Hi experts,
Please ignore my previous mail. Sent it without completion.
am trying to integrate Broadcom's OFDPA (OpenFlow Data Path Abstraction)
into my data-plane. As you all know, OFDPA supports multiple flow tables
for programming multiple flow entries.
when multiple flow tables are support
On Tue, Aug 04, 2015 at 04:05:58AM +, Wenyu Zhang wrote:
> In ipfix and sflow, the sampling rate can be set as 1, which means
> 100%. And in order to make 100% packets sampled, I work out a patch
> in datapath, which is for review on net-next. And the same change
> seems need to be added into
When sampling rate is 1, the sampling probability is UINT32_MAX. The packet
should be sampled even the prandom_u32() generate the number of UINT32_MAX.
And none packet need be sampled when the probability is 0.
Signed-off-by: Wenyu Zhang
---
net/openvswitch/actions.c |4 +++-
1 file changed,
Hello Ben,
Thanks for your comments.
In ipfix and sflow, the sampling rate can be set as 1, which means 100%. And
in order to make 100% packets sampled, I work out a patch in datapath, which is
for review on net-next.
And the same change seems need to be added into lib/odp-execute.c too. Do yo
When sampling rate is 1, the sampling probability is UINT32_MAX. The packet
should be sampled even the prandom_u32() generate the number of UINT32_MAX.
And none packet need be sampled when the probability is 0.
Signed-off-by: Wenyu Zhang
---
net/openvswitch/actions.c |4 +++-
1 file changed,
On 30 July 2015 at 11:21, Daniele Di Proietto wrote:
> Acked-by: Daniele Di Proietto
>
> On 30/07/2015 00:52, "Joe Stringer" wrote:
>
>>The initial sanity test only checked IPv4 without IP fragments. This patch
>>adds additional tests using IPv6 and VLANs with IP fragments and expands
>>the exis
Hey Russell,
With your add type/option change and Ben's tunnel-key work in, it is time
to rebase my ovn-controller-vtep series~~
So, now, I think I'd like to adopt your earlier suggestion, which requires
users
to specify this when creating logical port (for vtep switch):
Logical_Port
name:
On Mon, Aug 3, 2015 at 3:45 PM, Jarno Rajahalme wrote:
> With few suggestions and questions below:
>
> Acked-by: Jarno Rajahalme
Thanks for the comments - I had one follow up question below.
>> @@ -3014,11 +3014,25 @@ dp_netdev_upcall(struct dp_netdev_pmd_thread *pmd,
>> struct dp_packet *pack
On Mon, Aug 03, 2015 at 07:53:58PM -0400, Russell Bryant wrote:
> On 08/03/2015 07:47 PM, Ben Pfaff wrote:
> > On Mon, Aug 03, 2015 at 07:44:03PM -0400, Russell Bryant wrote:
> >> On 08/03/2015 07:42 PM, Ben Pfaff wrote:
> >>> Thank you for such a careful review!
> >>>
> >>> I incorporated all of t
On 08/03/2015 07:47 PM, Ben Pfaff wrote:
> On Mon, Aug 03, 2015 at 07:44:03PM -0400, Russell Bryant wrote:
>> On 08/03/2015 07:42 PM, Ben Pfaff wrote:
>>> Thank you for such a careful review!
>>>
>>> I incorporated all of these comments and applied this to master.
>>
>> Awesome work, Ben!
>
> Than
On Mon, Aug 03, 2015 at 07:48:28PM -0400, Russell Bryant wrote:
> On 08/03/2015 07:43 PM, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
>
> Acked-by: Russell Bryant
Thanks, applied.
> > diff --git a/datapath-windows/ovsext/User.c b/datapath-windows/ovsext/User.c
> > index 9f462cf..2ff2635 1006
On 08/03/2015 07:43 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Acked-by: Russell Bryant
> diff --git a/datapath-windows/ovsext/User.c b/datapath-windows/ovsext/User.c
> index 9f462cf..2ff2635 100644
> --- a/datapath-windows/ovsext/User.c
> +++ b/datapath-windows/ovsext/User.c
> @@ -836,7 +
On Mon, Aug 03, 2015 at 07:44:03PM -0400, Russell Bryant wrote:
> On 08/03/2015 07:42 PM, Ben Pfaff wrote:
> > Thank you for such a careful review!
> >
> > I incorporated all of these comments and applied this to master.
>
> Awesome work, Ben!
Thanks. It really needs testing, though--I haven't
Thanks, applied to master.
On Mon, Aug 03, 2015 at 02:26:09PM -0700, Alex Wang wrote:
> Thx for the refinement!~
>
> Acked-by: Alex Wang
>
> On Mon, Aug 3, 2015 at 2:13 PM, Ben Pfaff wrote:
>
> > In the commit message below, "match" on them, not "map" on them, sorry.
> >
> > On Mon, Aug 03, 2
Thanks Jarno, I applied this to master.
On Mon, Aug 03, 2015 at 02:31:29PM -0700, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme
>
> > On Aug 2, 2015, at 11:51 AM, Ben Pfaff wrote:
> >
> > Until now, flow translation has had to use try_ref to take a reference on
> > a rule, because a compe
On 08/03/2015 07:42 PM, Ben Pfaff wrote:
> Thank you for such a careful review!
>
> I incorporated all of these comments and applied this to master.
Awesome work, Ben!
--
Russell Bryant
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/m
Signed-off-by: Ben Pfaff
---
INSTALL.DPDK.md | 2 +-
datapath-windows/ovsext/User.c | 2 +-
datapath-windows/ovsext/Vport.c | 2 +-
datapath-windows/ovsext/Vxlan.c | 2 +-
lib/netdev-provider.h | 2 +-
lib/ovs-atomic.h| 2 +-
lib/packets.c
Thank you for such a careful review!
I incorporated all of these comments and applied this to master. Here's
the incremental (it omits the changes to this patch due to changes in
earlier patches, e.g. the Rule->Logical_Flow renaming):
diff --git a/ovn/controller/lflow.c b/ovn/controller/lflow.c
I applied most of your comments as obvious. A few deserved responses,
see below.
On Mon, Aug 03, 2015 at 01:19:01AM -0700, Justin Pettit wrote:
>
> > On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
> > +/* Translate logical table ID to physical table ID. */
> > +bool ingress = !st
With few suggestions and questions below:
Acked-by: Jarno Rajahalme
Jarno
> On Jul 29, 2015, at 8:09 PM, Jesse Gross wrote:
>
> The kernel implementation of Geneve options stores the TLV option
> data in the flow exactly as received, without any further parsing.
> This is then translated to
Acked-by: Jarno Rajahalme
> On Jul 29, 2015, at 8:09 PM, Jesse Gross wrote:
>
> If ofproto-dpif installs a flow into the userspace datapath that doesn't
> include a mask, we need to synthesize an exact match one. This is currently
> done using the metaflow infrastructure, iterating over each fi
Acked-by: Jarno Rajahalme
> On Aug 2, 2015, at 11:51 AM, Ben Pfaff wrote:
>
> Until now, flow translation has had to use try_ref to take a reference on
> a rule, because a competing thread might have released the last reference
> and done an RCU-postponed deletion. Since classifier versioning
Thx for the refinement!~
Acked-by: Alex Wang
On Mon, Aug 3, 2015 at 2:13 PM, Ben Pfaff wrote:
> In the commit message below, "match" on them, not "map" on them, sorry.
>
> On Mon, Aug 03, 2015 at 02:11:12PM -0700, Ben Pfaff wrote:
> > The text didn't say how to actually map on them. I took th
On Mon, Aug 03, 2015 at 01:55:04PM -0700, Justin Pettit wrote:
>
> > On Aug 3, 2015, at 1:48 PM, Ben Pfaff wrote:
> >
> > On Fri, Jul 31, 2015 at 03:12:38PM -0700, Justin Pettit wrote:
> >>> On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
> >>>if (br_int) {
> >>> +enum mf_fiel
In the commit message below, "match" on them, not "map" on them, sorry.
On Mon, Aug 03, 2015 at 02:11:12PM -0700, Ben Pfaff wrote:
> The text didn't say how to actually map on them. I took the liberty of
> massaging the text a little further, too.
>
> Suggested-by: Alex Wang
> Signed-off-by: Be
On Fri, Jul 31, 2015 at 11:08:19PM -0700, Alex Wang wrote:
> On Fri, Jul 31, 2015 at 10:15 PM, Ben Pfaff wrote:
>
> > On Fri, Jul 31, 2015 at 10:04:35PM -0700, Alex Wang wrote:
> > > Is there any documentation for the ovs side geneve negotiation?
> >
> > No. What kind of documentation would help
The text didn't say how to actually map on them. I took the liberty of
massaging the text a little further, too.
Suggested-by: Alex Wang
Signed-off-by: Ben Pfaff
---
utilities/ovs-ofctl.8.in | 47 ---
1 file changed, 24 insertions(+), 23 deletions(-)
From: Glenn Griffin
Date: Mon, 3 Aug 2015 09:56:54 -0700
> openvswitch modifies the L4 checksum of a packet when modifying
> the ip address. When an IP packet is fragmented only the first
> fragment contains an L4 header and checksum. Prior to this change
> openvswitch would modify all fragments,
> On Aug 3, 2015, at 1:48 PM, Ben Pfaff wrote:
>
> On Fri, Jul 31, 2015 at 03:12:38PM -0700, Justin Pettit wrote:
>>> On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
>>>if (br_int) {
>>> +enum mf_field_id mff_ovn_geneve = ofctrl_run(br_int);
>>> +
>>>struct hmap fl
On Fri, Jul 31, 2015 at 03:12:38PM -0700, Justin Pettit wrote:
> > On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
> > +static void
> > +recv_S_GENEVE_TABLE_MOD_SENT(const struct ofp_header *oh, enum ofptype
> > type)
> > +{
> > +if (oh->xid != xid && oh->xid != xid2) {
> > +ofctrl_recv
I forgot to give my ack:
Acked-by: Justin Pettit
--Justin
> On Jul 31, 2015, at 3:12 PM, Justin Pettit wrote:
>
>
>> On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
>>
>>
>> +static void
>> +recv_S_GENEVE_TABLE_MOD_SENT(const struct ofp_header *oh, enum ofptype type)
>> +{
>> +if (oh->
I applied this patch and all the previous ones to master, with that
naming change.
On Sun, Aug 02, 2015 at 12:12:48PM -0700, Ben Pfaff wrote:
> OK, I'll change this to use Logical_Flow for the name of the table.
>
> On Fri, Jul 31, 2015 at 03:40:19PM -0700, Alex Wang wrote:
> > Just to mention th
> On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
>
I missed this in my review yesterday on ovn-architure.7.xml:
> +OpenFlow tables 32 through 47 implement the output
> action
> +in the the logical ingress pipeline. Specifically, table 32 handles
This introduces two "the"s.
J
On Mon, Aug 3, 2015 at 11:18 AM, Pravin Shelar wrote:
> On Mon, Aug 3, 2015 at 12:11 AM, Wenyu Zhang wrote:
>> When sampling rate is 1, the sampling probability is UINT32_MAX. The packet
>> should be sampled even the prandom32() generate the number of UINT32_MAX.
>> And none packet need be sample
On Mon, Aug 3, 2015 at 12:11 AM, Wenyu Zhang wrote:
> When sampling rate is 1, the sampling probability is UINT32_MAX. The packet
> should be sampled even the prandom32() generate the number of UINT32_MAX.
> And none packet need be sampled when the probability is 0.
>
> Signed-off-by: Wenyu Zhang
Hello,
I am developing a new match option for openvswitch and I would like that it
runs through the datapath without need of user-space. I'm not really sure
if I missed something but this is what happens:
(1) First incomming packet is processed correctly (processed + stats
updated)
(2) Following
sorry for the confusion, my comments is for reminding myself, rather than
wanting to hold off the commit~ Still a lot to learn related to this patch~
Thanks,
Alex Wang,
On Mon, Aug 3, 2015 at 1:25 AM, Justin Pettit wrote:
>
> > On Aug 2, 2015, at 10:03 PM, Alex Wang wrote:
> >
> > Wow, this
> On Jul 20, 2015, at 12:00 PM, Sorin Vinturis
> wrote:
>
> Under certain circumstances, we might need to cancel a pending IRP
> that has been submitted and not yet responded. This might occur when
> the request takes too long to complete or when the process which
> initiated the request termina
> On Jul 20, 2015, at 12:00 PM, Sorin Vinturis
> wrote:
>
> In order to support IRP cancelling mechanism for pending IRPs, all
> tunnel filter requests, VXLAN create/delete tunnel, need to be
> processed iteratively.
>
> Signed-off-by: Sorin Vinturis
I reviewed the v1 of this patch. I’ll wait
> On Jul 15, 2015, at 7:50 AM, Sorin Vinturis
> wrote:
>
> If the extension was previously enabled and running, after issuing a
> restart, stop+start, the extension fails to be enabled. This happens
> because the extension's DeviceObject is not yet initialized before the
> FilterAttach routine i
It contains every change to the code. At this point, we really only put bug
fixes into 2.3. The commit messages should be pretty clear about what's
changing.
--Justin
> On Aug 3, 2015, at 9:58 AM, Karnataki, Gururaj wrote:
>
> Okay. Thanks a lot for the below information.
>
> I have a quer
On Mon, Aug 3, 2015 at 9:56 AM, Glenn Griffin wrote:
> openvswitch modifies the L4 checksum of a packet when modifying
> the ip address. When an IP packet is fragmented only the first
> fragment contains an L4 header and checksum. Prior to this change
> openvswitch would modify all fragments, modi
If that's an accurate description of the problem then that seems fine to
me. We could add a note to INSTALL.DPDK.md describing the issue I
suppose, if someone wants to write one up.
On Sat, Aug 01, 2015 at 02:40:16PM -0700, Ethan Jackson wrote:
> I personally am fine with waiting on this being fi
Okay. Thanks a lot for the below information.
I have a query here so does that mean each single commit id corresponds to
every single bug fix?
Regards,
Guru
On Mon, Aug 03, 2015 at 10:21 pm, Joe Stringer
mailto:joestrin...@nicira.com>> wrote:
Github also allows you to browse commits by tag,
Hi Saloni.
I'll be out of the office, traveling, from Aug. 4 to 18. If you want a
review of your code before then, you'll have to ask someone else.
Otherwise, I recommend rebasing and reposting your code around Aug. 18.
Thanks,
Ben.
___
dev mailing li
openvswitch modifies the L4 checksum of a packet when modifying
the ip address. When an IP packet is fragmented only the first
fragment contains an L4 header and checksum. Prior to this change
openvswitch would modify all fragments, modifying application data
in non-first fragments, causing checksu
hi Sorin,
Thanks for this patch. This is a functionality that is needed to complete the
tunnel filter async request story.
I have a few comments that are inlined.
How did you test this patch, BTW?
> On Jul 15, 2015, at 1:55 AM, Sorin Vinturis
> wrote:
>
> Under certain circumstances, we migh
On Mon, Aug 03, 2015 at 09:25:53AM -0700, Pravin Shelar wrote:
> On Sat, Aug 1, 2015 at 6:31 PM, Glenn Griffin
> wrote:
> > openvswitch modifies the L4 checksum of a packet when modifying
> > the ip address. When an IP packet is fragmented only the first
> > fragment contains an L4 header and che
Github also allows you to browse commits by tag, eg:
https://github.com/openvswitch/ovs/commits/v2.3.2
On 3 August 2015 at 09:38, Justin Pettit wrote:
> The best way to see what was fixed is to look at the commit history. Here's
> the github history for branch-2.3:
>
> https://github.c
The best way to see what was fixed is to look at the commit history. Here's
the github history for branch-2.3:
https://github.com/openvswitch/ovs/commits/branch-2.3
We tag every release, and you can see the commit id on this page:
https://github.com/openvswitch/ovs/tags
If you
On Sun, Aug 02, 2015 at 11:54:09PM -0700, Wenyu Zhang wrote:
> When sampling rate is 1, the sampling probability is UINT32_MAX. The packet
> should be sampled even the prandom32() generate the number of UINT32_MAX.
> And none packet need be sampled when the probability is 0.
If we have code that g
On Sat, Aug 1, 2015 at 6:31 PM, Glenn Griffin wrote:
> openvswitch modifies the L4 checksum of a packet when modifying
> the ip address. When an IP packet is fragmented only the first
> fragment contains an L4 header and checksum. Prior to this change
> openvswitch would modify all fragments, modi
> On Jul 15, 2015, at 1:55 AM, Sorin Vinturis
> wrote:
>
> In order to support IRP cancelling mechanism for pending IRPs, all
> tunnel filter requests, VXLAN tunnel create/delete, need to be
> processed iteratively.
>
> Signed-off-by: Sorin Vinturis
> ---
> datapath-windows/ovsext/TunnelFilte
> On Jul 3, 2015, at 8:36 AM, Sorin Vinturis
> wrote:
>
>
> +__inline BOOLEAN OvsBufferIsEmpty(char *buf, size_t size)
> +{
> +return (buf[0] == 0 && !memcmp(buf, buf + 1, size - 1));
> +}
> +
The reason 'instance->dumpState.ovsMsg’ was a pointer and not a structure
originally was to sav
> On Jul 2, 2015, at 4:39 AM, Sorin Vinturis
> wrote:
>
> This BSOD occurred in the context of a packet (NBL) with multiple
> NET_BUFFER(s) (NBs). The reason for the BSOD is due to the marking
> of NBLs created by OVS as being external and wrongly completing them.
>
> Signed-off-by: Sorin Vintu
> On Jul 1, 2015, at 10:28 AM, Sorin Vinturis
> wrote:
>
> OvsPrepareFlow() returns an error only when the new flow allocation
> fails. In this case HandleFlowPut() should return error without trying
> to free the flow, thus avoiding the BSOD.
>
> Signed-off-by: Sorin Vinturis
> Reported-by: S
On 08/03/2015 03:36 AM, Amitabha Biswas wrote:
> Hi,
>
> I think Gal is saying the following - ml2 allows the following:
>
> VM (10.0.0.2) --- Logical_Switch (10.0.0.2) LogicalRouter
> | |
> | |
> VM
On 07/31/2015 10:09 PM, Russell Bryant wrote:
> On 07/31/2015 07:07 PM, Ben Pfaff wrote:
>> I think that the code is intended to remove the patch ports when they
>> are not longer needed, but when I followed up with:
>>
>> ovs-vsctl remove open . external-ids ovn-bridge-mappings
>> ovs-vsct
On 08/03/2015 02:00 AM, Alex Wang wrote:
> Could we also add the explanation of "ovn-bridge-mappings" in ovn-controller
> manpage?
Sure, that's a good idea. Thanks.
--
Russell Bryant
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mail
> On Aug 2, 2015, at 9:33 PM, Sorin Vinturis
> wrote:
>
> Hi Nithin,
>
> Yes, the second check, from the 564 line, is caught by the case when the
> OvsInitExternalNBLContext() function is called with the NBL created by
> OvsPartialCopyNBL().
Right, so, do we still need your patch since L564
Hello Team,
Requesting you to please let me know for the confirmed list of bug fixes for
OVS 2.3.2 within earliest possible timeframe.
Thanking you all in advance !!!
Warm Regards,
Gururaj Karnataki
Product Engineer, Systems Integrator
NTT Communications
powered by Virtela
Desk +1-720-475-92
> On Aug 2, 2015, at 10:03 PM, Alex Wang wrote:
>
> Wow, this is really a big patch~ deserves multiple reviews~ I think I'll
> re-
> review it later~
Hi, Alex. I've got the patch mostly reviewed and will finish the final part in
the morning. It would be great to have another set of eyes re
> On Jul 28, 2015, at 8:44 AM, Ben Pfaff wrote:
>
Thanks, Ben. This is a great patch. I still owe you a review on the function
physical_run(), but it's fairly complicated, and my reviewing powers are
rapidly weakening. I'll finish the review in the morning, but wanted to give
you some ear
Hi Ben,
Greetings!
This is in reference to the patch series "Implementation of OpenFlow 1.4
Vacancy Events" which is under review.
1. [PATCH v1 1/3] : Implement Openflow 1.4 Vacancy Events for OFPT_TABLE_MOD.
2. [PATCH v1 2/3] : Implement Vacancy Events for OFPMP_TABLE_DESC.
3. [PATCH v1 3/3]
Hi,
I think Gal is saying the following - ml2 allows the following:
VM (10.0.0.2) --- Logical_Switch (10.0.0.2) LogicalRouter
| |
| |
VM (10.0.1.2) ++-(10.0.1.2) LogicalRouter
When sampling rate is 1, the sampling probability is UINT32_MAX. The packet
should be sampled even the prandom32() generate the number of UINT32_MAX.
And none packet need be sampled when the probability is 0.
Signed-off-by: Wenyu Zhang
---
net/openvswitch/actions.c |4 +++-
1 file changed, 3
66 matches
Mail list logo
