Hello.
I have some nodes with virtual machines and want that each owner can
view traffic from only owning virtual machines.
For example:
user1 have vm1_1 vm1_2 vm1_3
user2 have vm2_1 vm2_2 vm2_3
I want that user1 from vm1_1 and vm2_1 can view only traffic from this
three servers and can't see vm2_
On Wed, May 29, 2013 at 04:02:12PM -0700, Olof Johansson wrote:
> Hi,
>
> On Wed, May 29, 2013 at 3:54 PM, Simon Horman wrote:
> > On Wed, May 29, 2013 at 02:46:05PM -0700, Olof Johansson wrote:
> >> Sorry, that was not a great bug report.
> >>
> >> So, looks like it's the removal of NET_SKBUFF_D
Date: Wed, 29 May 2013 13:46:05 -0700
Subject: Re: [ovs-dev] [PATCH v2 1/3] ovs-vswitchd: An option to wait for
userspace > flow restore to complete. (Gurucharan Shetty)
From: shet...@nicira.com
To: ai_jing2...@hotmail.com
CC: dev@openvswitch.org
On Wed, May 29, 2013 at 1:23 PM, Jing Ai wrote:
This patch modifies the vsp_realdev_to_vlandev() function. Instead of taking
and returning datapath port number, the new implementation takes and returns
OpenFlow port number.
Signed-off-by: Alex Wang
---
ofproto/ofproto-dpif.c | 38 +-
1 file changed, 21 in
On Tue, May 28, 2013 at 10:56 PM, Maciej Galkiewicz
wrote:
> Package: openvswitch
> Version: 1.4.2+git20120612-9
> Severity: important
>
> # dkms install -m openvswitch -v 1.4.2+git20120612 -k 3.8-2-amd64
Open vSwitch 1.4 was released almost a year before Linux 3.8 so it is
not surprising that th
Thanks Ben for the suggestion, I'll adjust accordingly.
On Wed, May 29, 2013 at 4:58 PM, Ben Pfaff wrote:
> On Wed, May 29, 2013 at 04:57:11PM -0700, Alex Wang wrote:
> > This patch modifies the vsp_realdev_to_vlandev() function. Instead of
> taking
> > and returning datapath port number, the
On Wed, May 29, 2013 at 02:24:41PM -0700, Ethan Jackson wrote:
> This patch implements a new interface to xlate_actions which, aside
> from being simpler and more intuitive, achieves several goals. It
> pulls all of xlate_actions() results into a single structure which
> will be easier to cache an
On Wed, May 29, 2013 at 04:57:11PM -0700, Alex Wang wrote:
> This patch modifies the vsp_realdev_to_vlandev() function. Instead of taking
> and returning datapath port number, the new implementation takes and returns
> OpenFlow port number.
>
> Signed-off-by: Alex Wang
This needs a re-spin again
On Tue, May 28, 2013 at 02:02:17PM +, Gurucharan Shetty wrote:
> While upgrading openvswitch, it helps to restore openflow flows before
> starting packet processing. Typically we want to restart openvswitch,
> add the openflow flows and then start packet processing.
>
> To do this, we look fo
This patch modifies the vsp_realdev_to_vlandev() function. Instead of taking
and returning datapath port number, the new implementation takes and returns
OpenFlow port number.
Signed-off-by: Alex Wang
---
ofproto/ofproto-dpif.c | 34 +-
1 file changed, 17 insert
facet_lookup_valid() attempted to re-revalidate its facet after
it had been removed.
Found by inspection (by Justin).
Reported-by: Justin Pettit
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c |5 +
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/ofproto/ofproto-dp
On Wed, May 29, 2013 at 6:20 AM, Ben Pfaff wrote:
> On Fri, May 24, 2013 at 05:01:34PM -0700, Jesse Gross wrote:
>> Originally no rule existed for packets that did not match an
>> OpenFlow flow and therefore every packet with a rule could be
>> counted as a hit. However, newer versions of OVS have
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c | 10 --
1 file changed, 10 deletions(-)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index 40be3c8..580044e 100644
--- a/ofproto/ofproto-dpif.c
+++ b/ofproto/ofproto-dpif.c
@@ -397,13 +397,6 @@ struct subfacet {
Hi,
On Wed, May 29, 2013 at 3:54 PM, Simon Horman wrote:
> On Wed, May 29, 2013 at 02:46:05PM -0700, Olof Johansson wrote:
>> Sorry, that was not a great bug report.
>>
>> So, looks like it's the removal of NET_SKBUFF_DATA_USES_OFFSET that
>> does it for me.
>>
>> The devices I've seen it with ar
On Wed, May 29, 2013 at 02:46:05PM -0700, Olof Johansson wrote:
> Sorry, that was not a great bug report.
>
> So, looks like it's the removal of NET_SKBUFF_DATA_USES_OFFSET that
> does it for me.
>
> The devices I've seen it with are with asix usb-ethernet adapters (on
> Tegra seaboard) and with
> I'm not sure how likely the function "vsp_realdev_to_vlandev()" will be
> called with odp_port as argument in the future. I'll go with the "two
> functions" implementation first.
FWIW I suspect this is unlikely, and even if it wasn't. We can always
change it in future when we have more informat
Btw, this version is based on master, and is ready for review.
Thanks,
Ethan
On Wed, May 29, 2013 at 5:43 PM, Ethan Jackson wrote:
> The logic for updating statistics at the facet level had been
> spread through ofproto-dpif in a rather confusing manner. This
> patch consolidates as much of thi
The logic for updating statistics at the facet level had been
spread through ofproto-dpif in a rather confusing manner. This
patch consolidates as much of this logic as is reasonable into
facet_push_stats().
On a side note, I'd expect this patch to have a marginal positive
performance impact when
I saw you comment after sending the previous email. I'll have some time and
consider both. And I'll resubmit the patch later.
Thanks
On Wed, May 29, 2013 at 3:37 PM, Alex Wang wrote:
> Thanks Ben and Ethan,
>
> I'm okay with both ways.
>
> I'm not sure how likely the function "vsp_realdev_to_v
That would avoid an ofp->odp translation since compose_output_action__()
already has the ofp_port. It's probably a good idea. Alex, do you want
to do it that way?
On Wed, May 29, 2013 at 05:25:09PM -0500, Ethan Jackson wrote:
> I haven't looked at this patch carefully, so my next comment may or
Thanks Ben and Ethan,
I'm okay with both ways.
I'm not sure how likely the function "vsp_realdev_to_vlandev()" will be
called with odp_port as argument in the future. I'll go with the "two
functions" implementation first.
Hope to hear more comments from you.
Thanks
On Wed, May 29, 2013 at 3:
I haven't looked at this patch carefully, so my next comment may or
may not be worthwhile, but I'm wondering if it would be cleaner if we
kept a single vsp_realdev_to_vlandev, which took an ofp_port as an
argument, and then had compose_output_action__ pass in the openflow
port number instead of the
On Wed, May 29, 2013 at 03:01:43PM -0700, Alex Wang wrote:
> In ofproto-dpif.c, function vsp_realdev_to_vlandev() is called with both
> OpenFlow port and datapath port number as argument. This patch replaces
> the function vsp_realdev_to_vlandev() with vsp_ofp_realdev_to_vlandev() and
> vsp_odp_rea
On Wed, May 29, 2013 at 01:57:22PM -0700, Alex Wang wrote:
> This patch fixes a typo in the comment of the pktbuf_retrieve() function.
>
> Signed-off-by: Alex Wang
Looks good, thanks.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mail
Sorry, that was not a great bug report.
So, looks like it's the removal of NET_SKBUFF_DATA_USES_OFFSET that
does it for me.
The devices I've seen it with are with asix usb-ethernet adapters (on
Tegra seaboard) and with mv643xx_eth on cubox (dove).
I'll try to get a better stacktrace out of it. S
In ofproto-dpif.c, function vsp_realdev_to_vlandev() is called with both
OpenFlow port and datapath port number as argument. This patch replaces
the function vsp_realdev_to_vlandev() with vsp_ofp_realdev_to_vlandev() and
vsp_odp_realdev_to_vlandev(), which take OpenFlow port number and datapth
port
Before this patch, datapath keys with ODP_FIT_TO_LITTLE, would be
assigned subfacets and installed in the kernel with a SLOW_MATCH
slow path reason. This is problematic, because these flow keys
can't be reliable converted into a 'struct flow' thus breaking a
fundamental assumption of ofproto-dpif.
Before this patch, when in band control was enabled, every DHCP
packet had to be sent to userspace to calculate it's actions.
Those DHCP packets intended for the local port would have a special
action added to ensure they actually make it there. This
unnecessarily complicates the code, so this pat
This patch changes the name of action_xlate_ctx to xlate_ctx. Aside
from being a bit snappier, it fits more cleanly with structures
added in future patches.
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c | 183
1 file changed, 90 inserti
Used in future commits.
Signed-off-by: Ethan Jackson
---
lib/ofpbuf.h |6 ++
1 file changed, 6 insertions(+)
diff --git a/lib/ofpbuf.h b/lib/ofpbuf.h
index 8b03c7e..0c12162 100644
--- a/lib/ofpbuf.h
+++ b/lib/ofpbuf.h
@@ -107,6 +107,12 @@ static inline struct ofpbuf *ofpbuf_from_list(co
Suggested-by: Ben Pfaff
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c | 18 ++
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index e0c88a8..3a1484e 100644
--- a/ofproto/ofproto-dpif.c
+++ b/ofproto/ofpro
Upon close inspection, it appears that it's not possible for
actions to differ between subfacets belonging to a given facet.
Given this fact, it makes sense to move datapath actions from
subfacets to their parent facets. It's both conceptually more
straightforward, and necessary for future threadi
It's no longer possible for a single datapath flow to be slow
pathed for two different reasons. This patch updates the code to
reflect this fact (marginally simplifying it).
Signed-off-by: Ethan Jackson
---
lib/odp-util.c | 53 +++-
lib/odp-
handle_flow_miss_common() is intended to be called whether or not a
miss needs a facet. Found by inspection.
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index f74cfcb..e0c88a8 10
This is the backport of the recently merged xlate series to
branch-1.11. I'm sending it out to the list as a backport of this
size is slightly unusual, and therefore deserves some explanation.
Because of a potentially large performance benefit, we're planning
to backport userspace megaflows to 1.
Simon,
On Thu, May 23, 2013 at 11:51 PM, Simon Horman wrote:
> In order to mitigate ongoing incresase in the size of struct skbuff
> use 16 bit integer offsets rather than pointers for inner_*_headers.
>
> This appears to reduce the size of struct skbuff from 0xd0 to 0xc0
> bytes on x86_64 with t
On Wed, May 29, 2013 at 12:00 PM, Ben Pfaff wrote:
> On Tue, May 28, 2013 at 05:24:16PM -0700, Gurucharan Shetty wrote:
> > Reported-by: Maxime Brun
> > Signed-off-by: Gurucharan Shetty
>
> Looks good, thanks.
>
> Thank you. I pushed this to master, 1.11, 1.10 and 1.9.
I thought we had convert
On Wed, May 29, 2013 at 1:23 PM, Jing Ai wrote:
> > While upgrading openvswitch, it helps to restore openflow flows before
> > starting packet processing. Typically we want to restart openvswitch,
> > add the openflow flows and then start packet processing.
> >
> > To do this, we look for the oth
This patch fixes a typo in the comment of the pktbuf_retrieve() function.
Signed-off-by: Alex Wang
---
ofproto/pktbuf.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/ofproto/pktbuf.c b/ofproto/pktbuf.c
index 902b19d..2ec1f0d 100644
--- a/ofproto/pktbuf.c
+++ b/ofproto/pk
On Wed, May 29, 2013 at 01:22:00PM -0700, Ethan Jackson wrote:
> This incremental (applied on v3) should address your comments on the
> v2 version of the patch. I consider this series reviewed at this
> point, unless there's something you'd like to make additional comments
> on.
No, I'm happy, th
On Wed, May 29, 2013 at 03:06:39PM +0900, Simon Horman wrote:
> Add set skb_mark support to execute_set_action.
Now renamed odp_execute_actions().
> This also adds support for the user-space datapath
> to honour such actions if they occur before recirculation,
> which will be added by a subsequen
> While upgrading openvswitch, it helps to restore openflow flows before
> starting packet processing. Typically we want to restart openvswitch,
> add the openflow flows and then start packet processing.
>
> To do this, we look for the other_config:flow-restore-wait column
> in the Open_vSwitch t
This incremental (applied on v3) should address your comments on the v2 version
of the patch. I consider this series reviewed at this point, unless there's
something you'd like to make additional comments on.
Ethan
---
ofproto/ofproto-dpif.c | 43 +++
1
> Message: 3
> Date: Wed, 29 May 2013 10:21:28 -0700
> From: Ethan Jackson
> Subject: [ovs-dev] [PATCH] ofproto-dpif: handle_flow_miss_common()
> without facets.
> To: dev@openvswitch.org
> Message-ID: <1369848088-47098-1-git-send-email-et...@nicira.com>
>
> handle_flow_miss_common() is in
On Wed, May 29, 2013 at 03:06:38PM +0900, Simon Horman wrote:
> This moves generic action execution code out of lib/dpif-netedev.c
> and into a new file, lib/execute-actions.c.
>
> This is in preparation for using execute_set_action()
> in lib/odp-util.c to handle recirculation/
>
> Signed-off-by
Probably.
I need to backport the xlate series as well. I'll do it all in one go
when I merge that.
On Wed, May 29, 2013 at 2:02 PM, Ben Pfaff wrote:
> On Wed, May 29, 2013 at 10:21:28AM -0700, Ethan Jackson wrote:
>> handle_flow_miss_common() is intended to be called whether or not a
>> miss ne
OK, thanks.
On Wed, May 29, 2013 at 02:41:55PM -0500, Ethan Jackson wrote:
> Oops, this patch ends up deleteing, xlate_actions_for_side_effects(),
> but then I need it again in the xlate series. I think instead of
> basing this on master, I'll base it on the xlate series once it's
> merged. May
Oops, this patch ends up deleteing, xlate_actions_for_side_effects(),
but then I need it again in the xlate series. I think instead of
basing this on master, I'll base it on the xlate series once it's
merged. May as well hold off reviewing it until then.
Ethan
On Wed, May 29, 2013 at 2:38 PM, E
The logic for updating statistics at the facet level had been
spread through ofproto-dpif in a rather confusing manner. This
patch consolidates as much of this logic as is reasonable into
facet_push_stats().
On a side note, I'd expect this patch to have a marginal positive
performance impact when
On Wed, May 29, 2013 at 10:21:28AM -0700, Ethan Jackson wrote:
> handle_flow_miss_common() is intended to be called whether or not a
> miss needs a facet. Found by inspection.
>
> Signed-off-by: Ethan Jackson
Looks good, thanks.
Is it worth a backport?
_
On Tue, May 28, 2013 at 05:24:16PM -0700, Gurucharan Shetty wrote:
> Reported-by: Maxime Brun
> Signed-off-by: Gurucharan Shetty
Looks good, thanks.
I thought we had converted all this stuff to use ovs-lib.sh, but I guess
not.
___
dev mailing list
dev
On Tue, May 28, 2013 at 3:31 PM, Fan Du wrote:
> ovs_vport_cmd_new create a new vport with port num specified by user.
> when this port num is already occupied, it's more clear to return
> -EEXIST as the reason than -EBUSY.
>
> Signed-off-by: Fan Du
This error code is part of the userspace ABI s
handle_flow_miss_common() is intended to be called whether or not a
miss needs a facet. Found by inspection.
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index b547298..ae59eda 10
Hi OpenvSwitch Devs
Ubuntu Saucy development has been open for a while and the Ubuntu kernel
team are currently targeting either Linux 3.9 or 3.10 for this release.
Currently the 1.9.0 DKMS source is unhappy with 3.9 - I've hacked on it
for a bit and although it definitely possible to backpor
54 matches
Mail list logo
