs-dev] [PATCH v9 1/1] Separating OVN NB and SB
> database processes
>
>
> > On Mar 3, 20
There's a lot of work happening to improve ovsdb performance (both on the
client and server sides). There's testing happening in multiple
environments (physical and simulated) in the hundreds-of-hypervisors
range. Interestingly, most of the bottlenecks we're exposing are on the
client side.
We h
Hi Ben,
What's the current thinking around the OVSDB HA and scale solution?
Needless to say, the single SB DB to which all ovn-controllers connect
could be a liability in various production scenarios.
Cheers,
Dan
On Mar 4, 2016 00:48, "Ben Pfaff" wrote:
> Here's my OVN report for the week, sinc
- Original Message -
> From: "Lance Richardson"
> To: dev@openvswitch.org
> Sent: Wednesday, March 2, 2016 4:40:44 PM
> Subject: [ovs-dev] [RFC 0/4] testsuite code coverage usability fixes, lcov
>
> This series addresses usability issues when executing the testsuite
> on builds configured
On 3 March 2016 at 11:08, Darrell Ball wrote:
> Resending, as patches are visible in the mailing list archive, but not
> actually received by all
>
>
>
>
>
>
>
> On 2/12/16, 5:12 PM, "Darrell Ball" wrote:
>
> >Signed-off-by: Darrell Ball
>
Acked-by: Gurucharan Shetty
> >---
> > tests/ovn.at
On 12 February 2016 at 17:12, Darrell Ball wrote:
> Signed-off-by: Darrell Ball
> ---
> AUTHORS| 1 +
> ovn/controller-vtep/vtep.c | 207
> -
> 2 files changed, 186 insertions(+), 22 deletions(-)
>
> diff --git a/AUTHORS b/AUTHOR
On 04.03.2016 13:50, Wang, Zhihong wrote:
>
>
>> -Original Message-
>> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
>> Sent: Friday, March 4, 2016 6:00 PM
>> To: Wang, Zhihong ; dev@openvswitch.org
>> Cc: Flavio Leitner ; Traynor, Kevin
>> ;
>> Dyasly Sergey
>> Subject: Re: vhost
> -Original Message-
> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
> Sent: Friday, March 4, 2016 6:00 PM
> To: Wang, Zhihong ; dev@openvswitch.org
> Cc: Flavio Leitner ; Traynor, Kevin
> ;
> Dyasly Sergey
> Subject: Re: vhost-user invalid txqid cause discard of packets
>
> Hi,
This commit allows for vHost memory from QEMU, DPDK and OVS, as well
as the servicing PMD, to all come from the same socket.
DPDK v2.2 introduces a new configuration option:
CONFIG_RTE_LIBRTE_VHOST_NUMA. If enabled, DPDK detects the socket
from which a vhost device's memory has been allocated by Q
This RFC patch makes use of the CONFIG_RTE_LIBRTE_VHOST_NUMA option
available in DPDK v2.2 that relocates vhost device memory managed by
DPDK to the same NUMA node that QEMU device memory is located on.
In OVS, now the mempool and PMD servicing the port are relocated to this
NUMA node as well. Prev
Hi, Zhihong.
I can't reproduce this in my environment.
Could you please provide ovs-vswithd.log with VLOG_DBG enabled
for netdev-dpdk and outputs of following commands:
# ovs-vsctl show
# ovs-appctl dpctl/show
# ovs-appctl dpif-netdev/pmd-rxq-show
in 'good' and 'bad' states?
Also, are you sure tha
> -Original Message-
> From: Daniele Di Proietto [mailto:diproiet...@vmware.com]
> Sent: Friday, March 04, 2016 12:19 AM
> To: Stokes, Ian
> Cc: dev@openvswitch.org
> Subject: Re: [ovs-dev] [PATCH RFC 1/1] netdev-dpdk.c: Add ingress-
> policing functionality.
>
> Hi Ian,
>
> I haven't loo
On 03/03/2016 08:20 PM, b...@skyportsystems.com wrote:
From: Ben Warren
This patch set is a first stab at making a 'dev' distribution for OpenVswitch
The goal is to provide shared libraries and header files so that third party
applications can link in OVS rather than always using the command li
Maintain the minimal clauses tracked transaction ID.
Run lazy cleanup - while going over clauses tracked rows,
remove all rows that have a lower transaction than the minimal table
transaction ID.
Signed-off-by: Liran Schour
---
ovsdb/monitor.c | 99 +++
Insert changes to condition's clause changes tracking and compose
update notification according to these changes.
Signed-off-by: Liran Schour
---
ovsdb/monitor.c | 125 +++-
1 file changed, 106 insertions(+), 19 deletions(-)
diff --git a/ovsdb
This patch implements a data structure to hold all changed rows that matches any
condition's clause with function "==" in this ovsdb_monitor_table.
This data structure will be used only for sessions that have only
"=="/"includes"
or false function (which is the most common case in conditional moni
Add tests for conditional monitoring to IDL.
Signed-off-by: Liran Schour
---
v3->v4:
* Add test for multiple tables condtional monitoring
v2->v3:
* monitor_cond_update receives a single condition
* Use generated automatic *_add_clause_*() API
* Test multiple clauses
---
tests/ovsdb-idl.at | 1
ovsdb-server now accepts "monitor_cond_update" request. On conditions update
we compose update notification according to the current state of the
database (do not use of cahnges list). We assume that seesion is flushed when
composing update after conditon_update().
Sees ovsdb-server (1) man page fo
Conditional monitor Port_Binding Logical_Flow and Multicast_Group
tables. As a result ovn-controller will write flows only to datapath
that has a port added on this hypervisor.
Signed-off-by: Liran Schour
---
ovn/controller/binding.c| 93 +
ovn/con
This patch enables JSON cache of none conditional monitored tables to reduce
computation on sessions with mixed tables (true and none true conditions).
Signed-off-by: Liran Schour
---
ovsdb/monitor.c | 140 ++--
1 file changed, 86 insertions(+)
Change ovsdb_condition to be a 3-element json array or a boolean value.
Conditions utilities will be used later for conditional monitoring.
Signed-off-by: Liran Schour
---
v4->v5:
* Naming changes according to review
* Bug fix
v3->v4:
* Pass index_map only to ovsdb_condition_evaluate_or_datum()
This patch introduce an optimization for monitored tables with false
session's condition. (which should be common in conditional monitoring).
It returns an immediate NULL when trying to compose update notification
on a table with a false session's condition.
Signed-off-by: Liran Schour
---
ovsdb
Add API that allows the user to create condition iteratively
and send condition_update to the server.
Signed-off-by: Liran Schour
---
v3->v4:
* Accept condition with none single boolean clause
* Add remove_clause API
v2->v3:
* Simplify API to allow iteratively adding clauses to condition
befo
Python idl works now with "monitor_cond" method. Add test
for backward compatibility with old "monitor" method.
Signed-off-by: Liran Schour
---
v4->v5:
* Move back 3rd parameter of notify to be a Row with only updated columns
compliant with recently added unit tests
v3->v4:
* Change 3rd param
This mode indicates that this session has a table with only "=="/"includes"
conditions clauses (might have also a FALSE clause). And we clauses changes
tracking should be enabled.
Signed-off-by: Liran Schour
---
ovsdb/monitor.c | 31 +++
1 file changed, 31 insertions(
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 | 17 +++--
lib/ovsdb-id
The next few patchs implements a data structure to track rows according to
matched
condition's clauses. The goal is to reduce the computation of conditional
monitoring
when all clauses have "==" function (which is the most common case).
This patch adjust ovsdb_monitor_changes to hold rows either
This commit allows to add unmonitored columns to a monitored table
due to condition update.
It will be used to evaluate conditions on unmonitored columns.
Update notification includes only monitored columns.
Due to the limited number of columns We do not remove unused unmonitored
columns due to con
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 true condition only.
"monitor_cond" and "
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 | 29 +++
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 sess
45 matches
Mail list logo
