On current master, the exact match cache entry can keep reference to
'struct dp_netdev_flow' even after the flow is removed from the flow
table. This means the free of allocated memory of the flow is delayed
until the exact match cache entry is cleared or replaced.
If the allocated memory is ahea
Series merged to master,
Jarno
On Nov 14, 2014, at 3:53 PM, Jarno Rajahalme wrote:
>
> On Nov 14, 2014, at 2:04 PM, Ben Pfaff wrote:
>
>> On Thu, Nov 13, 2014 at 11:56:17AM -0800, Jarno Rajahalme wrote:
>>> This patch adds a new functions classifier_defer() and
>>> classifier_publish(), wh
Ben,
I merged this into the patch on which you commented on this, no need to review
this.
Jarno
On Nov 14, 2014, at 3:17 PM, Jarno Rajahalme wrote:
> ovs_router_flush() should call ovsrcu_postpone to free the entries it
> removes from the classifier.
>
> Suggested-by: Ben Pfaff
> Signed-
On Nov 14, 2014, at 2:04 PM, Ben Pfaff wrote:
> On Thu, Nov 13, 2014 at 11:56:17AM -0800, Jarno Rajahalme wrote:
>> This patch adds a new functions classifier_defer() and
>> classifier_publish(), which control when the classifier modifications
>> are made available to lookups. By default, all m
On Nov 14, 2014, at 1:42 PM, Ben Pfaff wrote:
> On Thu, Nov 13, 2014 at 11:56:16AM -0800, Jarno Rajahalme wrote:
>> classifier_remove() was recently changed to take a const struct
>> cls_rule *. Make the corresponding change to classifier_replace() and
>> classifier_insert(). This simplifies e
On Nov 13, 2014, at 5:11 PM, Ben Pfaff wrote:
> On Thu, Nov 13, 2014 at 11:56:14AM -0800, Jarno Rajahalme wrote:
>> Previously, accurate iteration required writers to be excluded during
>> iteration. This patch adds an rculist to struct cls_subtable, and a
>> corresponding list node to struct c
On Nov 13, 2014, at 3:50 PM, Ben Pfaff wrote:
> On Thu, Nov 13, 2014 at 11:56:13AM -0800, Jarno Rajahalme wrote:
>> There is no point in adding duplicate information into prefix tries.
>>
>> Also, since the lower-priority duplicate rules are not visible to
>> lookups, they do not need to be in
On Nov 14, 2014, at 2:50 PM, Daniele Venturino wrote:
> You're right, it's used only in rstp_port_set_admin_point_to_point_mac__()
> where it's set, if the old value change. The action on
> oper_point_to_point_mac it's inside that function.
>
Thanks! For now this can be as it is. Maybe you
On Nov 14, 2014, at 1:38 PM, Ben Pfaff wrote:
> On Thu, Nov 13, 2014 at 11:56:15AM -0800, Jarno Rajahalme wrote:
>> Almost all classifier users already exclude concurrent modifications,
>> or are single-threaded, hence the classifier internal mutex can be
>> removed. Due to this change, ovs-rou
ovs_router_flush() should call ovsrcu_postpone to free the entries it
removes from the classifier.
Suggested-by: Ben Pfaff
Signed-off-by: Jarno Rajahalme
---
lib/ovs-router.c |4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/lib/ovs-router.c b/lib/ovs-router.c
index ba51
You're right, it's used only in rstp_port_set_admin_point_to_point_mac__()
where it's set, if the old value change. The action on oper_point_to_point_mac
it's inside that function.
Maybe it would be clearer to have a function
"update_oper_point_to_point_mac__()"
that updates oper_point_to_point_ma
On Nov 14, 2014, at 11:09 AM, Daniele Venturino
wrote:
>
>
> 2014-11-14 19:46 GMT+01:00 Jarno Rajahalme :
>
> On Nov 14, 2014, at 9:36 AM, Daniele Venturino
> wrote:
>
>> A possible commit message:
>> "Admin-port-state is the Administrative Bridge Port state variable defined
>> in the 80
On Fri, Nov 14, 2014 at 06:58:47PM +, Sorin Vinturis wrote:
> Right now the gOvsSwitchContext pointer is checked against NULL
> in a lot of places of the OVS extension code. This check should
> be done only once to avoid wasteful checks. Thus I have added the
> check in the dispatch routine, be
On Fri, Nov 14, 2014 at 11:35:21AM +, Sorin Vinturis wrote:
> Right now the gOvsSwitchContext pointer is checked against NULL
> in a lot of places of the OVS extension code. This check should
> be done only once to avoid wasteful checks. Thus I have added the
> check in the dispatch routine, be
Series pushed to master up to this patch.
Thanks!
Jarno
On Nov 13, 2014, at 3:59 PM, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme
>
> Thanks!
>
> On Nov 6, 2014, at 7:31 AM, Daniele Venturino
> wrote:
>
>> A Config BPDU always conveys a Designated Port Role.
>>
>> Signed-of
This is my first attempt at patch submission, so please bear with me. Mostly
wanted to get this code out to the community for constructive criticism.
This patch adds set-field operations for nd_target, nd_sll, and nd_tll fields
using nicra extensions and OpenFlow 1.2 protocol. Unit tests are i
On Thu, Nov 13, 2014 at 11:56:17AM -0800, Jarno Rajahalme wrote:
> This patch adds a new functions classifier_defer() and
> classifier_publish(), which control when the classifier modifications
> are made available to lookups. By default, all modifications are made
> available to lookups immediate
On Nov 14, 2014, at 9:19 AM, Daniele Venturino wrote:
> Why lose the ‘_WITH_HASH’? It seems pointless to iterate over all the ports,
> when we just care of the port with the ‘port_number’?
>
> With the _WITH_HASH a test on learned addresses shifting fails. It seems like
> it doesn't even ente
On Thu, Nov 13, 2014 at 11:56:16AM -0800, Jarno Rajahalme wrote:
> classifier_remove() was recently changed to take a const struct
> cls_rule *. Make the corresponding change to classifier_replace() and
> classifier_insert(). This simplifies existing calling sites in
> ofproto.
>
> Signed-off-by
Verified the fix. Looks good.
Thanks.
On Fri, Nov 14, 2014 at 1:24 PM, Pravin B Shelar wrote:
> ovs-dpctl crashed due to uninitialized router classifier. To
> fix this issue move ovs router initialization to route table
> module.
>
> Reported-by: Madhu Challa
> Signed-off-by: Pravin B Shelar
That was quick. Thanks !
On Fri, Nov 14, 2014 at 1:25 PM, Pravin Shelar wrote:
> On Fri, Nov 14, 2014 at 10:23 AM, Madhu Challa
> wrote:
> > Looks like ovs_router_unixctl_register() isn't getting called in
> ovs-dpctl,
> > so we hit this crash. If I explicitly call this after main, it works
> f
On Thu, Nov 13, 2014 at 11:56:15AM -0800, Jarno Rajahalme wrote:
> Almost all classifier users already exclude concurrent modifications,
> or are single-threaded, hence the classifier internal mutex can be
> removed. Due to this change, ovs-router.c and tnl-ports.c need new
> mutexes, which are ad
On Fri, Nov 14, 2014 at 10:23 AM, Madhu Challa wrote:
> Looks like ovs_router_unixctl_register() isn't getting called in ovs-dpctl,
> so we hit this crash. If I explicitly call this after main, it works fine.
> I am not sure when this broke and am trying to figure that out.
>
> Thanks.
>
> Startin
ovs-dpctl crashed due to uninitialized router classifier. To
fix this issue move ovs router initialization to route table
module.
Reported-by: Madhu Challa
Signed-off-by: Pravin B Shelar
---
lib/ovs-router.c |2 +-
lib/ovs-router.h |2 +-
lib/route-table.c |1 +
ofp
On Fri, Nov 14, 2014 at 3:51 AM, Lorand Jakab wrote:
> diff --git a/datapath/datapath.c b/datapath/datapath.c
> index 3607170..3ecb3cc 100644
> --- a/datapath/datapath.c
> +++ b/datapath/datapath.c
> @@ -570,6 +558,23 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb,
> struct genl_info *i
On Fri, Nov 14, 2014 at 10:32 AM, Fabian Frederick wrote:
> Signed-off-by: Fabian Frederick
Acked-by: Pravin B Shelar
> ---
> net/openvswitch/flow_netlink.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/net/openvswitch/flow_netlink.c b/net/openvswitch/flow_netlink
Acked-by: Jarno Rajahalme
No need to submit a new version!
Thanks,
Jarno
On Nov 14, 2014, at 9:15 AM, Daniele Venturino wrote:
> Possible commit message:
> "Invoke move_rstp__() in rstp_port_set_administrative_bridge_port__() if port
> is
> not initializing.
> This was necessary in a tes
2014-11-14 19:46 GMT+01:00 Jarno Rajahalme :
>
> On Nov 14, 2014, at 9:36 AM, Daniele Venturino
> wrote:
>
> A possible commit message:
> "Admin-port-state is the Administrative Bridge Port state variable defined
> in the 802.1D-2004 standard.
> It can be set to include or exclude a port from the
Acked-by: Jarno Rajahalme
No need to submit a new version!
Thanks,
Jarno
On Nov 14, 2014, at 9:13 AM, Daniele Venturino wrote:
> The comment could be:
>
> +/* Finds a port which needs to flush its own MAC learning table. A NULL
> pointer
> + * is returned if no port needs to flush its M
Acked-by: Jarno Rajahalme
On Nov 14, 2014, at 9:09 AM, Daniele Venturino wrote:
> Possible commit message:
>
> "This patch was necessary in a test on the acceptable range values of max_age
> and forward_delay.
> Since rstp_set_bridge_max_age__() and rstp_set_bridge_forward_delay__() set
> t
Acked-by: Jarno Rajahalme
On Nov 14, 2014, at 9:07 AM, Daniele Venturino wrote:
> Commit message:
>
> "Fix Topology Change State Machine transition.
> Accordingly to the 802.1D-2004 standard, the transition from the
> TOPOLOGY_CHANGE_SM_NOTIFIED_TCN_EXEC state
> should go to the TOPOLOGY_CHA
Right now the gOvsSwitchContext pointer is checked against NULL
in a lot of places of the OVS extension code. This check should
be done only once to avoid wasteful checks. Thus I have added the
check in the dispatch routine, before doing any processing, and
removed all other checks from the rest of
On Nov 14, 2014, at 9:36 AM, Daniele Venturino wrote:
> A possible commit message:
> "Admin-port-state is the Administrative Bridge Port state variable defined in
> the 802.1D-2004 standard.
> It can be set to include or exclude a port from the active topology by
> management (see 7.4).
>
IM
Signed-off-by: Fabian Frederick
---
net/openvswitch/flow_netlink.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/net/openvswitch/flow_netlink.c b/net/openvswitch/flow_netlink.c
index c0d066d..c8fccdd 100644
--- a/net/openvswitch/flow_netlink.c
+++ b/net/openvswitch/flow_n
Looks like ovs_router_unixctl_register() isn't getting called in ovs-dpctl,
so we hit this crash. If I explicitly call this after main, it works fine.
I am not sure when this broke and am trying to figure that out.
Thanks.
Starting program: /usr/local/bin/ovs-dpctl show
warning: no loadable secti
The comment could be:
+/* Finds a port which needs to flush its own MAC learning table. A NULL
> pointer
> + * is returned if no port needs to flush its MAC learning table.
> + * '*port' needs to be NULL in the first call to start the iteration. If
> + * '*port' is passed as non-NULL, it must be t
Global transitions are highest priority transitions. When the condition
associated with a global transition is met, it supersedes all other exit
conditions including UCT.
Extracted from 802.1D-2004 standard (17.16):
A transition that is global in nature (i.e., a transition that occurs from any
of
Acked-by: Eitan Eliahu
Thanks!
Eitan
-Original Message-
From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Sorin Vinturis
Sent: Friday, November 14, 2014 3:35 AM
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH] datapath-windows: Removed all duplicate checking of
Right now th
A possible commit message:
"Admin-port-state is the Administrative Bridge Port state variable defined
in the 802.1D-2004 standard.
It can be set to include or exclude a port from the active topology by
management (see 7.4).
operPointToPointMAC and adminPointToPointMAC are a pair of parameters that
Commit message:
"Fix Topology Change State Machine transition.
Accordingly to the 802.1D-2004 standard, the transition from the
TOPOLOGY_CHANGE_SM_NOTIFIED_TCN_EXEC state
should go to the TOPOLOGY_CHANGE_SM_NOTIFIED_TC_EXEC state and not to the
TOPOLOGY_CHANGE_SM_ACTIVE.
See 17.31 of 802.1D-2004
Possible commit message:
"Invoke move_rstp__() in rstp_port_set_administrative_bridge_port__() if
port is
not initializing.
This was necessary in a test that check that a Port becoming
alternate/backup/disabled stops learning and forwarding.
The move_rstp__() call is necessary to immediatly disable
I'm sending a v2 of this patch with all the requested modifications.
Regards,
Daniele
2014-11-14 1:17 GMT+01:00 Jarno Rajahalme :
> Daniele,
>
> See my comments below,
>
> Thanks!
>
> Jarno
>
> On Nov 6, 2014, at 7:31 AM, Daniele Venturino
> wrote:
>
> > When the condition associated wit
Possible commit message:
"This patch was necessary in a test on the acceptable range values of
max_age and forward_delay.
Since rstp_set_bridge_max_age__() and rstp_set_bridge_forward_delay__() set
the rstp->bridge_max_age
and rstp->bridge_forward_delay variables, it was necessary to call
updt_rol
>
> Why lose the ‘_WITH_HASH’? It seems pointless to iterate over all the
> ports, when we just care of the port with the ‘port_number’?
With the _WITH_HASH a test on learned addresses shifting fails. It seems
like it doesn't even enter the FOR_EACH cycle when the port_number is not
1, maybe ther
The commit message and your modifications look good to me.
Daniele
2014-11-14 0:19 GMT+01:00 Jarno Rajahalme :
> Assuming this is a good commit message:
>
> “If the Unknown value of the Port Role parameter is received, the
> state machines should treat the RST BPDU as if it were a
> Configuratio
These actions will allow L2->L3 and L3->L2 switching, and are supposed
to be added to flows installed in the datapath transparently by
ovs-vswitchd.
Signed-off-by: Lorand Jakab
---
datapath/linux/compat/include/linux/openvswitch.h | 12 +
lib/dpif-netdev.c |
Implementation of the pop_eth and push_eth actions in the kernel, and
layer 3 flow support.
Signed-off-by: Lorand Jakab
---
datapath/actions.c| 87 ---
datapath/flow.c | 45 --
datapath/flow.h |
This commit relaxes the assumption that all packets have an Ethernet
header, and adds support for layer 3 flows. For each packet received on
the Linux kernel datapath the l2 and l3 members of struct ofpbuf are
intialized appropriately, and some functions now expect this (notably
flow_extract()), i
This is needed for sending a packet back to the datapath after a miss upcall
was processed. The presence of a layer 2 packet is signaled by adding
OVS_KEY_ATTR_ETHERNET to the packet metadata sent with the ovs_packet netlink
message.
Signed-off-by: Lorand Jakab
---
lib/flow.c | 1 +
lib/fl
This series implements support for layer 3 ports, of which we have one example
so far, the LISP vport. LISP support is currently implemented with a hack, by
adding/removing the Ethernet header within the datapath/vport-lisp.c file. By
removing the assumption that all packets/flows have Ethernet h
The presence of a layer 3 packet is signaled by the absence of
OVS_KEY_ATTR_ETHERNET in the packet metadata passed from user space.
Signed-off-by: Lorand Jakab
---
datapath/datapath.c | 29 +
datapath/flow.c | 19 +--
2 files changed, 34 insertions
I came around a BSOD that happened when trying to access pidHashLock
from the gOvsSwitchContext, which was NULL. The stop happened in
OvsAcquirePidHashLock function.
To reproduce this BSOD, make sure the extension is enabled and running,
disable it and, after that, execute 'ovs-dpctl.exe show'. Th
From: Dennis Flynn
Signed-off-by: Ludovic Beliveau
Signed-off-by: Dennis Flynn
diff --git a/lib/odp-util.h b/lib/odp-util.h
index 00dbf7b..96e7009 100644
--- a/lib/odp-util.h
+++ b/lib/odp-util.h
@@ -40,6 +40,7 @@ struct pkt_metadata;
SPR(SLOW_BFD,"bfd","Consists of BFD pa
From: Dennis Flynn
This patch sequence provides OVS support for the IETF Auto-Attach SPBM draft
standard. This standard describes a compact method of using IEEE 802.1AB Link
Layer Discovery Protocol (LLDP) together with a IEEE 802.1aq Shortest Path
Bridging (SPB) network to automatically attach n
From: Dennis Flynn
This is the final commit in the series of commits that deliver initial support
for Auto-Attach. Specifically this commit delivers auto-attach support to the
OVS bridge layer as well as the new auto-attach commands. The OVSDB schema is
modified to define the new auto-attach entr
Hi Eitan,
The BSOD reported in issue #53 is still reproducible with this patch. I tested
and manage to reproduce the BSOD. It is normal since the request that leads to
the BSOD described in issue #53 is a cleanup one and not a device control one.
In issue #53 a IRP_MJ_CLEANUP request is receive
Hi Sorin,
Thank you for working on this issue.
Can you please confirm that BSOD described in issue #53 does not happen with
your change applied?
Eitan
-Original Message-
From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Sorin Vinturis
Sent: Friday, November 14, 2014 3:35 AM
To:
On 11/11/14 at 03:11pm, Andy Zhou wrote:
> >
> > I would like to keep the option to share the same zones because I
> > can see a lot of value to match on CTs which have been created by
> > netfilter or vice versa.
> Is it sufficient if both ip_defrag and conntrack actions uses the same
> zone as th
Right now the gOvsSwitchContext pointer is checked against NULL
in a lot of places of the OVS extension code. This check should
be done only once to avoid wasteful checks. Thus I have added the
check in the dispatch routine, before doing any processing, and
removed all other checks from the rest of
On Thu, Nov 13, 2014 at 06:34:40PM -0800, Alex Wang wrote:
> Hey Flavio,
>
> I'm thinking of adding a kernel version dependency to the rpm spec file...
> Don't know if it is possible to specify something similar to this ?
>
> Requires: kernel >= 3.10.0-123.8.1.el7.x86_64
>
> Since that is the ve
60 matches
Mail list logo
