On Tue, Jun 12, 2012 at 09:42:10PM -0700, Ben Pfaff wrote:
> On Wed, Jun 13, 2012 at 01:40:19PM +0900, Simon Horman wrote:
> > On Tue, Jun 12, 2012 at 09:27:51PM -0700, Ben Pfaff wrote:
> > > On Wed, Jun 13, 2012 at 09:25:19AM +0900, Simon Horman wrote:
> > > > On Tue, Jun 12, 2012 at 12:32:14AM -0
On Wed, Jun 13, 2012 at 01:40:19PM +0900, Simon Horman wrote:
> On Tue, Jun 12, 2012 at 09:27:51PM -0700, Ben Pfaff wrote:
> > On Wed, Jun 13, 2012 at 09:25:19AM +0900, Simon Horman wrote:
> > > On Tue, Jun 12, 2012 at 12:32:14AM -0700, Ben Pfaff wrote:
> > > > "enum ofperr" is clearer than "int".
On Tue, Jun 12, 2012 at 09:27:51PM -0700, Ben Pfaff wrote:
> On Wed, Jun 13, 2012 at 09:25:19AM +0900, Simon Horman wrote:
> > On Tue, Jun 12, 2012 at 12:32:14AM -0700, Ben Pfaff wrote:
> > > "enum ofperr" is clearer than "int".
> > >
> > > Signed-off-by: Ben Pfaff
> >
> > Reviewed-by: Simon Hor
On Wed, Jun 13, 2012 at 09:25:19AM +0900, Simon Horman wrote:
> On Tue, Jun 12, 2012 at 12:32:14AM -0700, Ben Pfaff wrote:
> > "enum ofperr" is clearer than "int".
> >
> > Signed-off-by: Ben Pfaff
>
> Reviewed-by: Simon Horman
Thanks for all the reviews. I pushed patches 1 through 10 to maste
Thanks Pravin for the review comments. Inline
On Tue, Jun 12, 2012 at 5:31 PM, Pravin Shelar wrote:
> Hi Ravi,
>
> I see following issues with datapath mpls support:
>
> Actions:
> I do not see skb->csum updated when mpls action is performed on packet.
csum calculation is usually done in the k
On Tue, Jun 12, 2012 at 07:05:55PM -0700, Ben Pfaff wrote:
> On Wed, Jun 13, 2012 at 10:37:17AM +0900, Simon Horman wrote:
> > On Tue, Jun 12, 2012 at 06:25:40PM -0700, Ben Pfaff wrote:
> > > On Wed, Jun 13, 2012 at 09:35:14AM +0900, Simon Horman wrote:
> > > > On Tue, Jun 12, 2012 at 12:32:05AM -0
On Tue, Jun 12, 2012 at 12:32:24AM -0700, Ben Pfaff wrote:
> A work in progress. Doesn't build.
>
> I'm not really sure I want separate OFPUTIL_ constants for OF1.0 and OF1.1
> versions of each message. It's something I'm trying out as I write the
> code.
Here are some trivial build fixes. I am
On Tue, Jun 12, 2012 at 12:32:23AM -0700, Ben Pfaff wrote:
> Probably incomplete.
The following seems to be needed:
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index 439096f..625b809 100644
--- a/lib/ofp-print.c
+++ b/lib/ofp-print.c
@@ -701,7 +701,7 @@ ofp_print_flow_mod(struct ds *s, const s
On Wed, Jun 13, 2012 at 10:37:17AM +0900, Simon Horman wrote:
> On Tue, Jun 12, 2012 at 06:25:40PM -0700, Ben Pfaff wrote:
> > On Wed, Jun 13, 2012 at 09:35:14AM +0900, Simon Horman wrote:
> > > On Tue, Jun 12, 2012 at 12:32:05AM -0700, Ben Pfaff wrote:
> > > > OFPFW_DL_SRC and OFPFW_DL_DST don't e
On Tue, Jun 12, 2012 at 06:25:40PM -0700, Ben Pfaff wrote:
> On Wed, Jun 13, 2012 at 09:35:14AM +0900, Simon Horman wrote:
> > On Tue, Jun 12, 2012 at 12:32:05AM -0700, Ben Pfaff wrote:
> > > OFPFW_DL_SRC and OFPFW_DL_DST don't exist in OpenFlow 1.1. Replace them
> > > by the correct enums.
> > >
On Wed, Jun 13, 2012 at 09:35:14AM +0900, Simon Horman wrote:
> On Tue, Jun 12, 2012 at 12:32:05AM -0700, Ben Pfaff wrote:
> > OFPFW_DL_SRC and OFPFW_DL_DST don't exist in OpenFlow 1.1. Replace them
> > by the correct enums.
> >
> > Most of the change here is due to respacing since DL_VLAN_PCP is
On Tue, Jun 12, 2012 at 09:59:36AM -0700, Ben Pfaff wrote:
> On Mon, Jun 04, 2012 at 08:25:45AM +0900, Simon Horman wrote:
> > * The way that I have split the patchs is still somewhat arbitrary.
> > I wanted to avoid one very large patch to aid review. But a lot of the
> > chagnes are inter-re
On Tue, Jun 12, 2012 at 12:32:18AM -0700, Ben Pfaff wrote:
> For some time, Open vSwitch has combined both ofp_stats_msg and
> the data that follows it into single structures, because we found
> that this is usually easier to handle.than having them separated.
> But OpenFlow 1.1+ has a different st
Hello dear
Am Miss jasmine, sorry for my embarrassment to your email
I
was only interested and wanted to know if we can develop a good lovely
relationship. I wish if it is the will of God we can move from here. But
For better presentation kindly direct in my mail address
On Tue, Jun 12, 2012 at 12:32:05AM -0700, Ben Pfaff wrote:
> OFPFW_DL_SRC and OFPFW_DL_DST don't exist in OpenFlow 1.1. Replace them
> by the correct enums.
>
> Most of the change here is due to respacing since DL_VLAN_PCP is one
> character wider than any previous name.
>
> This doesn't fix a r
On Tue, Jun 12, 2012 at 12:32:19AM -0700, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> include/openflow/nicira-ext.h | 10 ++--
> include/openflow/openflow-1.0.h| 61 +++---
> include/openflow/openflow-common.h | 49 +-
Hi Ravi,
I see following issues with datapath mpls support:
Actions:
I do not see skb->csum updated when mpls action is performed on packet.
Push / Set_lse action:
You can compute mpls_lse value flow setup time in userspace. That will
save time in every packet switching plus it simplifies code.
On Tue, Jun 12, 2012 at 12:32:17AM -0700, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
Reviewed-by: Simon Horman
> lib/ofp-util.c | 52
> lib/ofp-util.h |8 +---
> 2 files changed, 1 insertions(+), 59 deletions(-)
On Tue, Jun 12, 2012 at 12:32:15AM -0700, Ben Pfaff wrote:
> Possibly the ofputil_decode_action() interface should be rethought now
> that ofputil_action_code has an invalid value.
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> lib/ofp-parse.c|3 +++
> lib/ofp-prin
On Tue, Jun 12, 2012 at 12:32:14AM -0700, Ben Pfaff wrote:
> "enum ofperr" is clearer than "int".
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> lib/ofp-util.c |2 +-
> lib/ofp-util.h |4 ++--
> 2 files changed, 3 insertions(+), 3 deletions(-)
_
On Tue, Jun 12, 2012 at 12:32:13AM -0700, Ben Pfaff wrote:
> So that I don't have to figure it out yet again.
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> include/openflow/openflow-1.0.h |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
__
On Tue, Jun 12, 2012 at 12:32:12AM -0700, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> lib/ofp-util.c | 26 --
> lib/ofp-util.h |4
> 2 files changed, 0 insertions(+), 30 deletions(-)
__
On Tue, Jun 12, 2012 at 12:32:11AM -0700, Ben Pfaff wrote:
> OF1.2 and later make these fields fully maskable so we might as well also.
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> NEWS |4 ++--
> include/openflow/nicira-ext.h |6 --
> li
On Tue, Jun 12, 2012 at 12:32:10AM -0700, Ben Pfaff wrote:
> OF1.1 and later make these fields fully maskable so we might as well also.
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> NEWS |8 +++-
> include/openflow/nicira-ext.h |6 --
>
On Tue, Jun 12, 2012 at 12:32:09AM -0700, Ben Pfaff wrote:
> This is another step toward OpenFlow 1.1 support. The change does not
> affect any outwardly visible OpenFlow behavior yet.
>
> Signed-off-by: Ben Pfaff
With the caveat that I am not particularly familiar with the OpenFlow 1.1
match f
On Tue, Jun 12, 2012 at 12:32:08AM -0700, Ben Pfaff wrote:
> We need these for OpenFlow 1.1 ofp_match support even if we don't support
> MPLS matching (which we don't, yet).
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> lib/packets.h |2 ++
> 1 files changed, 2 insertions
On Tue, Jun 12, 2012 at 12:32:07AM -0700, Ben Pfaff wrote:
> SUSv3 doesn't require IPPROTO_SCTP so some systems might not provide it.
>
> IPPROTO_SCTP isn't used in the tree yet so this doesn't fix a real bug.
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> lib/packets.h |
On Tue, Jun 12, 2012 at 12:32:06AM -0700, Ben Pfaff wrote:
> This better fits our general policy of adding a version number suffix
> to structures and constants whose values differ from one OpenFlow
> version to the next.
>
> Signed-off-by: Ben Pfaff
Reviewed-by: Simon Horman
> ---
> include/
On Tue, Jun 12, 2012 at 12:32:05AM -0700, Ben Pfaff wrote:
> OFPFW_DL_SRC and OFPFW_DL_DST don't exist in OpenFlow 1.1. Replace them
> by the correct enums.
>
> Most of the change here is due to respacing since DL_VLAN_PCP is one
> character wider than any previous name.
>
> This doesn't fix a r
Looks good, thanks.
Ethan
On Tue, Jun 12, 2012 at 4:45 PM, Ben Pfaff wrote:
> Found by valgrind.
>
> Signed-off-by: Ben Pfaff
> ---
> lib/vlog.c | 9 -
> 1 files changed, 8 insertions(+), 1 deletions(-)
>
> diff --git a/lib/vlog.c b/lib/vlog.c
> index a7d9e48..899072e 100644
> --- a
Found by valgrind.
Signed-off-by: Ben Pfaff
---
lib/vlog.c |9 -
1 files changed, 8 insertions(+), 1 deletions(-)
diff --git a/lib/vlog.c b/lib/vlog.c
index a7d9e48..899072e 100644
--- a/lib/vlog.c
+++ b/lib/vlog.c
@@ -477,6 +477,7 @@ vlog_unixctl_reopen(struct unixctl_conn *conn, i
I am facing an issue in flow matching for incoming vlan packets in ovs
userspace. In lib/flow.c and flow_extract function, eth->eth_type(line
no 363) is eth_type_ip. From logs, I see
Jun 12 14:43:43 rkerur-thinkpad-t500 ovs-vswitchd:
in_port(1),eth(src=00:27:13:67:b9:9b,dst=00:24:7e:68:ae:46),eth_
Accepted:
openvswitch-brcompat_1.4.2+git20120612-1_i386.deb
to main/o/openvswitch/openvswitch-brcompat_1.4.2+git20120612-1_i386.deb
openvswitch-common_1.4.2+git20120612-1_i386.deb
to main/o/openvswitch/openvswitch-common_1.4.2+git20120612-1_i386.deb
openvswitch-controller_1.4.2+git20120612-1
openvswitch_1.4.2+git20120612-1_i386.changes uploaded successfully to localhost
along with the files:
openvswitch_1.4.2+git20120612-1.dsc
openvswitch_1.4.2+git20120612.orig.tar.xz
openvswitch_1.4.2+git20120612-1.debian.tar.gz
openvswitch-common_1.4.2+git20120612-1_i386.deb
openvswitch-swi
I guess it's a relic from 2009 or so, when we actually passed
credentials over Unix domain sockets. When I removed that code I must
have missed this part.
On Tue, Jun 12, 2012 at 12:57:05PM -0700, Ethan Jackson wrote:
> Looks good thanks.
>
> Do you happen to know why punix_open() was passing tr
On Tue, Jun 12, 2012 at 12:13 PM, Ben Pfaff wrote:
> On Tue, Jun 12, 2012 at 11:40:22AM -0700, ravi kerur wrote:
>> On Tue, Jun 12, 2012 at 11:11 AM, Ben Pfaff wrote:
>> > Ravi, I'm more or less happy with the userspace code here. I mostly
>> > have style kinds of concerns. (I haven't fully scr
Looks good, thanks.
Ethan
On Tue, May 22, 2012 at 5:19 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> lib/fatal-signal.c | 4 ++--
> lib/process.c | 6 +++---
> lib/signals.c | 6 +++---
> lib/socket-util.c | 8
> lib/socket-util.h | 2 ++
> 5 files
Looks good thanks.
Do you happen to know why punix_open() was passing true for the parameter()?
Ethan
On Tue, May 22, 2012 at 5:19 PM, Ben Pfaff wrote:
> Nothing in the tree ever tries to send or receive credentials over a Unix
> domain socket so there's no point in configuring them to be recei
I am currently debugging this and had planned to send an email asking
for suggestions when code review was nearing completion. This bug has
been elusive for a long time. The packet format themselves looked fine
as I have found a way to test them and have successfully tested them.
The problem is th
On Tue, Jun 12, 2012 at 11:40:22AM -0700, ravi kerur wrote:
> On Tue, Jun 12, 2012 at 11:11 AM, Ben Pfaff wrote:
> > Ravi, I'm more or less happy with the userspace code here. I mostly
> > have style kinds of concerns. (I haven't fully scrutinized every line
> > though.)
> >
>
> is it coding s
On Mon, Jun 11, 2012 at 7:24 PM, Jesse Gross wrote:
> On Jun 12, 2012, at 10:13 AM, Pravin B Shelar wrote:
>
>> Fix return check typo.
>>
>> Signed-off-by: Pravin B Shelar
>>
>> Bug #11933
>> ---
>> datapath/datapath.c | 6 --
>> 1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff -
Thanks, I added that.
On Tue, Jun 12, 2012 at 11:17:24AM -0700, Ethan Jackson wrote:
> One thing I noticed btw, The case where the user has looped the
> network has often been reported as "my cpu usage is high". It may be
> worth calling that out explicitly.
>
> Ethan
>
> On Tue, Jun 12, 2012 a
One thing I noticed btw, The case where the user has looped the
network has often been reported as "my cpu usage is high". It may be
worth calling that out explicitly.
Ethan
On Tue, Jun 12, 2012 at 9:39 AM, Ben Pfaff wrote:
> On Mon, Jun 11, 2012 at 10:47:14PM -0700, Justin Pettit wrote:
>> Thi
LinkedIn
I'd like to add you to my professional network on LinkedIn.
- Nasir
Nasir Hussain
Telecommunications Professional
Prato Area, Italy
Confirm that you know Nasir Hussain:
https://www.linkedin.com/e/e2m9gr-h3daitd2-2i/isd/7463007830/YQNnPK_c/?hs=false&tok=2P6k73K2CCG5g1
--
Oh, I saw one testsuite failure also:
448: ofproto-dpif - controller FAILED
(ofproto-dpif.at:323)
Does this pass for you.
On Tue, Jun 12, 2012 at 11:11 AM, Ben Pfaff wrote:
> Ravi, I'm more or less happy with the userspace code here. I mostly
> have style kinds of concerns.
On Mon, Jun 04, 2012 at 08:25:45AM +0900, Simon Horman wrote:
> * The way that I have split the patchs is still somewhat arbitrary.
> I wanted to avoid one very large patch to aid review. But a lot of the
> chagnes are inter-related, so a bisectable split seems rather difficult.
> None the l
On Mon, Jun 11, 2012 at 10:47:14PM -0700, Justin Pettit wrote:
> This is great, Ben! Thank you for doing it!
>
> > + and as soon as I ran the "add-port" command I lost all connectivity
> > + through eth0. Help!
>
> I like the sense of urgency you've added. Maybe you missed your
> calling i
I reposted all of these patches starting here:
http://openvswitch.org/pipermail/dev/2012-June/017791.html
Please review them in the new thread.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
I reposted all of these patches starting here:
http://openvswitch.org/pipermail/dev/2012-June/017798.html
Please review them in the new thread.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
A work in progress. Doesn't build.
I'm not really sure I want separate OFPUTIL_ constants for OF1.0 and OF1.1
versions of each message. It's something I'm trying out as I write the
code.
Signed-off-by: Ben Pfaff
---
include/openflow/openflow-1.1.h| 17 ++--
include/openflow/openflow-com
Probably incomplete.
Signed-off-by: Ben Pfaff
---
include/openflow/nicira-ext.h |4 ++--
include/openflow/openflow-1.0.h| 19 ---
include/openflow/openflow-common.h | 14 ++
lib/ofp-print.c|4 ++--
lib/ofp-print.h
This is a start at extending the stats message helpers to support OF1.1
stats messages too. It isn't complete, doesn't compile, and I'm not fond
of the approach.
Signed-off-by: Ben Pfaff
---
lib/ofp-util.c | 74 ---
lib/ofp-util.h |7 +++
There's an upcoming need to decode yet another variant of OpenFlow
messages: OpenFlow 1.1 statistics requests and replies. The code for
decoding messages types is getting unmanageable. This commit refactors it
to make it easier to maintain and extend.
Before this is committed it needs the follow
This builds and doesn't break any existing unit tests, but nothing
calls it yet (and it doesn't have any unit tests of its own) so it
can hardly be called useful.
Left to do:
- Instructions other than Apply-Actions.
- Actions other than the ones that have identical semantics to
Ope
Signed-off-by: Ben Pfaff
---
include/openflow/nicira-ext.h | 10 ++--
include/openflow/openflow-1.0.h| 61 +++---
include/openflow/openflow-common.h | 49 +-
lib/ofp-print.c| 10 ++--
lib/ofp-util.c | 100 ++
For some time, Open vSwitch has combined both ofp_stats_msg and
the data that follows it into single structures, because we found
that this is usually easier to handle.than having them separated.
But OpenFlow 1.1+ has a different stats message header (it has 4
extra bytes of padding) whereas many o
Signed-off-by: Ben Pfaff
---
lib/ofp-util.c | 52
lib/ofp-util.h |8 +---
2 files changed, 1 insertions(+), 59 deletions(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 7ff46ab..e80fe18 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp
Possibly the ofputil_decode_action() interface should be rethought now
that ofputil_action_code has an invalid value.
Signed-off-by: Ben Pfaff
---
lib/ofp-parse.c|3 +++
lib/ofp-print.c|3 +++
lib/ofp-util.c |7 +++
lib/ofp-util.h |3 ++-
ofpro
"enum ofperr" is clearer than "int".
Signed-off-by: Ben Pfaff
---
lib/ofp-util.c |2 +-
lib/ofp-util.h |4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 07f885d..6d820b2 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -2318,
So that I don't have to figure it out yet again.
Signed-off-by: Ben Pfaff
---
include/openflow/openflow-1.0.h |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/include/openflow/openflow-1.0.h b/include/openflow/openflow-1.0.h
index 8afd9a1..831f6a9 100644
--- a/include/op
Signed-off-by: Ben Pfaff
---
lib/ofp-util.c | 26 --
lib/ofp-util.h |4
2 files changed, 0 insertions(+), 30 deletions(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 5bd220b..07f885d 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -3414,32 +3414,6 @@
OF1.2 and later make these fields fully maskable so we might as well also.
Signed-off-by: Ben Pfaff
---
NEWS |4 ++--
include/openflow/nicira-ext.h |6 --
lib/meta-flow.c | 25 +++--
lib/meta-flow.h |1 -
OF1.1 and later make these fields fully maskable so we might as well also.
Signed-off-by: Ben Pfaff
---
NEWS |8 +++-
include/openflow/nicira-ext.h |6 --
lib/meta-flow.c |8
lib/ofp-util.c| 17 ++--
This is another step toward OpenFlow 1.1 support. The change does not
affect any outwardly visible OpenFlow behavior yet.
Signed-off-by: Ben Pfaff
---
lib/ofp-util.c| 261 +
lib/ofp-util.h|7 +
tests/ovs-ofctl.at| 309 +++
We need these for OpenFlow 1.1 ofp_match support even if we don't support
MPLS matching (which we don't, yet).
Signed-off-by: Ben Pfaff
---
lib/packets.h |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/lib/packets.h b/lib/packets.h
index 19ddd9a..ad5631d 100644
--- a/li
SUSv3 doesn't require IPPROTO_SCTP so some systems might not provide it.
IPPROTO_SCTP isn't used in the tree yet so this doesn't fix a real bug.
Signed-off-by: Ben Pfaff
---
lib/packets.h |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/lib/packets.h b/lib/packets.h
i
This better fits our general policy of adding a version number suffix
to structures and constants whose values differ from one OpenFlow
version to the next.
Signed-off-by: Ben Pfaff
---
include/openflow/nicira-ext.h |6 +-
include/openflow/openflow-1.0.h | 58 +++--
lib/f
OFPFW_DL_SRC and OFPFW_DL_DST don't exist in OpenFlow 1.1. Replace them
by the correct enums.
Most of the change here is due to respacing since DL_VLAN_PCP is one
character wider than any previous name.
This doesn't fix a real bug because these constants didn't have any users
in the tree.
Signe
This is a rollup of all the outstanding, uncommitted patches I have
posted that relate to OF1.1 and OF1.2 support, plus a few more in
various states of completeness.
It begins with the ofp11_match v2 series most recently posted on
June 9. I believe that these are unchanged from the previous posti
69 matches
Mail list logo
