Thank you! I applied this to master.
On Thu, Jan 16, 2014 at 06:08:12PM -0800, Alex Wang wrote:
> Looks good to me~
>
>
> On Thu, Jan 16, 2014 at 5:22 PM, Ben Pfaff wrote:
>
> > Commit 73c85181d (netdev-linux: Read packet auxdata to obtain vlan_tid)
> > added #include to this file, to get th
On Thu, Jan 16, 2014 at 05:28:53PM -0800, Ben Pfaff wrote:
> On Fri, Jan 17, 2014 at 09:59:49AM +0900, Simon Horman wrote:
> > On Thu, Jan 16, 2014 at 04:46:23PM -0800, Ben Pfaff wrote:
> > > On Wed, Jan 15, 2014 at 04:13:25PM +0900, Simon Horman wrote:
> > > > The key is sourced from the datapath
Looks good to me~
On Thu, Jan 16, 2014 at 5:22 PM, Ben Pfaff wrote:
> Commit 73c85181d (netdev-linux: Read packet auxdata to obtain vlan_tid)
> added #include to this file, to get the definition
> of PACKET_AUXDATA and some other definitions, but on RHEL 6.1 this
> provoked compiler errors:
>
> 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 resume" test.
>>
>> Signed-off-by: YAMAMOTO Takashi
>
> Clang says:
>
>
On Mon, Dec 09, 2013 at 09:01:57AM +0100, Helmut Schaa wrote:
> On Sat, Dec 7, 2013 at 6:20 PM, Ben Pfaff wrote:
> > On Fri, Dec 06, 2013 at 04:18:42PM +0100, Helmut Schaa wrote:
> >> When adding a physical port to the main bridge the mac address
> >> of the bridge is updated. We can do the same f
On Fri, Jan 17, 2014 at 09:59:49AM +0900, Simon Horman wrote:
> On Thu, Jan 16, 2014 at 04:46:23PM -0800, Ben Pfaff wrote:
> > On Wed, Jan 15, 2014 at 04:13:25PM +0900, Simon Horman wrote:
> > > The key is sourced from the datapath so should not
> > > have more labels than it can handle.
> > >
> >
Commit 73c85181d (netdev-linux: Read packet auxdata to obtain vlan_tid)
added #include to this file, to get the definition
of PACKET_AUXDATA and some other definitions, but on RHEL 6.1 this
provoked compiler errors:
In file included from /usr/include/linux/rtnetlink.h:5,
from lib/netdev-linux
On Thu, Jan 16, 2014 at 04:50:45PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 04:13:17PM +0900, Simon Horman wrote:
> > This series provides some proposed improvements to your patchset which
> > includes "Implement OpenFlow support for MPLS, for up to 3 labels."
> >
> > This series is based
On Thu, Jan 16, 2014 at 04:46:23PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 04:13:25PM +0900, Simon Horman wrote:
> > The key is sourced from the datapath so should not
> > have more labels than it can handle.
> >
> > dpif-netdev supports as many LSEs as can fit in struct flow,
> > so it
On Wed, Jan 15, 2014 at 04:13:17PM +0900, Simon Horman wrote:
> This series provides some proposed improvements to your patchset which
> includes "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> This series is based on the blp/mpls branch of your tree on github. The
> head commit I
On Wed, Jan 15, 2014 at 04:13:25PM +0900, Simon Horman wrote:
> The key is sourced from the datapath so should not
> have more labels than it can handle.
>
> dpif-netdev supports as many LSEs as can fit in struct flow,
> so it is safe to pass SIZE_MAX as the limit there.
>
> This is an proposed e
I've just tested v4, and haven't seen any further kernel panics. Thanks!
On 16 January 2014 15:42, Jesse Gross wrote:
> On Wed, Jan 15, 2014 at 7:52 PM, Andy Zhou wrote:
> > Both mega flow mask's reference counter and per flow table mask list
> > should only be accessed when holding ovs_mutex(
On Thu, Jan 16, 2014 at 04:30:24PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 04:13:24PM +0900, Simon Horman wrote:
> > This is an proposed enhancement to
> > "Implement OpenFlow support for MPLS, for up to 3 labels."
> >
> > Signed-off-by: Simon Horman
> > ---
> > ofproto/ofproto-dpif-xl
On Thu, Jan 16, 2014 at 04:35:24PM -0800, Ben Pfaff wrote:
> On Fri, Jan 17, 2014 at 09:28:29AM +0900, Simon Horman wrote:
> > On Thu, Jan 16, 2014 at 04:10:43PM -0800, Ben Pfaff wrote:
> > > On Wed, Jan 15, 2014 at 04:13:23PM +0900, Simon Horman wrote:
> > > > This is an proposed enhancement to
>
On Fri, Jan 17, 2014 at 09:28:29AM +0900, Simon Horman wrote:
> On Thu, Jan 16, 2014 at 04:10:43PM -0800, Ben Pfaff wrote:
> > On Wed, Jan 15, 2014 at 04:13:23PM +0900, Simon Horman wrote:
> > > This is an proposed enhancement to
> > > "Implement OpenFlow support for MPLS, for up to 3 labels."
> >
Signed-off-by: Ben Pfaff
---
I intend to fold this into
"Implement OpenFlow support for MPLS, for up to 3 labels."
diff --git a/lib/odp-util.c b/lib/odp-util.c
index 520b314..97e681d 100644
--- a/lib/odp-util.c
+++ b/lib/odp-util.c
@@ -2908,6 +2908,9 @@ parse_l2_5_onward(const struct nlattr
*att
On Wed, Jan 15, 2014 at 04:13:24PM +0900, Simon Horman wrote:
> This is an proposed enhancement to
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> Signed-off-by: Simon Horman
> ---
> ofproto/ofproto-dpif-xlate.c | 26 +++---
> 1 file changed, 15 insertions(+)
On Thu, Jan 16, 2014 at 04:10:43PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 04:13:23PM +0900, Simon Horman wrote:
> > This is an proposed enhancement to
> > "Implement OpenFlow support for MPLS, for up to 3 labels."
> >
> > This makes the conservative assumption that the datapath should
>
On Wed, Jan 15, 2014 at 04:13:23PM +0900, Simon Horman wrote:
> This is an proposed enhancement to
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> This makes the conservative assumption that the datapath should
> not be asked to apply MPLS push actions such that the resulting
> pa
On Thu, Jan 16, 2014 at 03:09:04PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 05:17:02PM +0900, Simon Horman wrote:
> > If VLAN acceleration is used when the kernel receives a packet
> > then the outer-most VLAN tag will not be present in the packet
> > when it is received by netdev-linux. R
On Thu, Jan 16, 2014 at 03:50:31PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 04:13:21PM +0900, Simon Horman wrote:
> > This is an proposed enhancement to
> > "Implement OpenFlow support for MPLS, for up to 3 labels."
> >
> > It is in preparation for not adding flows to the datapath
> > whi
On Wed, Jan 15, 2014 at 04:13:22PM +0900, Simon Horman wrote:
> This is an proposed enhancement to
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> It is in preparation for not adding flows to the datapath
> which have more MPLS LSEs than the datapath supports.
>
> Signed-off-by:
On Wed, Jan 15, 2014 at 04:13:21PM +0900, Simon Horman wrote:
> This is an proposed enhancement to
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> It is in preparation for not adding flows to the datapath
> which have more MPLS LSEs than the datapath supports.
>
> Signed-off-by:
On Wed, Jan 15, 2014 at 7:52 PM, Andy Zhou wrote:
> 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
> Signe
On Wed, Jan 15, 2014 at 04:13:20PM +0900, Simon Horman wrote:
> The number of LSEs supported is ARRAY_SIZE(flow->mpls_lse) not
> ARRAY_SIZE(flow->mpls_lse) - 1.
>
> This corrects an inconsistency added my me to
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> Signed-off-by: Simon
On Wed, Jan 15, 2014 at 04:13:18PM +0900, Simon Horman wrote:
> This corrects what appears to be a minor error in
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> Signed-off-by: Simon Horman
Thanks, I'll squash this in.
___
dev mailing
On Wed, Jan 15, 2014 at 04:13:19PM +0900, Simon Horman wrote:
> The BoS bit of each LSE is masked by flow_count_mpls_labels()
>
> This corrects what appears to be a minor inconsistency in
> "Implement OpenFlow support for MPLS, for up to 3 labels."
>
> Signed-off-by: Simon Horman
Thanks, I'll f
On Wed, Jan 15, 2014 at 05:17:02PM +0900, Simon Horman wrote:
> If VLAN acceleration is used when the kernel receives a packet
> then the outer-most VLAN tag will not be present in the packet
> when it is received by netdev-linux. Rather, it will be present
> in auxdata.
>
> This patch uses recvms
On Thu, Jan 16, 2014 at 01:44:39PM -0800, Ben Pfaff wrote:
> On Thu, Jan 09, 2014 at 10:40:15PM -0200, Flavio Leitner wrote:
> > In order to get the testsuite log, I have tried again and it works now
> > commit fd2e50cd24ca18d8186a5c9211aa03c439cf
> > Author: James Page
> > Date: Thu Dec 5 1
On Wed, Jan 15, 2014 at 05:17:01PM +0900, Simon Horman wrote:
> This is in preparation for pushing vlan tags
> using the TPID provided by the kernel via auxdata.
>
> Signed-off-by: Simon Horman
Applied, thanks!
___
dev mailing list
dev@openvswitch.org
On Wed, Jan 15, 2014 at 05:17:00PM +0900, Simon Horman wrote:
> Update the netdev_class so that struct ofpbuf * is passed to rx_recv()
> to provide both the data and size of the data to read a packet into.
>
> On success, update struct ofpbuf size inside netdev_class rx_recv
> implementation and r
On Thu, Jan 09, 2014 at 10:40:15PM -0200, Flavio Leitner wrote:
> In order to get the testsuite log, I have tried again and it works now
> commit fd2e50cd24ca18d8186a5c9211aa03c439cf
> Author: James Page
> Date: Thu Dec 5 17:29:05 2013 +
>
> Add check for -latomic
>
> Later
On Fri, Jan 10, 2014 at 01:45:47PM -0800, Joe Stringer wrote:
> Fixes the following error on distcheck:-
>
> ERROR: files left in build directory after distclean:
> ./tests/test-controller.8
> make[1]: *** [distcleancheck] Error 1
>
> Signed-off-by: Joe Stringer
Applied, thanks.
___
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 resume" test.
>
> Signed-off-by: YAMAMOTO Takashi
Clang says:
../ofproto/con
Thank you.
I applied this to master.
On Thu, Jan 16, 2014 at 10:28:19AM -0800, Alex Wang wrote:
> Looks good to me, thx for the refine.
>
>
> On Thu, Jan 16, 2014 at 10:07 AM, Ben Pfaff wrote:
>
> > Nothing ever took monitor_rwlock's read lock, so it might as well be a
> > mutex.
> >
> > Sign
On Thu, Jan 16, 2014 at 08:15:18AM +0900, YAMAMOTO Takashi wrote:
> > This annotation would have caught the bug fixed by commit 491a67a0005347130
> > (ofproto-dpif-xlate: Avoid recursive acquisition of xlate_rwlock.).
>
> looks good to me.
>
> unless there's a more generic way to detect recursive
This OpenFlow action triggers the 'back_to_kernel' datapath action
to ask the datapath to return packets to the kernel.
Signed-off-by: Chris Luke
---
include/openflow/nicira-ext.h |1 +
lib/odp-execute.c |5 -
lib/ofp-actions.c | 30 +
This provides the 'back_to_kernel' datapath action which has one
of two effects:
- For packets originating from the kernel, sets a flag allowing
the kernel hook to return it to the kernel.
- For packets that came from userspace, re-inserts the packet
into the kernel input queue.
Signed-off-by
Jesse, all,
This patchset provides a method to return packets to the kernel for
handling as though OVS never touched them.
To support this, we extend the datapath on kernels since 2.6.39 to be
able to pass packets back by providing a 'back_to_kernel' action which
causes the network input hook to
If available, use netif_receive_skb_finish() to asynchronously
return packets to the network input path after the point where
they were handed to us when the flow action is back_to_kernel.
Signed-off-by: Chris Luke
---
acinclude.m4|3 +++
datapath/vport-netdev.c |6 +-
2
Signed-off-by: Chris Luke
---
AUTHORS |1 +
FAQ | 87 +++
NEWS|2 ++
3 files changed, 80 insertions(+), 10 deletions(-)
diff --git a/AUTHORS b/AUTHORS
index 218fe76..1f03203 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -21,6
Looks good to me, thx for the refine.
On Thu, Jan 16, 2014 at 10:07 AM, Ben Pfaff wrote:
> Nothing ever took monitor_rwlock's read lock, so it might as well be a
> mutex.
>
> Signed-off-by: Ben Pfaff
> ---
> ofproto/ofproto-dpif-monitor.c | 34 +-
> 1 file ch
Nothing ever took monitor_rwlock's read lock, so it might as well be a
mutex.
Signed-off-by: Ben Pfaff
---
ofproto/ofproto-dpif-monitor.c | 34 +-
1 file changed, 17 insertions(+), 17 deletions(-)
diff --git a/ofproto/ofproto-dpif-monitor.c b/ofproto/ofproto-dp
With glibc, rwlocks by default allow recursive read-locking even if a
thread is blocked waiting for the write-lock. POSIX allows such attempts
to deadlock, and it appears that the libc used in NetBSD, at least, does
deadlock. ofproto-dpif-xlate could take the ofgroup rwlock recursively if
an ofgr
These were useful back when we were trying to use the sparse lock balance
annotations, but we removed those in commit 47b52c71232c0 (sparse: Remove
support for thread-safety annotations.) and so they serve no purpose any
longer.
Signed-off-by: Ben Pfaff
---
lib/ovs-thread.h | 11 ---
1
glibc supports two kinds of rwlocks:
- The default kind of rwlock always allows recursive read-locks to
succeed, but threads blocked on acquiring the write-lock are treated
unfairly, causing them to be delayed indefinitely as long as new
readers continue to come along.
-
POSIX allows taking a rwlock recursively for read to deadlock if there
are threads waiting for the write lock. NetBSD does this. glibc gives one
a nasty tradeoff: either deadlock on recursive read lock attempts in the
presence of a writer, or lose fairness. Since we know that we can't avoid
dead
With glibc, rwlocks by default allow recursive read-locking even if a
thread is blocked waiting for the write-lock. POSIX allows such attempts
to deadlock, and it appears that the libc used in NetBSD, at least, does
deadlock. The netdev_class_rwlock is in fact acquired recursively in this
way, wh
48 matches
Mail list logo
