Applied to master, thanks!
On Wed, Feb 05, 2014 at 12:38:10PM -0500, Vasu Dasari wrote:
Ben,
The code looks good to me and it also works for me.
-Vasu
On Tue, Feb 4, 2014 at 12:05 PM, Ben Pfaff b...@nicira.com wrote:
Until now, the tcp_stream() code has ignored any TCP packets
On Tue, Feb 11, 2014 at 12:00:46AM +, Pritesh Kothari (pritkoth) wrote:
Hi Ben,
-static const struct tnl_match_pattern patterns[] = {
-{ false, false, IP_SRC_CFG }, /* remote_ip, local_ip, in_key. */
-{ false, false, IP_SRC_ANY }, /* remote_ip, in_key. */
-
On Mon, Feb 10, 2014 at 03:25:36PM -0800, Gurucharan Shetty wrote:
This is helpful if we want to wait either on 'fd' for POSIX or
events for Windows.
For Windows, if both 'fd' and 'wevent' is specified, we associate
that event with the 'fd' using WSAEventSelect() in poll_block().
So any
On Mon, Feb 10, 2014 at 03:25:37PM -0800, Gurucharan Shetty wrote:
For winsock2 functions, error number has to be converted to string
using FormatMessage().
Signed-off-by: Gurucharan Shetty gshe...@nicira.com
The ovs_strerror() comment is particularly helpful. Thanks.
Acked-by: Ben Pfaff
On Mon, Feb 10, 2014 at 03:25:38PM -0800, Gurucharan Shetty wrote:
This commit creates events and through poll_fd_wait_event()
associates them with socket file descriptors to get woken up
from poll_block().
Some other changes:
* Windows does not have sys/fcntl.h but has a fcntl.h
On
On Tue, Feb 11, 2014 at 8:07 AM, Ben Pfaff b...@nicira.com wrote:
On Mon, Feb 10, 2014 at 03:25:36PM -0800, Gurucharan Shetty wrote:
This is helpful if we want to wait either on 'fd' for POSIX or
events for Windows.
For Windows, if both 'fd' and 'wevent' is specified, we associate
that event
On Tue, Feb 11, 2014 at 8:17 AM, Ben Pfaff b...@nicira.com wrote:
On Mon, Feb 10, 2014 at 03:25:38PM -0800, Gurucharan Shetty wrote:
This commit creates events and through poll_fd_wait_event()
associates them with socket file descriptors to get woken up
from poll_block().
Some other changes:
Applied to master, thanks.
On Tue, Feb 11, 2014 at 10:42:43AM +0530, Arun Sharma wrote:
Verified looks good. Thanks Ben.
On 11/02/14 12:54 AM, Ben Pfaff b...@nicira.com wrote:
Reported-by: Arun Sharma arun.sha...@calsoftinc.com
Signed-off-by: Ben Pfaff b...@nicira.com
---
On Fri, Feb 07, 2014 at 04:39:54PM -0800, Joe Stringer wrote:
The userspace MPLS test case was checking the same things as the
drop test case, rather than checking to see that packets were being
sent to userspace. This patch makes the testsuite consistent with itself.
Signed-off-by: Joe
On Fri, Feb 07, 2014 at 04:39:55PM -0800, Joe Stringer wrote:
Signed-off-by: Joe Stringer joestrin...@nicira.com
---
This test currently fails, and for some reason we're seeing a VLAN
tagged packet arrive at the monitor. The expected result is not 100%
correct, but it's a lot closer than what
On Fri, Feb 07, 2014 at 10:46:41AM -0800, Alex Wang wrote:
Before this commit, ovs randomly selects a slave for unassigned
bond entry. If the selected slave is not enabled, the active slave
is chosen instead. In this commit, the slave is selected from the
list of all enabled slaves in a
On 11 February 2014 10:01, Ben Pfaff b...@nicira.com wrote:
On Fri, Feb 07, 2014 at 04:39:54PM -0800, Joe Stringer wrote:
The userspace MPLS test case was checking the same things as the
drop test case, rather than checking to see that packets were being
sent to userspace. This patch makes
On 11 February 2014 10:05, Ben Pfaff b...@nicira.com wrote:
On Fri, Feb 07, 2014 at 04:39:55PM -0800, Joe Stringer wrote:
Signed-off-by: Joe Stringer joestrin...@nicira.com
---
This test currently fails, and for some reason we're seeing a VLAN
tagged packet arrive at the monitor. The
looks good to me.
Acked-by: pritesh pritesh.koth...@cisco.com
On Feb 4, 2014, at 1:19 PM, Ben Pfaff b...@nicira.com wrote:
There are 12 possible ways to specify a tunnel (2 * 2 * 3 == 12):
- Specific in_key or flow-based (2 choices).
- Specific ip_dst or flow-based (2 choices).
On Feb 11, 2014, at 8:05 AM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 12:00:46AM +, Pritesh Kothari (pritkoth) wrote:
Hi Ben,
-static const struct tnl_match_pattern patterns[] = {
-{ false, false, IP_SRC_CFG }, /* remote_ip, local_ip, in_key. */
-{
Thanks for the improvement, this looks much better.
I should keep reminding myself about using the thread-safety annotations.
In the V2 patch, the writer of 'enabled_slaves' does not grab the mutex.
I add the
following change to comply with the comments and will apply it soon,
diff --git
The netdev-dummy unit test ran into the reconnect condition on Jarno's
machine. With his test environment, we found and fixed the bugs in
handling reconnect.
Co-authored-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Andy Zhou
netdev-dummy will mostly be used for testing and debugging over fairly
reliable connection. Reduce reconnect back off timeout in case the first
connect attempt failed.
Signed-off-by: Andy Zhou az...@nicira.com
---
lib/netdev-dummy.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
On Tue, Feb 11, 2014 at 11:41:49AM -0800, Andy Zhou wrote:
The netdev-dummy unit test ran into the reconnect condition on Jarno's
machine. With his test environment, we found and fixed the bugs in
handling reconnect.
Co-authored-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Jarno
On Tue, Feb 11, 2014 at 11:41:50AM -0800, Andy Zhou wrote:
netdev-dummy will mostly be used for testing and debugging over fairly
reliable connection. Reduce reconnect back off timeout in case the first
connect attempt failed.
Signed-off-by: Andy Zhou az...@nicira.com
Acked-by: Ben Pfaff
I see. Let me restructure this patch to avoid blocking in the main loop.
On Tue, Feb 11, 2014 at 11:55 AM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 11:41:49AM -0800, Andy Zhou wrote:
The netdev-dummy unit test ran into the reconnect condition on Jarno's
machine. With his
On Wed, Feb 05, 2014 at 07:04:05PM +0900, YAMAMOTO Takashi wrote:
Following OpenFlow 1.4 semantics, make barriers wait for
flow monitor replies. This should fix a race in
ofproto - flow monitoring pause and resume test.
Signed-off-by: YAMAMOTO Takashi yamam...@valinux.co.jp
I still don't
I've noticed that test 683 (ofproto-dpif - controller) fails
intermittently for me. The failure report is always something like
this:
../../tests/ofproto-dpif.at:893: cat ofctl_monitor.log
../../tests/ofproto-dpif.at:922: ovs-ofctl dump-flows br0 | ofctl_strip | sort
--- - 2014-02-11
The netdev-dummy unit test ran into the reconnect condition on Jarno's
machine. With his test environment, we found and fixed the bugs in
handling reconnect.
Co-authored-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Andy Zhou
netdev-dummy will mostly be used for testing and debugging over fairly
reliable connection. Reduce reconnect back off timeout in case the first
connect attempt failed.
Signed-off-by: Andy Zhou az...@nicira.com
---
lib/netdev-dummy.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
On Tue, Feb 11, 2014 at 01:15:02PM -0800, Andy Zhou wrote:
netdev-dummy will mostly be used for testing and debugging over fairly
reliable connection. Reduce reconnect back off timeout in case the first
connect attempt failed.
Signed-off-by: Andy Zhou az...@nicira.com
Acked-by: Ben Pfaff
On Tue, Feb 11, 2014 at 01:15:01PM -0800, Andy Zhou wrote:
The netdev-dummy unit test ran into the reconnect condition on Jarno's
machine. With his test environment, we found and fixed the bugs in
handling reconnect.
Co-authored-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Jarno
I'm about to post a few stats duplication bugfixes, but it's hard to say
whether they fix this particular issue.
On 11 February 2014 13:14, Ben Pfaff b...@nicira.com wrote:
I've noticed that test 683 (ofproto-dpif - controller) fails
intermittently for me. The failure report is always
I guess you (or I?) dropped the list by mistake, so I'm adding it back.
Definitely. With very few exceptions, we basically don't block anywhere
in vswitchd. If you do, it can delay everything else happening in that
thread.
On Tue, Feb 11, 2014 at 01:31:29PM -0800, Andy Zhou wrote:
Do you mean
Thanks. Will send out V3 soon. I dropped the list by mistake.
On Tue, Feb 11, 2014 at 1:37 PM, Ben Pfaff b...@nicira.com wrote:
I guess you (or I?) dropped the list by mistake, so I'm adding it back.
Definitely. With very few exceptions, we basically don't block anywhere
in vswitchd. If
netdev-dummy will mostly be used for testing and debugging over fairly
reliable connection. Reduce reconnect back off timeout in case the first
connect attempt failed.
Signed-off-by: Andy Zhou az...@nicira.com
---
lib/netdev-dummy.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
The netdev-dummy unit test ran into the reconnect condition on Jarno's
machine. With his test environment, we found and fixed the bugs in
handling reconnect.
Co-authored-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
Signed-off-by: Andy Zhou
Previously, we would clean up the ukeys whose flow was not seen in the
most recent dump, while leaving the flow in the datapath. In the
unlikely case that the datapath fails to dump a flow that still exists
in the datapath, this would cause double-counting of those flow stats.
This is currently
Basic test to check that the datapath reports the correct number of
packets seen, after a delay of 1 second.
Signed-off-by: Joe Stringer joestrin...@nicira.com
---
tests/ofproto-dpif.at | 21 +
1 file changed, 21 insertions(+)
diff --git a/tests/ofproto-dpif.at
This splits out functions for re-use by later patches, and compacts the
udump revalidation code.
Co-authored-by: Ethan Jackson et...@nicira.com
Signed-off-by: Joe Stringer joestrin...@nicira.com
---
ofproto/ofproto-dpif-upcall.c | 134 -
1 file changed,
Previously, we would delete all ukeys when changing the number of
threads, but leave all flows in the datapath. This would cause
double-counting of stats for any flows that remain in the datapath. This
patch fixes the issue by ensuring that all flows are deleted from the
datapath before changing
It is possible for a datapath to dump the same flow twice, for instance
if the flow is the last in a batch of flows to be dumped, then a new
flow is inserted into the same bucket before the flow dumper fetches
another batch.
In this case, datapath flow stats may be duplicated: The revalidator
Signed-off-by: Joe Stringer joestrin...@nicira.com
---
ofproto/ofproto-dpif-upcall.c |8 +---
ofproto/ofproto-provider.h|4
ofproto/ofproto.c |9 +
ofproto/ofproto.h |2 ++
vswitchd/bridge.c |2 ++
On Tue, Feb 04, 2014 at 04:36:18PM -0800, Andy Zhou wrote:
This patch adds the necessary infrastructure for the next patch series.
recirculation logic is still work in progress. It is not enabled. There
should
be no behavior change.
The need to always match on recirc_id makes it different
On Tue, Feb 11, 2014 at 06:48:30PM +, Pritesh Kothari (pritkoth) wrote:
looks good to me.
Acked-by: pritesh pritesh.koth...@cisco.com
Thanks, applied to master.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Tue, Feb 11, 2014 at 06:48:33PM +, Pritesh Kothari (pritkoth) wrote:
On Feb 11, 2014, at 8:05 AM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 12:00:46AM +, Pritesh Kothari (pritkoth) wrote:
any insight into this would be greatly appreciated, alternatively i was
Commit 96ed775f resizes all userspace metadata to be 8 bytes minimum.
Fix the upcall size checks accordingly.
Signed-off-by: Romain Lenglet rleng...@vmware.com
---
ofproto/ofproto-dpif-upcall.c |8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git
On Feb 11, 2014, at 3:20 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 06:48:33PM +, Pritesh Kothari (pritkoth) wrote:
On Feb 11, 2014, at 8:05 AM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 12:00:46AM +, Pritesh Kothari (pritkoth) wrote:
any insight
These work for me, and the code looks OK too, so (for both patches):
Acked-by: Jarno Rajahalme jrajaha...@nicira.com
On Feb 11, 2014, at 1:53 PM, Andy Zhou az...@nicira.com wrote:
netdev-dummy will mostly be used for testing and debugging over fairly
reliable connection. Reduce reconnect back
The 5-tuple optimization becomes unnecessary with a later per-NUMA
node stats patch. Remove it first to make the changes easier to
grasp.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c | 11 -
datapath/flow.c | 41
This series reduces kernel datapath lock contention. I have included
the NUMA stats patches, as the rest do not apply cleanly without them.
I have added netperf TCP_CRR results (connections/second) as the first
column in the list below. These are in a specific 2-socket server,
YMMV.
I have
Move most memory allocations away from the ovs_mutex critical
sections. vport allocations still happen while the lock is taken, as
changing that would require major refactoring. Also, vports are
created very rarely so it should not matter.
Change ovs_dp_cmd_get() now only takes the
Remove unnecessary locking from functions that are always called with
appropriate locking.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c | 13 +++--
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/datapath/datapath.c b/datapath/datapath.c
We already extract the TCP flags for the key, might as well use that
for stats.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/flow.c | 15 +++
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/datapath/flow.c b/datapath/flow.c
index 6ed5250..3a6d424
This helps reduce calls to ksize().
Also add support for statically alloced actions, which will be used by
a subsequent patch.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c |2 +-
datapath/flow.h |2 ++
datapath/flow_netlink.c | 20
Flow SET can set an empty set of actions by not passing any actions.
Previously, we assigned the flow's actions pointer to NULL in this
case, but we never check for the NULL pointer later on. This patch
modifies this case to allocate a valid but empty set of actions
instead.
Signed-off-by: Jarno
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c | 60 +++
1 file changed, 32 insertions(+), 28 deletions(-)
diff --git a/datapath/datapath.c b/datapath/datapath.c
index a46ceb0..90fdc60 100644
---
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c | 83 -
datapath/flow_table.c |2 +-
datapath/flow_table.h |2 +-
3 files changed, 43 insertions(+), 44 deletions(-)
diff --git a/datapath/datapath.c
Use netlink_has_listeners() and NLM_F_ECHO flag to determine if a
reply is needed or not.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c | 57 +++
1 file changed, 35 insertions(+), 22 deletions(-)
diff --git
Not returning the actions from flow delete allows the reply (if any) be
allocated before taking the lock, and failing out properly if no memory
is available.
It seems that userspace has no plausable use for the actions in a flow
delete reply, so this should be OK.
Signed-off-by: Jarno Rajahalme
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
datapath/datapath.c | 75 ++-
1 file changed, 39 insertions(+), 36 deletions(-)
diff --git a/datapath/datapath.c b/datapath/datapath.c
index 452b889..cbdd68f 100644
---
Rule's 'used' timestamp is updated at the same time with the other
stats. So far the 'used' has been updated without proper protection,
which may lead to 'tearing' in 32-bit architectures, resulting in an
incorrect 'used' timestamp. While in practice this is highly
improbable, it is still better
Reading the hmap count for determining if it is empty or not is thread
safe, so avoid locking when not necessary.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
ofproto/ofproto-dpif-xlate.c | 19 +--
ofproto/ofproto-dpif.c | 17 +++--
2 files
mf_set_flow_value() was not setting 'flow-nw_frag' properly.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/meta-flow.c |2 +-
tests/ofproto-dpif.at |4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/meta-flow.c b/lib/meta-flow.c
index
pthread_spin_* wrappers to be used for extremely small critical sections.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/ovs-thread.c | 56 --
lib/ovs-thread.h | 26 +
2 files changed, 63 insertions(+),
The code setting wc masks shows up in 'perf' reports. This patch streamlines
the code and avoids setting the mask bits for registers or metadata, as
those fields do not exist on datapaths.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/learn.c |4 +-
When the target flow exists already as intended, only the 'modified'
time needs to be updated. Allow modification without taking the
'ofproto_mutex' by always using the 'rule-mutex' when accessing the
'modified' time.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
I have included indicative performance test numbers from a netperf
TCP_CRR test, tested on top of the kernel side lock contention patches
sent out separately.
Jarno Rajahalme (7):
56.8k (datapath lock contention patches)
meta-flow: Fix setting MFF_IP_FRAG.
57kofproto: Lock for vlan
ovs_mutex showed high up in 'perf' stats, use a spinlock instead.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
ofproto/ofproto-dpif.c | 95 +++-
1 file changed, 54 insertions(+), 41 deletions(-)
diff --git a/ofproto/ofproto-dpif.c
On Tue, Feb 11, 2014 at 04:30:43PM -0800, Jarno Rajahalme wrote:
mf_set_flow_value() was not setting 'flow-nw_frag' properly.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
Will you backport this as far as necessary?
Acked-by: Ben Pfaff b...@nicira.com
On Feb 11, 2014, at 4:30 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 04:30:43PM -0800, Jarno Rajahalme wrote:
mf_set_flow_value() was not setting 'flow-nw_frag' properly.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
Will you backport this as far as necessary?
On Tue, Feb 11, 2014 at 04:30:44PM -0800, Jarno Rajahalme wrote:
Reading the hmap count for determining if it is empty or not is thread
safe, so avoid locking when not necessary.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
It might be worth adding an
On Tue, Feb 11, 2014 at 04:38:35PM -0800, Jarno Rajahalme wrote:
On Feb 11, 2014, at 4:30 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 04:30:43PM -0800, Jarno Rajahalme wrote:
mf_set_flow_value() was not setting 'flow-nw_frag' properly.
Signed-off-by: Jarno Rajahalme
Hi Joe,
On Fri, Feb 07, 2014 at 04:39:55PM -0800, Joe Stringer wrote:
Signed-off-by: Joe Stringer joestrin...@nicira.com
---
This test currently fails, and for some reason we're seeing a VLAN
tagged packet arrive at the monitor. The expected result is not 100%
correct, but it's a lot closer
On Tue, Feb 11, 2014 at 04:30:45PM -0800, Jarno Rajahalme wrote:
When the target flow exists already as intended, only the 'modified'
time needs to be updated. Allow modification without taking the
'ofproto_mutex' by always using the 'rule-mutex' when accessing the
'modified' time.
This is an interesting idea. I have a few questions in line:
On Tue, Feb 11, 2014 at 3:01 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 04, 2014 at 04:36:18PM -0800, Andy Zhou wrote:
This patch adds the necessary infrastructure for the next patch series.
recirculation logic is still
On 11 February 2014 16:53, Simon Horman ho...@verge.net.au wrote:
I believe that the reason you are seeing a VLAN tagged packet come out is
that you are sending a VLAN tagged packet in. This is because of
eth_type(0x8100), which I believe should be eth_type(0x8847).
I see, that makes sense.
On Tue, Feb 11, 2014 at 05:00:33PM -0800, Andy Zhou wrote:
On Tue, Feb 11, 2014 at 3:01 PM, Ben Pfaff b...@nicira.com wrote:
Here's another proposal. Take one OpenFlow table (say, table 254)
and use it for recirculation. Populate the metadata field (yes, it's 64
bits, but that's OK) with
Hi,
this series builds on a patch posted by Joe Stringer.
It aims to:
* Provide a working version of the MPLS push on MPLS test
originally posted by Joe and;
* Fix a bug and add a test-case for the case where an MPLS push on
an MPLS packet causes the ethernet type to change.
Joe Stringer
From: Joe Stringer joestrin...@nicira.com
Signed-off-by: Joe Stringer joestrin...@nicira.com
Co-Authored-By: Simon Horman ho...@verge.net.au
Signed-off-by: Simon Horman ho...@verge.net.au
---
v2 [Simon Horman]
* Ethtype of (MPLS) input packet should be an MPLS type rather
than a VLAN type
On Tue, Feb 11, 2014 at 5:12 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 05:00:33PM -0800, Andy Zhou wrote:
Bonding metaflow does not need it, but one can envision that
controller may want to manage (some subset of) recirc_ids in the
future. In those cases, table 254 is a
On Tue, Feb 11, 2014 at 05:22:55PM -0800, Jesse Gross wrote:
On Tue, Feb 11, 2014 at 5:12 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 05:00:33PM -0800, Andy Zhou wrote:
Bonding metaflow does not need it, but one can envision that
controller may want to manage (some subset
On Tue, Feb 11, 2014 at 04:30:48PM -0800, Jarno Rajahalme wrote:
ovs_mutex showed high up in 'perf' stats, use a spinlock instead.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
Do spinlocks perform better than adaptive mutexes, e.g. that one would
initialize by adding a function like
On Tue, Feb 11, 2014 at 05:06:53PM -0800, Joe Stringer wrote:
On 11 February 2014 16:53, Simon Horman ho...@verge.net.au wrote:
I believe that the reason you are seeing a VLAN tagged packet come out is
that you are sending a VLAN tagged packet in. This is because of
eth_type(0x8100),
On Tue, Feb 11, 2014 at 04:30:49PM -0800, Jarno Rajahalme wrote:
The code setting wc masks shows up in 'perf' reports. This patch streamlines
the code and avoids setting the mask bits for registers or metadata, as
those fields do not exist on datapaths.
Signed-off-by: Jarno Rajahalme
On Tue, Feb 11, 2014 at 5:12 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 05:00:33PM -0800, Andy Zhou wrote:
On Tue, Feb 11, 2014 at 3:01 PM, Ben Pfaff b...@nicira.com wrote:
Here's another proposal. Take one OpenFlow table (say, table 254)
and use it for recirculation.
On Tue, Feb 11, 2014 at 05:52:43PM -0800, Andy Zhou wrote:
On Tue, Feb 11, 2014 at 5:12 PM, Ben Pfaff b...@nicira.com wrote:
On Tue, Feb 11, 2014 at 05:00:33PM -0800, Andy Zhou wrote:
On Tue, Feb 11, 2014 at 3:01 PM, Ben Pfaff b...@nicira.com wrote:
Here's another proposal. Take one
On Wed, Feb 05, 2014 at 07:04:05PM +0900, YAMAMOTO Takashi wrote:
Following OpenFlow 1.4 semantics, make barriers wait for
flow monitor replies. This should fix a race in
ofproto - flow monitoring pause and resume test.
Signed-off-by: YAMAMOTO Takashi yamam...@valinux.co.jp
I still
ovs_mutex showed high up in 'perf' stats, use a spinlock instead.
is the mutex implementation on your platform non-adaptive?
YAMAMOTO Takash
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Allow datapath to recognize and extract MPLS labels into flow keys
and execute actions which push, pop, and set labels on packets.
Based heavily on work by Leo Alterman, Ravi K, Isaku Yamahata and Joe Stringer.
Cc: Ravi K rke...@gmail.com
Cc: Leo Alterman lalter...@nicira.com
Cc: Isaku Yamahata
Hi Jesse, Hi All,
As per the suggestion made by Ben in relation to this patch
I have updated it so that:
* The datapath rejects push MPLS actions in the presence of VLAN tags.
I have done this by blacklisting the following:
- ETH_P_8021Q (0x8100)
- ETH_P_8021AD (0x88A8)
- ETH_P_QINQ1
Make set_ethertype() static as it is not used outside of packet.c
Signed-off-by: Simon Horman ho...@verge.net.au
---
lib/packets.c | 2 +-
lib/packets.h | 2 --
2 files changed, 1 insertion(+), 3 deletions(-)
diff --git a/lib/packets.c b/lib/packets.c
index 43134dc..7238f42 100644
---
87 matches
Mail list logo