The original message was received at Thu, 17 Jul 2014 13:48:09 +0700
from voipon.co [74.127.153.167]
- The following addresses had permanent fatal errors -
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Thanks for the review. Both points are valid.
I have just posted another version of the re-factoring that should address
the issues.
On Wed, Jul 16, 2014 at 9:30 PM, Pravin Shelar wrote:
> On Tue, Jul 15, 2014 at 2:56 PM, Andy Zhou wrote:
>> The added get_dp_rcu() requires its caller to hold rc
Thx for the reviews, Ben and Joe,
Applied the patch to master, branch-2.[0,1,3]
Alex Wang,
On Wed, Jul 16, 2014 at 2:43 AM, Joe Stringer
wrote:
> I took another read through, minor nitpicks if you haven't pushed the
> patch yet:-
>
>
> On 16 July 2014 17:21, Alex Wang wrote:
>
>> + * FWD:
Remove get_dp() API, it becomes get_dp__() which is for internal
use only. Added the following APIs to make it clear of the locking
requirement.
o get_dp_rcu() requires its caller to hold rcu lock.
o get_dp_ovsl() requires its caller to hold ovs lock.
o get_dp_ovsl_rcu() requires its c
From: Stephen Hemminger
Date: Wed, 16 Jul 2014 11:25:52 -0700
> Generic netlink tables can be const.
>
> Signed-off-by: Stephen Hemminger
Applied, thanks Stephen.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 07/16/2014 08:04 PM, Ben Pfaff wrote:
> On Wed, Jul 16, 2014 at 02:53:37PM -0300, Flavio Leitner wrote:
>> On Wed, Jul 16, 2014 at 09:56:20AM -0700, Ben Pfaff wrote:
>>> On Wed, Jul 16, 2014 at 10:39:17AM -0300, Flavio Leitner wrote:
>>> There's more than one way to chroot. Maybe Eric is thinki
Hi All,
Does the openvswitch-1.10 supports IPsec+GRE and IPsec+VXLAN tunneling
protocols?
If not please point me to any of the patches for this functionality.
Does latest version of openvswitch 2.0 supports these tunneling protocols?
--
Thanks & Regards
Neelakantam Gaddam
On 12 July 2014 00:55, Thomas Graf wrote:
> On 07/11/14 at 11:29pm, Joe Stringer wrote:
> > I'm skeptical of taking the ovs_lock(). The current patch performs this
> > key/mask cache construction inside ovs_lock() as part of the critical
> > section for flow install. If we perform this during flo
The original message was received at Thu, 17 Jul 2014 10:29:58 +0530
from 73.28.108.230
- The following addresses had permanent fatal errors -
- Transcript of session follows -
... while talking to 217.137.70.153:
550 5.1.2 ... Host unknown (Name server: host not found)
On Tue, Jul 15, 2014 at 2:56 PM, Andy Zhou wrote:
> The added get_dp_rcu() requires its caller to hold rcu lock.
> get_dp() now requires caller to hold ovs lock.
>
> Sepearate APIs makes it clear on the intended use and the requirements
> of its caller.
>
> Signed-off-by: Andy Zhou
> ---
> datap
The following changes were made:
- Since we have two dpdk classes, we should split the initial operations needed
by both classes from the initialization needed by each class.
- The dpdk_ring class does not need an initialization function: it has been
removed. This also prevents many testcase f
I’ve decided to make more changes to netdev-dpdk.
I’m posting another patch with commit message "netdev-dpdk: Refactor
dpdk_class_init()”
On Jul 16, 2014, at 3:54 PM, Daniele Di Proietto wrote:
>
> On Jul 16, 2014, at 3:30 PM, Pravin Shelar wrote:
>
>> On Wed, Jul 16, 2014 at 2:47 PM, Daniel
On Wed, Jul 16, 2014 at 01:00:44PM +0200, Attila Czébán wrote:
> Dear Jean,
>
> I read the following documentation about the rate limiter API.
> http://openvswitch.org/pipermail/dev/2011-January/005639.html
> I have some questions about it.
>
> We have an SDN testbed at Budapest University of Tec
On Wed, Jul 16, 2014 at 11:25 AM, Stephen Hemminger
wrote:
> Generic netlink tables can be const.
>
OVS external module can not use const due to compatibility code. But
there is no reason we can not do it for in-kernel module.
Thanks.
> Signed-off-by: Stephen Hemminger
>
>
> --- a/net/openvswi
On Wed, Jul 16, 2014 at 3:05 PM, Ben Pfaff wrote:
> The only client for ofproto_controller_info was transforming the array of
> pairs into an smap anyway. It's easy for the code that fills in the array
> to generate it as an smap directly, and it's also easier to extend later.
>
> Signed-off-by:
On Wed, Jul 16, 2014 at 4:16 PM, Tom Gundersen wrote:
> On Thu, Jul 17, 2014 at 1:00 AM, Pravin Shelar wrote:
>> On Wed, Jul 16, 2014 at 3:43 PM, Tom Gundersen wrote:
>>> Signed-off-by: Tom Gundersen
>>> Cc: Pravin Shelar
>>> Cc: dev@openvswitch.org
>>> ---
>>>
>>> This patch goes on top of ne
On Thu, Jul 17, 2014 at 1:00 AM, Pravin Shelar wrote:
> On Wed, Jul 16, 2014 at 3:43 PM, Tom Gundersen wrote:
>> Signed-off-by: Tom Gundersen
>> Cc: Pravin Shelar
>> Cc: dev@openvswitch.org
>> ---
>>
>> This patch goes on top of net-next.
>>
>> net/openvswitch/datapath.c | 1 +
>> ne
On 17 July 2014 04:48, Ben Pfaff wrote:
> On Wed, Jul 16, 2014 at 02:50:12PM +1200, Joe Stringer wrote:
> > Converting the flow key and mask back into netlink format during flow
> > dump is fairly expensive. By caching the userspace-provided versions of
> > these during flow setup, and copying th
On Wed, Jul 16, 2014 at 3:43 PM, Tom Gundersen wrote:
> Signed-off-by: Tom Gundersen
> Cc: Pravin Shelar
> Cc: dev@openvswitch.org
> ---
>
> This patch goes on top of net-next.
>
> net/openvswitch/datapath.c | 1 +
> net/openvswitch/vport-internal_dev.c | 2 +-
> net/openvswitch/vport
On Wed, Jul 16, 2014 at 3:05 PM, Ben Pfaff wrote:
> 'n_txq' was initialized to 0 and never modified, so pinsched_count_txqlen()
> always returned 0. Instead, return the correct number.
>
> This only affected the results of "ovs-appctl memory/show", and only if
> controller rate limiting was turne
On Jul 16, 2014, at 3:30 PM, Pravin Shelar wrote:
> On Wed, Jul 16, 2014 at 2:47 PM, Daniele Di Proietto
> wrote:
>> If vswitchd is compiled with DPDK support but started without --dpdk
>> argument,
>> it shouldn't print any dpdk initialization messages.
>> OVS_VSWITCHD_START (tests/ofproto-ma
If the userspace wants to match on a flow with some tunnel attributesset to 0,
it simply omits them in the netlink attributes stream.
Since our wildcarding logic (when megaflows are disabled) is based on the
attributes in the netlink stream, we set our mask incorrectly.
This commit adds a check to
Signed-off-by: Tom Gundersen
Cc: Pravin Shelar
Cc: dev@openvswitch.org
---
This patch goes on top of net-next.
net/openvswitch/datapath.c | 1 +
net/openvswitch/vport-internal_dev.c | 2 +-
net/openvswitch/vport.h | 2 ++
3 files changed, 4 insertions(+), 1 deletion(-)
On Wed, Jul 16, 2014 at 3:05 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Acked-by: Gurucharan Shetty
> ---
> vswitchd/vswitch.xml |2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
> index 69b3567..ee2892b 100644
> --- a/v
On Wed, Jul 16, 2014 at 2:47 PM, Daniele Di Proietto
wrote:
> If vswitchd is compiled with DPDK support but started without --dpdk argument,
> it shouldn't print any dpdk initialization messages.
> OVS_VSWITCHD_START (tests/ofproto-macros.at:54) is pretty strict about what
> the
> daemon prints w
Signed-off-by: Ben Pfaff
---
NEWS |1 +
ofproto/connmgr.c| 26 +++-
ofproto/pinsched.c | 17 +---
ofproto/pinsched.h | 11 -
vswitchd/vswitch.xml | 112 +++---
5 files changed, 135 insertions(+), 32 del
An upcoming commit will add a caller that needs to format both the key and
the value. That isn't cleanly possible with the current interface.
Signed-off-by: Ben Pfaff
---
lib/smap.c | 11 ++-
lib/smap.h |3 ++-
2 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/lib/sm
The only client for ofproto_controller_info was transforming the array of
pairs into an smap anyway. It's easy for the code that fills in the array
to generate it as an smap directly, and it's also easier to extend later.
Signed-off-by: Ben Pfaff
---
ofproto/connmgr.c | 27 ++-
I received a request to report statistics on Open vSwitch rate-limiting of
packet-ins to OpenFlow controllers. This series implements that feature.
Ben Pfaff (5):
vswitch.xml: Fix typo in documentation.
pinsched: Report queued packet count correctly.
Simplify ofproto_controller_info by usin
'n_txq' was initialized to 0 and never modified, so pinsched_count_txqlen()
always returned 0. Instead, return the correct number.
This only affected the results of "ovs-appctl memory/show", and only if
controller rate limiting was turned on, so it is not a serious bug.
Signed-off-by: Ben Pfaff
Signed-off-by: Ben Pfaff
---
vswitchd/vswitch.xml |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
index 69b3567..ee2892b 100644
--- a/vswitchd/vswitch.xml
+++ b/vswitchd/vswitch.xml
@@ -3324,7 +3324,7 @@
Equivalent to o
The controller setup in my personal test environment has been broken for a
while. I figured that it wasn't anything important, though, because no one
else had reported similar problems. Anyway, it turns out that enabling
OpenFlow 1.3 by default broke test-controller because OpenFlow 1.3 doesn't
s
If vswitchd is compiled with DPDK support but started without --dpdk argument,
it shouldn't print any dpdk initialization messages.
OVS_VSWITCHD_START (tests/ofproto-macros.at:54) is pretty strict about what the
daemon prints when starting.
The testcases failures were introduced with commit 95fb79
Generic netlink tables can be const.
Signed-off-by: Stephen Hemminger
--- a/net/openvswitch/datapath.c2014-07-09 03:06:11.400291765 -0700
+++ b/net/openvswitch/datapath.c2014-07-09 03:09:35.688297236 -0700
@@ -66,16 +66,16 @@ static struct genl_family dp_packet_genl
static stru
On Wed, Jul 16, 2014 at 10:46:20AM -0700, Daniele Di Proietto wrote:
> cmap_next_position() didn't update the node pointer while iterating through a
> list of nodes with the same hash.
> This commit fixes the bug and improve test-cmap to detect it.
>
> Signed-off-by: Daniele Di Proietto
Applied,
On Wed, Jul 16, 2014 at 02:53:37PM -0300, Flavio Leitner wrote:
> On Wed, Jul 16, 2014 at 09:56:20AM -0700, Ben Pfaff wrote:
> > On Wed, Jul 16, 2014 at 10:39:17AM -0300, Flavio Leitner wrote:
> > There's more than one way to chroot. Maybe Eric is thinking of a
> > model where one chroots to an em
cmap_next_position() didn't update the node pointer while iterating through a
list of nodes with the same hash.
This commit fixes the bug and improve test-cmap to detect it.
Signed-off-by: Daniele Di Proietto
---
lib/cmap.c| 2 +-
tests/test-cmap.c | 20 +++-
2 files cha
On Wed, Jul 16, 2014 at 09:56:20AM -0700, Ben Pfaff wrote:
> On Wed, Jul 16, 2014 at 10:39:17AM -0300, Flavio Leitner wrote:
> > The main_loop refactoring is a nice thing even without the series, I find
> > the code easier to understand.
>
> That's reasonable, so I applied that patch.
Thanks!
>
I’m running from the test suite, with GCC 4.9
It was slow because on my dev machine I compile without optimization (my bad).
With optimizations:
- without testing cmap_next_position() : 8.6s
- if we always test also cmap_next_position() always (like you suggested): 18.8s
So I’m posting v2 with
There's a couple of reasons that test-cmap can be slow.
First, if you run it without a numerical argument, it iterates 100
times. That takes a long time, and it's only meant for cases where
you're doing actual development and want high confidence. The
testsuite gives it an argument of 1, so that
Thanks,
The reason I didn’t do that in the first place is that test-cmap is already
taking 55s to run on my machine, and I thought that adding other code to
check_cmap()
might increase runtime significantly (we call check_cmap() a lot and
cmap_next_position()
is slow).
After applying your sugge
git am warning:
Applying: Extend OVS IPFIX exporter to export tunnel headers
/home/pravin/ovs/w8/.git/rebase-apply/patch:143: trailing whitespace.
if (out_tun_key)
/home/pravin/ovs/w8/.git/rebase-apply/patch:236: space before tab in indent.
const struct udphdr * udp_hdr = tph;
/home
On Wed, Jul 16, 2014 at 10:39:17AM -0300, Flavio Leitner wrote:
> The main_loop refactoring is a nice thing even without the series, I find
> the code easier to understand.
That's reasonable, so I applied that patch.
> I don't think there is a need to pre-open files in /dev since they usually
> a
On Wed, Jul 16, 2014 at 05:14:27PM +1200, Joe Stringer wrote:
> On 16 July 2014 05:58, Ben Pfaff wrote:
>
> > Commit 0791315e4d (netlink-socket: Work around kernel Netlink dump thread
> > races.) introduced a simple workaround for Linux kernel races in Netlink
> > dumps. However, the code remain
On Wed, Jul 16, 2014 at 02:50:12PM +1200, Joe Stringer wrote:
> Converting the flow key and mask back into netlink format during flow
> dump is fairly expensive. By caching the userspace-provided versions of
> these during flow setup, and copying the memory directly during flow
> dump, we are able
On Fri, Jul 11, 2014 at 5:37 AM, maryam.tahhan wrote:
> Shared memory ring patch
>
> This patch enables the client dpdk rings within the netdev-dpdk. It adds
> a new dpdk device called dpdkr (other naming suggestions?). This allows
> for the use of shared memory to communicate with other dpdk ap
On Wed, Jul 16, 2014 at 12:15:40PM +1200, Joe Stringer wrote:
> Side comment: OVS_KEY_ATTR_ETHERTYPE appears twice in this list. I guess
> it's not harmful to reserve a bit of extra space, but does anyone know
> why it's there twice?
It's because OVS_KEY_ATTR_ETHERTYPE appears twice if there's a V
On Wed, Jul 16, 2014 at 05:12:58PM +1200, Joe Stringer wrote:
> On 16 July 2014 05:55, Ben Pfaff wrote:
> >
> > The kernel shouldn't ever return EAGAIN if there's more data to dump.
> > When dumping is single-threaded (as in branch-1.11) or serialized (as
> > post-commit 0791315e4d), the kernel im
On Wed, Jul 16, 2014 at 02:18:30PM +0900, YAMAMOTO Takashi wrote:
> > /* A flow classifier. */
> > struct classifier {
> > -struct cls_classifier *cls;
> > +struct ovs_mutex mutex;
> > +int n_rules OVS_GUARDED;/* Total number of rules. */
> > +uint8_t n_flow_segments;
> >
On Tue, Jul 15, 2014 at 09:57:55PM -0700, Daniele Di Proietto wrote:
> cmap_next_position() didn't update the node pointer while iterating through a
> list of nodes with the same hash.
> This commit fixes the bug and improve test-cmap to detect it.
>
> Signed-off-by: Daniele Di Proietto
Good cat
On Fri, Jul 11, 2014 at 01:24:06PM +0200, Eric Sesterhenn wrote:
> commit b5431b8dcc8e02c335c388afaee4b606d3253204
> Author: Eric Sesterhenn
> Date: Fri Jul 11 03:54:48 2014 -0500
>
> Factor the ovsdb-server main loop into a new function
>
> This refactors the ovsdb-server main loop in
Thanks, applied!
On Mon, Jul 14, 2014 at 12:41:21PM -0700, Andy Zhou wrote:
> LGTM
>
> Acked-by: Andy Zhou
>
> On Tue, Jun 24, 2014 at 11:45 AM, Ben Pfaff wrote:
> > Suggested-by: Gerald Rogers
> > Signed-off-by: Ben Pfaff
> > ---
> > AUTHORS |1 +
> > CodingStyle |7 +++
> >
On Fri, Jul 11, 2014 at 01:24:01PM +0200, Eric Sesterhenn wrote:
> Hi,
>
> on my debian installation, the ovsdb-server is running as root. Since I
> prefer to add additional mitigations for running services, I was looking
> into putting the ovsdb-server into a chroot and implemented it in the
> fo
Dear Jean,
I read the following documentation about the rate limiter API.
http://openvswitch.org/pipermail/dev/2011-January/005639.html
I have some questions about it.
We have an SDN testbed at Budapest University of Technology. We have
HP-E3500yl switches with K.15.15.0006 firmware and openvswit
I took another read through, minor nitpicks if you haven't pushed the patch
yet:-
On 16 July 2014 17:21, Alex Wang wrote:
>
> + * FWD: the port should forward any incomging non-stp-bpdu
> + * packets.
>
incomging -> incoming
bpdu -> BPDU
+ * TX_BPDU/RX_BPDU: th
55 matches
Mail list logo
