On Tue, Jun 26, 2012 at 11:53:43PM -0700, Ben Pfaff wrote:
> The first five patches here are ready for review. The rest need
> work. I've indicated changes from v2 in the individual patches.
>
> This is also available from git://benpfaff.org/openvswitch in the
> of1.1 branch.
I pushed refinemen
Never mind, I figured it out. Thanks for the inputs.
On Wed, Jun 27, 2012 at 11:11 AM, ravi kerur wrote:
> Thanks Ethan. Core file generation is enabled. However, I do not see
> any core files under failed test directory i.e
> tests/testsuite.dir/0449. Is the generated core file available in that
On Thu, Jun 28, 2012 at 01:47:42PM +0900, Isaku Yamahata wrote:
> Define NXAST_ACTION instead of listing all OFPUTIL_NXAST_xxx.
>
> Signed-off-by: Isaku Yamahata
That's a good idea, thank you.
I folded this in to "ofp-actions: Add decoding and encoding OF1.1
instructions and actions." and trans
On Thu, Jun 28, 2012 at 05:43:28PM +1200, Joe Stringer wrote:
> On 28 June 2012 17:32, Ben Pfaff wrote:
> > On Thu, Jun 28, 2012 at 05:30:33PM +1200, Joe Stringer wrote:
> >> Reviewed-by: Joe Stringer
> >> Signed-off-by: Ben Pfaff
> >> ---
> >> v4: Fix build errors, test failures.
> >
> > Er, ye
On 28 June 2012 17:32, Ben Pfaff wrote:
> On Thu, Jun 28, 2012 at 05:30:33PM +1200, Joe Stringer wrote:
>> Reviewed-by: Joe Stringer
>> Signed-off-by: Ben Pfaff
>> ---
>> v4: Fix build errors, test failures.
>
> Er, yeah, oops. I had already applied the same change locally, sorry
> for wasting
On Thu, Jun 28, 2012 at 05:30:33PM +1200, Joe Stringer wrote:
> Reviewed-by: Joe Stringer
> Signed-off-by: Ben Pfaff
> ---
> v4: Fix build errors, test failures.
Er, yeah, oops. I had already applied the same change locally, sorry
for wasting your time.
_
Reviewed-by: Joe Stringer
Signed-off-by: Ben Pfaff
---
v4: Fix build errors, test failures.
---
include/openflow/openflow-1.0.h|4 +---
include/openflow/openflow-1.1.h|4 +---
include/openflow/openflow-common.h |4
lib/ofp-util.c | 10 +-
4
What do you mean by hands-on? From NOX? I don't think it speaks OVSDB. I'd
probably start off by using ovs-vsctl and ovs-ofctl to manually configure what
you want. Translating from ovs-ofctl to OpenFlow should be pretty
straight-forward, since the commands map pretty directly to OpenFlow. F
Thank you very much Mr. Petit for the prompt reply.
Do you know any hand-on example to implement that?
Trinh Minh Tri
-
PhD Student, Department of Electrical Engineering
Chulalongkorn University, Bangkok 10330, Thailand.
Phone:0824566602
Email: no.
The configuration of QoS is done through the OVSDB protocol, which handle
switch-level configuration. You can then use OpenFlow to direct flows to
particular queues that you've configured through OVSDB. There are C and Python
bindings to OVSDB (but I don't think NOX makes use of them) or you c
The logic of ofpacts_from_openflow11() should be same to
ofpacts_from_openflow10(), so consolidate them. Thus the issue is fixed.
Signed-off-by: Isaku Yamahata
---
lib/ofp-actions.c | 44 +++-
1 files changed, 15 insertions(+), 29 deletions(-)
diff --gi
Define NXAST_ACTION instead of listing all OFPUTIL_NXAST_xxx.
Signed-off-by: Isaku Yamahata
---
lib/ofp-actions.c | 45 +
1 files changed, 5 insertions(+), 40 deletions(-)
diff --git a/lib/ofp-actions.c b/lib/ofp-actions.c
index d527dcd..ba9e61c 100
Dear openvswitch dev!
I'm trying to implement a system which can dynamically control the flow rate
limiting of openvswitch. And would like control that rate from NOX controller
via openflow protocol.
Could you suggest me how to implement that system and is there any example
about that system t
On Wed, Jun 27, 2012 at 07:28:46AM -0700, Ben Pfaff wrote:
> On Wed, Jun 27, 2012 at 05:19:36PM +0900, Simon Horman wrote:
> > this series is based on v2 of your Open Flow 1.1 work.
> > Please let me know if you feel that I should rebase on top of v3.
>
> It's probably not necessary, because my v3
On Wed, Jun 27, 2012 at 7:15 PM, Learner Study wrote:
> Hello,
>
> I'm using ovs and would like to use my own algorithm for hash related
> operations in ovs code base.
>
> I have identified that ovs_flow_hash() makes calls to jhash2()
> function inside Linux kernel. Would it sufficient to replace
Related -- I've been looking into implementing OFPIT_WRITE_METADATA on
top of your work at git://benpfaff.org/openvswitch on the of1.1
branch, and I have some related thoughts.
So, with the above patch, I have used reg_load to load (fully-masked)
metadata. Am I correct in understanding that reg_lo
Hello,
I'm using ovs and would like to use my own algorithm for hash related
operations in ovs code base.
I have identified that ovs_flow_hash() makes calls to jhash2()
function inside Linux kernel. Would it sufficient to replace jhash2
function w/ my own hash function (with exactly same semantic
On Wed, Jun 27, 2012 at 02:25:55PM -0700, Pravin Shelar wrote:
> >> I am seeing this test failure on my machine.
> >
> > OK, what's in the test's testsuite.log?
> >
> > I've reposted these two patches now.
>
> I do not see any failure with new patches.
Oh, I see. Thanks, I'll push the new versio
You mentioned test failures before, but I didn't actually change
anything in this version. Do you still get test failures?
Thanks,
Ben.
On Wed, Jun 27, 2012 at 02:26:13PM -0700, Pravin Shelar wrote:
> Looks good.
>
> On Wed, Jun 27, 2012 at 10:05 AM, Ben Pfaff wrote:
> > Until now, Open vSwit
Looks good.
Thanks,
Pravin.
On Wed, Jun 27, 2012 at 10:05 AM, Ben Pfaff wrote:
> dump_stats_transaction() ignored errors and other non-stats replies to
> its request and would continue to wait forever. This fixes the problem.
>
> Signed-off-by: Ben Pfaff
> ---
> v1 was here: http://openvswitch
Looks good.
On Wed, Jun 27, 2012 at 10:05 AM, Ben Pfaff wrote:
> Until now, Open vSwitch has ignored missing ports and queues in most cases
> in queue stats requests, simply returning an empty set of statistics.
> It seems that it is better to report an error, so this commit does this.
>
> Report
On Wed, Jun 27, 2012 at 10:05 AM, Ben Pfaff wrote:
> On Tue, Jun 26, 2012 at 04:03:32PM -0700, Pravin Shelar wrote:
>> On Mon, Jun 11, 2012 at 11:26 AM, Ben Pfaff wrote:
>> > Until now, Open vSwitch has ignored missing ports and queues in most cases
>> > in queue stats requests, simply returning
I found one patch that I was able to apply cleanly to master and
push (I followed up separately to that one). If there are any other
patches that are well suited for this (or that could easily be made
suited for it) let me know, so that we can pick off a little of this
stuff early.
Thanks,
Ben.
On Wed, Jun 27, 2012 at 05:19:52PM +0900, Simon Horman wrote:
> Signed-off-by: Simon Horman
I pushed this one to master as follows:
--8<--cut here-->8--
From: Simon Horman
Date: Wed, 27 Jun 2012 17:19:52 +0900
Subject: [PATCH] openflow: Add Open
Thanks, looks good.
Ethan
On Wed, Jun 27, 2012 at 11:25 AM, Ben Pfaff wrote:
> Writes to regular files under Unix-like kernels, including Linux, typically
> block until the write is complete, regardless of O_NONBLOCK. When the I/O
> subsystem is busy, this can cause indefinite delays. We have
All of this makes sense, thanks.
Ethan
On Wed, Jun 27, 2012 at 11:25 AM, Ben Pfaff wrote:
> ovs-vswitchd is effectively a "soft real-time" process, because flows that
> do not get set up quickly lead to packet loss or retransmission. We've
> done our best to keep it from blocking unnecessarily,
On Wed, Jun 27, 2012 at 01:44:00PM -0700, Ethan Jackson wrote:
> It's useful to know how long each system call took, and at what
> time each system call happened. In addition this patch causes
> strace to print strings more fully allowing log messages to be seen
> in the output.
>
> Signed-off-by
It's useful to know how long each system call took, and at what
time each system call happened. In addition this patch causes
strace to print strings more fully allowing log messages to be seen
in the output.
Signed-off-by: Ethan Jackson
---
utilities/ovs-lib.in |2 +-
1 file changed, 1 ins
The stats gathering can be time-consuming in some cases, so it's better to
do it in the worker process.
Signed-off-by: Ben Pfaff
---
vswitchd/bridge.c | 59 -
vswitchd/system-stats.c | 160 +--
vswitchd/system-stats.h |6 +-
"smap" is now the appropriate data structure for a string-to-string map.
Also changes ovsdb_datum_from_shash() into ovsdb_datum_from_smap() since
system-stats related code was the only client.
Signed-off-by: Ben Pfaff
---
lib/ovsdb-data.c| 19 ---
lib/ovsdb-data.h|
An upcoming commit will add a caller.
Signed-off-by: Ben Pfaff
---
lib/smap.c | 12
lib/smap.h |3 ++-
2 files changed, 14 insertions(+), 1 deletions(-)
diff --git a/lib/smap.c b/lib/smap.c
index e612ac7..ff78598 100644
--- a/lib/smap.c
+++ b/lib/smap.c
@@ -125,6 +125,18 @@ s
These will acquire a caller in an upcoming commit.
Signed-off-by: Ben Pfaff
---
lib/smap.c | 36
lib/smap.h |5 +
2 files changed, 41 insertions(+), 0 deletions(-)
diff --git a/lib/smap.c b/lib/smap.c
index 8665321..e612ac7 100644
--- a/lib/smap.c
Writes to regular files under Unix-like kernels, including Linux, typically
block until the write is complete, regardless of O_NONBLOCK. When the I/O
subsystem is busy, this can cause indefinite delays. We have actually
observed "write" calls sleep for 5 seconds or more for this reason.
Delegati
This will be easier to deal with in an upcoming commit that will use a
subprocess to do asynchronous log I/O.
Signed-off-by: Ben Pfaff
---
lib/vlog.c | 24
1 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/lib/vlog.c b/lib/vlog.c
index 108c84e..bcfec7b 1
ovs-vswitchd is effectively a "soft real-time" process, because flows that
do not get set up quickly lead to packet loss or retransmission. We've
done our best to keep it from blocking unnecessarily, but some operations
unavoidably block. This new library allows a daemon to break itself up
into a
Whereas VLOG_FATAL() eventually calls exit(1), VLOG_ABORT()
eventually calls abort(). The key difference is that abort()
will cause a "monitor" process to restart, where exit(1) will
cause it to exit along with the monitored process.
Signed-off-by: Ben Pfaff
---
lib/util.c |9 +++--
lib
This code will have another user in an upcoming commit.
Signed-off-by: Ben Pfaff
---
lib/daemon.c | 27 ++-
lib/daemon.h |1 +
2 files changed, 23 insertions(+), 5 deletions(-)
diff --git a/lib/daemon.c b/lib/daemon.c
index a2eff07..84ed614 100644
--- a/lib/daemon.
This function will be useful in an upcoming commit.
Signed-off-by: Ben Pfaff
---
lib/daemon.c | 33 +++--
lib/daemon.h |2 ++
2 files changed, 29 insertions(+), 6 deletions(-)
diff --git a/lib/daemon.c b/lib/daemon.c
index ecca606..a2eff07 100644
--- a/lib/daem
This will be more useful later when we introduces "worker" subprocesses.
I don't have any current plans to introduce threading, but I can't
think of a disadvantage to wording this in a general manner.
Signed-off-by: Ben Pfaff
---
NEWS |4 +++-
lib/daemon.c |
These will be used in upcoming commits.
This commit also adds corresponding definitions to the "sparse" header,
so that sparse still works.
Signed-off-by: Ben Pfaff
---
include/sparse/sys/socket.h | 31 +
lib/socket-util.c | 255 +++
lib/
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 changed, 18 insertions(+), 8 deletions(-)
diff --git a/lib/fatal-signal.c b/lib/fatal-signal.
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 received.
Signed-off-by: Ben Pfaff
---
lib/socket-util.c| 15 ++-
lib/socket-util.h|4 ++--
lib/stream-unix.c|4 ++--
ovs-vswitchd occasionally needs to do some activities that may block
in the kernel unavoidably. This patch series changes ovs-vswitchd to
do these activities in a separate "worker process".
This second version of the patch series applies feedback from Ethan
and revises patch 11 and 12 into patche
Thanks Ethan. Core file generation is enabled. However, I do not see
any core files under failed test directory i.e
tests/testsuite.dir/0449. Is the generated core file available in that
directory or some place else? or it could be that ovs-vswitchd didn't
segfault at all?
openvswitch# ulimit -a
c
This patch and patch 12 need some revision for the new "smap" data
structure. I'll send out revised versions in a bit.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Tue, Jun 26, 2012 at 06:31:33PM -0700, Ethan Jackson wrote:
> Looks good to me.
>
> I would prefer vlog_async_inited be initialized to false, though I'm
> not sure if the C sped requires it. You would know better than me.
It's not required (static variables are initialized to zero) but I
don'
On Tue, Jun 26, 2012 at 06:17:17PM -0700, Ethan Jackson wrote:
> Looks good to me, couple of questions.
>
> From worker_start() it looks to me like worker_main() never exists?
> What if a user wants to ask ovs-vswitchd to exit cleanly using appctl?
> Will the child process simply be killed?
work
On Tue, Jun 26, 2012 at 04:03:32PM -0700, Pravin Shelar wrote:
> On Mon, Jun 11, 2012 at 11:26 AM, Ben Pfaff wrote:
> > Until now, Open vSwitch has ignored missing ports and queues in most cases
> > in queue stats requests, simply returning an empty set of statistics.
> > It seems that it is bette
Until now, Open vSwitch has ignored missing ports and queues in most cases
in queue stats requests, simply returning an empty set of statistics.
It seems that it is better to report an error, so this commit does this.
Reported-by: Prabina Pattnaik
Signed-off-by: Ben Pfaff
---
v1 was here: http:/
dump_stats_transaction() ignored errors and other non-stats replies to
its request and would continue to wait forever. This fixes the problem.
Signed-off-by: Ben Pfaff
---
v1 was here: http://openvswitch.org/pipermail/dev/2012-June/017779.html
v2: Fix dumb assumption that all replies are queue s
On Tue, Jun 26, 2012 at 04:00:00PM -0700, Pravin Shelar wrote:
> On Mon, Jun 11, 2012 at 11:26 AM, Ben Pfaff wrote:
> > dump_stats_transaction() ignored errors and other non-stats replies to
> > its request and would continue to wait forever. This fixes the problem.
> >
> > Signed-off-by: Ben Pfa
Thanks, pushed to master.
Let me know if you need it on any other branches.
On Wed, Jun 27, 2012 at 05:30:13PM +0100, Dominic Curran wrote:
> Rob wrote it.
> I read through it and OK'd it.
>
> Thanks
> dom
>
> > -Original Message-
> > From: Ben Pfaff [mailto:b...@nicira.com]
> > Sent: 2
Rob wrote it.
I read through it and OK'd it.
Thanks
dom
> -Original Message-
> From: Ben Pfaff [mailto:b...@nicira.com]
> Sent: 27 June 2012 17:26
> To: Dominic Curran
> Cc: dev@openvswitch.org; Rob Hoes
> Subject: Re: [ovs-dev] [PATCH 1/1] xenserver: Improve efficiency of
> code by using
On Wed, Jun 27, 2012 at 04:14:21PM +0100, Dominic Curran wrote:
> Replace the get_record() for network references which caused as many
> slave-to-master calls as there are Network records plus one.
> The get_all_records_where() call gets exactly what is needed with a single
> call.
>
> Signed-off-
Replace the get_record() for network references which caused as many
slave-to-master calls as there are Network records plus one.
The get_all_records_where() call gets exactly what is needed with a single
call.
Signed-off-by: Rob Hoes
Acked-by: Dominic Curran
---
.../usr_share_openvswitch_scrip
On Wed, Jun 27, 2012 at 05:19:36PM +0900, Simon Horman wrote:
> this series is based on v2 of your Open Flow 1.1 work.
> Please let me know if you feel that I should rebase on top of v3.
It's probably not necessary, because my v3 is so preliminary, and the
patches that are close to being ready are
On Wed, Jun 27, 2012 at 04:21:32PM +0900, Isaku Yamahata wrote:
> Introduce a macro instead for
> With this macro, the code is a bit reduced.
> test: compile-tested and unit tests passed.
>
> Signed-off-by: Isaku Yamahata
Thanks. I made some adjustments to suit my idea of style and pushed
this.
Thanks, I pushed this one.
On Wed, Jun 27, 2012 at 08:25:12PM +1200, Joe Stringer wrote:
> Seems that I missed this one. Looks good.
>
> On 27 June 2012 18:53, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
> > ---
> > v3: New patch.
> > ---
> > include/openflow/nicira-ext.h | 2 +-
> > 1 fi
On Wed, Jun 27, 2012 at 05:19:36PM +0900, Simon Horman wrote:
> Hi Ben,
>
> this series is based on v2 of your Open Flow 1.1 work.
> Please let me know if you feel that I should rebase on top of v3.
>
> Also, I apologise for not having reviewed v3 of your patches yet,
> I will try to get to it be
* request OFP12_VERSION
* Reset ofconn to OF1.2
* Allow OF1.2 to be used as a dump protocol
---
lib/ofp-util.c| 1 +
lib/vconn.c | 14 +++---
ofproto/connmgr.c | 2 +-
3 files changed, 9 insertions(+), 8 deletions(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index fe3d911..
Signed-off-by: Simon Horman
---
v5
* Initial post
---
lib/ofp-util.c | 4
1 file changed, 4 insertions(+)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 99d4269..fe3d911 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -966,6 +966,8 @@ static const struct ofputil_msg_type ofputil_m
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
utilities/ovs-ofctl.c | 41 -
1 file changed, 32 insertions(+), 9 deletions(-)
diff --git a/utilities/ovs-ofctl.c b/utilities/ovs-ofctl.c
index 252d94f..f26356e 100644
--- a/utilities/ovs-ofctl.c
+++
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
lib/ofp-print.c | 63 ++---
1 file changed, 47 insertions(+), 16 deletions(-)
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index e80c597..ac905a7 100644
--- a/lib/ofp-print.c
+++ b/lib/o
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
lib/ofp-print.c | 35 ---
1 file changed, 32 insertions(+), 3 deletions(-)
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index ac905a7..356ade3 100644
--- a/lib/ofp-print.c
+++ b/lib/ofp-print.c
@@ -1351,13
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
lib/ofp-print.c | 85 +
1 file changed, 62 insertions(+), 23 deletions(-)
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index 9c40efe..8473a03 100644
--- a/lib/ofp-print.c
+++ b/lib/o
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
lib/ofp-util.c| 4
ofproto/ofproto.c | 33 +
2 files changed, 33 insertions(+), 4 deletions(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index f430dc3..99d4269 100644
--- a/lib/ofp-util.c
+++ b/l
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
ofproto/ofproto.c | 40 ++--
1 file changed, 38 insertions(+), 2 deletions(-)
diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c
index 6d54b5a..4c5fe9b 100644
--- a/ofproto/ofproto.c
+++ b/ofproto/ofproto
Signed-off-by: Simon Horman
---
v5
* Initial post
---
lib/ofp-util.c | 4
1 file changed, 4 insertions(+)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index e8d96ff..f430dc3 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -960,6 +960,8 @@ static const struct ofputil_msg_type ofputil_m
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
lib/ofp-util.c| 4
ofproto/ofproto.c | 27 ---
2 files changed, 28 insertions(+), 3 deletions(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index fdded88..e8d96ff 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp
Pass vconn to fetch_port_by_features() and callers.
In some cases this will reduce the number of connections
that ovs-ofputil sets up.
It should not alter the behaviour of ovs-ofputil.
Signed-off-by: Simon Horman
---
v5
* Initial post
---
utilities/ovs-ofctl.c | 41 ++-
Signed-off-by: Simon Horman
---
v5
* Initial Post
---
lib/ofp-print.c | 19 +--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index 8473a03..e80c597 100644
--- a/lib/ofp-print.c
+++ b/lib/ofp-print.c
@@ -1074,8 +1074,23 @@ stati
Open Flow 1.1 and 1.2 make use of 32 bit ports,
however Open vSwtich maps them to 16 bits. Make
ovs-ofputl aware of this.
Also, only accept ports that fit into 16 bits
for Open Flow 1.0.
Signed-off-by: Simon Horman
---
v5
* Initial post
---
utilities/ovs-ofctl.c | 38 +
Signed-off-by: Simon Horman
---
v5
* No change
v4
* Initial post
---
ofproto/ofproto.c | 64 +--
1 file changed, 53 insertions(+), 11 deletions(-)
diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c
index d70208d..c71ec14 100644
--- a/ofproto
Signed-off-by: Simon Horman
---
v5
* Manual rebase
* Add OFPST_TABLE entry for Open Flow 1.1 and 1.2 to ofputil_msg_types,
this wires-up decoding of table statistics messages.
v4
* Initial post
Fix dump-tables
---
include/openflow/openflow-1.1.h | 4 ++
include/openflow/openflow-1.2.h |
I am unsure of the overall merit if this change,
but it does allow to avoid a small amount of code duplication.
Signed-off-by: Simon Horman
---
v5
* Manual rebase
v4
* Manual rebase
* Use an extra level of indirection for the msg argument of
ofputil_encode_aggregate_stats_reply__ as the valu
Signed-off-by: Simon Horman
---
v5
* Initial post
---
lib/ofp-print.c | 162 +++-
1 file changed, 148 insertions(+), 14 deletions(-)
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index cb1ff3a..7a373e7 100644
--- a/lib/ofp-print.c
+++ b/lib/
Signed-off-by: Simon Horman
---
v5
* Initial post
---
utilities/ovs-ofctl.c | 29 +
1 file changed, 25 insertions(+), 4 deletions(-)
diff --git a/utilities/ovs-ofctl.c b/utilities/ovs-ofctl.c
index a4361bd..146522c 100644
--- a/utilities/ovs-ofctl.c
+++ b/utilities/
Signed-off-by: Simon Horman
---
v5
* Manual rebase
* Place OFPST12_REPLY(OFPST11_AGGREGATE,... below
OFPST12_REPLY(OFPST11_FLOW...
This is consistent with the ordering for Open Flow 1.0.
v4
* Use OFPUTIL_OFPST11_AGGREGATE_REPLY as code for Open Flow 1.1 and 1.2
messages.
* Don't explicitl
Allow decoding of Open Flow 1.1 and 1.2 flow and aggregate flow statistics
request messages.
Signed-off-by: Simon Horman
---
v5
* Manual rebase
v4
* Use OFPG11_ANY in place of OFPG_ANY.
The value is the same, but it seems to make sense to
use the Open Flow 1.1 constant when working with Op
Enable the decoding of some Open Flow 1.2 and 1.1 messages which
have the same wire-format and thus decoders as Open Flow 1.0.
Signed-off-by: Simon Horman
---
v5
* Add OFPST_PORT_DESC reply for Open Flow 1.1 and 1.2
v4
* Consistently use OFPT12 to initialise elements of
ofputil_msg_types for
Open Flow 1.0 uses a different alignment of members to
Open Flow 1.1 and 1.2. For this reason I would like to
add a struct ofp11_aggregate_stats_reply in a subsequent patch.
In preparation for that rename ofp_aggregate_stats_reply.
Signed-off-by: Simon Horman
---
v5
* No change
v4
* No change
Signed-off-by: Simon Horman
---
v5
* Manual rebase
* Use padded_match_len parameter of ofputil_pull_ofp12_match()
to obtain the length of a match. Using the remaining
length of the message only works for the last flow.
v4
* Don't explicitly pill the pad of the header of the message.
This
Signed-off-by: Simon Horman
---
v5
* No change
v4
* Do not manually add header pad, it is handled by
ofputil_postappend_stats_reply()
v3
* Initial post
---
lib/ofp-util.c| 24 ++--
lib/ofp-util.h| 3 ++-
ofproto/ofproto.c | 2 +-
3 files changed, 25 insertions(
Signed-off-by: Simon Horman
---
v5
* No change
v4
* Manual rebase
v3
* Add protocol variable to do_probe().
Previously this was added by a separate patch.
v2
* No change
---
lib/learning-switch.c | 4 ++--
lib/ofp-util.c| 30 ++
lib/ofp-util.h|
Signed-off-by: Simon Horman
---
v5
* No change
v4
* No change
v3
* Differentiate versions using oh->version rather than relying
on a separate code for each version of the Packet Out message.
v2
* No change
---
lib/ofp-print.c | 4 ++--
lib/ofp-util.c| 35 +++-
Signed-off-by: Simon Horman
---
v5
* Initial Post
* Broken out of "ofp-util: Allow decoding of Open Flow 1.1 & 1.2 Flow Aggregate
Statistics Response Messages"
* Do not modify packet data
ofp_print_ofpst_aggregate_reply
---
lib/ofp-print.c | 33 +++--
1 file change
Signed-off-by: Simon Horman
---
v5
* Manual rebase
v4
* No change
v3
* Add OF1.2 entry to to ofputil_msg_types
* Make OF1.2 the first option in the if statement in ofputil_decode_packet_in()
* Rename o11pi as opi, which is consistent with the style used elsewhere
* Rebase
v2
* No change
---
Signed-off-by: Simon Horman
---
v5
* No change
v4
* Manual rebase
v3
* Correct title: this patch relates to Packet Out not Packet In
* Correct decoding of buffer_id, it is 32bits wide not 16bits wide
* Add decoding of stats reply messages
v2
* No change
---
lib/ofp-util.c | 10 +-
1
Signed-off-by: Simon Horman
---
v5
* No change
v4
* No change
v3
* Add protocol parameter to ofputil_encode_packet_in().
This allows packet_in_format to be ignored for Open Flow 1.2.
v2
* Update for new ofputil_put_match() implementation
---
lib/ofp-util.c| 76 +
This may be used by both ofputil_decode_flow_mod() and
ofputil_decode_packet_in()
Signed-off-by: Simon Horman
---
v5
* Add padded_match_len parameter to
ofputil_pull_ofp12_match() and __ofputil_pull_ofp11_match().
This will be used when decoding messages that need
to know the match length
Signed-off-by: Simon Horman
---
v5
* No change
v4
* No change
v3
* Manual rebase
* Correct title: This patch relates to Packet Out not Packet In
* Use OFPT_PACKET_OUT instead of OFPT11_PACKET_OUT, it seems
easier to use a single code and differentiate using the version
in the ofp_header
*
There are capabilities which are present in one, two and three
of Open Flow 1.0, 1.1 and 1.2. Update OFPC_COMMON to only include
capabilities that are prsent in all three Open Flow versions and
add ofputil_capabilities_mask() to return the mask of capabilities
for each version.
This does not cover
Open Flow 1.2, like Open Flow 1.1, has OFPC11_GROUP_STATS.
All other Open Flow 1.2 capabilities are handled by
ofputil_capabilities_mask() which is used by both
ofputil_decode_switch_features() and ofputil_encode_switch_features().
No update to ofputil_encode_switch_features() is required
to supp
Signed-off-by: Simon Horman
---
v5
* No change
v4
* No change
v3
* No change
v2
* No change
---
lib/ofp-util.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 1b56981..ae89225 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -83
Signed-off-by: Simon Horman
---
v5
* Manual rebase
v4
* No change
v3
* No change
*v3
Initial post
---
lib/ofp-util.c | 28 +++-
1 file changed, 27 insertions(+), 1 deletion(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 6fb21c7..71e5cd9 100644
--- a/lib/ofp-ut
Signed-off-by: Simon Horman
---
v5
* No change
v4
* Manual rebase
v3
* No change
v2
* No change
---
lib/ofp-util.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 92bcc84..1b56981 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -3
Signed-off-by: Simon Horman
---
v5
* No change
v4
* Manual rebase
v3
* No change
v2
* Use ofputil_put_match()
---
lib/ofp-util.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index e915292..6fb21c7 100644
--- a/lib/ofp-util.c
+++ b/lib
Signed-off-by: Simon Horman
---
v5
* Manual rebase
v4
* No change
v3
* No change
v2
* No change
---
lib/ofp-util.c | 17 -
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 7caf096..e915292 100644
--- a/lib/ofp-util.c
+++ b/li
* OFPUTIL_C_ARP_MATCH_IP is not present in Open Flow 1.2.
* OFPUTIL_C_PORT_BLOCKED is new in Open Flow 1.2.
Otherwise the Open Flow 1.2 capabilities are the same as those of
Open Flow 1.1.
Signed-off-by: Simon Horman
---
v5
* No change
v4
* No change
v3
* No change
v2
* No change
---
lib/o
Empirically it appears that process_packet_in() should call
ofputil_encode_flow_mod() with protocol as defived from rconn
rather than sw->protocol.
Signed-off-by: Simon Horman
---
v5
* Actually retrieve protocol from rconn.
This patch was dependent on a subsequent patch.
v4
* No change
v3
*
1 - 100 of 124 matches
Mail list logo
