On Fri, Mar 28, 2014 at 10:34:08AM +0200, Daniel Baluta wrote:
> On Tue, Mar 25, 2014 at 6:39 PM, Alexandru Copot
> wrote:
> > This series adds the basic communication handlers for the bundles
> > functionality. The actual message pre-validation and commits are
> > not implemented.
> >
> > Signed
On Fri, Mar 28, 2014 at 3:19 PM, Gurucharan Shetty wrote:
> Windows does not have a sleep(seconds). But it does have
> a Sleep(milliseconds). Sleep() in windows does not have a
> return value. Since we are not using the return value for xsleep()
> anywhere as of now, don't return any.
>
> Introduc
Windows does not have a sleep(seconds). But it does have
a Sleep(milliseconds). Sleep() in windows does not have a
return value. Since we are not using the return value for xsleep()
anywhere as of now, don't return any.
Introduced by commit 275eebb9 (utils: Introduce xsleep for RCU quiescent state
On Thu, Mar 27, 2014 at 11:17 AM, Kyle Mestery
wrote:
> Add support for building the in-tree kernel datapath for Linux kernel 3.13.
> There were some changes in the netlink area which required adding new
> compatibility code for this layer.
>
> Signed-off-by: Kyle Mestery
I think we probably als
On Sun, Mar 23, 2014 at 12:22 PM, Monam Agarwal
wrote:
> This patch replaces rcu_assign_pointer(x, NULL) with RCU_INIT_POINTER(x, NULL)
>
> The rcu_assign_pointer() ensures that the initialization of a structure
> is carried out before storing a pointer to that structure.
> And in the case of the
On Fri, Mar 28, 2014 at 1:42 PM, Andy Zhou wrote:
> Jesse helped to clarify how to maintain the ABI. Making the
> adjustment accordingly and add some comments.
>
> Signed-off-by: Andy Zhou
Acked-by: Jesse Gross
___
dev mailing list
dev@openvswitch.org
On Thu, Mar 27, 2014 at 2:56 PM, Wei Zhang wrote:
> diff --git a/net/openvswitch/vport-gre.c b/net/openvswitch/vport-gre.c
> index a3d6951..d64c639 100644
> --- a/net/openvswitch/vport-gre.c
> +++ b/net/openvswitch/vport-gre.c
> @@ -110,6 +110,12 @@ static int gre_rcv(struct sk_buff *skb,
>
On Fri, Mar 28, 2014 at 8:50 AM, Jarno Rajahalme wrote:
>
>> On Mar 27, 2014, at 3:59 PM, Jesse Gross wrote:
>>
>>> On Fri, Feb 21, 2014 at 11:41 AM, Jarno Rajahalme
>>> wrote:
>>> Minimize padding in sw_flow_key and move 'tp' top the main struct.
>>> These changes simplify code when accessing
On Fri, Mar 28, 2014 at 12:20:00PM -0700, Pritesh Kothari wrote:
> sparse emits the following warning:
> lib/dpif-netdev.c:1755:15: warning: Initializer entry defined twice
> lib/dpif-netdev.c:1755:15: also defined here
> due to a bug in sparse which doesn't like inlined functions which
> expands
On Fri, Mar 28, 2014 at 09:24:42PM +0200, Alexandru Copot wrote:
> On Wed, Mar 26, 2014 at 7:11 AM, Ben Pfaff wrote:
>
> > OpenFlow 1.4 Role Status Message
> >
> >
> > OpenFlow 1.4 section 7.4.4 ``Controller Role Status Message''
> > defines a new message sent by
Jesse helped to clarify how to maintain the ABI. Making the
adjustment accordingly and add some comments.
Signed-off-by: Andy Zhou
---
include/linux/openvswitch.h |7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/include/linux/openvswitch.h b/include/linux/openvswitch
From: Flavio Leitner
Date: Thu, 27 Mar 2014 11:05:34 -0300
> There are two problematic situations.
>
> A deadlock can happen when is_percpu is false because it can get
> interrupted while holding the spinlock. Then it executes
> ovs_flow_stats_update() in softirq context which tries to get
> the
From: Flavio Leitner
Date: Thu, 27 Mar 2014 09:53:56 -0300
> There are two problematic situations.
>
> A deadlock can happen when is_percpu is false because it can get
> interrupted while holding the spinlock. Then it executes
> ovs_flow_stats_update() in softirq context which tries to get
> the
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> Following patch will be easier to reason about with separate
> ovs_flow_cmd_new() and ovs_flow_cmd_set() functions.
>
> Signed-off-by: Jarno Rajahalme
> ---
> datapath/datapath.c | 160
> +
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> Signed-off-by: Jarno Rajahalme
> ---
> v5: Make functionality after ovs_unlock() case specific to make the common
> case more optimal.
>
> datapath/datapath.c | 135
> ++-
> 1 file cha
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> A datapath pointer is available only when holding a lock. Change
> ovs_flow_cmd_fill_info() and ovs_flow_cmd_build_info() to take a
> dp_ifindex directly, rather than a datapath pointer that is then
> (only) used to get the dp_ifindex. Th
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> Masks are inserted when flows are inserted to the table, so it is
> logical to correspondingly remove masks when flows are removed from
> the table, in ovs_flow_table_remove().
>
> This allows later patches to relax the locking requirements
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> ovs_flow_cmd_del() now allocates reply (if needed) after the flow has
> already been removed from the flow table. If the reply allocation
> fails, a netlink error is signaled with netlink_set_err(), as is
> already done in ovs_flow_cmd_new
On Mar 28, 2014, at 1:26 PM, Pravin Shelar wrote:
> On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme
> wrote:
>> Following patch will be easier to reason about with separate
>> ovs_flow_cmd_new() and ovs_flow_cmd_set() functions.
>>
>> Signed-off-by: Jarno Rajahalme
>> —
...
>> -
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> ovs_flow_cmd_alloc_info() does not need the flow, but the actions.
> Taking actions directly allows further optimization in a subsequent
> patch.
>
This patch would change if you agree with earlier comment.
> Signed-off-by: Jarno Rajahalme
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> 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 m
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> This allows hlist_unhashed() to be used to find out if a flow has been
> removed from the flow tables. Both nodes need to be checked, though.
>
I commented on next patch since that patch using this functionality.
> Signed-off-by: Jarno Ra
On Tue, Mar 25, 2014 at 2:35 PM, Jarno Rajahalme wrote:
> Use netlink_has_listeners() and NLM_F_ECHO flag to determine if a
> reply is needed or not for OVS_FLOW_CMD_NEW, OVS_FLOW_CMD_SET, or
> OVS_FLOW_CMD_DEL. Currently, OVS userspace does not request a reply
> for OVS_FLOW_CMD_NEW, but usually
On Wed, Mar 26, 2014 at 7:11 AM, Ben Pfaff wrote:
> OpenFlow 1.4 Role Status Message
>
>
> OpenFlow 1.4 section 7.4.4 ``Controller Role Status Message''
> defines a new message sent by a switch to notify the controller that
> its role (whether it is a master or a
sparse emits the following warning:
lib/dpif-netdev.c:1755:15: warning: Initializer entry defined twice
lib/dpif-netdev.c:1755:15: also defined here
due to a bug in sparse which doesn't like inlined functions which
expands a #define within it. This commit removes inline to make
sparse happy.
Sig
This fix restores the order of include path such that the local include/ comes
before the system /usr/include in the #include path. Thus by making sure that
include/linux/types.h and include/linux/openvswitch.h take precedence over the
similar files in /usr/include/ directory.
Signed-off-by: Prite
On Thu, Mar 27, 2014 at 09:50:35PM -0700, Andy Zhou wrote:
> From: Andy Zhou
>
> Number of bytes in 2 words should be 8, instead of 4 bytes,
> to better follow the mhash_finish() API. It is unlikely the fix
> will improve the quality of hashing results.
>
> Signed-off-by: Andy Zhou
I doubt it'
On Fri, Mar 28, 2014 at 9:58 AM, Ben Pfaff wrote:
> On Fri, Mar 28, 2014 at 09:10:45AM -0700, Gurucharan Shetty wrote:
>> For Windows, use a kernel assigned localhost TCP port to listen for
>> runtime management connections and then write it into a file
>> so that a client can read it and then mak
All,
While working on unit tests for qinq patch, I noticed this utility,
ovs-vlan-bug-workaround which seems to be compiled but not linked. It is
a work around for bugs in earlier upstream kernels in thehandling vlan
tag offloading (HW_ACCEL) that caused 802.1q single tagged vlans to break.
On Fri, Mar 28, 2014 at 10:49:29AM +0100, Thomas Graf wrote:
> Obviously we don't want to change behaviour of ovs-vsctl show in
> terms of exit code. So how about a new command 'ovs-vsctl status'
> which acts as a dashboard. It would be based on 'show' but without
> port options to keep it simple.
On Fri, Mar 28, 2014 at 09:10:44AM -0700, Gurucharan Shetty wrote:
> We should use closesocket() while closing sockets so that
> closing sockets work fine on both POSIX and Windows.
> (In POSIX, we #define closesocket close)
>
> Signed-off-by: Gurucharan Shetty
Acked-by: Ben Pfaff
_
Great. Thanks Andy and Thomas, I applied these two patches to master.
I added a NEWS item:
- ovs-vsctl now reports when ovs-vswitchd fails to create a new port or
bridge.
On Fri, Mar 28, 2014 at 12:13:45AM -0700, Andy Zhou wrote:
> Not sure if I suppose to ack this patch since I am one
On Fri, Mar 28, 2014 at 09:10:46AM -0700, Gurucharan Shetty wrote:
> Signed-off-by: Gurucharan Shetty
Acked-by: Ben Pfaff
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
For Windows, use a kernel assigned localhost TCP port to listen for
runtime management connections and then write it into a file
so that a client can read it and then make a TCP connection.
Since we do not have the infrastructure to create pidfiles on
windows as of now, we create the *.ctl file wi
Got reviewed by Ethan offline, so pushed to master,
On Tue, Mar 25, 2014 at 3:18 PM, Alex Wang wrote:
> Hey Ben,
>
> Could you review this?
>
> Thanks,
> Alex Wang,
>
>
> On Tue, Mar 25, 2014 at 2:24 PM, Alex Wang wrote:
>
>> Commit 61057e884ca9c(ofproto-dpif-upcall: Slightly simplify
>> udpif
Franck BAUDIN wrote:
- It requires OVS to link against the library. We don't want to link
against third-party libraries, and I don't think this will work for most
distributions anyway, unless you're planning to upstream a library to the
various Linux distributions.
The DPI library itsel
On Fri, Mar 28, 2014 at 09:10:45AM -0700, Gurucharan Shetty wrote:
> For Windows, use a kernel assigned localhost TCP port to listen for
> runtime management connections and then write it into a file
> so that a client can read it and then make a TCP connection.
>
> Since we do not have the infras
On Thu, Mar 27, 2014 at 10:41 PM, Neelakantam Gaddam
wrote:
> Hi All,
>
> Are there any design/implementation guide available for openvswitch
> userspace and kernel space components ?
Most of the documentation is in the header files (i.e *.h files).
There are some files like PORTING, DESIGN data
On Fri, Mar 28, 2014 at 12:16:29PM -0400, Thomas F Herbert wrote:
> While working on unit tests for qinq patch, I noticed this utility,
> ovs-vlan-bug-workaround which seems to be compiled but not linked.
> It is a work around for bugs in earlier upstream kernels in
> thehandling vlan tag offloadin
On Fri, Mar 28, 2014 at 12:28:58AM -0700, Andy Zhou wrote:
> On Thu, Mar 27, 2014 at 8:29 PM, Ben Pfaff wrote:
> > I don't think it's a good idea to actually exit with an error code,
> > because it's a change in behavior that could break scripts. For
> > example, I bet that some people use "ovs-v
> On Mar 27, 2014, at 3:59 PM, Jesse Gross wrote:
>
>> On Fri, Feb 21, 2014 at 11:41 AM, Jarno Rajahalme
>> wrote:
>> Minimize padding in sw_flow_key and move 'tp' top the main struct.
>> These changes simplify code when accessing the transport port numbers
>> and the tcp flags, and makes the
On Fri, Mar 28, 2014 at 09:10:43AM -0700, Gurucharan Shetty wrote:
> When we send a wrong command to a Open vSwitch daemon, it returns
> an error. In that case, we should gracefully close the connection
> to the daemon.
>
> Signed-off-by: Gurucharan Shetty
At root, I don't think this does anythi
We should use closesocket() while closing sockets so that
closing sockets work fine on both POSIX and Windows.
(In POSIX, we #define closesocket close)
Signed-off-by: Gurucharan Shetty
---
lib/stream-tcp.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/stream-tcp.c b/
Signed-off-by: Gurucharan Shetty
---
BUILD.Windows |3 ++-
utilities/ovs-appctl.8.in |9 +
utilities/ovs-appctl.c|7 +++
3 files changed, 18 insertions(+), 1 deletion(-)
diff --git a/BUILD.Windows b/BUILD.Windows
index 119334b..227822f 100644
--- a/BUILD.W
When we send a wrong command to a Open vSwitch daemon, it returns
an error. In that case, we should gracefully close the connection
to the daemon.
Signed-off-by: Gurucharan Shetty
---
utilities/ovs-appctl.c |1 +
1 file changed, 1 insertion(+)
diff --git a/utilities/ovs-appctl.c b/utilities
On Thu, Mar 27, 2014 at 09:50:36PM -0700, Andy Zhou wrote:
> From: Andy Zhou
>
> Add hash_uint64() and apply it when appropriate.
>
> Signed-off-by: Andy Zhou
I don't think the casts in hash_uint64() are necessary, but they don't
hurt.
Acked-by: Ben Pfaff
> On Mar 17, 2014, at 6:57 PM, Joe Stringer wrote:
>
>> On Fri, Mar 7, 2014 at 2:45 PM, Ethan Jackson wrote:
>> I'm surprised its only 10%, do you have a sense of why? Perhaps when
>> running TCP_CRR we delete the flow the first time we see it, and we
>> need to do xlate_actions() to generate
On Thu, Mar 27, 2014 at 09:50:37PM -0700, Andy Zhou wrote:
> From: Andy Zhou
>
> Add hash_mac() and apply it when appropriate.
>
> Signed-off-by: Andy Zhou
Same comment as previous patch, but for hash_uint64_basis().
Acked-by: Ben Pfaff
___
dev mai
On 03/28/2014 05:02 AM, Ben Pfaff wrote:
From: Andy Zhou
ovs-vsctl is a command-line interface to the Open vSwitch database,
and as such it just modifies the desired Open vSwitch configuration in
the database. ovs-vswitchd, on the other hand, monitors the database
and implements the actual con
On Tue, Mar 25, 2014 at 6:39 PM, Alexandru Copot wrote:
> This series adds the basic communication handlers for the bundles
> functionality. The actual message pre-validation and commits are
> not implemented.
>
> Signed-off-by: Alexandru Copot
> Cc: Daniel Baluta
We found some bugs in this pat
On Thu, Mar 27, 2014 at 8:29 PM, Ben Pfaff wrote:
> On Thu, Mar 27, 2014 at 05:10:31PM +0100, Thomas Graf wrote:
>> From: Andy Zhou
>>
>> ovs-vsctl is a command-line interface to the Open vSwitch database,
>> and as such it just modifies the desired Open vSwitch configuration in
>> the database.
Not sure if I suppose to ack this patch since I am one of authors.
I did review v4 in detail and tested it .
Just in case it is useful, here is my ack.
Acked-by: Andy Zhou
On Thu, Mar 27, 2014 at 9:02 PM, Ben Pfaff wrote:
> From: Andy Zhou
>
> ovs-vsctl is a command-line interface to the Op
I don't know the exact syntex and semitics of this file, but judging
from the fact the 2nd patch compiles and works -- I have tested,
these changes must be fine.
Acked-by: Andy Zhou
On Thu, Mar 27, 2014 at 9:02 PM, Ben Pfaff wrote:
> This allows a client to obtain the IDL version of a row given
53 matches
Mail list logo
