On Thu, Aug 29, 2013 at 02:00:51PM -0700, Ben Pfaff wrote:
> From: Neil Zhu
>
> This doesn't include a dpif implementation of groups functionality. In its
> current form, it is untested. Before this is committed, it needs some
> more comments and an ovs-ofctl manpage update.
>
> Signed-off-by:
Signed-off-by: Simon Horman
---
Ben, feel free to squash this into 'Implement OpenFlow 1.1+ "groups" protocol'.
---
utilities/ovs-ofctl.8.in | 108 +++
1 file changed, 108 insertions(+)
diff --git a/utilities/ovs-ofctl.8.in b/utilities/ovs-ofctl.8.in
Signed-off-by: Ed Maste
---
On FreeBSD pthread_set_name_np's prototype is provided by pthread_np.h.
As I believe it is the only platform to provide the "set_name" (with an
underscore) variant I hope it's fine to use the existing autoconf macro
HAVE_PTHREAD_SET_NAME_NP.
lib/util.c | 3 +++
1 file
[redirecting to ovs-dev]
On Fri, Aug 30, 2013 at 06:17:52AM -0700, Gurucharan Shetty wrote:
> On Aug 29, 2013, at 9:57 PM, Ben Pfaff wrote:
>
> >Unclaimed Tasks (Small, Required)
> >-
> >
> >The following tasks could use volunteers. These tasks are all
> >require
On Thu, Aug 29, 2013 at 03:05:05PM -0700, Ben Pfaff wrote:
> James pointed out that branch-1.9 has version 1.9.2 and branch-1.10
> has version 1.10.1, but that none of those has any tarball or tag or
> announcement. I can do the tags, but you usually do the rest. Also,
> each of those branches ha
On Fri, Aug 30, 2013 at 05:42:54PM +0900, Simon Horman wrote:
> Signed-off-by: Simon Horman
>
> ---
>
> Ben, feel free to squash this into 'Implement OpenFlow 1.1+ "groups"
> protocol'.
Thanks, done. I added your sign-off and gave you co-authorship also
(I really like to encourage documentatio
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
On 30/08/13 17:02, Ben Pfaff wrote:
> On Thu, Aug 29, 2013 at 03:05:05PM -0700, Ben Pfaff wrote:
>>> James pointed out that branch-1.9 has version 1.9.2 and
>>> branch-1.10 has version 1.10.1, but that none of those has any
>>> tarball or tag or anno
I will send a v2 soon based on Andy's feedback.
Thanks
Guolin
On Thu, Aug 29, 2013 at 8:43 PM, Ben Pfaff wrote:
> Andy, thanks for the review.
>
> Guolin, are you willing to submit a v2?
>
> Thanks,
>
> Ben.
>
> On Thu, Aug 29, 2013 at 05:21:03PM -0700, Andy Zhou wrote:
> > Good catch. The log
On Fri, Aug 30, 2013 at 04:30:39PM +0900, Simon Horman wrote:
> On Thu, Aug 29, 2013 at 02:00:51PM -0700, Ben Pfaff wrote:
> > From: Neil Zhu
> >
> > This doesn't include a dpif implementation of groups functionality. In its
> > current form, it is untested. Before this is committed, it needs s
Signed-off-by: Ben Pfaff
---
FAQ | 34 ++
1 file changed, 34 insertions(+)
diff --git a/FAQ b/FAQ
index 75d9007..5406e84 100644
--- a/FAQ
+++ b/FAQ
@@ -250,6 +250,40 @@ A: The following commands configure br0 with eth0 and tap0
as trunk
ovs-vsctl clea
From: Guolin Yang
In odp_flow_key_from_flow__(), when converting ICMPv6 flow/mask
to flow/mask key, we should always use flow to check for whether
ND informaition is present or not. In mask case, both type and code
field should be masked, otherwise ND fields can be masked.
Similarly in reverse c
On Fri, Aug 30, 2013 at 10:56:08AM -0400, Ed Maste wrote:
> Signed-off-by: Ed Maste
Thanks, applied to master and branch-2.0.
> ---
> On FreeBSD pthread_set_name_np's prototype is provided by pthread_np.h.
> As I believe it is the only platform to provide the "set_name" (with an
> underscore) va
Looks good to me. Thanks for the revision.
Acked-by: Andy Zhou
On Fri, Aug 30, 2013 at 9:57 AM, wrote:
> From: Guolin Yang
>
> In odp_flow_key_from_flow__(), when converting ICMPv6 flow/mask
> to flow/mask key, we should always use flow to check for whether
> ND informaition is present or no
netdev_features_to_bps() returns a speed in bps, but struct
ofputil_phy_port's curr_speed and max_speed are in kbps, so a conversion
is necessary. This commit fixes the problem.
Reported-by: Benjamin Lunsky
Signed-off-by: Ben Pfaff
---
ofproto/ofproto.c |4 ++--
1 file changed, 2 insertion
ack, looks good to me.
-pritesh
On Aug 30, 2013, at 10:02 AM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> FAQ | 34 ++
> 1 file changed, 34 insertions(+)
>
> diff --git a/FAQ b/FAQ
> index 75d9007..5406e84 100644
> --- a/FAQ
> +++ b/FAQ
> @@ -250,6 +250,
sw_flow_key alignment was declared as " __aligned(__alignof__(long))"
However, this breaks on m68k architecture where long is 32 bit in size
but 16 bit aligned by default. Use __aligned(sizeof(long) instead.
Reported by: Fengguang Wu
Signed-off-by: Andy Zhou
---
net/openvswitch/flow.c |4 +
It doesn't apply directly to branch-1.11. Can you take a look and see
whether there's some other commit we're missing there?
Thanks,
Ben.
On Fri, Aug 30, 2013 at 11:09:21AM -0700, Andy Zhou wrote:
> Yes, this patch fixes a bug.
>
>
> On Fri, Aug 30, 2013 at 11:06 AM, Ben Pfaff wrote:
>
> >
On Fri, Aug 30, 2013 at 11:03:24AM -0700, Andy Zhou wrote:
> On Fri, Aug 30, 2013 at 11:01 AM, Ben Pfaff wrote:
>
> > On Fri, Aug 30, 2013 at 09:57:13AM -0700, gy...@nicira.com wrote:
> > > @@ -2981,7 +2986,8 @@ parse_l2_5_onward(const struct nlattr
> > *attrs[OVS_KEY_ATTR_MAX + 1],
> > >
On Fri, Aug 30, 2013 at 09:57:13AM -0700, gy...@nicira.com wrote:
> @@ -2981,7 +2986,8 @@ parse_l2_5_onward(const struct nlattr
> *attrs[OVS_KEY_ATTR_MAX + 1],
> if (is_mask) {
> if (!is_all_zeros((const uint8_t *) nd_key,
>
On Fri, Aug 30, 2013 at 11:01 AM, Ben Pfaff wrote:
> On Fri, Aug 30, 2013 at 09:57:13AM -0700, gy...@nicira.com wrote:
> > @@ -2981,7 +2986,8 @@ parse_l2_5_onward(const struct nlattr
> *attrs[OVS_KEY_ATTR_MAX + 1],
> > if (is_mask) {
> > if (!is_all_z
Yes, this patch fixes a bug.
On Fri, Aug 30, 2013 at 11:06 AM, Ben Pfaff wrote:
> On Fri, Aug 30, 2013 at 11:03:24AM -0700, Andy Zhou wrote:
> > On Fri, Aug 30, 2013 at 11:01 AM, Ben Pfaff wrote:
> >
> > > On Fri, Aug 30, 2013 at 09:57:13AM -0700, gy...@nicira.com wrote:
> > > > @@ -2981,7 +29
> After this patch, coverage_read() seems halfway between two designs.
> My goal was to avoid holding coverage_mutex while constructing the log
> message, to keep the hold time low, by copying out the totals and then
> building the message with the copies. Your version retains the
> copy-out for t
On Fri, Aug 30, 2013 at 10:32 AM, Andy Zhou wrote:
> diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
> index ad1aeeb..fe7524c4 100644
> --- a/net/openvswitch/flow.c
> +++ b/net/openvswitch/flow.c
> int ovs_flow_init(void)
> {
> BUILD_BUG_ON(sizeof(struct sw_flow_key) % sizeo
On Tue, Aug 27, 2013 at 09:25:10PM -0700, Alex Wang wrote:
> This commit changes the "ovs-appctl coverage/show" command to show the
> per-second, per-minute and per-hour rates of function invocation. More
> importantly, this makes using coverage counter an easy way to monitor
> the execution of sp
On Fri, Aug 30, 2013 at 11:20:54AM -0700, Alex Wang wrote:
> > After this patch, coverage_read() seems halfway between two designs.
> > My goal was to avoid holding coverage_mutex while constructing the log
> > message, to keep the hold time low, by copying out the totals and then
> > building the
>
> Ah. I understand now. But I'm not sure that this is actually a good
> way to do it, because it makes the output harder to interpret. If I
> see that the rate over the last 5 seconds is 123/s, for the last
> minute is 4920/min, and for the last hour is 19680/h, then I have a
> hard time figur
+ Conseguí los mejores descuentos
+ Elegí entre las mejores empresas
+ Congelamos tu cuota semestralmente !!!
Porque... más atención y más beneficios es mucho más que un seguro!.
abSeguro.com/cotizar___
dev mailing list
dev@openvswitch.org
http://openv
From: Jesse Gross
Date: Fri, 30 Aug 2013 11:22:11 -0700
> The suggestion that Geert made was to just drop this entirely and
> rely on the natural alignment from these values.
Indeed, Geert's patch was 1,000 times superior to this one.
___
dev mailing l
This declares a new API extension between ofproto and the ofproto-provider
to
support Openflow forwarding groups as introduced in the 1.1 spec. This patch
depends on an earlier submission that defines 'struct ofpbucket.' There is
no
executable code in this patch.
The API is modeled after that for
May be it caused by missing commit 4a2216156e6e168 ?
On Fri, Aug 30, 2013 at 11:11 AM, Ben Pfaff wrote:
> It doesn't apply directly to branch-1.11. Can you take a look and see
> whether there's some other commit we're missing there?
>
> Thanks,
>
> Ben.
>
> On Fri, Aug 30, 2013 at 11:09:21AM -
I still get conflicts:
blp@sigsegv:~/ovs$ git checkout openvswitch/branch-1.11
HEAD is now at 37f9d3a... Set release date for 1.11.0.
blp@sigsegv:~/ovs$ git cherry-pick 4a2216156e6e168
error: could not apply 4a22161... odp-util: New function
odp_flow_key_to_mask().
hint: after
datapath support for older kernel is been broken for kernel
older than 2.6.32 for some time now. There have not been any
complains about this. So it makes sense to get rid of the code.
This make datapath module much close to upstream datapath and
make code easy to understand.
To make review easy I
cleanup various header file.
Signed-off-by: Pravin B Shelar
---
datapath/linux/compat/include/linux/if_ether.h | 13
datapath/linux/compat/include/linux/kernel.h| 31 -
datapath/linux/compat/include/linux/netdevice.h | 81 ---
datapath/linux/compat/inc
Signed-off-by: Pravin B Shelar
---
datapath/Modules.mk |1 -
datapath/datapath.c | 10 --
datapath/datapath.h |5 ---
datapath/linux/compat/netdevice.c |4 --
datapath/linux/compat/vxlan.c | 10 +-
datapath/vlan.c
Signed-off-by: Pravin B Shelar
---
acinclude.m4 |3 -
datapath/linux/Modules.mk | 16 -
datapath/linux/compat/addrconf_core-openvswitch.c | 82
datapath/linux/compat/include/linux/dmi.h | 114 -
datapath/li
It will be simpler if we just fix the bug instead of pulling in more
patches. I will derive a patch and send it out.
On Fri, Aug 30, 2013 at 1:19 PM, Ben Pfaff wrote:
> I still get conflicts:
>
> blp@sigsegv:~/ovs$ git checkout openvswitch/branch-1.11
> HEAD is now at 37f9d3a... Set rel
That's perfect, thanks a lot!
On Fri, Aug 30, 2013 at 01:26:44PM -0700, Andy Zhou wrote:
> It will be simpler if we just fix the bug instead of pulling in more
> patches. I will derive a patch and send it out.
>
>
> On Fri, Aug 30, 2013 at 1:19 PM, Ben Pfaff wrote:
>
> > I still get conflicts:
Signed-off-by: Pravin B Shelar
---
datapath/actions.c |2 +-
datapath/compat.h | 39 ---
datapath/datapath.c| 10 ++
datapath/flow.c|9 ++---
datapath/vport-gre.c |2 +-
datapath/vport-lisp.c |2 +-
datapa
Signed-off-by: Pravin B Shelar
---
acinclude.m4|1 -
datapath/compat.h | 26
datapath/datapath.c | 55 ---
3 files changed, 0 insertions(+), 82 deletions(-)
diff --git a/acinclude.m4 b/acinclude.m4
index e
Signed-off-by: Pravin B Shelar
---
.../linux/compat/include/linux/reciprocal_div.h| 40
datapath/linux/compat/reciprocal_div.c |1 +
2 files changed, 1 insertions(+), 40 deletions(-)
delete mode 100644 datapath/linux/compat/include/linux/reciprocal_div.
Signed-off-by: Pravin B Shelar
---
datapath/Modules.mk |2 -
datapath/actions.c| 15 +-
datapath/checksum.c | 271 -
datapath/checksum.h | 173
Patch removes genl, netlink, rtnl compat code.
Signed-off-by: Pravin B Shelar
---
datapath/compat.h |8 --
datapath/datapath.c | 12 +-
datapath/dp_notify.c|2 +-
datapath/linux/compat/genetlink-openv
Rather than having compile time check in datapath.c, its better
to check kernel version at configuration step.
This patch also removes support for older kernel.
Signed-off-by: Pravin B Shelar
---
acinclude.m4| 12 ++--
datapath/datapath.c |5 -
2 files changed, 10 inser
Signed-off-by: Pravin B Shelar
---
datapath/linux/compat/include/linux/if.h |7 ---
datapath/linux/compat/include/net/ip.h |6 --
datapath/vport-internal_dev.c| 13 -
datapath/vport-lisp.c|4
datapath/vport-netdev.c
Signed-off-by: Pravin B Shelar
---
datapath/linux/Modules.mk |1 -
datapath/linux/compat/include/net/genetlink.h | 18 -
datapath/linux/compat/include/net/net_namespace.h | 80 +
datapath/linux/compat/include/net/netns/generic.h | 12 -
The "groups" container is protected with a rwlock, and each group is
protected with their own rwlock. Whenever groups are looked up, the
container's lock is taken. If a group is found, the group is locked
before the container is released, as otherwise the group might be
deleted while the lock is
Also, since all kinds of groups can refer to other groups, there is no
reason to delete indirect groups first.
Signed-off-by: Jarno Rajahalme
---
ofproto/ofproto.c | 20 +---
1 file changed, 9 insertions(+), 11 deletions(-)
diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c
in
Indirect groups can have at most one bucket.
Signed-off-by: Jarno Rajahalme
---
lib/ofp-parse.c | 14 +++---
1 file changed, 11 insertions(+), 3 deletions(-)
diff --git a/lib/ofp-parse.c b/lib/ofp-parse.c
index 9098467..f17364f 100644
--- a/lib/ofp-parse.c
+++ b/lib/ofp-parse.c
@@ -20
These should apply on the groups 2/2 v3.
Jarno Rajahalme (3):
Check the number of buckets for indirect groups.
Delete groups in ofproto_destruct.
Locking for groups.
lib/ofp-parse.c| 14 ++-
ofproto/ofproto-provider.h | 23 -
ofproto/ofproto.c | 220 ++
Patch for branch 1.11 is attached.
On Fri, Aug 30, 2013 at 1:27 PM, Ben Pfaff wrote:
> That's perfect, thanks a lot!
>
> On Fri, Aug 30, 2013 at 01:26:44PM -0700, Andy Zhou wrote:
> > It will be simpler if we just fix the bug instead of pulling in more
> > patches. I will derive a patch and sen
This commit changes the "ovs-appctl coverage/show" command to show the
the averaged per-second rates for the last few seconds, the last minute
and the last hour, and the total counts of all of the coverage counters.
Signed-off-by: Alex Wang
---
v2 -> v3:
- fix typo.
- provide the averaged per-s
This commit removes the CACHE_TIME scheme from timeval module. This
is for eliminating the lock contention over the read/write lock of
the cached time. To get the time, the thread now will directly do
the system call 'clock_gettime()'.
As a side effect, time can only be warpped after timer is st
It's probably worth mentioning in NEWS and updating FAQ, SubmittingPatches, and
README to remove references to 2.6.18.
--Justin
On Aug 30, 2013, at 1:22 PM, Pravin B Shelar wrote:
> datapath support for older kernel is been broken for kernel
> older than 2.6.32 for some time now. There have n
On Fri, Aug 30, 2013 at 11:42 AM, David Miller wrote:
> From: Jesse Gross
> Date: Fri, 30 Aug 2013 11:22:11 -0700
>
>> The suggestion that Geert made was to just drop this entirely and
>> rely on the natural alignment from these values.
>
> Indeed, Geert's patch was 1,000 times superior to this o
On Wed, Aug 28, 2013 at 1:57 AM, Viresh Kumar wrote:
> On 28 August 2013 05:37, Jesse Gross wrote:
>> On Tue, Aug 27, 2013 at 3:51 AM, Viresh Kumar
>> wrote:
>>> On 26 August 2013 21:59, Jesse Gross wrote:
>>>
This is compatibility code for older kernels so by definition it
doesn't e
On Wed, Aug 28, 2013 at 2:57 AM, Viresh Kumar wrote:
> On 28 August 2013 05:48, Jesse Gross wrote:
>
>> The implementation is actually pretty much exactly the same as before.
>> The only reason why there are no longer separate process/interrupt
>> counters is because we started disabling bottom h
On Fri, Aug 30, 2013 at 09:15:19AM -0700, Ben Pfaff wrote:
> On Fri, Aug 30, 2013 at 05:42:54PM +0900, Simon Horman wrote:
> > Signed-off-by: Simon Horman
> >
> > ---
> >
> > Ben, feel free to squash this into 'Implement OpenFlow 1.1+ "groups"
> > protocol'.
>
> Thanks, done. I added your sign
From: Jesse Gross
Date: Fri, 30 Aug 2013 16:20:25 -0700
> I looked through the struct definition and I think that the idea of
> manually padding as Geert did in his patch will be difficult to
> maintain over time (and actually there are a few that he missed) since
> there are a number of differen
On Thu, Aug 29, 2013 at 09:52:07PM -0700, Ben Pfaff wrote:
> It appears to me that OFPST_TABLE is correct for OF1.1+, so I'm removing
> that.
>
> DESIGN describes OF1.2 flow_mod, so I'm removing that.
>
> Port and queue stats include duration these days. The proposed patch to
> implement groups
On Fri, Aug 30, 2013 at 01:40:16PM -0700, Jarno Rajahalme wrote:
> Indirect groups can have at most one bucket.
>
> Signed-off-by: Jarno Rajahalme
Reviewed-by: Simon Horman
> ---
> lib/ofp-parse.c | 14 +++---
> 1 file changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/lib/
On Fri, Aug 30, 2013 at 01:40:17PM -0700, Jarno Rajahalme wrote:
> Also, since all kinds of groups can refer to other groups, there is no
> reason to delete indirect groups first.
This seems reasonable to me, but its not clear to me
how the code prevents the deletion of groups that
are still refer
On Wed, Aug 28, 2013 at 05:45:28PM +0900, Simon Horman wrote:
> Hi Ethan,
>
> On Wed, Aug 28, 2013 at 12:25:31AM -0700, Ethan Jackson wrote:
> > > In short, it would make my life much easier if facets could be looked up
> > > execute_flow_miss(). And I am interested to know if you have any plans i
On Wed, Aug 28, 2013 at 8:20 AM, Nicolas Dichtel
wrote:
> This argument is not used, let's remove it.
>
> Signed-off-by: Nicolas Dichtel
> ---
> drivers/net/vxlan.c | 3 +--
> include/net/ip_tunnels.h| 3 +--
> net/ipv4/ip_tunnel.c| 3 +--
> net/ipv4/ip_tunnel_core.c | 3 +-
63 matches
Mail list logo
