hi Sorin,
Thanks for the patient review.
The document should not rever to ovs-wind anymore since the model of adding
vport is changed to be "friendly-name" based. I'll clean it up for that.
I have addressed all your comments as you suggested except the ones below.
> Due to these differences, it
In this patch, we update the design document to reflect the netlink
based kernel-userspace interface implementation and a few other changes.
I have covered at a high level.
Please feel free to extend the document with more details that you think
got missed out.
Signed-off-by: Nithin Raju
Plain code motion. No additional cleanup or refactoring peformed.
I'll be doing some code refactoring in subsequent patches.
Signed-off-by: Nithin Raju
---
datapath-windows/ovsext/Datapath.c | 750 +---
datapath-windows/ovsext/Vport.c|
GBI&e=
>NetEventSwitchActivate indicates that the extensible switch has completed
>activation so we can now safely query the switch itself.
>
>Also we are not forwarding the PNP event to the overlaying drivers unless
>we succeeded in the operation, this issue has been fixed also.
>
Hi Alin,
Thanks for the change.
Only suggestion I had is to add the new functions to Checksum.c. If you
are concerned about the name Checksum.c, we could rename it to Offloads.c.
BufferMgmt[ch] is for NBL clone/copy/split etc.
Thanks,
-- Nithin
-Original Message-
From: dev on behalf of
Hi Sorin,
Thanks for incorporating the comments from previous iteration. I had a few
cosmetic comments to improve readability, but looks good otherwise.
Acked-by: Nithin Raju
Thanks,
-- Nithin
-Original Message-
From: dev on behalf of Sorin Vinturis
Date: Tuesday, January 19, 2016 at
sent down, have the same treatment as #1, #2, #3. Only
thing that changes is avoiding logging.
I am ok with the implementation you have in the meantime.
Acked-by: Nithin Raju
Thanks,
-- Nithin
-Original Message-
From: dev on behalf of Sorin Vinturis
Date: Tuesday, January 19, 2016 at 12
Acked-by: Nithin Raju
-Original Message-
From: dev on behalf of Sorin Vinturis
Date: Tuesday, February 2, 2016 at 2:43 AM
To: "dev@openvswitch.org"
Subject: [ovs-dev] [PATCH] datapath-windows: Updated FAQ regarding
Hyper-V MPLS.
>Updated FAQ document to reflect MPLS supp
Let’s drop it, and use 8.1. Thanks for following this up.
Thanks,
-- Nithin
-Original Message-
From: dev on behalf of Alin Serdean
Date: Tuesday, February 2, 2016 at 3:20 PM
To: Ben Pfaff
Cc: "dev@openvswitch.org"
Subject: Re: [ovs-dev] [PATCH] Update windows datapath build system
>N
Thanks for taking care of the comments.
Acked-by: Nithin Raju
-Original Message-
From: dev on behalf of Alin Serdean
Date: Wednesday, February 10, 2016 at 5:38 PM
To: "dev@openvswitch.org"
Subject: [ovs-dev] [PATCH] datapath-windows: Refactor sofware
offloadsa
hi Sorin,
Thanks for the patch. It is great that you are addressing many issues that have
been pending for sometime.
I don’t know if there’s a whole lot of technical advantage to using an separate
ioctl v/s using a netlink command, but I can see that it makes the
get_sock_pid_from_kernel() func
>
> Signed-off-by: Sorin Vinturis
>
hi Sorin,
Thanks for addressing the review comments.
I just had one super minor comment. Looks good otherwise.
Acked-by: Nithin Raju
> static NTSTATUS HandleDpTransactionCommon(
> POVS_USER_PARAMS_CONTEXT usrParamsCtx, UINT
The documentation to run the executables has fallen behind quite
a bit. In this patch, we make a few updates. There's more to come.
The patch assumes the user has not run 'make install' or does not
have access to an installer.
Signed-off-by: Nithin Raju
---
INSTALL.W
d add is the location of the log file, and also
the steps to stop the daemon(s) using app-ctl. This can happen in a subsequent
commit if you wish.
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
oks like you have the steps to stop services covered. In indicated in the
previous review that it might be good to mention the appctl command. Not sure
if they are necessary now :)
patch looks good.
Acked-by: Nithin Raju
___
dev mailing
The documentation to run the executables has fallen behind quite
a bit. In this patch, we make a few updates. There's more to come.
The patch assumes the user has not run 'make install' or does not
have access to an installer.
Signed-off-by: Nithin Raju
---
v2: updated the ste
hi Sorin,
Thanks for making this change. Apologies of the delay in the review. In
defense, it was not a small patch :)
The change looks good in general. I had the following comments though:
The WFP (tunnel filter) state machine is quite confusing. Would it be possible
to write a diagram/sentenc
> On Mar 12, 2015, at 5:42 PM, Sorin Vinturis
> wrote:
>
> Modified 'nl_sock_transact_multiple__' function to send asynchronous
> I/O requests to the kernel driver.
>
> Signed-off-by: Sorin Vinturis
> Reported-by: Alin Gabriel Serdean
> Reported-at:
> https://urldefense.proofpoint.com/v2/ur
Sorin,
In general, ‘gOvsControlLock’ is there to protect ‘gOvsSwitchContext’.
Basically, while we are accessing ‘gOvsSwitchContext’ from the netlink path or
even during an NDIS switch callback, we want to make sure ‘gOvsSwitchContext’
doesn’t get deallocated under the rug in another context.
Al
by: Gurucharan Shetty
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
> On Apr 9, 2015, at 1:47 PM, Gurucharan Shetty wrote:
>
>> Looks good. The only thing I’d add is the location of the log file, and also
>> the steps to stop the daemon(s) using app-ctl. This can happen in a
>> subsequent commit if you wish.
>>
>> Acked-
> On Apr 9, 2015, at 5:20 AM, Sorin Vinturis
> wrote:
>
> The extension failed to be activated during booting due to the failure to
> initialize tunnel filter. This happened because the Base Filtering Engine
> (BFE) is not started and no session to the engine could be acquired.
>
> The solution
Added the commands to be able to install an unsigned driver.
---
INSTALL.Windows.md | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/INSTALL.Windows.md b/INSTALL.Windows.md
index 0029500..0dd5cda 100644
--- a/INSTALL.Windows.md
+++ b/INSTALL.Windows.md
@@ -162,7 +162,12 @@
Added the commands to be able to install an unsigned driver.
Acked-by: Nithin Raju
---
v2: Added Acked-by.
INSTALL.Windows.md | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/INSTALL.Windows.md b/INSTALL.Windows.md
index 0029500..0dd5cda 100644
--- a
> On Apr 13, 2015, at 1:10 PM, Gurucharan Shetty wrote:
>
> On Mon, Apr 13, 2015 at 1:08 PM, Nithin Raju wrote:
>> Added the commands to be able to install an unsigned driver.
>>
>> Acked-by: Nithin Raju
>> ---
>>
>> v2: Added Acked-by.
>
>
> -----Original Message-
> From: Nithin Raju [mailto:nit...@vmware.com]
> Sent: Monday, 13 April, 2015 20:55
> To: Sorin Vinturis
> Cc: dev@openvswitch.org
> Subject: Re: [ovs-dev] [PATCH] datapath-windows: extension fails to be enabled
>
>> On Apr 9, 2015, at
> On Apr 14, 2015, at 1:41 PM, Ben Pfaff wrote:
>
> On Tue, Apr 14, 2015 at 08:25:59PM +, Sorin Vinturis wrote:
>> The nl_sock_transact_multiple function enters in an infinite loop,
>> when invalid error, EINVAL, is returned by nl_sock_transact_multiple__.
>> EINVAL is the error returned by t
> On Apr 14, 2015, at 2:19 PM, Alin Serdean
> wrote:
>
> In nl_sock_transact_multiple__ we do the following:
>
> if (!DeviceIoControl(sock->handle, OVS_IOCTL_TRANSACT,
> txn->request->data,
> txn->request->size,
>
> On Apr 14, 2015, at 2:39 PM, Ben Pfaff wrote:
>
>>
>> What is the genetlink semantics for returning an error from the OVS
>> module in Linux. I looked at the OVS code, and it returns -EINVAL if
>> it runs into a condition of invalid input (eg. flow with key & UFID
>> missing). If OVS module re
> On Apr 14, 2015, at 1:25 PM, Sorin Vinturis
> wrote:
>
> The nl_sock_transact_multiple function enters in an infinite loop,
> when invalid error, EINVAL, is returned by nl_sock_transact_multiple__.
> EINVAL is the error returned by the latter function when a driver
> request fails.
>
> Signed
Sorin,
I’ll put a breakpoint on OvsExtNetPnPEvent() like you suggested and see how
this function is being used. Your patch is good, but I also feel we are only
getting rid of a symptom rather than a problem. I feel you should also be
removing the “switchActive == TRUE” check in the if() conditio
menting ‘*done’? ‘done’, IMO, represents the number of
transactions that got executed successfully. If the transaction failed (due to
a transport error), then it did not get executed successfully. So, IMO, we
should not be incrementing ‘done’.
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
> On Apr 15, 2015, at 1:49 PM, Eitan Eliahu wrote:
>
> If we remove the increment of the transaction than the whole sequence of
> transaction will be lost.
This is not true. We increment ‘*done’ after each successful transaction:
48 /* Count the number of successful transactions. */
49
e enter an
> infinite loop.
> Thanks,
> Eitan
>
> -Original Message-
> From: Sorin Vinturis [mailto:svintu...@cloudbasesolutions.com]
> Sent: Wednesday, April 15, 2015 9:10 PM
> To: Nithin Raju; Eitan Eliahu
> Cc: dev@openvswitch.org
> Subject: RE: [ovs-dev] [P
> On Apr 16, 2015, at 9:13 AM, Eitan Eliahu wrote:
>
> Nithin, the only issue I have with that is that currently the driver returns
> an I/O error in some cases where the I/O succeeds but the transaction fails
> (e.g. when OVS sends down a port delete NL message and there is no such a
> port).
Fixing a minor typo introduced in a previous commit.
Signed-off-by: Nithin Raju
---
INSTALL.Windows.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/INSTALL.Windows.md b/INSTALL.Windows.md
index 0dd5cda..78af0a1 100644
--- a/INSTALL.Windows.md
+++ b/INSTALL.Windows.md
h cases, netlink functions are to return
success with a 'struct nlmsgerr' populated in the output
buffer.
This patch implements these semantics.
Signed-off-by: Nithin Raju
Reported-at: https://github.com/openvswitch/ovs-issues/issues/72
---
datapath-windows/ovsext/Datapath.c | 60
hi Sorin,
I looked at your replies, and responded to the ones that required a response.
I’ll await the next version of the patch. In general, we want to make sure the
add/deletion workflow follows the pattern:
a. During OVS_VPORT_CMD_NEW, tunnel vport does not “exist” until
OvsTunnelVportPendin
This is a hard to hit corner case, because currently we recommend that
all handles to the kernel datapath be closed before trying to unload the
OVS extension.
Signed-off-by: Nithin Raju
---
datapath-windows/ovsext/Switch.c | 11 ---
1 file changed, 8 insertions(+), 3 deletions(-)
diff
.
I’ll ack this now. You can address the comments and send out a v3.
Acked-by: Nithin Raju
> On Apr 15, 2015, at 1:28 AM, Sorin Vinturis
> wrote:
>
> There is no need to use gOvsCtrlLock spinlock to guard the switch
> context, as there is now the switch context's refere
hi Sorin,
Thanks so much for the review.
Let me address your comments and then send out the updated patch. If I missed
anything, pls. let me know.
> UINT8 version;
> -UINT8 pad;
> +UINT8 pad1;
> UINT16 maxAttr;
> +UINT16 pad2;
> SV: Do we need the new 'pad' members, 'pad1' &
h cases, netlink functions are to return
success with a 'struct nlmsgerr' populated in the output
buffer.
This patch implements these semantics.
Signed-off-by: Nithin Raju
Acked-by: Sorin Vinturis
Reported-at: https://github.com/openvswitch/ovs-issues/issues/72
---
v2: addressed Sorin&#
6 deletions(-)
>
> diff --git a/datapath-windows/ovsext/User.c b/datapath-windows/ovsext/User.c
> index 03f0377..76a9ed1 100644
> --- a/datapath-windows/ovsext/User.c
> +++ b/datapath-windows/ovsext/User.c
Thanks for the fix.
Acked-by: Nithin Raju
h cases, netlink functions are to return
success with a 'struct nlmsgerr' populated in the output
buffer.
This patch implements these semantics.
Signed-off-by: Nithin Raju
Acked-by: Sorin Vinturis
Reported-at: https://github.com/openvswitch/ovs-issues/issues/72
---
v2: addressed Sorin&
> On Apr 29, 2015, at 7:27 AM, Gurucharan Shetty wrote:
>
> On Tue, Apr 28, 2015 at 11:49 PM, Sorin Vinturis
> wrote:
>> Instance PID is already deleted in the OvsCleanupPacketQueue function.
>>
>> Signed-off-by: Sorin Vinturis
>> Acked-by: Nithin Raju
&
Thanks for the review.
> On Apr 29, 2015, at 6:06 AM, Eitan Eliahu wrote:
>
>
> Hi Nithin,
> Sorry, I come late on this one, change looks good.
>
> We need to check if the IRP output buffer is big enough to fit the
> transaction message before copying it. Otherwise, we need to return error an
> On Apr 29, 2015, at 7:45 AM, Ben Pfaff wrote:
>
> On Tue, Apr 28, 2015 at 02:35:37PM -0700, Nithin Raju wrote:
>> In this patch, we make changes to usersapce as well as
>> kernel datapath on hyperv to make it more netlink socket
>> like. Previously, the kernel d
>> I agree with your comment.
>>
>> The error handling was in place since we were doing the whole “netlink
>> emulation” with any OS support unlike Linux, and we wanted to catch issues
>> in release builds also.
>>
>> If you are not very particular, can I let the code be? Once the code matures
>> As you know, the Windows kernel is synchronous in terms of netlink
>> messages. Transaction semantics are implemented in one call that
>> includes both the “request” and the “reply” in one shot. So, if
>> there’s a mismatch, it implies the kernel bungled the ‘nlmsg_seq’. So,
>> the assert was ad
> On Apr 29, 2015, at 2:52 PM, Eitan Eliahu wrote:
>
>
> Hi Alin,
> We would like to run on Windows Server too. In this case the installer should
> install the Win 8.0 driver.
> I don’t recall why we bumped the minor version. Unless we use 6.40 NDIS both
> 8.0 and 8.1 should register as 6.30 f
> On Apr 29, 2015, at 3:04 PM, Eitan Eliahu wrote:
>
> Actually, both version should be available for the installer. It should query
> the OS type and install the appropriate driver.
>
> But, unless there is a good reason for registering as 6.40 we could register
> as 6.30 with the 8.1 driver.
> On Apr 23, 2015, at 2:05 PM, Sorin Vinturis
> wrote:
>
> Signed-off-by: Sorin Vinturis
hi Sorin,
Thanks for these patches. They are super important. This one looks good, but
for a couple of minor comments. I’ll review the rest of them too.
Acked-by: Nithin Raju
> @@
> On Apr 23, 2015, at 2:05 PM, Sorin Vinturis
> wrote:
>
> OvsPartialCopyToMultipleNBLs function failed to correctly link the newly
> created NBL with single NB to the multiple NBLs list.
>
> Signed-off-by: Sorin Vinturis
> Co-authored-by: Alin Gabriel Serdean
> On May 6, 2015, at 11:57 PM, Sorin Vinturis
> wrote:
>
> Hi Nithin,
>
> Thanks for your review on my two patches. Please ignore the fourth patch in
> series, ' [PATCH 4/5] datapath-windows: Multiple NBLs support for userspace
> forwarded packets', because it considers that OvsTcpSegmentNBL
hi Sorin,
I looked at this patch. I had a few comments.
> +static NTSTATUS
> +OvsCreateNewNBLsFromMultipleNBs(POVS_SWITCH_CONTEXT switchContext,
> +PNET_BUFFER_LIST *curNbl,
> +PNET_BUFFER_LIST *nextNbl)
> +{
> +NTSTATUS status =
> On Apr 24, 2015, at 3:03 AM, Sorin Vinturis
> wrote:
>
> Added support for handling multiple NBLs with only one NB when
> encapsulating VXLAN packets.
>
> Signed-off-by: Sorin Vinturis
hi Sorin,
Was this patch necessary?
Here’s my reasoning for the question. If a NBL contains an NB that c
> On May 18, 2015, at 2:01 AM, Sorin Vinturis
> wrote:
>
> Hi Nithin,
>
> I agree with you. We can drop this patch then.
Sounds good. So, I presume you’ll send out a v2 or v3 with the comments
addressed and only including the required patches?
I’m reviewing the other patches.
thanks,
-- Nit
hi Sorin,
The patch had some minor merge issues in Datapath.c for me. You might want to
fix them.
I have a few relatively minor comments. Pls. address them and the patch should
be good to go.
> +_Use_decl_annotations_
> +VOID
> +OvsTunnelFilterThreadProc(PVOID context)
> +{
> +NTSTATUS
;r=pNHQcdr7B40b4h6Yb7FIedI1dnBsxdDuTLBYD3JqV80&m=f010yynSWCc8PLXnlLZW4FecGrSjZTYsvHaWf08jYok&s=SvgV-P3xJ1aU0rWB6Og1-nM3g05kapVCtr6SmnV-N5I&e=
>
> Acked-by: Eitan Eliahu
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
> On May 20, 2015, at 7:40 AM, Gurucharan Shetty wrote:
>
> On Mon, May 11, 2015 at 5:42 AM, Sorin Vinturis
> wrote:
>> At the moment the OVS extension supports only one VXLAN tunnel that
>> is cached in the extension switch context. Replaced the latter
>> cached pointer with an array list that
hi Sorin,
> While I’m ok with using do-while loop as an alternative to using labels, you
> are using 2 do-while looks here. Does it add more clarity than using labels?
> :) I am ok either way.
> SV: I am not in favor of using labels. For me, it makes the code very hard to
> follow and maintain
ows: document OVS tunnel filter
> callout
>
> Signed-off-by: Sorin Vinturis
> ---
> datapath-windows/ovsext/TunnelFilter.c | 225 -
> 1 file changed, 221 insertions(+), 4 deletions(-)
Acked-by: Nithin Raju
__
Sorin/Eitan,
Do we need this patch or not?
Eitan tells me that if return status is STATUS_PENDING, then the context is
blocked in the kernel. If that is true, we probably don’t need this patch since
control won’t return to userspace at all, till the IRP is complete. This is
especially true sinc
> On May 20, 2015, at 11:24 AM, Eitan Eliahu wrote:
> If a NULL sent rather than an overlapped structure, the thread which calls
> DeviceIOControl is blocked until the I/O is completed by the driver.
Agree. So, looks like we don’t need handling for STATUS_PENDING when overlapped
structure is NU
>> +static NTSTATUS
>> +OvsCreateNewNBLsFromMultipleNBs(POVS_SWITCH_CONTEXT switchContext,
>> +PNET_BUFFER_LIST *curNbl,
>> +PNET_BUFFER_LIST *nextNbl) {
>> +NTSTATUS status = STATUS_SUCCESS;
>> +PNET_BUFFER_LIST newNbls = NULL
hi Sorin,
Thanks for the respin.
I had 2 comments in the previous review regarding access to a freed up ‘vport’
in OvsDeleteVportCmdHandler() and OvsTunnelVportPendingUninit().
It looks like you addressed the issue with OvsTunnelVportPendingUninit(), by
making sure that the Vport is not dealloc
> On May 21, 2015, at 4:30 AM, Sorin Vinturis
> wrote:
>
> Modified 'nl_sock_transact_multiple__' function to send asynchronous
> I/O requests to the kernel driver.
>
> Signed-off-by: Sorin Vinturis
> Reported-by: Alin Gabriel Serdean
> Reported-at:
> https://urldefense.proofpoint.com/v2/url
> On May 21, 2015, at 12:54 PM, Sorin Vinturis
> wrote:
>
> Nithin,
>
> I thought I have addressed your comment. My understanding was that was an
> issue in OvsDeleteVportCmdHandler() after the call to
> OvsRemoveAndDeleteVport(). The latter function might deallocate the vport,
> which woul
hi Eitan,
Thanks for this patch. Looks good overall. I had a few minor comments which I
have inlined.
Also, please update he instructions in INSTALL.Windows.md about the
configuration steps for setting STT tunnels and testing them.
thanks,
-- Nithin
> On May 7, 2015, at 6:08 PM, Eitan Eliah
Signed-off-by: Nithin Raju
---
datapath-windows/ovsext/NetProto.h | 3 ---
datapath-windows/ovsext/Vxlan.c| 2 +-
2 files changed, 1 insertion(+), 4 deletions(-)
diff --git a/datapath-windows/ovsext/NetProto.h
b/datapath-windows/ovsext/NetProto.h
index a364869..1a81cc3 100644
--- a
There's some code in PacketIO.c that is supported in WDK 8.1 only.
The variable declarations for that code must also be WDK 8.1 only.
Signed-off-by: Nithin Raju
---
datapath-windows/ovsext/PacketIO.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/datapath-wi
hi Sorin,
This patch looks much better. Like I mentioned the review to the previous
revision, a duplicate vport delete request for the same VXLAN port can mess
things up, including accessing a freed up ‘vport’ structure. But, that is a
corner case, and we can address it later.
I just found a n
> On May 25, 2015, at 11:50 PM, Nithin Raju wrote:
>
>
> hi Sorin,
> This patch looks much better. Like I mentioned the review to the previous
> revision, a duplicate vport delete request for the same VXLAN port can mess
> things up, including accessing a freed up ‘
> On May 27, 2015, at 9:06 AM, Sorin Vinturis
> wrote:
>
> Nithin,
>
> I will address your comment in the next version of this patch.
Sure.
> Regarding the duplicate vport delete request, I have not found a way to
> generate such request from userspace. When I use 'ovs-vsctl' application to
. OvsTunnelFilterUninitialize function closes the
> global engine handle used by most of the WFP calls, including filter
> removal.
>
> Signed-off-by: Sorin Vinturis
> Reported-by: Alin Gabriel Serdean
Thanks for your patience!
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
nnection through ping works. File transfer.
>
> Signed-off-by: Eitan Eliahu
> Co-authored-by: Saurabh Shah
> Signed-off-by: Saurabh Shah
Thanks for addressing all the comments.
Acked-by: Nithin Raju
___
dev mailing list
dev@openvsw
> On May 27, 2015, at 9:30 AM, Ben Pfaff wrote:
>
> Hi Sorin, I'd like to apply this but I get a patch failure:
>
>Applying: datapath-windows: Support for custom VXLAN tunnel port
>/home/blp/nicira/ovs/.git/rebase-apply/patch:917: trailing whitespace.
>
>/home/blp/nicira/ovs/.git/re
.proofpoint.com/v2/url?u=https-3A__github.com_openvswitch_ovs-2Dissues_issues_2&d=AwIGaQ&c=Sqcl0Ez6M0X8aeM67LKIiDJAXVeAw-YihVMNtXt-uEs&r=pNHQcdr7B40b4h6Yb7FIedI1dnBsxdDuTLBYD3JqV80&m=y4y3UuLaGUsnNvkV7isVjJjW2X9jaIPMz8W-Yj5zS5E&s=D8zmIwV4t9MSl1rSRvq4G72
datapath-windows\ovsext\PacketIO.c
>
> Signed-off-by: Alin Gabriel Serdean
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
or multiple (per TCP port) STT ports
>
> Testing: link layer connection through ping works. File transfer.
>
> Signed-off-by: Eitan Eliahu
> Co-authored-by: Saurabh Shah
> Signed-off-by: Saurabh Shah
Acked-by: Nithin Raju
_
he fix. I am curious how this was not caught during testing for the
previous patch.
Acked-by: Nithin Raju
thanks,
-- Nithin
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
> On Jun 12, 2015, at 5:27 AM, Sorin Vinturis
> wrote:
>
> Nithin,
>
> The previous patch you are referring was one of a series I had pushed when
> working at the "Multiple NBLs" issue. At that time I had some trouble when
> generating packets large enough for the a NBL with multiple NBs to b
> On Jun 12, 2015, at 8:56 AM, Gurucharan Shetty wrote:
>
> I haven't done any research here. A quick google says that Hyper-V
> does support vagrant. So a way to test kernel module would be to get a
> vagrant infrastructure for Hyper-V too. We currently have such a
> infrastructure for Linux ins
DJAXVeAw-YihVMNtXt-uEs&r=pNHQcdr7B40b4h6Yb7FIedI1dnBsxdDuTLBYD3JqV80&m=rbO3fZK9znzO5lzyAsa-ysHnSltAFG7n3MO1cKmTR9Q&s=Vy-jaJqDgXO2rECjvh4JgxFajSNAkX46YorJbxE3bCo&e=
>>
Too late for the ACK, but doing it nevertheless since I reviewed this.
Acked-by: Nithin Raju
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
> On Jun 18, 2015, at 10:49 AM, Alin Serdean
> wrote:
>
> Acked-by: Alin Gabriel Serdean
>
>
>
> -Mesaj original-
> De la: dev [mailto:dev-boun...@openvswitch.org] În numele Sorin Vinturis
> Trimis: Thursday, June 18, 2015 8:48 PM
> Către: dev@openvswitch.org
> Subiect: [ovs-dev] [P
Sorin/Alin,
I worked on a patch for this issue, and I am testing it today. I’ve also
refactored some of the code in Vport.c. Would it be ok if I send out my patch
instead of this?
IIRC, we had agreed that VMware folks would work on this issue.
thanks,
-- Nithin
_
> On Jun 18, 2015, at 11:52 AM, Sorin Vinturis
> wrote:
>
> Disconnecting the NIC from HV switch and then unloading the driver hits an
> ASSERT (ASSERT(switchContext->numPhysicalNics);).
>
> Also if the external or internal port is deleted from the userspace, the
> latter ports would not be re-
> On Jun 18, 2015, at 1:57 PM, Alin Serdean
> wrote:
>
> I have no issue with it if it is tested.
>
> Yes Nithin you are right we agreed that you will work on it but it was in the
> backlog for some time and it is a show stopper because it also
> fixes(https://urldefense.proofpoint.com/v2/url
A previous commit used the wrong DST port in the UDP header during Vxlan
Tx which caused Vxlan tunneling to break. Fixing it here..
Testing done: Ping across Vxlan tunnel and Stt tunnel.
Signed-off-by: Nithin Raju
Reported-by: Eitan Eliahu
---
datapath-windows/ovsext/Actions.c | 9
done: Ping across Vxlan tunnel and Stt tunnel.
Signed-off-by: Nithin Raju
Reported-by: Eitan Eliahu
---
v2: updated the commit message
---
datapath-windows/ovsext/Actions.c | 9 -
datapath-windows/ovsext/Stt.c | 4 ++--
datapath-windows/ovsext/Vxlan.c
These logs were being logged very frequently on the debugger.
Signed-off-by: Nithin Raju
---
datapath-windows/ovsext/Flow.c| 2 +-
datapath-windows/ovsext/Netlink/Netlink.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/datapath-windows/ovsext/Flow.c b
letion
- Ping on Vxlan/Stt tunnels
- Ovs Extension load/unload. There's an unrelated issue I found that is
reported in: https://github.com/openvswitch/ovs-issues/issues/86
Signed-off-by: Nithin Raju V
Reported-at: https://github.com/openvswitch/ovs-issues/issues/79
Reported-by: Alin Gabriel Ser
ould have
isPresentOnHv == TRUE.
In this patch, we rename the variable to reflect its meaning.
vport->isAbsentOnHv is TRUE iff:
- vport is bridge internal port
- vport is tunnel port
- vport was added from Hyper-V and also from OVS, but got deleted from
Hyper-V
Signed-off-by: Nithin Raju
---
v2:
letion
- Ping on Vxlan/Stt tunnels
- Ovs Extension load/unload. There's an unrelated issue I found that is
reported in: https://github.com/openvswitch/ovs-issues/issues/86
Signed-off-by: Nithin Raju V
Reported-at: https://github.com/openvswitch/ovs-issues/issues/79
Reported-by: Alin Gabriel Serdea
Pls. discard this patch, since I made it part of a series now.
> On Jun 18, 2015, at 9:57 PM, Nithin Raju wrote:
>
> In this patch, there a couple of fixes and some code refactoring:
> 1. During deletion of "internal" and "external" in
> OvsRemoveAn
Alin,
Thanks for the patch. A general comment I have is that, once you call into
OvsTunnelPortTx(), the caller has lost ownership of ovsFwdCtx->curNbl.
ovsFwdCtx->fwdDetail is actual a pointer to the fwdDetail within
ovsFwdCtx->curNbl. So, using ovsFwdCtx->fwdDetail would not be valid after the
nternal port
> Delete external port
> Test connectivity
> Add internal port
> Test connectivity
> Add external port
> Test connectivity
>
> Uninstall extension verify port deletion.
>
> Alin.
>
> -Mesaj original-
> De la: dev [mailto:dev-boun...@
I actually meant to ask about the validation for this patch.
Was any simulation/synthetic test done to validate the new code being
introduced in the patch?
thanks,
-- Nithin
> On Jun 18, 2015, at 5:19 PM, Alin Serdean
> wrote:
>
> Acked-by: Alin Gabriel Serdean
>
>
> -Mesaj original--
> On Jun 19, 2015, at 8:47 AM, Alin Serdean
> wrote:
>
> Just a personal preference maybe we could drop OvsIpHlprCbVxlan all togheter.
>
> Rest looks fine with me.
>
> Acked-by: Alin Gabriel Serdean
I don’t have any issues removing OvsIpHlprCbVxlan() since it is unused, and can
be resurrect
o 1, before 'gOvsSwitchContext =
switchContext’, since the gating factor in OvsDeviceControl() to begin
processing is: gOvsSwitchContext != NULL. Just my personal opinion though. I am
ok with this change too.
Acked-by: Nithin Raju
___
dev ma
701 - 800 of 936 matches
Mail list logo
