On Fri, Jan 13, 2012 at 2:11 AM, Ben Pfaff wrote:
>
> On Thu, Jan 12, 2012 at 04:02:05PM +0900, Yoshi Tamura wrote:
> > I noticed a weird behavior in openvswitch 1.3.0 that when I install a
> > flow with mod_dl_dst, the packet gets corrupted. I created a deb
> > package from the official 1.3.0 re
On Tue, Jan 10, 2012 at 5:25 PM, Pravin B Shelar wrote:
> diff --git a/datapath/compat.h b/datapath/compat.h
> index efad6a0..f9ec590 100644
> --- a/datapath/compat.h
> +++ b/datapath/compat.h
> +#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,32)
> +#define SET_NETNSOK
> +#else
> +#define SET_NETNSOK
Fixed according to comments from Ben.
v1-v2:
- Fixed man page
- Added invalid-ttl arg for ofctl-monitor.
- Added invalid-ttl packet-in test case.
--8<--cut here-->8-
Following patch implements dec_ttl as vendor action with si
> Yes lacp_negotiated is true if and only if LACP has decided at least
> one slave may be enabled. I orginally explicitly enforced this in
> bond_choose_output_slave(), but it's unnecessary. Perhaps I should
> bring that patch back for clarity though? It may be safer than
> relying on the LACP co
List available: US businesses
- 4 million emails
Price until Friday: $99
CLICK HERE TO ORDER
File
Available for download in excel format
Some of the business
categories
> Is this commit complete? It looks like the following code remains in
> bond_check_admissibility():
>
> case BM_TCP:
> /* TCP balancing has degraded to SLB (otherwise the
> * bond->lacp_negotiated check above would have processed this).
> *
> * Fall through.
On Thu, Jan 12, 2012 at 05:10:42PM -0800, Ethan Jackson wrote:
> > Should we commit this to earlier branches as a bug fix? ?It looks low
> > risk to me.
>
> I'm not even sure if it's really a bug because whenever the
> lacp_negotiated status changes, presumably the may_enable status of
> one of th
> Should we commit this to earlier branches as a bug fix? It looks low
> risk to me.
I'm not even sure if it's really a bug because whenever the
lacp_negotiated status changes, presumably the may_enable status of
one of the links will change as well. It's so edge case-ish that I
wanted to put it
On Thu, Jan 12, 2012 at 05:09:02PM -0800, Ethan Jackson wrote:
> > I would document in vswitch.xml that the LACP system ID defaults to
> > the bridge's Ethernet address if it is not specified.
>
> I updated the system-id other_config column to the following:
>
>
> The LACP syst
Thank you, I pushed this to master.
On Thu, Jan 12, 2012 at 04:24:31PM -0800, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Wed, Nov 30, 2011 at 15:08, Ben Pfaff wrote:
> > At one point in the past, there were three separate queues between the
> > kernel module and OVS userspace, each of
> I would document in vswitch.xml that the LACP system ID defaults to
> the bridge's Ethernet address if it is not specified.
I updated the system-id other_config column to the following:
The LACP system ID of this . The system ID of a
LACP bond is used to identify i
On Thu, Jan 12, 2012 at 05:04:04PM -0800, Ethan Jackson wrote:
> That's fine with me. Here's the code without renaming the tests.
Looks good to me. Thank you.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
That's fine with me. Here's the code without renaming the tests.
---
tests/lacp.at | 22 +++---
1 files changed, 19 insertions(+), 3 deletions(-)
diff --git a/tests/lacp.at b/tests/lacp.at
index de7a3a7..16daa3d 100644
--- a/tests/lacp.at
+++ b/tests/lacp.at
@@ -6,8 +6,7 @@ OV
Thanks, pushed to master, branch-1.4, branch-1.3. I think it's also
broken on 1.2, but there are merge conflicts there and it's such a
minor bug that I don't really care enough to resolve them.
On Thu, Jan 12, 2012 at 04:18:33PM -0800, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Tue, De
On Thu, Jan 12, 2012 at 11:40:04AM -0800, Ethan Jackson wrote:
> As promised, post version 1.4 the default bond_mode is changing to
> active-backup with this commit.
>
> Signed-off-by: Ethan Jackson
Thanks.
There is something about the NEWS entry that my English parser doesn't
like. How about
On Thu, Jan 12, 2012 at 04:43:41PM -0800, Justin Pettit wrote:
> On Jan 12, 2012, at 4:14 PM, Ben Pfaff wrote:
>
> > As long as I'm suggesting changes, I know that I suggested using xxx
> > in dates earlier, because we don't know when we're going to release in
> > advance, but it makes Debian pack
On Thu, Jan 12, 2012 at 11:40:03AM -0800, Ethan Jackson wrote:
> In the original Open vSwitch LACP implementation, when no slaves
> found a LACP partner, the LACP module would attach all of them.
> This allowed the LACP bond to fall back to a standard bond when
> partnered with a non-LACP switch.
On Jan 12, 2012, at 4:14 PM, Ben Pfaff wrote:
> As long as I'm suggesting changes, I know that I suggested using xxx
> in dates earlier, because we don't know when we're going to release in
> advance, but it makes Debian package builders unhappy, so I now think
> that it would be best to put real
On Thu, Jan 12, 2012 at 11:40:02AM -0800, Ethan Jackson wrote:
> Typically, when the lacp_negotiated flag changes, the result of
> bond_is_tcp_hash() will change as well. However, for certain bond
> types (SLB, Active-Backup, etc) this may not be true. Since the
> lacp_negotiated flag affects the
On Thu, Jan 12, 2012 at 04:25:33PM -0800, Ethan Jackson wrote:
> > Hmm. ?The "bond - single port" test is a bit paradoxical seeing as a
> > single port isn't a bond. ?I guess it illustrates that LACP works with
> > a single port too.
>
> Yes, I wanted to verify that a bond doesn't get created in t
> Hmm. The "bond - single port" test is a bit paradoxical seeing as a
> single port isn't a bond. I guess it illustrates that LACP works with
> a single port too.
Yes, I wanted to verify that a bond doesn't get created in this case.
Would you prefer I rename it, or change the test in some way?
On Thu, Jan 12, 2012 at 11:40:01AM -0800, Ethan Jackson wrote:
> With this patch, when a user attempts to configure LACP with an
> invalid system ID, OVS will fail to create the bond and warn.
> This behavior seems safer then defaulting to the bridge Ethernet
> Address which may surprise users.
>
Looks good.
Ethan
On Wed, Nov 30, 2011 at 15:08, Ben Pfaff wrote:
> At one point in the past, there were three separate queues between the
> kernel module and OVS userspace, each of which corresponded to a Netlink
> socket (or, before that, to a character device). It made sense to allow
> each
On Thu, Jan 12, 2012 at 11:40:00AM -0800, Ethan Jackson wrote:
> This commit makes the LACP unit tests more general by adding
> ovs-appctl bond/show output.
>
> Signed-off-by: Ethan Jackson
Hmm. The "bond - single port" test is a bit paradoxical seeing as a
single port isn't a bond. I guess it
Looks good.
Ethan
On Tue, Dec 27, 2011 at 13:37, Ben Pfaff wrote:
> The previous calculation was wrong when n_links was a power of 2.
>
> Reported-by: Paul Ingram
> Signed-off-by: Ben Pfaff
> ---
> lib/multipath.c | 2 +-
> lib/util.c | 8
> lib/util.h | 1 +
> 3 f
On Thu, Jan 12, 2012 at 04:11:20PM -0800, Justin Pettit wrote:
> On Jan 12, 2012, at 4:06 PM, Ben Pfaff wrote:
>
> > Also looks good, thanks.
> >
> > Should we also update debian/changelog at the same time? I see that
> > has not been our practice in the past, but I don't know why we
> > shouldn
On Jan 12, 2012, at 4:06 PM, Ben Pfaff wrote:
> Also looks good, thanks.
>
> Should we also update debian/changelog at the same time? I see that
> has not been our practice in the past, but I don't know why we
> shouldn't.
How about the following incremental?
diff --git a/debian/changelog b/de
On Thu, Jan 12, 2012 at 11:39:59AM -0800, Ethan Jackson wrote:
> This will simplify unit tests which appear in future commits.
>
> Signed-off-by: Ethan Jackson
Looks good, thank you.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailm
On Thu, Jan 12, 2012 at 11:39:58AM -0800, Ethan Jackson wrote:
> Similarly to lacp/show and cfm/show, the bond/show ovs-appctl
> command now shows all bonds when given no arguments.
>
> Signed-off-by: Ethan Jackson
Looks good, thank you.
___
dev mailin
Also looks good, thanks.
Should we also update debian/changelog at the same time? I see that
has not been our practice in the past, but I don't know why we
shouldn't.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
---
NEWS |4
configure.ac |2 +-
2 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/NEWS b/NEWS
index c678e8e..b628e29 100644
--- a/NEWS
+++ b/NEWS
@@ -1,3 +1,7 @@
+post-v1.5.0
+
+
+
v1.5.0 - xx xxx
- Ope
Looks good to me.
On Thu, Jan 12, 2012 at 03:54:21PM -0800, Justin Pettit wrote:
> ---
> NEWS |2 +-
> configure.ac |2 +-
> debian/changelog | 17 +
> 3 files changed, 19 insertions(+), 2 deletions(-)
>
> diff --git a/NEWS b/NEWS
> index 1128af0..c678e8
---
NEWS |2 +-
configure.ac |2 +-
debian/changelog | 17 +
3 files changed, 19 insertions(+), 2 deletions(-)
diff --git a/NEWS b/NEWS
index 1128af0..c678e8e 100644
--- a/NEWS
+++ b/NEWS
@@ -1,4 +1,4 @@
-post-v1.4.0
+v1.5.0 - xx xxx
---
All of this sounds reasonable to me. Thanks for writing this up.
Ethan
On Thu, Jan 12, 2012 at 15:47, Ben Pfaff wrote:
> On Thu, Jan 12, 2012 at 03:17:56PM -0800, Ethan Jackson wrote:
>> My compiler complains about this patch:
>>
>> ofproto/ofproto.c: In function ?ofoperation_complete?:
>> ofpr
On Thu, Jan 12, 2012 at 03:17:56PM -0800, Ethan Jackson wrote:
> My compiler complains about this patch:
>
> ofproto/ofproto.c: In function ?ofoperation_complete?:
> ofproto/ofproto.c:3157:5: error: comparison of unsigned expression >=
> 0 is always true [-Werror=type-limits]
I guess that's the l
On Thu, Jan 12, 2012 at 3:09 AM, Benoit ML wrote:
> Hello,
>
> Thanks a lot for the patch.
>
> With 1.3.0 compilation seems to work (not tested any much).
>
> But with 1.2.2 compilation failed :(
> ==
> /usr/local/src/openvswitch-1.2.2/datapath/l
On Thu, Jan 12, 2012 at 9:21 AM, Ben Pfaff wrote:
> On Wed, Jan 11, 2012 at 03:12:16PM -0800, Jesse Gross wrote:
>> On kernels 2.6.37 and above IFF_OVS_DATAPATH has a unique value upstream,
>> on 2.6.36 we define it to be IFF_BRIDGE_DATAPATH and below to 0. There
>> isn't really a good reason to
On Wed, Jan 11, 2012 at 01:43:08PM -0800, Ethan Jackson wrote:
> > +choose_port(struct dpif *dpif, struct netdev *netdev)
> > {
> > struct dp_netdev *dp = get_dp_netdev(dpif);
> > int port_no;
> >
> > +if (dpif->dpif_class == &dpif_dummy_class) {
> > +/* If the port name contai
Are you looking for
lists for your email campaign?
We have what you need.
Healthcare Lists including:
*Physicians
*Chirporactors
*Dentists
*Veterinarians
On Wed, Jan 11, 2012 at 07:25:52PM -0800, Pravin B Shelar wrote:
> Following patch implements dec_ttl as vendor action with similar
> semantics as OpenFlow 1.2. If TTL reaches zero while procession actions
> in current table, the remaining actions in previous tables are
> processed. An configuratio
Argh, this breaks "make install".
I'll work on a v3.
On Thu, Jan 12, 2012 at 10:25:11AM -0800, Romain Lenglet wrote:
> Thanks, it looks good!
>
> On Thu, 12 Jan 2012 10:01:39 -0800, Ben Pfaff wrote:
> > Thanks to Romain Lenglet for reminding me that there are standard
> Makefile
> > targets for
As promised, post version 1.4 the default bond_mode is changing to
active-backup with this commit.
Signed-off-by: Ethan Jackson
---
NEWS |3 +++
tests/bond.at|1 -
vswitchd/bridge.c| 11 +--
vswitchd/vswitch.xml |2 +-
4 files changed, 5 insertio
In the original Open vSwitch LACP implementation, when no slaves
found a LACP partner, the LACP module would attach all of them.
This allowed the LACP bond to fall back to a standard bond when
partnered with a non-LACP switch. In practice, this has caused
confusion with marginal benefit, so this f
Typically, when the lacp_negotiated flag changes, the result of
bond_is_tcp_hash() will change as well. However, for certain bond
types (SLB, Active-Backup, etc) this may not be true. Since the
lacp_negotiated flag affects the results of
bond_check_admissibility() revalidation needs to happen whe
With this patch, when a user attempts to configure LACP with an
invalid system ID, OVS will fail to create the bond and warn.
This behavior seems safer then defaulting to the bridge Ethernet
Address which may surprise users.
Bug #8710.
Signed-off-by: Ethan Jackson
---
vswitchd/bridge.c | 11 ++
This commit makes the LACP unit tests more general by adding
ovs-appctl bond/show output.
Signed-off-by: Ethan Jackson
---
tests/automake.mk |2 +-
tests/{lacp.at => bond.at} | 26 +-
tests/testsuite.at |2 +-
3 files changed, 23 insertions(+),
This will simplify unit tests which appear in future commits.
Signed-off-by: Ethan Jackson
---
lib/bond.c | 13 +
1 files changed, 13 insertions(+), 0 deletions(-)
diff --git a/lib/bond.c b/lib/bond.c
index 0daecf0..79fc18c 100644
--- a/lib/bond.c
+++ b/lib/bond.c
@@ -32,6 +32,7 @
Similarly to lacp/show and cfm/show, the bond/show ovs-appctl
command now shows all bonds when given no arguments.
Signed-off-by: Ethan Jackson
---
lib/bond.c | 67 +++
vswitchd/ovs-vswitchd.8.in | 18 ++--
2 files changed, 51 i
Thank you. I pushed patches 1 and 2.
On Wed, Jan 11, 2012 at 02:46:16PM -0800, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Wed, Dec 14, 2011 at 10:01, Ben Pfaff wrote:
> > This makes it possible to add entries for decoding OpenFlow messages with
> > newer versions, e.g. OpenFlow 1.1 or
Thanks, it looks good!
On Thu, 12 Jan 2012 10:01:39 -0800, Ben Pfaff wrote:
> Thanks to Romain Lenglet for reminding me that there are standard
Makefile
> targets for generating and installing PDF documentation.
>
> CC: Romain Lenglet
> Reported-by: Alan Shieh
> Bug #8153.
> ---
> v1->v2: Use
You're right.
I sent out a v2.
On Tue, Jan 10, 2012 at 10:44:24PM -0800, Romain Lenglet wrote:
> Why do you use a custom target instead of the standard "pdf" target?
> http://www.gnu.org/prep/standards/standards.html#Standard-Targets
>
> The standard place to install PDF documentation (only by t
Thanks to Romain Lenglet for reminding me that there are standard Makefile
targets for generating and installing PDF documentation.
CC: Romain Lenglet
Reported-by: Alan Shieh
Bug #8153.
---
v1->v2: Use standard 'pdf' target instead of invented 'pdf-man' target.
Add 'install-pdf' implementation.
On Wed, Jan 11, 2012 at 03:12:16PM -0800, Jesse Gross wrote:
> On kernels 2.6.37 and above IFF_OVS_DATAPATH has a unique value upstream,
> on 2.6.36 we define it to be IFF_BRIDGE_DATAPATH and below to 0. There
> isn't really a good reason to use IFF_BRIDGE_DATAPATH on 2.6.36 (and it's
> perhaps ba
On Thu, Jan 12, 2012 at 04:02:05PM +0900, Yoshi Tamura wrote:
> I noticed a weird behavior in openvswitch 1.3.0 that when I install a
> flow with mod_dl_dst, the packet gets corrupted. I created a deb
> package from the official 1.3.0 release and installed to my machine.
Thank you for the report.
54 matches
Mail list logo
