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-by: Joe Stringer
Signed-off-by: Simon Horman
---
v15
* Add additional flow
[ CC Ethan Jackson.
Ethan, this makes changes to ofproto-dpif and ofproto-dpif-xlate.
We would appreciate some feedback with how these changes fit with
your ongoing refactoring in this area. ]
Recirculation is a technique to allow a frame to re-enter
frame processing. This is intended to be
Signed-off-by: Joe Stringer
---
tests/ofproto-dpif.at | 21 +
1 file changed, 21 insertions(+)
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at
index 5e2afb9..7204195 100644
--- a/tests/ofproto-dpif.at
+++ b/tests/ofproto-dpif.at
@@ -265,6 +265,7 @@ cookie=0x6 tab
"if_cfg_queue" is not a member of "struct bridge" in bridge.c
Signed-off-by: Min Chen
---
vswitchd/bridge.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/vswitchd/bridge.c b/vswitchd/bridge.c
index 686eb93..608bc5c 100644
--- a/vswitchd/bridge.c
+++ b/vswitchd/bridg
On 10 July 2013 19:03, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Acked-by: Ed Maste
(I rebased on master and had a trivial conflict to resolve.)
X-CudaMail-Whitelist-To: dev@openvswitch.org
___
dev mailing list
dev@openvswitch.org
http://openvswitc
On 10 July 2013 19:03, Ben Pfaff wrote:
> This should make the "timeval" module thread-safe, except for its
> calls into vlog (because vlog is not yet thread-safe).
>
> Signed-off-by: Ben Pfaff
Acked-by: Ed Maste
X-CudaMail-Whitelist-To: dev@openvswitch.org
_
On Thu, Jul 11, 2013 at 6:55 PM, Simon Horman wrote:
> On Thu, Jul 11, 2013 at 06:26:32PM -0700, Jesse Gross wrote:
>> On Wed, Jul 10, 2013 at 5:16 PM, Simon Horman wrote:
>> > On Mon, Jul 08, 2013 at 11:51:18PM -0700, Jesse Gross wrote:
>> >> On Tue, Jul 2, 2013 at 6:31 PM, Simon Horman wrote:
On Fri, Jul 12, 2013 at 04:58:13PM +0900, Joe Stringer wrote:
> Signed-off-by: Joe Stringer
Thanks Joe, I applied this to master.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
I don't see anything that guaranteed that ovs-ofctl would receive and print
the packets before it exited. This commit inserts explicit waits, to avoid
the problem.
Found by inspection.
Signed-off-by: Ben Pfaff
---
tests/ofproto-dpif.at | 46 +++---
1 f
On Fri, Jul 12, 2013 at 05:59:34PM +0800, Min Chen wrote:
> "if_cfg_queue" is not a member of "struct bridge" in bridge.c
>
> Signed-off-by: Min Chen
I tried to apply your patch but failed:
Applying: vswitchd: Fix typo in comment
fatal: corrupt patch at line 10
Patch failed at 0001 vswitchd: Fi
On Jul 12, 2013, at 9:41 AM, Ben Pfaff wrote:
> On Fri, Jul 12, 2013 at 05:59:34PM +0800, Min Chen wrote:
>> "if_cfg_queue" is not a member of "struct bridge" in bridge.c
>>
>> Signed-off-by: Min Chen
>
> I tried to apply your patch but failed:
>
> Applying: vswitchd: Fix typo in comment
> fa
The OVS MAC address set function was removed in favor of the version
in the kernel but the function pointer for older kernels was not
updated.
Reported-by: Cali Ente
Signed-off-by: Jesse Gross
---
datapath/vport-internal_dev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/
On Fri, Jul 12, 2013 at 10:15 AM, Jesse Gross wrote:
> The OVS MAC address set function was removed in favor of the version
> in the kernel but the function pointer for older kernels was not
> updated.
>
> Reported-by: Cali Ente
> Signed-off-by: Jesse Gross
> ---
looks good.
Acked-by: Pravin B
On Fri, Jul 12, 2013 at 11:06 AM, Pravin Shelar wrote:
> On Fri, Jul 12, 2013 at 10:15 AM, Jesse Gross wrote:
>> The OVS MAC address set function was removed in favor of the version
>> in the kernel but the function pointer for older kernels was not
>> updated.
>>
>> Reported-by: Cali Ente
>> Si
On Fri, Jul 12, 2013 at 09:16:10AM -0400, Ed Maste wrote:
> On 10 July 2013 19:03, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
>
> Acked-by: Ed Maste
Thanks, I'll apply to master soon.
> (I rebased on master and had a trivial conflict to resolve.)
Oh, sorry, I forgot to update my github "t
On Fri, Jul 12, 2013 at 09:19:39AM -0400, Ed Maste wrote:
> On 10 July 2013 19:03, Ben Pfaff wrote:
> > This should make the "timeval" module thread-safe, except for its
> > calls into vlog (because vlog is not yet thread-safe).
> >
> > Signed-off-by: Ben Pfaff
>
> Acked-by: Ed Maste
Thanks, I
Thanks for getting this together. Comments below.
Major stuff:
When anything changes about the configuration of the decay feature,
the decay_detect_time needs to be reset and we should switch back to
the original min_rx. To be safe, I'd also reset it when the min_rx
changes.
I wouldn't bother
On Wed, Jul 10, 2013 at 4:03 PM, Ben Pfaff wrote:
> The worker process implementation isn't thread-safe and, once OVS
> itself is threaded, it doesn't make much sense to have a worker
> process anyway.
>
> Signed-off-by: Ben Pfaff
>
Looks good.
> ---
> vswitchd/system-stats.c | 145
>
This makes "top" and "ps" output more readable on FreeBSD at least, and
the names are also visible in debuggers.
Suggested-by: Ed Maste
Signed-off-by: Ben Pfaff
---
One reason that I didn't know about these functions was that my main
dev box runs a distro so old that its glibc doesn't even have
On Thu, Jul 11, 2013 at 05:16:33PM -0700, Ben Pfaff wrote:
> On Thu, Jul 11, 2013 at 09:40:57AM -0400, Ed Maste wrote:
> > On 10 July 2013 19:04, Ben Pfaff wrote:
> > > Signed-off-by: Ben Pfaff
> >
> > > +void
> > > +set_subprogram_name(const char *name)
> > > +{
> > > +free(subprogram_name_
On Thu, Jul 11, 2013 at 09:26:58AM -0400, Ed Maste wrote:
> On 10 July 2013 19:04, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
>
> Acked-by: Ed Maste
Thanks, I'll apply this soon.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mail
On Thu, Jul 11, 2013 at 10:03:07AM -0400, Ed Maste wrote:
> On 10 July 2013 19:04, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
>
> Acked-by: Ed Maste
Thanks, I'll apply this soon.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mail
On Thu, Jul 11, 2013 at 09:56:43AM -0400, Ed Maste wrote:
> On 10 July 2013 19:04, Ben Pfaff wrote:
> > This is harder to implement once threads are introduced.
> >
> > Signed-off-by: Ben Pfaff
>
> Acked-by: Ed Maste
Thanks, I'll apply this soon.
___
On Wed, Jul 10, 2013 at 4:03 PM, Ben Pfaff wrote:
> It had no remaining users.
>
> Signed-off-by: Ben Pfaff
>
Looks good to me.
> ---
> Makefile.am |2 +-
> lib/automake.mk |4 +-
> lib/util.c |3 +-
> lib/worker.c| 472
> -
On Fri, Jul 12, 2013 at 02:37:34PM -0700, Gurucharan Shetty wrote:
> On Wed, Jul 10, 2013 at 4:03 PM, Ben Pfaff wrote:
>
> > It had no remaining users.
> >
> > Signed-off-by: Ben Pfaff
> >
> Looks good to me.
Thanks, applied.
___
dev mailing list
dev@
On Fri, Jul 12, 2013 at 02:18:41PM -0700, Gurucharan Shetty wrote:
> On Wed, Jul 10, 2013 at 4:03 PM, Ben Pfaff wrote:
>
> > The worker process implementation isn't thread-safe and, once OVS
> > itself is threaded, it doesn't make much sense to have a worker
> > process anyway.
> >
> > Signed-off
On Thu, Jul 11, 2013 at 08:59:22AM -0400, Ed Maste wrote:
> On 10 July 2013 19:04, Ben Pfaff wrote:
> > static void format_log_message(const struct vlog_module *, enum vlog_level,
> > - enum vlog_facility, unsigned int msg_num,
> > + enu
Many patches have been committed since v1. Of the remaining patches,
I think only "vlog: Make thread-safe." has changed, and only to fix
a PRINTF_FORMAT annotation.
Ben Pfaff (13):
stress: Remove essentially unused library.
netlink-socket: Simplify use of transactions and dumps.
netlink-soc
The "stress" library was introduced years ago. We intended at the time to
start using it to provoke errors in testing, to make sure that Open vSwitch
was resilient against those errors. The intention was good, but there were
few actual implementations of stress options, and the testing never
mate
The uses of vlog in this module are not thread-safe, because vlog itself
is not yet thread-safe.
Signed-off-by: Ben Pfaff
---
lib/netlink-socket.c | 24 +++-
lib/netlink-socket.h |6 ++
2 files changed, 25 insertions(+), 5 deletions(-)
diff --git a/lib/netlink-sock
The simplification comes from dropping support for canceling a
poll_waiter, which was a feature that was never used. The speedup
comes from avoiding a malloc() for every call to poll_fd_wait().
(I doubt that this significantly improves performance.)
This prepares for making the polling structures
Signed-off-by: Ben Pfaff
---
lib/automake.mk |2 +
lib/latch.c | 72 +++
lib/latch.h | 37
3 files changed, 111 insertions(+), 0 deletions(-)
create mode 100644 lib/latch.c
create mode 100644 lib/la
Signed-off-by: Ben Pfaff
---
lib/poll-loop.c | 107 +++
lib/poll-loop.h | 14 +--
2 files changed, 78 insertions(+), 43 deletions(-)
diff --git a/lib/poll-loop.c b/lib/poll-loop.c
index 567c19f..97fc806 100644
--- a/lib/poll-loop.c
+++ b
This disentangles "struct nl_dump" from "struct nl_sock", clearing the way
to make the use of either one thread-safe in an obviously correct manner.
Signed-off-by: Ben Pfaff
---
lib/dpif-linux.c | 20 ++---
lib/netdev-linux.c | 18 +
lib/netlink-socket.c | 201
Signed-off-by: Ben Pfaff
---
vswitchd/system-stats.c | 105 +-
1 files changed, 84 insertions(+), 21 deletions(-)
diff --git a/vswitchd/system-stats.c b/vswitchd/system-stats.c
index e7c1d73..ed63899 100644
--- a/vswitchd/system-stats.c
+++ b/vswitchd
Signed-off-by: Ben Pfaff
---
lib/byteq.c | 26 +++---
lib/byteq.h | 10 --
lib/jsonrpc.c |5 +++--
3 files changed, 22 insertions(+), 19 deletions(-)
diff --git a/lib/byteq.c b/lib/byteq.c
index 2ee4a65..3f865cf 100644
--- a/lib/byteq.c
+++ b/lib/byteq.c
@
This will be hooked into the vlog library in an upcoming commit.
Signed-off-by: Ben Pfaff
---
configure.ac|1 +
lib/async-append-aio.c | 178 +++
lib/async-append-sync.c | 62
lib/async-append.h | 67
Signed-off-by: Ben Pfaff
---
lib/util.h| 19 +++
tests/library.at |2 +
tests/test-util.c | 65 +
3 files changed, 86 insertions(+), 0 deletions(-)
diff --git a/lib/util.h b/lib/util.h
index ae8bfd7..2159594 100644
Signed-off-by: Ben Pfaff
---
lib/vlog.c | 182 +++-
lib/vlog.h |9 +++
2 files changed, 116 insertions(+), 75 deletions(-)
diff --git a/lib/vlog.c b/lib/vlog.c
index d134c29..074225e 100644
--- a/lib/vlog.c
+++ b/lib/vlog.c
@@ -33,6 +3
Signed-off-by: Ben Pfaff
---
lib/vlog.c| 12 +++-
vswitchd/bridge.c |3 +++
2 files changed, 14 insertions(+), 1 deletions(-)
diff --git a/lib/vlog.c b/lib/vlog.c
index ded434f..d134c29 100644
--- a/lib/vlog.c
+++ b/lib/vlog.c
@@ -28,6 +28,7 @@
#include
#include
#inclu
Of course, the code it calls into isn't thread-safe itself yet.
Signed-off-by: Ben Pfaff
---
lib/vlog.h | 14 +++---
1 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/lib/vlog.h b/lib/vlog.h
index 3466d96..eeec5fc 100644
--- a/lib/vlog.h
+++ b/lib/vlog.h
@@ -231,13 +231,13
Thanks Ethan, for the comments,
On Fri, Jul 12, 2013 at 2:15 PM, Ethan Jackson wrote:
> In bfd_run() the final if statement should run whether or not decay is
> enabled. It will have to become a bit more sophisticated, but
> assuming my suggestion in the previous patch works, it should be
> st
RCUfy dp-dump operation which is already read-only. This
makes all ovs dump operations lockless.
Signed-off-by: Pravin B Shelar
---
datapath/datapath.c |7 ---
1 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/datapath/datapath.c b/datapath/datapath.c
index 2f02f71..09367cf
On Fri, Jul 12, 2013 at 2:54 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> vswitchd/system-stats.c | 105
> +-
> 1 files changed, 84 insertions(+), 21 deletions(-)
>
> diff --git a/vswitchd/system-stats.c b/vswitchd/system-stats.c
> index e
When deleting subfacets, subfacet_uninstall() asserts that the
subfacet's counters are zero to make sure we don't lose counters. We
have seen cases where a subfacet cannot be installed, but the counters
have values. This should not happen, since we shouldn't create a
subfacet if the datapath has
On Fri, Jul 12, 2013 at 2:54 PM, Ben Pfaff wrote:
> The "stress" library was introduced years ago. We intended at the time to
> start using it to provoke errors in testing, to make sure that Open vSwitch
> was resilient against those errors. The intention was good, but there were
> few actual i
One thing that isn't entirely clear from the commit message and comment is
whether we believe this message, when it appears, indicates that there is a
bug somewhere. I think the intention is that the message indicates a bug,
I'm just saying this to check that I'm right.
The code change definitely
Don't install a flow if it's the result of the "userspace" action and it
contains a slow-path action. This can occur when a datapath flow with
wildcards has a "userspace" action and flows sent to userspace result in
a different subfacet, which will then be rejected as overlapping by the
datapath.
This makes me a bit uncomfortable because we're going to have facet's floating
around which have never been installed. I would expect the rest of ofproto-dpif
to be unhappy about that, though I can't say exactly how. At any rate, how
about an alternative approach. If a packet came in as DPIF_UC_
On Fri, Jul 12, 2013 at 4:37 PM, Pravin B Shelar wrote:
> diff --git a/datapath/datapath.c b/datapath/datapath.c
> index 2f02f71..09367cf 100644
> --- a/datapath/datapath.c
> +++ b/datapath/datapath.c
> @@ -1902,7 +1903,7 @@ static int ovs_dp_cmd_dump(struct sk_buff *skb, struct
> netlink_callbac
When there is no incoming data traffic at the interface for a period,
BFD decay allows the bfd session to increase the min_rx. This is
helpful in that some interfaces usually idle for long time. And cpu
consumption can be reduced by processing fewer bfd control packets.
Signed-off-by: Alex Wang
-
Hey Ethan,
Sorry, I didn't add the version number. Should be [PATCH 1/2 V2]
One clarification:
"""
+bfd->decay_detect_time = (bfd->decay_min_rx < 2000 ?
+ 2000 : bfd->decay_min_rx) + time_msec();
"""
The decay_detect_time should be greater than 2000, since
"facet_
Okay, I made it more explicit in both and pushed to affected branches. Thanks
for the review and strategy discussion.
--Justin
On Jul 12, 2013, at 5:21 PM, Ben Pfaff wrote:
> One thing that isn't entirely clear from the commit message and comment is
> whether we believe this message, when i
The install guide here:
http://n40lab.wordpress.com/2013/06/03/centos-6-4-openvswitch-installation/
shows unexpected startup errors:
#/etc/init.d/openvswitch start
/usr/share/openvswitch/scripts/ovs-lib: line 49:
/var/log/openvswitch/ovs-ctl.log: No such file or directory
tee: /
On Fri, Jul 12, 2013 at 7:31 PM, Ben Pfaff wrote:
> The install guide here:
>
>
> http://n40lab.wordpress.com/2013/06/03/centos-6-4-openvswitch-installation/
>
> shows unexpected startup errors:
>
> #/etc/init.d/openvswitch start
> /usr/share/openvswitch/scripts/ovs-lib: line 49:
> /var/l
During installation create the /var/log/openvswitch directory
so that openvswitch startup script is able to write the ovs-ctl.log
Signed-off-by: Gurucharan Shetty
---
rhel/openvswitch-fedora.spec.in |2 ++
rhel/openvswitch.spec.in |2 ++
xenserver/openvswitch-xen.spec.in |
Thanks, this looks good. But I am surprised that "make install" does not
create the log directory. Perhaps that is an underlying problem that we
should fix instead of adding a directory creation call to every packager.
On Jul 12, 2013 9:29 PM, "Gurucharan Shetty" wrote:
> During installation crea
57 matches
Mail list logo
