From: Daniele Di Proietto
commit 43d4be9cb55f3bac5253e9289996fd9d735531db (openvswitch: Allow user space
to announce ability to accept unaligned Netlink messages) introduced
OVS_DP_ATTR_USER_FEATURES netlink attribute in datapath responses,
but the attribute size was not taken into account in ovs
From: Pravin B Shelar
ovs_flow_free() is not called under ovs-lock during packet
execute path (ovs_packet_cmd_execute()). Since packet execute
does not touch flow->mask, there is no need to take that
lock either. So move assert in case where flow->mask is checked.
Found by code inspection.
Sign
A handful of bug fixes for net/3.14. High level fixes are:
* Regressions introduced by the zerocopy changes, particularly with
old userspaces.
* A few bugs lingering from the introduction of megaflows.
* Overly zealous error checking that is now being triggered frequently
in common cases.
From: Andy Zhou
With subfacets, we'd expect megaflow updates message to carry
the original micro flow. If not, EINVAL is returned and kernel
logs an error message. Now that the user space subfacet layer is
removed, it is expected that flow updates can arrive with a
micro flow other than the orig
From: Thomas Graf
While the zerocopy method is correctly omitted if user space
does not support unaligned Netlink messages. The attribute is
still not padded correctly as skb_zerocopy() will not ensure
padding and the attribute size is no longer pre calculated
though nla_reserve() which ensured p
From: Andy Zhou
Both mega flow mask's reference counter and per flow table mask list
should only be accessed when holding ovs_mutex() lock. However
this is not true with ovs_flow_table_flush(). The patch fixes this bug.
Reported-by: Joe Stringer
Signed-off-by: Andy Zhou
Signed-off-by: Jesse Gr
From: Daniele Di Proietto
commit 43d4be9cb55f3bac5253e9289996fd9d735531db (openvswitch: Allow user space
to announce ability to accept unaligned Netlink messages) introduced
OVS_DP_ATTR_USER_FEATURES netlink attribute in datapath responses,
but the attribute size was not taken into account in ovs
On Tue, Feb 4, 2014 at 10:40 PM, Jesse Gross wrote:
> A handful of bug fixes for net/3.14. High level fixes are:
> * Regressions introduced by the zerocopy changes, particularly with
>old userspaces.
> * A few bugs lingering from the introduction of megaflows.
> * Overly zealous error check
From: Andy Zhou
Both mega flow mask's reference counter and per flow table mask list
should only be accessed when holding ovs_mutex() lock. However
this is not true with ovs_flow_table_flush(). The patch fixes this bug.
Reported-by: Joe Stringer
Signed-off-by: Andy Zhou
Signed-off-by: Jesse Gr
From: Pravin B Shelar
ovs_flow_free() is not called under ovs-lock during packet
execute path (ovs_packet_cmd_execute()). Since packet execute
does not touch flow->mask, there is no need to take that
lock either. So move assert in case where flow->mask is checked.
Found by code inspection.
Sign
From: Andy Zhou
With subfacets, we'd expect megaflow updates message to carry
the original micro flow. If not, EINVAL is returned and kernel
logs an error message. Now that the user space subfacet layer is
removed, it is expected that flow updates can arrive with a
micro flow other than the orig
A handful of bug fixes for net/3.14. High level fixes are:
* Regressions introduced by the zerocopy changes, particularly with
old userspaces.
* A few bugs lingering from the introduction of megaflows.
* Overly zealous error checking that is now being triggered frequently
in common cases.
From: Thomas Graf
While the zerocopy method is correctly omitted if user space
does not support unaligned Netlink messages. The attribute is
still not padded correctly as skb_zerocopy() will not ensure
padding and the attribute size is no longer pre calculated
though nla_reserve() which ensured p
On Mon, Feb 3, 2014 at 8:43 AM, Sergei Shtylyov
wrote:
> Hello.
>
>
> On 03-02-2014 5:08, Jesse Gross wrote:
>
>> From: Thomas Graf
>
>
>> While the zerocopy method is correctly omitted if user space
>> does not support unaligned Netlink messages. The attribute is
>> still not padded correctly as
From: Fernando Luis Vázquez Cao
Date: Tue, 04 Feb 2014 19:35:02 +0900
> We should check whether rtnetlink link operations
> are defined before calling get_slave_size().
>
> Without this, the following oops can occur when
> adding a tap device to OVS.
...
> Signed-off-by: Fernando Luis Vazquez C
> On Wed, Jan 15, 2014 at 12:41:20PM +0900, YAMAMOTO Takashi wrote:
>> Add a comment about implicit synchronization which
>> fat-rwlock seems to rely on.
>>
>> Signed-off-by: YAMAMOTO Takashi
>> ---
>> lib/ovs-thread.h | 8
>> 1 file changed, 8 insertions(+)
>>
>> diff --git a/lib/ovs-
> On Fri, Jan 17, 2014 at 11:04:14AM +0900, YAMAMOTO Takashi wrote:
>> > On Wed, Jan 08, 2014 at 07:31:23PM +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
On Tue, Feb 04, 2014 at 10:45:03AM -0800, Ben Pfaff wrote:
> On Mon, Feb 03, 2014 at 04:53:45PM -0800, Ben Pfaff wrote:
> > On Tue, Feb 04, 2014 at 09:52:40AM +0900, Simon Horman wrote:
> > > On Mon, Feb 03, 2014 at 03:27:16PM -0800, Ben Pfaff wrote:
> > > > Earlier today, I applied your 2-patch se
Originally part of "odp: Allow VLAN actions after MPLS actions"
by Joe Stringer.
Co-authored-by: Joe Stringer
Signed-off-by: Simon Horman
---
tests/ofproto-dpif.at | 389 ++
1 file changed, 389 insertions(+)
diff --git a/tests/ofproto-dpif.at b/t
On Tue, Feb 04, 2014 at 12:39:37PM -0800, Ben Pfaff wrote:
> OpenFlow 1.1 and 1.2 always inserted MPLS labels after VLAN tags.
> OpenFlow 1.3 and 1.4 insert MPLS labels before VLAN tags.
> OpenFlow 1.3.4 and 1.5, both in preparation, recognize that the change in
> 1.3 was an error and revert it. T
This is the first patch series to add megaflow support for bond interface
using recirculation.
The final goal is that the bonding port selection will be made within the
datapath, using recirculation. Currently, the bonding port selection is
made in the user space, preventing possible wider mega f
Signed-off-by: Andy Zhou
---
include/linux/openvswitch.h | 26 ++
lib/dpif-netdev.c |2 ++
lib/dpif.c |2 ++
lib/odp-execute.c |2 ++
lib/odp-util.c | 13 +
5 files changed, 45 insertions(+)
di
Added recirculation support for bond.
A unique per datapath recirculation id is allocated whenever a bond
is configured with balancing mode other than active/backup. This
recirculation ID is freed back when bond configuration changes to
active/backup mode or when the bond is no longer configured.
Recirculation ID needs to be unique per datapath. Its usage will be
tracked by the backer that corresponds to the datapath.
In theory, Recirculation ID can be any uint32_t value, except 0. This
implementation limits to a smaller range just for ease of debugging.
Make the range size 0 effectively d
Signed-off-by: Andy Zhou
---
lib/match.c | 26 +
lib/match.h |6 ++
lib/meta-flow.c | 58 +++
lib/meta-flow.h |2 ++
4 files changed, 92 insertions(+)
diff --git a/lib/match.c b/lib/match.c
index
Signed-off-by: Andy Zhou
---
ofproto/bond.c |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/ofproto/bond.c b/ofproto/bond.c
index b100e19..2905464 100644
--- a/ofproto/bond.c
+++ b/ofproto/bond.c
@@ -1094,15 +1094,15 @@ bond_unixctl_list(struct unixctl_conn *conn,
Signed-off-by: Andy Zhou
---
lib/flow.c |2 +-
lib/flow.h | 12 +---
lib/match.c |4 +++-
lib/nx-match.c |2 +-
lib/ofp-util.c |2 +-
ofproto/ofproto-dpif-xlate.c |2 +-
6 files changed
'cls' is no longer a parameter to this function.
Signed-off-by: Joe Stringer
---
ofproto/ofproto.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c
index f2a88bb..a9cf221 100644
--- a/ofproto/ofproto.c
+++ b/ofproto/ofproto.c
@@ -1217,7
On Tue, Feb 4, 2014 at 1:25 PM, Pravin Shelar wrote:
> On Tue, Feb 4, 2014 at 11:10 AM, Jarno Rajahalme
> wrote:
>> diff --git a/datapath/datapath.c b/datapath/datapath.c
>> index 5f1b34c..7992330 100644
>> --- a/datapath/datapath.c
>> +++ b/datapath/datapath.c
>> @@ -116,20 +116,19 @@ static in
On Mon, Jan 27, 2014 at 11:07 AM, Linda Sun wrote:
> ---
> lib/automake.mk |9 +-
> lib/stream-fd-windows.c | 278
> +++
> lib/stream.c|4 +
> 3 files changed, 288 insertions(+), 3 deletions(-)
> create mode 100644 lib/str
On Thu, Jan 23, 2014 at 4:55 PM, Ben Pfaff wrote:
> On Thu, Jan 23, 2014 at 04:45:34PM -0800, Luigi Rizzo wrote:
>> On Thu, Jan 23, 2014 at 4:43 PM, Ben Pfaff wrote:
>>
>> > On Fri, Jan 24, 2014 at 01:29:32AM +0100, Daniele Di Proietto wrote:
>> > > I'll just attach the patch (just for a quick pr
On Tue, Feb 4, 2014 at 11:10 AM, Jarno Rajahalme wrote:
> Change ovs_dp_cmd_get() and ovs_flow_cmd_get() to only take the
> rcu_read_lock(), instead of ovs_lock(), as nothing need to be changed.
> This was done by ovs_vport_cmd_get() already.
>
ovs_dp_cmd_get() and ovs_flow_cmd_get() does blocking
On Wed, Jan 15, 2014 at 12:41:20PM +0900, YAMAMOTO Takashi wrote:
> Add a comment about implicit synchronization which
> fat-rwlock seems to rely on.
>
> Signed-off-by: YAMAMOTO Takashi
> ---
> lib/ovs-thread.h | 8
> 1 file changed, 8 insertions(+)
>
> diff --git a/lib/ovs-thread.h b/
On Fri, Jan 17, 2014 at 11:04:14AM +0900, YAMAMOTO Takashi wrote:
> > On Wed, Jan 08, 2014 at 07:31:23PM +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 resum
On Tue, Feb 4, 2014 at 11:10 AM, Jarno Rajahalme wrote:
> ovs_vport_cmd_dump() did rcu_read_lock() only after getting the
> datapath, which could have been deleted in between. Resolved by
> taking rcu_read_lock() before the get_dp() call.
>
Good catch.
> Remove unnecessary locking from functions
On Mon, Feb 03, 2014 at 09:36:11AM -0800, Ben Pfaff wrote:
> I'll look this over a bit and properly submit it.
Done: http://openvswitch.org/pipermail/dev/2014-February/036419.html
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/li
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).
- Specific ip_src, wildcarded, or flow-based (3 choices).
Until now, only 6 of the 12 possibilities have been supported. We
ha
OpenFlow 1.1 and 1.2 always inserted MPLS labels after VLAN tags.
OpenFlow 1.3 and 1.4 insert MPLS labels before VLAN tags.
OpenFlow 1.3.4 and 1.5, both in preparation, recognize that the change in
1.3 was an error and revert it. This commit implements that reversion
in Open vSwitch.
EXT-457.
Sig
ovs_vport_cmd_dump() did rcu_read_lock() only after getting the
datapath, which could have been deleted in between. Resolved by
taking rcu_read_lock() before the get_dp() call.
Remove unnecessary locking from functions that are always called with
appropriate locking (get_dp(), ovs_dp_cmd_fill_inf
Change ovs_dp_cmd_get() and ovs_flow_cmd_get() to only take the
rcu_read_lock(), instead of ovs_lock(), as nothing need to be changed.
This was done by ovs_vport_cmd_get() already.
Signed-off-by: Jarno Rajahalme
---
datapath/datapath.c | 17 ++---
1 file changed, 6 insertions(+), 1
Keep kernel flow stats for each NUMA node rather than each (logical)
CPU. This avoids using the per-CPU allocator and removes most of the
kernel-side OVS locking overhead otherwise on the top of perf report
and allows OVS to scale better with higher number of threads.
With 9 handlers and 4 revali
On Fri, Jan 24, 2014 at 10:31:14AM +0900, YAMAMOTO Takashi wrote:
> And add a comment.
>
> Signed-off-by: YAMAMOTO Takashi
All applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Fri, Jan 24, 2014 at 03:50:41PM +0900, YAMAMOTO Takashi wrote:
> Signed-off-by: YAMAMOTO Takashi
Thanks, applied to master and branches for 2.1, 2.0, 1.9.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Mon, Feb 03, 2014 at 04:53:45PM -0800, Ben Pfaff wrote:
> On Tue, Feb 04, 2014 at 09:52:40AM +0900, Simon Horman wrote:
> > On Mon, Feb 03, 2014 at 03:27:16PM -0800, Ben Pfaff wrote:
> > > Earlier today, I applied your 2-patch series to the mpls branch
> > > earlier.
> > >
> > > Then, I squashe
On Tue, Feb 4, 2014 at 9:19 AM, Jarno Rajahalme wrote:
>
> On Feb 3, 2014, at 9:56 AM, Jesse Gross wrote:
>
>> On Fri, Jan 24, 2014 at 2:58 PM, Jarno Rajahalme
>> wrote:
>>> Userspace already sets NLM_F_ECHO when it wants a reply, make use of
>>> it.
>>>
>>> For set and del, OVS userspace only
On Tue, Feb 4, 2014 at 9:03 AM, Jarno Rajahalme wrote:
> Pravin,
>
> Thanks for your thorough review. Some comments below,
>
> Jarno
>
> On Feb 3, 2014, at 9:14 AM, Pravin Shelar wrote:
>
> On Thu, Jan 23, 2014 at 4:43 PM, Jarno Rajahalme
> wrote:
>
>
> ...
>
> +static void get_skb_stats(struc
On Feb 3, 2014, at 9:56 AM, Jesse Gross wrote:
> On Fri, Jan 24, 2014 at 2:58 PM, Jarno Rajahalme
> wrote:
>> Userspace already sets NLM_F_ECHO when it wants a reply, make use of
>> it.
>>
>> For set and del, OVS userspace only sets the NLM_F_ECHO if stats are
>> needed back. So, we minimize
On Feb 3, 2014, at 10:38 AM, Jesse Gross wrote:
> On Fri, Jan 24, 2014 at 2:58 PM, Jarno Rajahalme
> wrote:
>> diff --git a/datapath/datapath.c b/datapath/datapath.c
>> index 92ae66a..5b12a5d 100644
>> --- a/datapath/datapath.c
>> +++ b/datapath/datapath.c
>> @@ -817,11 +840,27 @@ static int o
Pravin,
Thanks for your thorough review. Some comments below,
Jarno
On Feb 3, 2014, at 9:14 AM, Pravin Shelar wrote:
> On Thu, Jan 23, 2014 at 4:43 PM, Jarno Rajahalme
> wrote:
>>
...
>> +static void get_skb_stats(struct flow_stats *stat, const struct sw_flow
>> *flow,
>> +
Until now, the tcp_stream() code has ignored any TCP packets that don't
correspond to a stream that began with a TCP SYN. This commit changes the
code so that any TCP packet that contains a payload starts a new stream.
Signed-off-by: Ben Pfaff
Reported-by: Vasu Dasari
---
AUTHORS |
On Tue, Feb 04, 2014 at 08:29:41AM -0800, Gurucharan Shetty wrote:
> >
> > That's a nice start, let's go wiht that fo rnow.
>
> Thanks for the review. I pulled in changes from patch2 of the series
> (as you suggested) to this patch and committed them.
Great! I hope that we are making progress on
>
> That's a nice start, let's go wiht that fo rnow.
Thanks for the review. I pulled in changes from patch2 of the series
(as you suggested) to this patch and committed them.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinf
Hi,
Are you interested in acquiring *VMware* contact data that will help you to
reach out to all the key decision makers?
*Below are few targeted lists:-*
Autonomy
Mobile iron
Checkpoint
Blue coat
NetApp
CSA
CA Clarity
PTC-CREO
Commvault
Citrix
QlikView
Dassault
DB2
Right Now
i
Tue, Feb 04, 2014 at 11:35:02AM CET, fernando...@lab.ntt.co.jp wrote:
>We should check whether rtnetlink link operations
>are defined before calling get_slave_size().
>
>Without this, the following oops can occur when
>adding a tap device to OVS.
>
>[ 87.839553] BUG: unable to handle kernel NULL
We should check whether rtnetlink link operations
are defined before calling get_slave_size().
Without this, the following oops can occur when
adding a tap device to OVS.
[ 87.839553] BUG: unable to handle kernel NULL pointer dereference at
00a8
[ 87.839595] IP: [] if_nlmsg_size+
55 matches
Mail list logo
