+1
> On 20.11.2013, at 00:17, "Ethan Jackson" wrote:
>
> Ben's the authority on this, but FWIW a macro sounds fine to me.
> We've already got PRI_U32 etc, adding a PRI_ZU seems reasonable.
>
> Ethan
>
>
>
> On Tue, Nov 19, 2013 at 1:57 PM, Alessandro Pilotti
> wrote:
>> Talking about %zu,
On Wed, Nov 20, 2013 at 06:09:09AM +, Vivek Kumar (EPG) wrote:
> It would be very helpful if you could please point me to the right
> documentation(capwap configuration) for this, as I am not able to find the
> correct document.
ovs-vswitchd.conf.db(5)
___
Hi Ben,
Thanks a lot for the quick response.
It would be very helpful if you could please point me to the right
documentation(capwap configuration) for this, as I am not able to find the
correct document.
Thanks,
Vivek
-Original Message-
From: Ben Pfaff [mailto:b...@nicira.com]
Sent: W
On Wed, Nov 20, 2013 at 05:53:34AM +, Vivek Kumar (EPG) wrote:
> We are trying to evaluate "openvswitch" performance. Could you please
> help us with configuration the openvswitch. How to configure capwap
> in openvswitch.
CAPWAP support was removed in OVS 1.10. To configure CAPWAP in earlie
HI
We are trying to evaluate "openvswitch" performance. Could you please help us
with configuration the openvswitch. How to configure capwap in openvswitch.
Thanks,
Vivek
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listi
On Tue, Nov 19, 2013 at 6:49 PM, Nithin Nayak Sujir wrote:
>
>>> I see these errors which don't appear with the inbox openvswitch driver.
>>>
>>> [ 4198.153985] gre_compat_init: cannot register gre_cisco protocol
>>> handler
>>
>>
>> This indicates that the ip_gre module is loaded when OVS is tryi
I see these errors which don't appear with the inbox openvswitch driver.
[ 4198.153985] gre_compat_init: cannot register gre_cisco protocol handler
This indicates that the ip_gre module is loaded when OVS is trying to
register its handler.
Doesn't appear to be.
18:23 dl6-fed-19:~/openvswi
Hey Ben and Ethan,
Could you review this RFC when you are available?
This patch is pretty complete. I'll still do some experiments to find the
best UPCALL_QUEUE_LENGTH and UPCALL_QUEUES value.
Thanks,
Alex Wang,
On Tue, Nov 19, 2013 at 6:09 PM, Alex Wang wrote:
> This commit improves the up
This commit improves the upcall dispatching fairness by
introduing a 2-stage scheme. And the two stages are run in two
threads, 'dispatcher' and 'distributor', respectively.
At the first stage, the dispatcher thread will read upcalls from
kernel and put the upcalls into the fair queues based on t
All of these stack traces seem completely different to me (and some
don't even involve OVS) so it's pretty difficult to tell what's going
on. However, if I had to guess, I suspect that you have a loop with
your GRE configuration.
On Tue, Nov 19, 2013 at 12:17 AM, nicolas prochazka
wrote:
> hello,
On Nov 19, 2013, at 2:49 PM, Ben Pfaff wrote:
> On Tue, Nov 19, 2013 at 11:20:37AM -0800, Jarno Rajahalme wrote:
>> Signed-off-by: Jarno Rajahalme
>
> miniflow_get_map_in_range() is longer than I would ordinarily write as
> inline in a header file. Do you have a special reason to do that?
>
On Mon, Nov 18, 2013 at 5:50 PM, Nithin Nayak Sujir wrote:
>>> However, I'm not able to ping vm-vm. Digging into this, I see it's
>>> because
>>> ARP is not resolved. On the destination system, I see the GRE
>>> encapsulated
>>> ARP packets being received at p4p1 and br0. But they don't get forwar
Allow TCP flags match specification with symbolic flag names. TCP
flags are optionally specified as a string of flag names, each
preceded by '+' when the flag must be one, or '-' when the flag must
be zero. Any flags not explicitly included are wildcarded. The
existing hex syntax is still allowe
On Tue, Nov 19, 2013 at 11:42:14AM -0800, Ben Pfaff wrote:
> Please forgive me for not answering point-by-point. I think that
> there is some basic confusion here that I can clear up quickly, and
> then I think the rest will follow from that.
>
> In OpenFlow 1.0 through 1.2, a "no match" packet-i
On Tue, Nov 19, 2013 at 03:33:30PM -0800, Saurabh Shah wrote:
> The Makefile rule would probably do the job in this case, but for a
> more generic use case we probably need some sort of a pre-checkin
> build sanitizer. For e.g. how will we prevent unix system calls from
> being introduced (assuming
On Tue, Nov 19, 2013 at 02:16:43PM -0800, Ethan Jackson wrote:
Ben's the authority on this, but FWIW a macro sounds fine to me.
We've already got PRI_U32 etc, adding a PRI_ZU seems reasonable.
I don't object to that, I just want to make sure that there's some
Makefile rule to blacklist %zu so tha
On Tue, Nov 19, 2013 at 11:17:37AM -0800, Romain Lenglet wrote:
> Guard any access to an IPFIX row referenced from
> Flow_Sample_Collector_Set by a test that the reference is not NULL.
> Factorized that check into macro __VALID_FSCS.
>
> Signed-off-by: Romain Lenglet
The macros in this series ar
On Tue, Nov 19, 2013 at 11:20:37AM -0800, Jarno Rajahalme wrote:
> Signed-off-by: Jarno Rajahalme
miniflow_get_map_in_range() is longer than I would ordinarily write as
inline in a header file. Do you have a special reason to do that?
I spent some time studying find_match_wc(). I think that th
On Tue, Nov 19, 2013 at 02:16:43PM -0800, Ethan Jackson wrote:
> Ben's the authority on this, but FWIW a macro sounds fine to me.
> We've already got PRI_U32 etc, adding a PRI_ZU seems reasonable.
I don't object to that, I just want to make sure that there's some
Makefile rule to blacklist %zu so
Ben's the authority on this, but FWIW a macro sounds fine to me.
We've already got PRI_U32 etc, adding a PRI_ZU seems reasonable.
Ethan
On Tue, Nov 19, 2013 at 1:57 PM, Alessandro Pilotti
wrote:
> Talking about %zu, we solved it with a brutal macro. Do you have any
> suggestion for a better a
On Tue, Nov 19, 2013 at 09:57:43PM +, Alessandro Pilotti wrote:
> Talking about %zu, we solved it with a brutal macro. Do you have any
> suggestion for a better approach?
I have thought of two viable approaches:
1. Use a macro plus add a makefile rule to scan the tree for use of
'z
Talking about %zu, we solved it with a brutal macro. Do you have any suggestion
for a better approach?
On 19 Nov 2013, at 18:51 , Gurucharan Shetty wrote:
> On Tue, Nov 19, 2013 at 8:37 AM, Ben Pfaff wrote:
>> On Tue, Nov 19, 2013 at 12:22:40PM +, Alessandro Pilotti wrote:
>>> The freshl
Rather than using complete flow hash, we can use skb->rxhash for
calculating source port. Similar calculation is done by vxlan.
Signed-off-by: Pravin B Shelar
---
datapath/vport-lisp.c |9 +++--
1 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/datapath/vport-lisp.c b/datapa
Following patch improves rxhash calculation, It is taken from
upstream Linux kernel code.
>From kernel 3.8, skb_get_rxhash() can handle hardware generated
l4-rxhash. Therefore compat skb_get_rxhash() is not used on
kernel 3.8 or new.
Signed-off-by: Pravin B Shelar
---
datapath/linux/Modules.mk
Some distribution kernel has backported skb->rxhash (e.g. RHEL)
Following patch allows use precalculated rxhash.
Signed-off-by: Pravin B Shelar
---
acinclude.m4 |1 +
datapath/linux/compat/flow_dissector.c |2 +-
datapath/linux/compat/include/linux/s
Thanks Thomas and Jesse for the reviews. I applied it to master.
On Tue, Nov 19, 2013 at 3:38 AM, Thomas Graf wrote:
> On 11/13/2013 10:29 PM, Andy Zhou wrote:
>
>> API changes only for code readability. No functional chnages.
>>
>> This patch removes the underscored version. Added a new API
>>
One can only hope :).
Kind Regards,
Alin.
From: Saurabh Shah [ssaur...@vmware.com]
Sent: Tuesday, November 19, 2013 8:19 PM
To: Alin Serdean; Gurucharan Shetty; Ben Pfaff
Cc: dev@openvswitch.org; Saurabh Shah; Gurucharan Shetty
Subject: Re: [ovs-dev] Window
Please forgive me for not answering point-by-point. I think that
there is some basic confusion here that I can clear up quickly, and
then I think the rest will follow from that.
In OpenFlow 1.0 through 1.2, a "no match" packet-in meant exactly
that: no match was found in a table. If any match wa
Guard any access to an IPFIX row referenced from
Flow_Sample_Collector_Set by a test that the reference is not NULL.
Factorized that check into macro __VALID_FSCS.
Signed-off-by: Romain Lenglet
---
vswitchd/bridge.c |8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/v
The "targets" column in IPFIX had a min=1 constraints, so OVSDB
implicitly adds an empty string "" into that column if no value is
given. No connection can be opened to a target with address "", so
the whole IPFIX exporter for that row was disabled until that ""
target was removed by users. That
Signed-off-by: Jarno Rajahalme
---
v3: Simplify _range() functions by using 64-bit maps and by
consolidating setup + Rebase, updated comments, and more
rigorous testing.
v2: Properly finish hashes, more flexible configuration.
lib/classifier.c | 156
On Tue, Nov 19, 2013 at 11:03:04AM +0200, Lori Jakab wrote:
> On 11/19/13 12:51 AM, Ben Pfaff wrote:
> >On Mon, Nov 18, 2013 at 02:45:46PM -0800, Ben Pfaff wrote:
> >>On Mon, Nov 18, 2013 at 10:27:48PM +0200, Lorand Jakab wrote:
> >>>Add member is_layer3 to struct ofport_dpif to mark layer 3 ports.
On Tue, Nov 19, 2013 at 11:06:44AM +0200, Lorand Jakab wrote:
> Add member is_layer3 to struct ofport_dpif to mark layer 3 ports. Set
> it to "true" for the only layer 3 port we support for now: lisp.
>
> Additionally, prevent flooding to layer 3 ports. A later patch will
> also prevent MAC lear
I'm glad to hear that! Thanks, I applied this to master.
On Tue, Nov 19, 2013 at 11:01:03AM -0800, Andy Zhou wrote:
> Ben thanks for fixing this. It works properly in my environment with this
> fix.
>
> Acked-by: Andy Zhou
>
>
> On Mon, Nov 18, 2013 at 1:26 PM, Ben Pfaff wrote:
>
> > Other
Ben thanks for fixing this. It works properly in my environment with this
fix.
Acked-by: Andy Zhou
On Mon, Nov 18, 2013 at 1:26 PM, Ben Pfaff wrote:
> Otherwise running "configure" twice with different --prefix (etc.) will
> fail to update ovs-lib, so that "make install" installs an ovs-lib
On Tue, Nov 19, 2013 at 05:13:00PM +0900, YAMAMOTO Takashi wrote:
> This avoids a conflict with NetBSD's strings.h/libc.
> (http://netbsd.gw.com/cgi-bin/man-cgi?popcount++NetBSD-current)
>
> The new name is suggested by Ben Pfaff.
>
> Signed-off-by: YAMAMOTO Takashi
Thanks, I'll apply this soon
We haven't moved immediately to VS2013, because it doesn't impact us
immediately. The changes for the C99 features added by microsoft were mostly
made prior to the VS2013 release, I think.
Also, one of the commenters on the blog(who works for MSFT) mentioned -
"It may be possible to sneak in sn
On Tue, Nov 19, 2013 at 10:01 AM, James Page wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA256
>
> On 19/11/13 17:46, Pravin Shelar wrote:
>> From 3.12 kernel, ovs can start using kernel API for gre and vxlan
>> tunneling. Can you turn off compat tunnel code for kernel 3.12 and
>> newer?
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
On 19/11/13 17:46, Pravin Shelar wrote:
> From 3.12 kernel, ovs can start using kernel API for gre and vxlan
> tunneling. Can you turn off compat tunnel code for kernel 3.12 and
> newer?
How does LISP tunnelling work? Will disabling the compat tun
On Tue, Nov 19, 2013 at 5:39 AM, James Page wrote:
> Bump kernel support for datapath module to include 3.12.
>
> Signed-off-by: James Page
> ---
> FAQ | 2 +-
> NEWS | 2 +-
> acinclude.m4 | 4 ++--
> 3 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/FAQ b/FAQ
>
We started working on this project using Visual Studio 2013 RC and made the
finishing touches the moment the final release got out.
A lot of C99 was added but yes unfortunately %zu still has not yet been
implemented .
A more detailed list of things added can be found on the following link:
htt
Acked-by: Jarno Rajahalme
On Nov 19, 2013, at 12:13 AM, YAMAMOTO Takashi wrote:
> This avoids a conflict with NetBSD's strings.h/libc.
> (http://netbsd.gw.com/cgi-bin/man-cgi?popcount++NetBSD-current)
>
> The new name is suggested by Ben Pfaff.
>
> Signed-off-by: YAMAMOTO Takashi
> ---
> lib
sorry please ignore the previous email... didn't intend to broadcast...
sorry for the spam,
On Tue, Nov 19, 2013 at 9:07 AM, Alex Wang wrote:
> Every other day reminder~, could you review this patch~? thx
>
>
> On Fri, Nov 15, 2013 at 1:58 PM, Alex Wang wrote:
>
>> This commit adds a forwardi
Every other day reminder~, could you review this patch~? thx
On Fri, Nov 15, 2013 at 1:58 PM, Alex Wang wrote:
> This commit adds a forwarding flag to "struct bfd". And
> bfd_forwarding__() will update this flag at each invocation.
>
> Signed-off-by: Alex Wang
> ---
> lib/bfd.c | 30 ++
On Tue, Nov 19, 2013 at 8:37 AM, Ben Pfaff wrote:
> On Tue, Nov 19, 2013 at 12:22:40PM +, Alessandro Pilotti wrote:
>> The freshly released Visual Studio 2013 is the minimum supported
>> version due to its improved C99 support that avoided us a few
>> headaches.
>
> Guru, have you been using t
On Tue, Nov 19, 2013 at 12:22:40PM +, Alessandro Pilotti wrote:
> The freshly released Visual Studio 2013 is the minimum supported
> version due to its improved C99 support that avoided us a few
> headaches.
Guru, have you been using the compiler from VS2013? I see that it still
does not impl
Bump kernel support for datapath module to include 3.12.
Signed-off-by: James Page
---
FAQ | 2 +-
NEWS | 2 +-
acinclude.m4 | 4 ++--
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/FAQ b/FAQ
index 4a8a5be..9b3a900 100644
--- a/FAQ
+++ b/FAQ
@@ -148,7 +148,7 @@ A
Hi guys,
My company, Cloudbase Solutions, as part of our effort in integrating OpenStack
with Windows and Hyper-V is currently involved in porting Open vSwitch to
Windows.
We are currently revieweing the first Windows patches that have been posted to
this ML and we will gladly start to contrib
On 11/13/2013 10:29 PM, Andy Zhou wrote:
API changes only for code readability. No functional chnages.
This patch removes the underscored version. Added a new API
ovs_flow_tbl_lookup_stats() that
returns the n_mask_hits.
Reported by: Ben Pfaff
Signed-off-by: Andy Zhou
---
v1->v2:
Ad
On 11/19/2013 10:06 AM, Lorand Jakab wrote:
Add member is_layer3 to struct ofport_dpif to mark layer 3 ports. Set
it to "true" for the only layer 3 port we support for now: lisp.
Additionally, prevent flooding to layer 3 ports. A later patch will
also prevent MAC learning.
Signed-off-by: Lora
Add member is_layer3 to struct ofport_dpif to mark layer 3 ports. Set
it to "true" for the only layer 3 port we support for now: lisp.
Additionally, prevent flooding to layer 3 ports. A later patch will
also prevent MAC learning.
Signed-off-by: Lorand Jakab
---
v3:
* Fix failing unit tests
Hi,
I am somewhat confused about this patch.
I became confused while trying to make use of pin->generated_by_table_miss
to drop controller actions in connmgr_send_packet_in()
for OpenFlow1.3 if its default table-miss policy is in effect.
I have successfully implemented a prototype pf that logic.
On 11/19/13 12:51 AM, Ben Pfaff wrote:
On Mon, Nov 18, 2013 at 02:45:46PM -0800, Ben Pfaff wrote:
On Mon, Nov 18, 2013 at 10:27:48PM +0200, Lorand Jakab wrote:
Add member is_layer3 to struct ofport_dpif to mark layer 3 ports. Set
it to "true" for the only layer 3 port we support for now: lisp.
hello,
It's 100% reproductible on physical server ( dell , network card
driver bnx2 ) , with a virtual vm i can't reproducte it.
it's near this description :
http://permalink.gmane.org/gmane.linux.network/269580
More information :
GRE tunnel are active but huge traffic is on standard port (not gre
> On Tue, Nov 19, 2013 at 02:49:53PM +0900, YAMAMOTO Takashi wrote:
>> > On Tue, Nov 19, 2013 at 11:20:44AM +0900, YAMAMOTO Takashi wrote:
>> >> > Having a single function that can do popcount() on any integer type is
>> >> > easier for callers to get right. The implementation is probably slower
>
This avoids a conflict with NetBSD's strings.h/libc.
(http://netbsd.gw.com/cgi-bin/man-cgi?popcount++NetBSD-current)
The new name is suggested by Ben Pfaff.
Signed-off-by: YAMAMOTO Takashi
---
lib/flow.c| 5 +++--
lib/util.c| 16
lib/util.h| 2 +-
test
56 matches
Mail list logo
