look good in general. a few small comments in line.
Acked-by: Andy Zhou
On Wed, Sep 17, 2014 at 6:58 PM, Pravin B Shelar wrote:
> OVS keeps pointer to packet key in skb->cb, but the packet key is
> store on stack. This could make code bit tricky. So it is better to
> get rid of the pointer.
>
>
This fixes numerous testsuite failures of the form "SSL_connect:
error:0D0C50A1:asn1 encoding routines:ASN1_item_verify:unknown message
digest algorithm" on systems that disable MD5 in OpenSSL. Centos 7 is one
example. Presumably it increase security as well for anyone who generates
certificates
Thx a lot,
Applied to master!
On Thu, Sep 18, 2014 at 5:21 PM, Daniele Di Proietto wrote:
> Thanks for the fix Alex!
>
> Acked-by: Daniele Di Proietto
>
> On 9/18/14, 5:09 PM, "Alex Wang" wrote:
>
> >Since dpdk_do_tx_copy() will be called by both pmd and
> >non-pmd thread, it should take the
Thanks for the fix Alex!
Acked-by: Daniele Di Proietto
On 9/18/14, 5:09 PM, "Alex Wang" wrote:
>Since dpdk_do_tx_copy() will be called by both pmd and
>non-pmd thread, it should take the queue id as input.
>The current ovs always uses NON_PMD_THREAD_TX_QUEUE
>as queue id, which causes unprotec
Since dpdk_do_tx_copy() will be called by both pmd and
non-pmd thread, it should take the queue id as input.
The current ovs always uses NON_PMD_THREAD_TX_QUEUE
as queue id, which causes unprotected multi-access
to the same queue.
This commit fixes the issue by passing the queue id
to dpdk_do_tx_c
hi Sam,
Pls. find my replies inline.
On Sep 14, 2014, at 9:12 PM, Samuel Ghinet
wrote:
> Hello Nithin,
>
> Thanks a lot again for the review!
Welcome. As you said, the code is intricate. It is fine though. Thanks for
working on this one.
> -
>> More than that I was a
On Wed, Sep 17, 2014 at 9:11 PM, Pravin Shelar wrote:
> On Wed, Sep 17, 2014 at 9:02 PM, Joe Stringer wrote:
>>
>>
>> On 18 September 2014 13:02, Joe Stringer wrote:
> @@ -229,17 +244,19 @@ static bool match_validate(const struct
> sw_flow_match *match,
>
>> Signed-off-by: Samuel Ghinet
>
> This didn't apply for me:
>
>Applying: datapath-windows: fix OVS_VPORT_TYPE
>error: patch failed: datapath-windows/include/OvsDpInterfaceExt.h:80
>error: datapath-windows/include/OvsDpInterfaceExt.h: patch does not apply
>error: patch failed: d
hi Samuel,
Thanks for sending out the changes.
Some high level comments are:
1. You might have to rebase your change when Eitan's change gets checked in. He
has coalesced the declarations of the command handlers.
2. We'll have to get rid of the OvsDumpVportIoctl() sometime. We need not do it
in
Ah, great. I don't think of that as "rate limiting", hence my
confusion.
I'll review the series.
Thank you!
On Thu, Sep 18, 2014 at 04:04:18PM -0700, Alex Wang wrote:
> Hey Ben,
>
> I sent out a new series of patches here:
> http://openvswitch.org/pipermail/dev/2014-September/045949.html
> whi
Hey Ben,
I sent out a new series of patches here:
http://openvswitch.org/pipermail/dev/2014-September/045949.html
which does exactly what you said.
Forgot to mark this thread as invalid.
Thanks,
Alex Wang,
On Thu, Sep 18, 2014 at 4:00 PM, Ben Pfaff wrote:
> On Thu, Sep 18, 2014 at 11:37:08AM
On Wed, Sep 17, 2014 at 01:06:43PM +, Samuel Ghinet wrote:
> The windows ovs kernel uses an OVS_VPORT_TYPE enum that is incompatible with
> the userspace counterpart (enum ovs_vport_type from openvswitch.h). We must
> use
> the same enum type - enum ovs_vport_type - for the netlink communicati
On Thu, Sep 18, 2014 at 11:37:08AM -0700, Alex Wang wrote:
> When ovs is running with large topology (e.g. large number of
> interfaces), the stats and status update to ovsdb become huge and
> normally require multiple run of ovsdb jsonrpc message processing
> loop to consume. Also, this could cau
On Wed, Sep 17, 2014 at 11:12:05PM -0700, Eitan Eliahu wrote:
> This code handles an event notification subscription for a user mode thread
> which joins an MC group. The event wait handler queues an IRP which is
> completed upon change in a port state.
>
> Signed-off-by: Eitan Eliahu
I fixed a
This commit adds a warning message to warn the excessive backlog
for jsonrpc.
Signed-off-by: Alex Wang
---
lib/jsonrpc.c |6 ++
1 file changed, 6 insertions(+)
diff --git a/lib/jsonrpc.c b/lib/jsonrpc.c
index 0841ad8..c0f80bc 100644
--- a/lib/jsonrpc.c
+++ b/lib/jsonrpc.c
@@ -259,6 +259
This commit refactors the stats and status update in bridge_run()
by moving the corresponding code to separate functions. This
makes the code more organized.
Signed-off-by: Alex Wang
---
vswitchd/bridge.c | 211 -
1 file changed, 112 insertio
When ovs is running with large topology (e.g. large number of
interfaces), the stats update to ovsdb becomes huge and normally
requires multiple run of ovsdb jsonrpc message processing loop to
consume.
To prevent the periodic stats update from backlogging in the
jsonrpc sending queue, this commit
When testing ovs with 10K or more tunnels with cfm/bfd enabled,
it can be observed that the memory utilization of ovs-vswitchd
process keeps growing. After investigation, it is found that
the periodic statistics update to ovsdb becomes huge when there
are that many interfaces, and would take long
On Sep 17, 2014, at 6:06 AM, Samuel Ghinet
wrote:
> The windows ovs kernel uses an OVS_VPORT_TYPE enum that is incompatible with
> the userspace counterpart (enum ovs_vport_type from openvswitch.h). We must
> use
> the same enum type - enum ovs_vport_type - for the netlink communication to
>
On Thu, Sep 18, 2014 at 2:48 PM, Ben Pfaff wrote:
> On Thu, Sep 18, 2014 at 05:24:12PM -0300, Flavio Leitner wrote:
>> On Thu, Sep 18, 2014 at 01:18:57PM -0700, Ethan Jackson wrote:
>> > In our experience, most people actually expect this number to be
>> > fairly high. For example, our traffic ge
On Thu, Sep 18, 2014 at 05:24:12PM -0300, Flavio Leitner wrote:
> On Thu, Sep 18, 2014 at 01:18:57PM -0700, Ethan Jackson wrote:
> > In our experience, most people actually expect this number to be
> > fairly high. For example, our traffic generator runs a little bit of
> > traffic to populate cac
As I said before, I would enable eviction if it is configured through
the database or through the OpenFlow connection. This does not mean
that enabling eviction through OpenFlow would modify the database.
I don't think you are looking at the current codebase, on the master
branch. The current co
Hi,
I am getting packets dropped when going out of an OVS bridge. The TX dropped is
increased in ovs-dpctl but not at the physical interface level (ifconfig).
Can anyone tell me how to enable some OVS debugging so I can find the reason to
fix that problem?
thanks.
sam
On Thu, Sep 18, 2014 at 02:17:47PM +0530, Neha Joshi wrote:
> I want to check statistics of openvswitch in terms of CPU, load average and
> memory.
> I found that i need to set *other_config:enable-statistics* to *true*
> Can anyone help, how to do this? I mean is there any command or such?
ovs-v
On Fri, Sep 19, 2014 at 12:24:38AM +0900, YAMAMOTO Takashi wrote:
> It seems that the behaviour is not so intuitive.
> cf. https://bugs.launchpad.net/neutron/+bug/1346861
>
> Signed-off-by: YAMAMOTO Takashi
I am not sure that I understand the bug report there. It might be
reporting an actual bu
On Sep 18, 2014, at 4:35 AM, Samuel Ghinet
wrote:
> Hi Nithin,
>
> Good catch!
>
> I would normally prefer the validation in a function to be done like:
> ret = TRUE;
>
> if (fail_cond_1) {
> log();
> ret = FALSE;
> goto done;
> }
>
> if (fail_cond_2) {
> log();
> ret = FALSE;
> goto
On Thu, Sep 18, 2014 at 01:20:30PM -0700, Alex Wang wrote:
> Yes, that is correct,
Perfect, thanks!
fbl
>
> Thanks,
>
> On Thu, Sep 18, 2014 at 1:15 PM, Flavio Leitner wrote:
>
> > On Thu, Sep 18, 2014 at 01:10:00PM -0700, Alex Wang wrote:
> > > On Thu, Sep 18, 2014 at 1:07 PM, Flavio Leitner
On Thu, Sep 18, 2014 at 01:18:57PM -0700, Ethan Jackson wrote:
> In our experience, most people actually expect this number to be
> fairly high. For example, our traffic generator runs a little bit of
> traffic to populate caches, pauses for a couple of seconds, before
> finally starting the main
Yes, that is correct,
Thanks,
On Thu, Sep 18, 2014 at 1:15 PM, Flavio Leitner wrote:
> On Thu, Sep 18, 2014 at 01:10:00PM -0700, Alex Wang wrote:
> > On Thu, Sep 18, 2014 at 1:07 PM, Flavio Leitner wrote:
> >
> > > On Thu, Sep 18, 2014 at 11:37:08AM -0700, Alex Wang wrote:
> > > > When ovs is
In our experience, most people actually expect this number to be
fairly high. For example, our traffic generator runs a little bit of
traffic to populate caches, pauses for a couple of seconds, before
finally starting the main test. On top of that, we've had some minor
complaints from people mana
On Thu, Sep 18, 2014 at 01:10:00PM -0700, Alex Wang wrote:
> On Thu, Sep 18, 2014 at 1:07 PM, Flavio Leitner wrote:
>
> > On Thu, Sep 18, 2014 at 11:37:08AM -0700, Alex Wang wrote:
> > > When ovs is running with large topology (e.g. large number of
> > > interfaces), the stats and status update t
On Thu, Sep 18, 2014 at 1:07 PM, Flavio Leitner wrote:
> On Thu, Sep 18, 2014 at 11:37:08AM -0700, Alex Wang wrote:
> > When ovs is running with large topology (e.g. large number of
> > interfaces), the stats and status update to ovsdb become huge and
> > normally require multiple run of ovsdb js
On Thu, Sep 18, 2014 at 11:37:08AM -0700, Alex Wang wrote:
> When ovs is running with large topology (e.g. large number of
> interfaces), the stats and status update to ovsdb become huge and
> normally require multiple run of ovsdb jsonrpc message processing
> loop to consume. Also, this could cau
Original New Cisco Switches,Modules and Firewalls available for sale now.
Hello dear,
Good day !
Following is Original New Cisco Switches,Modules and Firewalls available for
sale now.
Any needs of them, please feel free to contact me.
Skype: elaine.hu87
Part Number Unit Price Stock Qty
WS-
When ovs is running with large topology (e.g. large number of
interfaces), the stats and status update to ovsdb become huge and
normally require multiple run of ovsdb jsonrpc message processing
loop to consume. Also, this could cause jsonrpc messages backlogged
in ovs.
This commit adds a warning
On Thu, Sep 18, 2014 at 1:31 AM, Nicolas Dichtel
wrote:
> From: Samuel Gauthier
>
> Since commit fb5d1e9e127a ("openvswitch: Build flow cmd netlink reply only if
> needed."),
> the new flows are not notified to the listeners of OVS_FLOW_MCGROUP.
>
> This commit fixes the problem by using the gen
On Thu, Sep 18, 2014 at 1:31 AM, Nicolas Dichtel
wrote:
> This function is the counterpart of the function netlink_has_listeners().
>
> Signed-off-by: Nicolas Dichtel
Looks good.
Acked-by: Pravin B Shelar
> ---
>
> v2: add patch 1/2
>
> include/net/genetlink.h | 8
> 1 file changed,
It seems that the behaviour is not so intuitive.
cf. https://bugs.launchpad.net/neutron/+bug/1346861
Signed-off-by: YAMAMOTO Takashi
---
FAQ | 21 +
1 file changed, 21 insertions(+)
diff --git a/FAQ b/FAQ
index df5ac0e..d9ccc6e 100644
--- a/FAQ
+++ b/FAQ
@@ -731,6 +731,27 @@
On Fri, Sep 12, 2014 at 10:03:56AM +1200, Joe Stringer wrote:
> The datapath max_idle value determines how long to wait before deleting
> an idle datapath flow when operating below the flow_limit. This patch
> increases the max_idle to 10 seconds, which allows datapath flows to be
> remain cached e
Thanks for the suggestion and we have done more investigation on code.
We are thinking of implementing eviction on the basis of importance as
done in current ovs code for lifetime by following the same algorithm as
mentioned in conf.db(5)(Page 33).
As suggested by you we can enable eviction by
Port datapath to work with kernrels up to 3.17 and use 3.16.2 as
the new kernel for CI testing.
Tested with 3.14, 3.16.2, and net-next (3.17).
Signed-off-by: Thomas Graf
Co-authored-by: Madhu Challa
---
v2:
- Swapped ignore_df local_df compat direction
- Provide alloc_netdev() compat macro to
Hi Sam, all we need to do is to protect critical variable in the Cleanup
handler. In this case specifically we can store the dump state variables in the
per process instance (rather than allocating memory) so it will avoid the
possible race condition in freeing dynamic memory.
Everything else wi
Thanks Eitan for clarifications!
I have checked the userspace code, there is no call for DuplicateHandle.
I am not sure whether we should put checks in kernel code now for a possible
future usage of this, or simply leave it to be handled later - if / when such a
case will appear.
Sam
_
Looks good to me!
Acked-by: Samuel Ghinet
Date: Wed, 17 Sep 2014 23:12:05 -0700
From: Eitan Eliahu
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH] datapath-windows: NetLink kernel side,
Event subscription and notification
Message-ID: <141102072
Hey Nithin,
I think this could work refactored a bit (such as to better separate commands),
but we should postpone refactor for a later time.
Acked-by: Samuel Ghinet
Date: Tue, 16 Sep 2014 19:06:11 -0700
From: Nithin Raju
To: dev@openvswitch.org, sghi.
Acked-by: Samuel Ghinet
Date: Tue, 16 Sep 2014 19:06:10 -0700
From: Nithin Raju
To: dev@openvswitch.org, sghi...@cloudbasesolutions.com,
elia...@vmware.com, ankursha...@vmware.com
Subject: [ovs-dev] [PATCH 4/5 v1] extract-odp-netlink-windows-dp-h:
Hi Nithin,
Good catch!
I would normally prefer the validation in a function to be done like:
ret = TRUE;
if (fail_cond_1) {
log();
ret = FALSE;
goto done;
}
if (fail_cond_2) {
log();
ret = FALSE;
goto done;
}
done:
return ret;
It would be a bit more text, but I think it improves c
No Sam, Duplication of handles is done through the calling DuplicateHandle(().
I am not sure if OVS UM calls it. The general case if that one handle
corresponds to a single file object.
Thanks,
Eitan
From: Samuel Ghinet [mailto:sghi...@cloudbasesolutions.com]
Sent: Thursday, September 18, 2014 4
Hi Eitan,
By duplication of HANDLES, you mean assignment (like, "HANDLE a =
CreateFile(params); HANDLE b = a;")?
We do have cases in the vswitchd where we keep multiple duplicates / instances
of the same HANDLE, or where we create children processes that may inherit the
parent handle?
Thanks,
Hi Sam,
As far as I understand when a process calls CreatFile the system will create a
file handle and return it to the process. The file handle (in UM) is associated
with a file object in (KM). Both objects are valid until the process closes the
handle or terminates.
This means that if the devi
I meant to send it to ML.
From: Samuel Ghinet
Sent: Wednesday, September 17, 2014 7:46 PM
To: Eitan Eliahu
Subject: RE: [PATCH v2] datapath-windows: Netlink command: vport dump
Hi Eitan,
Yes, you're right, I forgot to check the return values from the NlMsg
Eitan,
> In case that a handle is duplicated in user mode there will be multiple file
> handles per file object associated with the handle returned by the system
> when the device was opened.
As I remember when working with userspace HANDLEs and FILE_OBJECT ptrs in
kernel, even though only one
Hii all,
I want to check statistics of openvswitch in terms of CPU, load average and
memory.
I found that i need to set *other_config:enable-statistics* to *true*
Can anyone help, how to do this? I mean is there any command or such?
Thanks,
neha
___
de
This function is the counterpart of the function netlink_has_listeners().
Signed-off-by: Nicolas Dichtel
---
v2: add patch 1/2
include/net/genetlink.h | 8
1 file changed, 8 insertions(+)
diff --git a/include/net/genetlink.h b/include/net/genetlink.h
index 93695f0e22a5..af10c2cf8a1d
From: Samuel Gauthier
Since commit fb5d1e9e127a ("openvswitch: Build flow cmd netlink reply only if
needed."),
the new flows are not notified to the listeners of OVS_FLOW_MCGROUP.
This commit fixes the problem by using the genl function, ie
genl_has_listerners() instead of netlink_has_listeners
55 matches
Mail list logo
