Hi Alex,
On Wed, 2014-02-12 at 20:24 -0800, Alex Wang wrote:
>
>
> I'd very much like to get a full review from you. Are you planning to do
> that?
>
>
> Also want to ask about your timetable, do you think you could finish review
> this week?
>
Yes, I will try to finish review by this wee
This reworks lookup of rules for both table 0 and table action translation.
The result is that Table Mod settings, which can alter the miss-behaviour
of tables, including table 0, on a per-table basis may be honoured.
Previous patches proposed by myself which build on earlier merged patches
by And
Hey Han,
I'd very much like to get a full review from you. Are you planning to do
that?
Also want to ask about your timetable, do you think you could finish review
this week?
Appreciate your help, ;D
Alex Wang,
On Wed, Feb 12, 2014 at 6:37 AM, Zhou, Han wrote:
> Hi Alex,
>
> On Fri, 2014-02
Signed-off-by: Romain Lenglet
---
ofproto/ofproto-dpif-ipfix.c | 120 ++
1 file changed, 87 insertions(+), 33 deletions(-)
diff --git a/ofproto/ofproto-dpif-ipfix.c b/ofproto/ofproto-dpif-ipfix.c
index 2500efd..a529884 100644
--- a/ofproto/ofproto-dpif-ip
On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme wrote:
> 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 reports
> and allows OVS to scale b
On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme wrote:
> Signed-off-by: Jarno Rajahalme
> ---
> datapath/datapath.c | 60
> +++
> 1 file changed, 32 insertions(+), 28 deletions(-)
>
> diff --git a/datapath/datapath.c b/datapath/datapath.c
> ind
On Tue, Feb 11, 2014 at 4:07 PM, 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.
>
> Signed-off-by: Jarno Rajahalme
Looks good.
Acked-by:
Signed-off-by: Jarno Rajahalme
---
lib/ovs-thread.c | 11 +++
lib/ovs-thread.h |8
2 files changed, 19 insertions(+)
diff --git a/lib/ovs-thread.c b/lib/ovs-thread.c
index a20b2fd..f0b1e9e 100644
--- a/lib/ovs-thread.c
+++ b/lib/ovs-thread.c
@@ -168,6 +168,17 @@ ovs_mutex_
Stats critical sections are held for very short time, so it makes
sense to spin on them for a while rather than sleeping immediately.
Signed-off-by: Jarno Rajahalme
---
ofproto/ofproto-dpif.c |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/ofproto/ofproto-dpif.c b/of
Use send/recv for socket stream instead of read/write.
Use event handle for polling on socket stream.
Check windows specific return code.
Signed-off-by: Linda Sun
---
lib/automake.mk |7 +-
lib/latch-unix.c| 87 +++
lib/latch.c | 87 ---
Thanks Han for the review, ;D
I misunderstood how rcu works.
The access of "vport->upcall_pids" is a "single writer, multiple readers"
problem.
And it has already been protected by the 'ovs_lock()' in callers.
I'll remove the rcu read critical section in the V3 patch.
Alex Wang,
On Wed, Feb 1
Signed-off-by: Joe Stringer
---
I've heard term in the context of using a switch to "bless" traffic.
Adding this wording makes it easier to search for the relevant FAQ
entry.
---
FAQ |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/FAQ b/FAQ
index a08d65c..6c79efe 1006
Recent changes to the flow_dump_next() interface have made it the
responsibility of the dpif implementation to track error status over a
flow dump operation.
This patch removes status tracking from 'struct dpif_flow_dump', allowing
multiple threads to call dpif_flow_dump_next() and track their sta
This new function allows callers to determine whether previously
returned keys will be modified or reallocated on the next call to
dpif_flow_dump_next(). This will be used in a future commit to allow
batched flow deletion by revalidator threads.
Signed-off-by: Joe Stringer
---
v3: Renamed functio
This patch separates the structures for thread-local flow dump state
("state") from the shared flow dump state ("iter") in dpif-linux and
dpif-netdev. Future patches will make use of this to allow multiple
threads to dump flows from the same flow dump operation.
Signed-off-by: Joe Stringer
---
v3
This patch modifies 'struct nl_dump' and nl_dump_next() to allow
multiple threads to share the same nl_dump. These changes are targeted
around synchronizing dump status between multiple callers, and
allowing callers to fully process their existing buffers before
determining whether to stop fetching
This patch makes it the caller's responsibility to initialize a
per-thread 'state' object and pass it down to the dpif_flow_dump_next()
implementation. The implementation can expect to be called from multiple
threads with the same 'iter' and different 'state' objects.
When flow_dump_next() returns
From: Ethan Jackson
Previously, we had a separate flow_dumper thread that fetched flows from
the datapath to distribute to revalidator threads. This patch takes the
logic for dumping and pushes it into the revalidator threads, resulting
in simpler code with similar performance to the current code
Currently, we have a single thread (the 'flow_dumper') that fetches the flows
from the datapath and distributes them to revalidator threads. To clean up the
code and support more flows in the datapath, we are looking at removing this
flow_dumper thread, and have each revalidator fetch flows from th
This patch makes all of the users of 'struct nl_dump' allocate their own
buffers to pass down to nl_dump_next(). This paves the way for allowing
multithreaded flow dumping.
Signed-off-by: Joe Stringer
---
v3: Rebase.
v2: Add NL_DUMP_BUFSIZE.
Add buffer to dpif_linux_flow_state.
Fold "netl
Thanks for the review and testing. Pushed.
On Tue, Feb 11, 2014 at 3:31 PM, Jarno Rajahalme wrote:
> These work for me, and the code looks OK too, so (for both patches):
>
> Acked-by: Jarno Rajahalme
>
> On Feb 11, 2014, at 1:53 PM, Andy Zhou wrote:
>
> > netdev-dummy will mostly be used for t
On Wed, Feb 12, 2014 at 08:59:18AM -0800, Jarno Rajahalme wrote:
>
> On Feb 11, 2014, at 4:40 PM, Ben Pfaff wrote:
>
> > On Tue, Feb 11, 2014 at 04:30:44PM -0800, Jarno Rajahalme wrote:
> >> Reading the hmap count for determining if it is empty or not is thread
> >> safe, so avoid locking when n
On Wed, Feb 12, 2014 at 08:54:46AM -0800, Jarno Rajahalme wrote:
>
> On Feb 11, 2014, at 4:45 PM, Ben Pfaff wrote:
>
> > On Tue, Feb 11, 2014 at 04:38:35PM -0800, Jarno Rajahalme wrote:
> >>
> >> On Feb 11, 2014, at 4:30 PM, Ben Pfaff wrote:
> >>
> >>> On Tue, Feb 11, 2014 at 04:30:43PM -0800
On Wed, Feb 12, 2014 at 05:24:30PM +0900, Simon Horman wrote:
> On Mon, Feb 03, 2014 at 05:12:12PM -0800, Ben Pfaff wrote:
> > On Wed, Jan 29, 2014 at 12:53:03PM +0900, Simon Horman wrote:
> > > This reworks lookup of rules for both table 0 and table action
> > > translation.
> > > The result is t
On Feb 11, 2014, at 4:40 PM, Ben Pfaff wrote:
> On Tue, Feb 11, 2014 at 04:30:44PM -0800, Jarno Rajahalme wrote:
>> Reading the hmap count for determining if it is empty or not is thread
>> safe, so avoid locking when not necessary.
>>
>> Signed-off-by: Jarno Rajahalme
>
> It might be worth a
On Feb 11, 2014, at 4:45 PM, Ben Pfaff wrote:
> On Tue, Feb 11, 2014 at 04:38:35PM -0800, Jarno Rajahalme wrote:
>>
>> On Feb 11, 2014, at 4:30 PM, Ben Pfaff wrote:
>>
>>> On Tue, Feb 11, 2014 at 04:30:43PM -0800, Jarno Rajahalme wrote:
mf_set_flow_value() was not setting 'flow->nw_frag'
Hi Alex,
On Fri, 2014-02-07 at 17:17 -0800, Alex Wang wrote:
> This commit removes the 'dispatcher' thread by allowing 'handler'
> threads to read upcalls directly from dpif. vport in dpif will
> open netlink sockets for each handler and will use the 5-tuple
> hash from the missed packet to choos
On Mon, Feb 03, 2014 at 05:12:12PM -0800, Ben Pfaff wrote:
> On Wed, Jan 29, 2014 at 12:53:03PM +0900, Simon Horman wrote:
> > This reworks lookup of rules for both table 0 and table action translation.
> > The result is that Table Mod settings, which can alter the miss-behaviour
> > of tables, inc
On Tue, Feb 11, 2014 at 03:01:27PM -0800, Ben Pfaff wrote:
> On Tue, Feb 04, 2014 at 04:36:18PM -0800, Andy Zhou wrote:
> > This patch adds the necessary infrastructure for the next patch series.
> > recirculation logic is still work in progress. It is not enabled. There
> > should
> > be no beha
This is in preparation for adding recirculation where
it will be used to test if load, move and set_field actions
require recirculation to be performed first.
Signed-off-by: Simon Horman
---
lib/meta-flow.h | 8
1 file changed, 8 insertions(+)
diff --git a/lib/meta-flow.h b/lib/meta-fl
A quick hack to allow the learn test to pass.
I need to investigate the root-cause of its failure.
Signed-off-by: Simon Horman
---
tests/ofproto-dpif.at | 1 +
1 file changed, 1 insertion(+)
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at
index bca44de..cecf5ac 100644
--- a/tests/ofp
I am posting this share my implementation with Andy Zhou.
I hope this will help us to work towards a recirculation
implementation which works for both bridging (the target of his work)
and MPLS (the target of this work.
Although working in parts intended an RFC and has several limitations:
- The l
This makes maintaining the controller test slightly easier
by splitting it in two.
Based on a similar patch by Joe Stringer.
Cc: Joe Stringer
Signed-off-by: Simon Horman
---
tests/ofproto-dpif.at | 352 ++
1 file changed, 185 insertions(+), 167 d
Add count_mpls_labels() which counts LSEs present in an array.
This is a more general form of flow_count_mpls_labels().
This is in preparation for using count_mpls_labels()
in parse_odp_key_mask_attr() where no struct flow is available.
Signed-off-by: Simon Horman
---
lib/flow.c | 36 ++
Signed-off-by: Simon Horman
---
ofproto/ofproto-dpif.c | 26 +-
ofproto/ofproto-dpif.h | 4
2 files changed, 21 insertions(+), 9 deletions(-)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index 7b3e1eb..8606226 100644
--- a/ofproto/ofproto-dpif.c
+++
Hi Andy,
On Tue, Feb 04, 2014 at 04:36:22PM -0800, Andy Zhou wrote:
>
> Signed-off-by: Andy Zhou
> ---
> include/linux/openvswitch.h | 26 ++
> lib/dpif-netdev.c |2 ++
> lib/dpif.c |2 ++
> lib/odp-execute.c |2 ++
> li
36 matches
Mail list logo
