On Thu, Sep 29, 2016 at 1:09 AM, Simon Horman
wrote:
> Hi Or,
>
> On Wed, Sep 28, 2016 at 04:54:40PM +0300, Or Gerlitz wrote:
>> On Wed, Sep 28, 2016 at 3:42 PM, Simon Horman
>> wrote:
>>
>> > A different approach, not implemented by this patch-set, is for user-space
>> > to program flows into ha
Also, add the keyword bundle_action to all the tests in bundle.at,
distinguishing it from OF bundles.
It came to my attention recently that bundle_load will load 0x in case all
the slaves are down, as bundle_execute will return OFPP_NONE.
As I noticed this was not explicitly tested, not even
On Fri, Sep 30, 2016 at 03:31:03PM -0300, Thadeu Lima de Souza Cascardo wrote:
> Some of these references are not valid anymore, as things were moved to
> either meta-flow.h or ofp-actions.c.
>
> Signed-off-by: Thadeu Lima de Souza Cascardo
Thanks, applied to master.
On Fri, Sep 30, 2016 at 1:29 AM, Ben Pfaff wrote:
> What a nice introduction!
>
Agreed!
> Russell, do you want to review this?
Yes. The patch looks good. I trimmed the introduction from the commit
message, added Gabe to the AUTHORS file, and applied this to master and
branch-2.6.
Thanks!
Some of these references are not valid anymore, as things were moved to
either meta-flow.h or ofp-actions.c.
Signed-off-by: Thadeu Lima de Souza Cascardo
---
lib/bundle.h | 2 +-
lib/learn.h | 2 +-
lib/meta-flow.c | 2 +-
tests/learn.at | 2 +-
tutorial/Tutorial.md |
Applied to master, thanks
2016-09-30 0:48 GMT-07:00 Ilya Maximets :
> It is not necessary to touch the physical device each time, if the
> configuration has not been changed. Also, few style issues fixed.
>
> Thread-safety annotation added to 'dpdk_set_rxq_config()'. It was
> missed while previou
Applied to master, thanks
2016-09-30 6:50 GMT-07:00 Ciara Loftus :
> The 'options:n_rxq_desc' and 'n_txq_desc' fields allow the number of rx
> and tx descriptors for dpdk ports to be modified. By default the values
> are set to 2048, but can be modified to an integer between 1 and 4096
> that is
Petr,
While deleting all the rules before deleting al the meters is more efficient, I
think we should make the code more future proof by making sure we never delete
a meter without deleting the rules referring to it.
Would you try the following patch to see if it fixes the problem you face:
When deleting a bridge it is currently possible to delete a mater
without deleting the rules using the meter first. Fix this by moving
the meter's rule deletion to meter_delete().
Signed-off-by: Jarno Rajahalme
Reported-by: Petr Machata
---
ofproto/ofproto.c | 35 +++---
This fixes test failures when the tests are run under a shell like ash or
dash.
Signed-off-by: Ben Pfaff
---
tests/ovn.at | 62 ++--
1 file changed, 31 insertions(+), 31 deletions(-)
diff --git a/tests/ovn.at b/tests/ovn.at
index f40940f..
Thanks for your tips, I fixed the memory leak(via free(ips)).
The patch follows the style of the original code. So the 'const' is not used
with
'may_exist', 'add_duplicate' and 'must_exist'. You can find that in files, such
as
'utilities/ovs-vsctl.c', 'ovn/utilities/ovn-sbctl.c' and
'ovn/utiliti
As reported in [1], the ovs python idl mutations cause empty
column data to be returned by getattr after a mutation is
done the column. This can be recreated via the following
steps:
# Good column values
column_values = getattr(row, column, [])
# New value added via mutation
row.addvalue(column, n
On Thu, Sep 29, 2016 at 05:56:00PM -0400, Flavio Fernandes wrote:
>
> > On Sep 29, 2016, at 5:41 PM, Ben Pfaff wrote:
> >
> > Reported-by: Stuart Cardall
> > Reported-at:
> > http://openvswitch.org/pipermail/discuss/2016-September/022803.html
> > Signed-off-by: Ben Pfaff
>
>
> Acked-by: Fla
On Fri, Sep 30, 2016 at 04:45:22PM +0500, Valentine Sinitsyn wrote:
> Hi Ben,
>
> On 30.09.2016 02:55, Ben Pfaff wrote:
> >On Thu, Sep 29, 2016 at 12:15:36PM -0700, Han Zhou wrote:
> >>On Thu, Sep 29, 2016 at 11:31 AM, Ben Pfaff wrote:
> >>>
> >>>When a VM sends an ARP or an ND NS for its own IP
Hi all,
On 23.09.2016 17:59, Valentine Sinitsyn wrote:
Hi,
Suppose we want to collect per-port IPFIX statistics in an OVN-managed
virtual network.
Looks like the most straightforward way to do this is to add a
Flow_Sample_Collector_Set entry and a flow with sample() action to match
the given p
On Fri, Sep 30, 2016 at 10:00:34AM +0300, Ilya Maximets wrote:
> People in OVS again and again tries to break log messages in parts.
> This breaks ability to find this lines in code while debugging.
> It's required to document this thing directly in CodingStyle.md to
> avoid such situations in the
The 'options:n_rxq_desc' and 'n_txq_desc' fields allow the number of rx
and tx descriptors for dpdk ports to be modified. By default the values
are set to 2048, but can be modified to an integer between 1 and 4096
that is a power of two. The values can be modified at runtime, however
require the NI
Regards
_Sugesh
> -Original Message-
> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
> Sent: Friday, September 30, 2016 8:49 AM
> To: dev@openvswitch.org; Daniele Di Proietto ;
> Chandran, Sugesh
> Cc: Dyasly Sergey ; Heetae Ahn
> ; Bodireddy, Bhanuprakash
> ; Loftus, Ciara
> ; I
> The 'options:n_rxq_desc' and 'n_txq_desc' fields allow the number of rx
> and tx descriptors for dpdk ports to be modified. By default the values
> are set to 2048, but can be modified to an integer between 1 and 4096
> that is a power of two. The values can be modified at runtime, however
> requ
On 30.09.2016 16:36, Aaron Conole wrote:
> Ilya Maximets writes:
>
>> Double word removed. Error message rewritten to avoid breaking of the
>> user-visible string.
>>
>> CC: Mark Kavanagh
>> Fixes: 58be5c0eecf3 ("netdev-dpdk: Fix coding style")
>> Signed-off-by: Ilya Maximets
>> ---
>> lib/net
Ilya Maximets writes:
> Double word removed. Error message rewritten to avoid breaking of the
> user-visible string.
>
> CC: Mark Kavanagh
> Fixes: 58be5c0eecf3 ("netdev-dpdk: Fix coding style")
> Signed-off-by: Ilya Maximets
> ---
> lib/netdev-dpdk.c | 7 ---
> 1 file changed, 4 insertion
Ich brauche Ihre Unterstützung in Ihrem Land zu investieren, mit mir bitte für
weitere Informationen zu kontaktieren !!!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Hi Ben,
On 30.09.2016 02:55, Ben Pfaff wrote:
On Thu, Sep 29, 2016 at 12:15:36PM -0700, Han Zhou wrote:
On Thu, Sep 29, 2016 at 11:31 AM, Ben Pfaff wrote:
When a VM sends an ARP or an ND NS for its own IP address, it is trying to
check for a duplicate address in the network. OVN needs to su
>
> 2016-09-29 3:28 GMT-07:00 Loftus, Ciara :
> >
> > Hi Ciara,
> > thanks for the patch, it looks good to me.
> > I only have a minor comment:
> > I'd like the requested values to depend only on the current database
> > state. With the current patch when a value is invalid (not pow2 or bigger
>
The 'options:n_rxq_desc' and 'n_txq_desc' fields allow the number of rx
and tx descriptors for dpdk ports to be modified. By default the values
are set to 2048, but can be modified to an integer between 1 and 4096
that is a power of two. The values can be modified at runtime, however
require the NI
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
>
>On 30.09.2016 11:49, Kavanagh, Mark B wrote:
>>>
>>> People in OVS again and again tries to break log messages in parts.
>>> This breaks ability to find this lines in code while debugging.
>>
>> should be 'these lines', or 'this line'
>
>Oh, I meant 'these lines'. Thanks for catching this typo
On 30.09.2016 11:49, Kavanagh, Mark B wrote:
>>
>> People in OVS again and again tries to break log messages in parts.
>> This breaks ability to find this lines in code while debugging.
>
> should be 'these lines', or 'this line'
Oh, I meant 'these lines'. Thanks for catching this typo.
I thin
>
>People in OVS again and again tries to break log messages in parts.
>This breaks ability to find this lines in code while debugging.
should be 'these lines', or 'this line'
Other than that, LGTM - thanks for making this explicit.
Acked-by: Mark Kavanagh
>It's required to document this t
>
>Double word removed. Error message rewritten to avoid breaking of the
>user-visible string.
>
>CC: Mark Kavanagh
>Fixes: 58be5c0eecf3 ("netdev-dpdk: Fix coding style")
>Signed-off-by: Ilya Maximets
>---
> lib/netdev-dpdk.c | 7 ---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
>diff -
It is not necessary to touch the physical device each time, if the
configuration has not been changed. Also, few style issues fixed.
Thread-safety annotation added to 'dpdk_set_rxq_config()'. It was
missed while previous refactoring of the flow control configuration.
Signed-off-by: Ilya Maximets
Hi,
Currently ingress and egress policing was supported on dpdk ports.
Is there a plan to support egress shaping by creating user-space queues ?
Or currently is there a way to do egress shaping on dpdk netdev interfaces
?
It seems egress shaping does not work on tap/internal-ports/patch port
On 29.09.2016 02:52, Daniele Di Proietto wrote:
> Why are the variables uint8_t instead of bool?
>
> I think we shouldn't assume that converting to bool always returns 0 or 1,
> but the return value of smap_get_bool() is always 0 or 1 (as we always go
> through ! or !=). I would remove the ternar
Double word removed. Error message rewritten to avoid breaking of the
user-visible string.
CC: Mark Kavanagh
Fixes: 58be5c0eecf3 ("netdev-dpdk: Fix coding style")
Signed-off-by: Ilya Maximets
---
lib/netdev-dpdk.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/lib/ne
Ilya Maximets (2):
CodingStyle.md: Clarify limit for line length.
netdev-dpdk: Fix issues introduced by recent style fix.
CodingStyle.md| 6 +-
lib/netdev-dpdk.c | 7 ---
2 files changed, 9 insertions(+), 4 deletions(-)
--
2.7.4
___
d
People in OVS again and again tries to break log messages in parts.
This breaks ability to find this lines in code while debugging.
It's required to document this thing directly in CodingStyle.md to
avoid such situations in the future.
Text actually ported from the Linux kernel coding style.
Sign
36 matches
Mail list logo
