Fix following sparse warnings.
drivers/net/vxlan.c:238:44: warning: incorrect type in argument 3 (different
base types)
drivers/net/vxlan.c:238:44:expected restricted __be32 [usertype] value
drivers/net/vxlan.c:238:44:got unsigned int const [unsigned] [usertype]
remote_vni
drivers/net/vxl
Following patch adds data and priority fields to vxlan handlers
and export vxlan handler api.
vh->data is required to store private data per vxlan handler.
vh->priority allows ovs to assign lower priority for ovs vxlan
handler. So that vxlan device modules gets to look at vxlan
packet before ovs.
Following patch adds basic multiple vxlan protocol handlers.
This does not change any functionality. This is required for
openvswitch vxlan support.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 212 +--
1 files changed, 137 insertions(
First two patches fixes vxlan issues. Next two patches
extends vxlan so that openvswitch can share vxlan udp port
with vxlan module. Rest of patches refactors vxlan data
plane so that ovs can share that code with vxlan module.
Last patch adds vxlan-vport to openvswitch.
Pravin B Shelar (8):
vxl
Rather than having static headroom calculation, adjust headroom
according to target device.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 13 +
1 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 6ca2632..e53c947
Following patch allows more code sharing between vxlan and ovs-vxlan.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 97 ++-
include/net/vxlan.h |7
2 files changed, 64 insertions(+), 40 deletions(-)
diff --git a/drivers/net/v
Following patch allows transmit side vlan offload for vxlan
devices.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 15 ++-
1 files changed, 14 insertions(+), 1 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index e53c947..a7feba4 100644
--- a/driver
Delete vxlan-sock list entry from list before freeing the memory.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 284c6c0..173705f 100644
--- a/drivers/net/vxlan.c
+
Following patch adds vxlan vport type for openvswitch using
vxlan api.
This patch adds vxlan dependency for openvswitch.
Signed-off-by: Pravin B Shelar
---
include/uapi/linux/openvswitch.h | 11 ++
net/openvswitch/Kconfig |1 +
net/openvswitch/Makefile |3 +-
net/openv
Hello.
On 20-06-2013 11:26, Pravin B Shelar wrote:
Following patch allows transmit side vlan offload for vxlan
devices.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 15 ++-
1 files changed, 14 insertions(+), 1 deletions(-)
diff --git a/drivers/net/vxlan.c b/d
> From: Pravin B Shelar
> @@ -1205,13 +1222,13 @@ static netdev_tx_t vxlan_xmit(struct sk_buff
> *skb, struct net_device *dev)
>
>skb1 = skb_clone(skb, GFP_ATOMIC);
>if (skb1) {
> - rc1 = vxlan_xmit_one(skb1, dev, rdst, did_rsc);
> + rc1 = __vxlan_xmit(skb1, dev
Hi,
The bug still exists and the added "ifup -a" helps, but if the interface
has not been set up before you need to set the link up manually :
# ip link set eth1 up
I added this to my /etc/network/interfaces to fix this at boot :
auto eth1
iface eth1 inet manual
pre-up ip link set
Signed-off-by: Jarno Rajahalme
---
lib/ofp-actions.c | 34 +++---
lib/ofp-actions.h |3 ++-
2 files changed, 13 insertions(+), 24 deletions(-)
diff --git a/lib/ofp-actions.c b/lib/ofp-actions.c
index c9e000f..7a2851e 100644
--- a/lib/ofp-actions.c
+++ b/lib/ofp
Signed-off-by: Jarno Rajahalme
---
lib/ofp-util.h |2 --
1 file changed, 2 deletions(-)
diff --git a/lib/ofp-util.h b/lib/ofp-util.h
index 1c8d6cd..431521c 100644
--- a/lib/ofp-util.h
+++ b/lib/ofp-util.h
@@ -676,8 +676,6 @@ void *ofputil_put_action(enum ofputil_action_code, struct
ofpbuf
Signed-off-by: Jarno Rajahalme
---
lib/ofp-actions.c | 35 +++-
lib/ofp-actions.h | 14 ++
lib/rconn.c| 14 +-
ofproto/ofproto-dpif.c | 33
ofproto/ofproto-provider.h | 24 +++
ofproto/ofproto.c | 467
This series adds basic support for OpenFlow 1.3 meters. As meters need to
act on possibly every packet seen by a datapath, the meters must be
implemented in the datapath(s). This series includes a crude
implementation on the userspace datapath, which is yet to be fully tested.
Since a datapath ca
Signed-off-by: Jarno Rajahalme
---
include/openflow/openflow-1.3.h|2 +-
include/openflow/openflow-common.h |1 +
lib/ofp-actions.c | 32 +++-
lib/ofp-actions.h | 16 +-
lib/ofp-msgs.h |6 +-
lib/ofp-parse.c
Signed-off-by: Jarno Rajahalme
---
lib/dpif-netdev.c | 273 +
1 file changed, 236 insertions(+), 37 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 1f57917..32a0b7a 100644
--- a/lib/dpif-netdev.c
+++ b/lib/dpif-netdev.c
Adds commands add-meter, mod-meter, del-meter, del-meters, dump-meter,
dump-meters, meter-stats, and meter-features.
Syntax is as follows:
add-meter meter= (kbps|pktps) [burst] [stats] bands=
type=(drop|dscp_remark) rate= [burst_size=] [prec_level=]
mod-meter meter= (kbps|pktps) [burst] [sta
Signed-off-by: Jarno Rajahalme
---
include/linux/openvswitch.h |3 ++
lib/dpif-linux.c | 40 +++
lib/dpif-netdev.c| 43 -
lib/dpif-provider.h | 28 ++
lib/dpif.c | 87 +
Signed-off-by: Jarno Rajahalme
---
datapath/actions.c | 16 +++-
datapath/datapath.c |5 +
datapath/datapath.h |2 ++
3 files changed, 22 insertions(+), 1 deletion(-)
diff --git a/datapath/actions.c b/datapath/actions.c
index 09d0c3f..8c5e6d5 100644
--- a/datapath/act
On Thu, 20 Jun 2013 00:26:25 -0700
Pravin B Shelar wrote:
> Delete vxlan-sock list entry from list before freeing the memory.
>
> Signed-off-by: Pravin B Shelar
> ---
> drivers/net/vxlan.c |4 +++-
> 1 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/net/vxlan.c b/d
On Thu, Jun 20, 2013 at 09:47:38AM +0900, Simon Horman wrote:
> This adds support for specifying flow miss handling behaviour at
> runtime, through a new "other-config" option in the Open_vSwitch table.
> This is an extension to flow-eviction-threshold.
>
> By default, the behaviour is the same as
On Thu, 20 Jun 2013 00:26:15 -0700
Pravin B Shelar wrote:
> First two patches fixes vxlan issues. Next two patches
> extends vxlan so that openvswitch can share vxlan udp port
> with vxlan module. Rest of patches refactors vxlan data
> plane so that ovs can share that code with vxlan module.
> L
On Thu, Jun 20, 2013 at 09:25:30AM +0900, Joe Stringer wrote:
> Patch 27a88d1373cbfcceac6d901bbf1c17051aa7845f caused the vswitchd
> documentation and the code to digress. This brings them back in line.
>
> Signed-off-by: Joe Stringer
Thanks, applied to master, branch-1.11, branch-1.10.
On Thu, Jun 20, 2013 at 09:47:41AM +0900, Simon Horman wrote:
> From: Joe Stringer
>
> This patch introduces a python script to generate about 1500 tests for
> permutations of mpls_push,mpls_pop,dec_mpls_ttl,dec_ttl where
> recirculation occurs up to (and including) three times.
>
> Signed-off-b
On Wed, Jun 19, 2013 at 11:14 PM, Andy Zhou wrote:
> This patch seems to work fine. Please feel free to push.
Thanks, I applied the combined patch.
> I have two concerns mainly on the style side:
>
> 1. In the same file flow.c, we are mixing the _deferred_ functions with the
> style of passing
On Thu, Jun 20, 2013 at 8:58 AM, Stephen Hemminger
wrote:
> On Thu, 20 Jun 2013 00:26:15 -0700
> Pravin B Shelar wrote:
>
>> First two patches fixes vxlan issues. Next two patches
>> extends vxlan so that openvswitch can share vxlan udp port
>> with vxlan module. Rest of patches refactors vxlan
Looks good. Do we need to filter the warnings out from comments if they
are not to be used in the code?
Not specific to this patch, Why do we assume git is being used? it seems we
can make the Makefile.am more portable otherwise.
--andy
___
dev mailing
On Tue, Jun 18, 2013 at 04:15:10PM -0700, Andy Zhou wrote:
> Added support to allow mega flow specified and displayed. ovs-dpctl tool
> is mainly used as debugging tool.
>
> This patch also implements the low level user space routines to send
> and receive mega flow netlink messages. Those netlink
On Thu, Jun 20, 2013 at 10:07 AM, Jesse Gross wrote:
> On Wed, Jun 19, 2013 at 11:14 PM, Andy Zhou wrote:
> > This patch seems to work fine. Please feel free to push.
>
> Thanks, I applied the combined patch.
>
> Sounds good.
> > I have two concerns mainly on the style side:
> >
> > 1. In the
On Ubuntu Saucy based desktops, upstart runs with user sessions
enabled which means that the init process under which a daemon
might run is not always pid = 1.
Check for a process with cmd name of 'init' instead.
Signed-off-by: James Page
---
tests/daemon-py.at | 10 ++
tests/daemon.at
On Wed, Jun 19, 2013 at 04:58:44PM -0700, Alex Wang wrote:
> Currently datapath ports and openflow ports are both represented by unsigned
> integers of various sizes. With implicit casts, etc. it is easy to mix them
> up and use one where the other is expected. This commit creates two typedefs
> of
Looks good.
On Wed, Jun 19, 2013 at 1:17 PM, Ben Pfaff wrote:
> This ensures that attempts to use them cause sparse to complain.
>
> Signed-off-by: Ben Pfaff
> ---
> include/sparse/math.h |5 +
> include/sparse/netinet/in.h |3 +--
> 2 files changed, 2 insertions(+), 6 delet
Thanks for the updates and pushing it in.
On Thu, Jun 20, 2013 at 10:34 AM, Ben Pfaff wrote:
> On Tue, Jun 18, 2013 at 04:15:10PM -0700, Andy Zhou wrote:
> > Added support to allow mega flow specified and displayed. ovs-dpctl tool
> > is mainly used as debugging tool.
> >
> > This patch also im
Meh - I need to test this better; not working reliably everywhere.
On 20/06/13 18:44, James Page wrote:
On Ubuntu Saucy based desktops, upstart runs with user sessions
enabled which means that the init process under which a daemon
might run is not always pid = 1.
Check for a process with cmd na
I think it might be sufficient to check that the parent pid changed,
rather than being picky about the new parent pid or the name of the
process associated with that new parent pid.
On Thu, Jun 20, 2013 at 07:11:51PM +0100, James Page wrote:
> Meh - I need to test this better; not working reliably
On Thu, Jun 20, 2013 at 12:27 AM, Pravin B Shelar wrote:
> diff --git a/net/openvswitch/vport-vxlan.c b/net/openvswitch/vport-vxlan.c
> new file mode 100644
> index 000..be22134
> --- /dev/null
> +++ b/net/openvswitch/vport-vxlan.c
> @@ -0,0 +1,220 @@
> +/*
> + * Copyright (c) 2013 Nicira, Inc
Thanks, applied.
On Thu, Jun 20, 2013 at 10:54:44AM -0700, Andy Zhou wrote:
> Looks good.
>
>
> On Wed, Jun 19, 2013 at 1:17 PM, Ben Pfaff wrote:
>
> > This ensures that attempts to use them cause sparse to complain.
> >
> > Signed-off-by: Ben Pfaff
> > ---
> > include/sparse/math.h |
On Thu, Jun 20, 2013 at 10:27:01AM -0700, Andy Zhou wrote:
> Looks good. Do we need to filter the warnings out from comments if they
> are not to be used in the code?
It's not easy to reliably filter out warnings from comments. It's
usually easy to avoid mentioning these functions in comments.
On Thu, Jun 20, 2013 at 11:23:38AM -0700, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 10:27:01AM -0700, Andy Zhou wrote:
> > Looks good. Do we need to filter the warnings out from comments if they
> > are not to be used in the code?
>
> It's not easy to reliably filter out warnings from comments.
On Wed, Jun 19, 2013 at 5:56 PM, Simon Horman wrote:
> On Tue, Jun 18, 2013 at 04:06:49PM +0900, Simon Horman wrote:
>> 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,
Make sense!
On Thu, Jun 20, 2013 at 12:25 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 11:23:38AM -0700, Ben Pfaff wrote:
> > On Thu, Jun 20, 2013 at 10:27:01AM -0700, Andy Zhou wrote:
> > > Looks good. Do we need to filter the warnings out from comments if
> they
> > > are not to be used in
Thanks very much, Ben,
And sorry, I should have put you as co-author in the last commit.
On Thu, Jun 20, 2013 at 10:53 AM, Ben Pfaff wrote:
> On Wed, Jun 19, 2013 at 04:58:44PM -0700, Alex Wang wrote:
> > Currently datapath ports and openflow ports are both represented by
> unsigned
> > intege
On Thu, Jun 20, 2013 at 12:35:57PM -0700, Alex Wang wrote:
> And sorry, I should have put you as co-author in the last commit.
I don't think it was warranted, you did most of the work.
> On Thu, Jun 20, 2013 at 10:53 AM, Ben Pfaff wrote:
>
> > On Wed, Jun 19, 2013 at 04:58:44PM -0700, Alex Wang
On Wed, Jun 19, 2013 at 01:17:03PM -0700, Ben Pfaff wrote:
> POSIX defines a portable pthread_key_t API for per-thread data. GCC and
> C11 have two different forms of per-thread data that are generally faster
> than the POSIX API, where they are available. This commit adds a
> macro-based wrapper
On Thu, Jun 20, 2013 at 05:26:16PM +0300, Jarno Rajahalme wrote:
>
> Signed-off-by: Jarno Rajahalme
Applied, thanks.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Thu, Jun 20, 2013 at 05:26:17PM +0300, Jarno Rajahalme wrote:
>
> Signed-off-by: Jarno Rajahalme
I have mixed feelings here. On one hand, it is nice to avoid a big
copy in some cases (rare cases really). But on the other hand I am
not fond of interfaces that temporarily modify data, even if
On Thu, Jun 20, 2013 at 10:13 AM, Stephen Hemminger
wrote:
> On Thu, 20 Jun 2013 08:58:42 -0700
> Stephen Hemminger wrote:
>
>> On Thu, 20 Jun 2013 00:26:15 -0700
>> Pravin B Shelar wrote:
>>
>> > First two patches fixes vxlan issues. Next two patches
>> > extends vxlan so that openvswitch can
On Ubuntu Saucy based desktops, upstart runs with user sessions
enabled which means that the init process under which a daemon
might run is not always pid = 1.
Instead of checking for pid = 1, check to ensure that the parent
pid of the monitor is not the pid of the shell that started it.
Signed-o
OK - resubmitted with suggested approach.
Cheers
James
On 20/06/13 19:14, Ben Pfaff wrote:
I think it might be sufficient to check that the parent pid changed,
rather than being picky about the new parent pid or the name of the
process associated with that new parent pid.
On Thu, Jun 20, 2013
Thanks for the explanation,
On Thu, Jun 20, 2013 at 12:47 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 12:35:57PM -0700, Alex Wang wrote:
> > And sorry, I should have put you as co-author in the last commit.
>
> I don't think it was warranted, you did most of the work.
>
> > On Thu, Jun 20, 2
It is more comprehensible to use UINT16_MAX in mask assignment than OFPP_NONE.
Signed-off-by: Alex Wang
---
tests/test-classifier.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/test-classifier.c b/tests/test-classifier.c
index cd081a5..601aaf8 100644
--- a/tests/t
Signed-off-by: Ben Pfaff
---
FAQ |5 +
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/FAQ b/FAQ
index 1f0f841..7df4e90 100644
--- a/FAQ
+++ b/FAQ
@@ -940,6 +940,11 @@ A: Open vSwitch 1.9 and earlier support only OpenFlow 1.0
(plus
ovs-vsctl set bridge br0 protoc
If a flow cannot be installed in the datapath, we should notice
this and not treat it as installed. This becomes an issue with
megaflows, since a batch of unique flows may come in that generate
a single new datapath megaflow that covers them. Since userspace
doesn't know whether the datapath supp
Add a new function for converting a mask into a set of
OVS_KEY_ATTR* attributes. This will be useful in a future commit.
Signed-off-by: Justin Pettit
---
lib/odp-util.c | 139 ---
lib/odp-util.h |2 +
2 files changed, 93 insertions(+), 48
When debugging the system, it's useful to not just see the key but
also the mask.
Signed-off-by: Justin Pettit
---
lib/dpif.c | 15 +--
ofproto/ofproto-dpif.c |6 +++---
2 files changed, 12 insertions(+), 9 deletions(-)
diff --git a/lib/dpif.c b/lib/dpif.c
index fa
The commit configures the masks generated from megaflows and pushes
them through the dpif layer.
With this commit and a wildcard supporting OVS kernel module,
ovs-vswitchd's flow setup rate is very close to that of the Linux
bridge.
Signed-off-by: Justin Pettit
---
NEWS |4
On Thu, Jun 20, 2013 at 03:10:04PM -0700, Alex Wang wrote:
> It is more comprehensible to use UINT16_MAX in mask assignment than OFPP_NONE.
>
> Signed-off-by: Alex Wang
Applied, thanks.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/ma
Looks good.
Acked-by: Justin Pettit
--Justin
On Jun 20, 2013, at 3:16 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> FAQ |5 +
> 1 files changed, 5 insertions(+), 0 deletions(-)
>
> diff --git a/FAQ b/FAQ
> index 1f0f841..7df4e90 100644
> --- a/FAQ
> +++ b/FAQ
> @@ -940,6 +
On Thu, Jun 20, 2013 at 10:31:52PM +0100, James Page wrote:
> On Ubuntu Saucy based desktops, upstart runs with user sessions
> enabled which means that the init process under which a daemon
> might run is not always pid = 1.
>
> Instead of checking for pid = 1, check to ensure that the parent
> p
Applied to master, thanks.
Oops, I forgot to add the ack. Sorry about that, I'll try harder in
the future.
On Thu, Jun 20, 2013 at 03:43:57PM -0700, Justin Pettit wrote:
> Looks good.
>
> Acked-by: Justin Pettit
>
> --Justin
>
>
> On Jun 20, 2013, at 3:16 PM, Ben Pfaff wrote:
>
> > Signed
On Wed, Jun 12, 2013 at 5:46 PM, Pravin B Shelar wrote:
> diff --git a/datapath/linux/compat/gre.c b/datapath/linux/compat/gre.c
> new file mode 100644
> index 000..fbb9fb9
> --- /dev/null
> +++ b/datapath/linux/compat/gre.c
> +struct sk_buff *gre_handle_offloads(struct sk_buff *skb, bool gre_
On Thu, Jun 20, 2013 at 03:16:42PM -0700, Justin Pettit wrote:
> Add a new function for converting a mask into a set of
> OVS_KEY_ATTR* attributes. This will be useful in a future commit.
>
> Signed-off-by: Justin Pettit
"sparse" says:
../lib/odp-util.c:2427:47: warning: incorrect type in
On Thu, Jun 20, 2013 at 03:16:43PM -0700, Justin Pettit wrote:
> If a flow cannot be installed in the datapath, we should notice
> this and not treat it as installed. This becomes an issue with
> megaflows, since a batch of unique flows may come in that generate
> a single new datapath megaflow th
The commit allows a user to add a database file to a
ovsdb-server during run time. One can also remove a
database file from ovsdb-server's control.
Feature #14595.
Signed-off-by: Gurucharan Shetty
---
ovsdb/jsonrpc-server.c | 10 ++
ovsdb/jsonrpc-server.h |2 +
ovsdb/ovsdb-server.1.in |
An upcoming commit provides the ability to add and remove databases.
Having the databases in a shash instead of an array makes it easier
to add and remove databases.
Feature #14595.
Signed-off-by: Gurucharan Shetty
---
ovsdb/ovsdb-server.c | 186 +
On Jun 20, 2013, at 4:06 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 03:16:43PM -0700, Justin Pettit wrote:
>> If a flow cannot be installed in the datapath, we should notice
>> this and not treat it as installed. This becomes an issue with
>> megaflows, since a batch of unique flows may com
On Thu, Jun 20, 2013 at 03:16:44PM -0700, Justin Pettit wrote:
> When debugging the system, it's useful to not just see the key but
> also the mask.
>
> Signed-off-by: Justin Pettit
Acked-by: Ben Pfaff
___
dev mailing list
dev@openvswitch.org
http://o
On Thu, Jun 20, 2013 at 04:20:16PM -0700, Justin Pettit wrote:
> On Jun 20, 2013, at 4:06 PM, Ben Pfaff wrote:
> > On Thu, Jun 20, 2013 at 03:16:43PM -0700, Justin Pettit wrote:
> > The following is a little worrisome because EINVAL is very generic and
> > can indicate many kinds of errors. If we
On Jun 20, 2013, at 3:59 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 03:16:42PM -0700, Justin Pettit wrote:
>> Add a new function for converting a mask into a set of
>> OVS_KEY_ATTR* attributes. This will be useful in a future commit.
>>
>> Signed-off-by: Justin Pettit
>
> "sparse" says:
>
On Thu, Jun 20, 2013 at 04:32:39PM -0700, Justin Pettit wrote:
> On Jun 20, 2013, at 3:59 PM, Ben Pfaff wrote:
>
> > On Thu, Jun 20, 2013 at 03:16:42PM -0700, Justin Pettit wrote:
> >> Add a new function for converting a mask into a set of
> >> OVS_KEY_ATTR* attributes. This will be useful in a
Flow update still requires unmasked key to match. If not,
return EINVAL.
CC: Justin Pettit
Signed-off-by: Andy Zhou
---
datapath/datapath.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/datapath/datapath.c b/datapath/datapath.c
index ee3d5e4..a514e74 100644
-
On Jun 20, 2013, at 4:33 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 04:32:39PM -0700, Justin Pettit wrote:
>> On Jun 20, 2013, at 3:59 PM, Ben Pfaff wrote:
>>
>>> On Thu, Jun 20, 2013 at 03:16:42PM -0700, Justin Pettit wrote:
Add a new function for converting a mask into a set of
On Thu, Jun 20, 2013 at 03:16:45PM -0700, Justin Pettit wrote:
> The commit configures the masks generated from megaflows and pushes
> them through the dpif layer.
>
> With this commit and a wildcard supporting OVS kernel module,
> ovs-vswitchd's flow setup rate is very close to that of the Linux
On Thu, Jun 20, 2013 at 04:35:46PM -0700, Justin Pettit wrote:
> On Jun 20, 2013, at 4:33 PM, Ben Pfaff wrote:
>
> > On Thu, Jun 20, 2013 at 04:32:39PM -0700, Justin Pettit wrote:
> >> On Jun 20, 2013, at 3:59 PM, Ben Pfaff wrote:
> >>
> >>> On Thu, Jun 20, 2013 at 03:16:42PM -0700, Justin Pett
On Thu, Jun 20, 2013 at 4:36 PM, Andy Zhou wrote:
> Flow update still requires unmasked key to match. If not,
> return EINVAL.
>
> CC: Justin Pettit
> Signed-off-by: Andy Zhou
Applied, thanks.
___
dev mailing list
dev@openvswitch.org
http://openvswitc
On 20 June 2013 16:18, Ben Pfaff wrote:
> On Wed, Jun 19, 2013 at 01:17:03PM -0700, Ben Pfaff wrote:
>> POSIX defines a portable pthread_key_t API for per-thread data. GCC and
>> C11 have two different forms of per-thread data that are generally faster
>> than the POSIX API, where they are availa
On Mon, Jun 17, 2013 at 11:19 AM, Lori Jakab wrote:
> On 6/11/13 1:09 AM, Jesse Gross wrote:
>>
>> On Wed, Jun 5, 2013 at 2:56 PM, Lori Jakab wrote:
>>>
>>> At a high level, we would introduce layer 3 (tunnel) vports, and LISP
>>> would be such a vport. Whenever a packet that ingressed on a L2 v
Add a new function for converting a mask into a set of
OVS_KEY_ATTR* attributes. This will be useful in a future commit.
Signed-off-by: Justin Pettit
---
lib/odp-util.c | 139 ---
lib/odp-util.h |2 +
2 files changed, 93 insertions(+), 48
When debugging the system, it's useful to not just see the key but
also the mask.
Signed-off-by: Justin Pettit
Acked-by: Ben Pfaff
---
lib/dpif.c | 15 +--
ofproto/ofproto-dpif.c |6 +++---
2 files changed, 12 insertions(+), 9 deletions(-)
diff --git a/lib/dpif.c
If a flow cannot be installed in the datapath, we should notice
this and not treat it as installed. This becomes an issue with
megaflows, since a batch of unique flows may come in that generate
a single new datapath megaflow that covers them. Since userspace
doesn't know whether the datapath supp
The commit configures the masks generated from megaflows and pushes
them through the dpif layer.
With this commit and a wildcard supporting OVS kernel module,
ovs-vswitchd's flow setup rate is very close to that of the Linux
bridge.
Signed-off-by: Justin Pettit
---
NEWS |4
On Jun 20, 2013, at 4:38 PM, Ben Pfaff wrote:
> "sparse" says:
>
>../ofproto/ofproto-dpif.c:5063:42: warning: incorrect type in argument 4
> (different base types)
>../ofproto/ofproto-dpif.c:5063:42:expected restricted odp_port_t
> [usertype] odp_in_port
>../ofproto/ofproto-dpi
On Jun 20, 2013, at 4:31 PM, Ben Pfaff wrote:
> The other thing that worries me is that you mentioned that you saw
> multiple installations of a single subfacet in a single batch. If so,
> that worries me because it seems like this could happen:
>
>handle_flow_miss_with_facet() queues u
On Thu, Jun 20, 2013 at 04:51:53PM -0700, Justin Pettit wrote:
> Add a new function for converting a mask into a set of
> OVS_KEY_ATTR* attributes. This will be useful in a future commit.
>
> Signed-off-by: Justin Pettit
Acked-by: Ben Pfaff
___
dev m
On Thu, Jun 20, 2013 at 04:51:54PM -0700, Justin Pettit wrote:
> If a flow cannot be installed in the datapath, we should notice
> this and not treat it as installed. This becomes an issue with
> megaflows, since a batch of unique flows may come in that generate
> a single new datapath megaflow th
On Thu, Jun 20, 2013 at 04:51:56PM -0700, Justin Pettit wrote:
> The commit configures the masks generated from megaflows and pushes
> them through the dpif layer.
>
> With this commit and a wildcard supporting OVS kernel module,
> ovs-vswitchd's flow setup rate is very close to that of the Linux
On Thu, Jun 20, 2013 at 07:49:18PM -0400, Ed Maste wrote:
> On 20 June 2013 16:18, Ben Pfaff wrote:
> > On Wed, Jun 19, 2013 at 01:17:03PM -0700, Ben Pfaff wrote:
> >> POSIX defines a portable pthread_key_t API for per-thread data. GCC and
> >> C11 have two different forms of per-thread data that
On Jun 20, 2013, at 5:04 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 04:51:54PM -0700, Justin Pettit wrote:
>> If a flow cannot be installed in the datapath, we should notice
>> this and not treat it as installed. This becomes an issue with
>> megaflows, since a batch of unique flows may com
When parsing flow Netlink messages we currently have arrays to hold the
attribute pointers for both values and masks. This results in a large
stack, which some compilers warn about. It's not actually necessary
to have both arrays at the same time, so we can collapse this to a
single array.
Reporte
Following patch restructures ovs tunneling and gre vport
implementation to make ovs tunneling more in sync with
upstream kernel tunneling. Doing this tunneling code is
simplified as most of protocol processing on send and
recv is pushed to kernel tunneling. For external ovs
module the code is mov
Looks good. Just curious, which compiler warn about this? What's the
warning message?
On Thu, Jun 20, 2013 at 5:15 PM, Jesse Gross wrote:
> When parsing flow Netlink messages we currently have arrays to hold the
> attribute pointers for both values and masks. This results in a large
> stack, w
Thanks, I applied it.
It was Pravin's compiler, not mine, so I'm not exactly sure but I
believe that it is an older version of GCC. The message is that the
stack size exceeds 1024 bytes (which is not automatically a problem
but given that the kernel has a fixed stack size it can be dangerous
if yo
On Thu, Jun 20, 2013 at 12:28:32PM -0700, Jesse Gross wrote:
> On Wed, Jun 19, 2013 at 5:56 PM, Simon Horman wrote:
> > On Tue, Jun 18, 2013 at 04:06:49PM +0900, Simon Horman wrote:
> >> Allow datapath to recognize and extract MPLS labels into flow keys
> >> and execute actions which push, pop, an
On Thu, Jun 20, 2013 at 5:19 PM, Pravin B Shelar wrote:
> diff --git a/datapath/linux/compat/skbuff-openvswitch.c
> b/datapath/linux/compat/skbuff-openvswitch.c
> index ef43ba9..720bc18 100644
> --- a/datapath/linux/compat/skbuff-openvswitch.c
> +++ b/datapath/linux/compat/skbuff-openvswitch.c
>
On Jun 20, 2013, at 5:08 PM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 04:51:56PM -0700, Justin Pettit wrote:
>> The commit configures the masks generated from megaflows and pushes
>> them through the dpif layer.
>>
>> With this commit and a wildcard supporting OVS kernel module,
>> ovs-vswitch
Sure. Is there a preference for where the string parsing code goes --
ofproto.c or bridge.c? The configuration applies globally, but it seems
that similar configuration is parsed in the bridge code.
On Fri, Jun 21, 2013 at 12:57 AM, Ben Pfaff wrote:
> On Thu, Jun 20, 2013 at 09:47:38AM +0900, S
This code modularizes ofproto-dpif's mirror code by moving it to
ofproto-dpif-mirror. Not only does this shorten ofproto-dpif and
hide complexity, but its also necessary for future patches which
modularize ofproto-dpif-xlate in preparation for multi-threading.
Signed-off-by: Ethan Jackson
---
o
Before this patch, the mirroring code would retroactively insert
actions for destination mirrors after actions were translated.
This relied on converting datapath output actions into ofports
which doesn't work for tunnels and patch ports. This patch
refactors the code to handle destination mirrors
1 - 100 of 103 matches
Mail list logo
