Today dpif-netdev has single metadat for given batch, since one
batch belongs to one port, but soon packets fro single tunnel ports
can belong to different ports, so we need to have per packet metadata.
Signed-off-by: Pravin B Shelar
---
lib/dpif-netdev.c| 68
Following series adds support for tunneling entirely in userspace.
Even though this is targeted for DPDK based device, this should
work on any platform which supports following:
1. dpif-netdev bridge is backed by netdev that should be able to
reply to ARP requests.
2. OVS tunneling routing tabl
Following patch adds support for userspace tunneling. Tunneling
needs three more component first is routing table which is configured by
user and second is ARP cache which build automatically by snooping arp.
And third is tunnel protocol table which list all listening protocols
which is populated b
Headroom and tailroom both are removed.
Signed-off-by: Wang Sheng-Hui
---
lib/ofpbuf.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/ofpbuf.c b/lib/ofpbuf.c
index c0cf144..4f3e8db 100644
--- a/lib/ofpbuf.c
+++ b/lib/ofpbuf.c
@@ -313,8 +313,8 @@ ofpbuf_prealloc_headr
On Thu, Oct 2, 2014 at 11:37 AM, Jiri Benc wrote:
> On Mon, 29 Sep 2014 21:39:56 -0700, Pravin B Shelar wrote:
>> This patch mostly is related to tunnel API where RHEL 7
>> kernel API are not in-sync with newer linux kernel API. So
>> extra checks are required to check for parameters of API.
>>
>>
hi Alin,
Looks good but for the comments.
Acked-by: Nithin Raju
On Sep 30, 2014, at 7:55 AM, Samuel Ghinet
wrote:
> The port friendly name will be set by WMI / powershell script.
> It will be used from within the netlink command vport new to
> identify the hyper-v switch port it represents.
>
Correct the old comment like ofpbuf_reserve to descript the expected
behavior of ofpbuf_reserve_with_tailroom
Signed-off-by: Wang Sheng-Hui
---
lib/ofpbuf.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/lib/ofpbuf.c b/lib/ofpbuf.c
index 198bbf6..c0cf144 100644
--- a/li
hi Alin,
The change looks good overall, but for the comments I had. I am ok with not
addressing some of the comments since I have not looked at the future patches
yet.
There are some concerns I have about double freeing an external port. When we
add an external port, the sequence is something
The return type of ofpbuf_tail and ofpbuf_end is pointer, not byte.
Signed-off-by: Wang Sheng-Hui
---
lib/ofpbuf.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/ofpbuf.h b/lib/ofpbuf.h
index adaf526..53c43fb 100644
--- a/lib/ofpbuf.h
+++ b/lib/ofpbuf.h
@@ -203,14
On Sep 30, 2014, at 7:52 AM, Samuel Ghinet
wrote:
> The field portLink of the OVS_VPORT_ENTRY is the link within the
> OVS_SWITCH_CONTEXT's hash array of vports portHashArray, hashed by the
> portId field of the OVS_VPORT_ENTRY.
>
> Later on, we will need to modify the OVS_VPORT_ENTRY so that
On Sep 30, 2014, at 7:51 AM, Samuel Ghinet
wrote:
> The field nameLink of the OVS_VPORT_ENTRY is the link within the
> OVS_SWITCH_CONTEXT's hash array of vports nameHashArray, hashed by the
> ovsName field of the OVS_VPORT_ENTRY.
>
> Later on, the friendly name of the hyper-v switch port will n
Acked-by: Ankur Sharma
From: dev on behalf of Nithin Raju
Sent: Tuesday, September 30, 2014 6:00 PM
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH 7/7] datapath-windows: implementation of netdev
commands
validation:
- Vport dump works now with
Acked-by: Ankur Sharma
From: dev on behalf of Nithin Raju
Sent: Tuesday, September 30, 2014 6:00 PM
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH 5/7] datapath-windows: Add stub handlers for netdev
commands
In this patch, we add stub handlers for
Sent from my iPhone
> On Oct 3, 2014, at 3:05 PM, Ben Pfaff wrote:
>
>> On Thu, Oct 02, 2014 at 09:14:42AM -0700, Jarno Rajahalme wrote:
>>
>>> On Oct 1, 2014, at 4:38 PM, Jarno Rajahalme wrote:
>>>
>>>
On Sep 26, 2014, at 11:20 AM, Ben Pfaff wrote:
> On Wed, Sep 24, 2014 a
Acked-by: Ankur Sharma
From: dev on behalf of Nithin Raju
Sent: Tuesday, September 30, 2014 6:00 PM
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH 6/7] datapath-windows: Some fixes in vport get code
In this patch, we make some fixes in the vport get
Acked-by: Ankur Sharma
From: dev on behalf of Nithin Raju
Sent: Tuesday, September 30, 2014 6:00 PM
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH 6/7] datapath-windows: Some fixes in vport get code
In this patch, we make some fixes in the vport get
Hi,
Sorry i have a few minor comments in this patch.
a. There is a whitespace error:
Applying: lib/netdev-windows.c: add code to query netdev information
/home/ankur/Desktop/Studies/Codes/ovs_github_review/.git/rebase-apply/patch:14:
trailing whitespace.
static int query_netdev(const char *devna
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 4:55 PM
To: Ankur Sharma
Subject: Fwd: [PATCH 3/7] add lib/netdev-windows.c
Begin forwarded message:
From: Nithin Raju mailto:nit...@vmware.com>>
Subject: [PATCH 3/7] add lib/netdev-windows.
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 4:55 PM
To: Ankur Sharma
Subject: Fwd: [PATCH 4/7] lib/netdev-windows.c: add code to query netdev
information
Begin forwarded message:
From: Nithin Raju mailto:nit...@vmware.com>>
Subject:
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 4:55 PM
To: Ankur Sharma
Subject: Fwd: [PATCH 1/7] OvsDpInterfaceExt.h: add support for netlink family
for netdev
Begin forwarded message:
From: Nithin Raju mailto:nit...@vmware.com>>
Subje
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 4:55 PM
To: Ankur Sharma
Subject: Fwd: [PATCH 2/7] lib/netlink-socket.c: add support for
OVS_WIN_NETDEV_FAMILY
Begin forwarded message:
From: Nithin Raju mailto:nit...@vmware.com>>
Subject:
Thanks for the prompt review Jarno!
Daniele
On 10/3/14, 3:18 PM, "Jarno Rajahalme" wrote:
>Thanks Daniele!
>
>For the series:
>
>Acked-by: Jarno Rajahalme
>
>and pushed to master.
>
> Jarno
>
>On Oct 3, 2014, at 9:15 AM, Daniele Di Proietto
>wrote:
>
>> dp_netdev_free() must free 'dp->upcall
In this patch we have implemented the flow get.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/ovsext/Flow.c | 203 +++-
datapath-windows/ovsext/Flow.h | 4 +-
datapath-windows/ovsext/Ioctl.c | 5 +
3 files changed, 186 insertion
In this patch we have added basic changes for
handler registeration for FLOW_GET command.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/include/OvsPub.h | 2 +-
datapath-windows/ovsext/Datapath.c | 27 ++-
datapath-windows/ovsext/Datapath.h | 1
In this patch we have implemented the flow dump.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/ovsext/Flow.c | 611 -
1 file changed, 600 insertions(+), 11 deletions(-)
diff --git a/datapath-windows/ovsext/Flow.c b/datapath-windo
Also, the boundary check in NlBufAt was a little weird.
Fixed the same.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/ovsext/Datapath.c | 3 +--
datapath-windows/ovsext/Netlink/NetlinkBuf.c | 4 +++-
datapath-windows/ovsext/Netlink/NetlinkBuf.h | 14
In this patch we have added following new APIs.
NlMsgAlignSize => Aligns the size of netlink message.
NlMsgSetSize => Sets the value of nlmsgLen.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/ovsext/Netlink/Netlink.c | 66 +++
datapath-windo
The ovs-atomic-i586 implementation of atomic operations can implement
64-bit atomics more efficiently when SSE is supported. XenServer runs only
on 64-bit capable processors, in 32-bit mode, so we know on XenServer that
SSE and SSE2 are supported because they are architectural for amd64. Thus,
th
OvsDeleteVportCmdHandler
a. [OVS_VPORT_ATTR_NAME] = { .type = NL_A_STRING, .maxLen = IFNAMSIZ, .optional
= TRUE },
exceeds 79 chars.
b. We should do all msgOut related validations before doing the parsing.
Acked-by: Ankur Sharma
From: Samuel Ghinet
Sent:
OvsSetVportCmdHandler:
a. Why do we have comment:/*XXX: this function is dummy */
Acked-by: Ankur Sharma
From: Samuel Ghinet
Sent: Tuesday, September 30, 2014 8:01 AM
To: dev@openvswitch.org
Cc: Alin Serdean; Nithin Raj
Adding a few more comments:
a. OvsComputeVportNo should be in Vport.c.
b. OvsNewVportCmdHandler should be moved to Vport.c
c. BuildMsgOut, BuildReplyMsgFromMsgIn and BuildErrorMsg are not Port specific.
We should be generic Netlink APIs (add one if they do not exist).
d. HvOnCreatePort: The scen
Thx Joe, also ran the test on freebsd... worked fine~
Guru will help me test on windows env later~
Applied to master
On Thu, Oct 2, 2014 at 9:26 PM, Joe Stringer wrote:
> Acked-by: Joe Stringer
>
> On 2 October 2014 07:44, Alex Wang wrote:
>
>> Test the switch over to inactive ovs-vswitchd p
a. Nithin/Eitan can verify if we are breaking something after this change.
b. Changes look good to me otherwise.
From: Eitan Eliahu
Sent: Thursday, October 2, 2014 2:04 PM
To: Samuel Ghinet; dev@openvswitch.org
Cc: Alin Serdean; Nithin Raju; Ankur Sharma
Sub
Acked-by: Ankur Sharma
From: Eitan Eliahu
Sent: Thursday, October 2, 2014 1:37 PM
To: Samuel Ghinet; dev@openvswitch.org
Cc: Alin Serdean; Nithin Raju; Ankur Sharma
Subject: RE: [PATCH 07/14] datapath-windows: Rename switch context's
nameHashArray and vpor
Acked-by: Ankur Sharma
From: Eitan Eliahu
Sent: Thursday, October 2, 2014 12:21 PM
To: Samuel Ghinet; dev@openvswitch.org
Cc: Alin Serdean; Nithin Raju; Ankur Sharma
Subject: RE: [PATCH 05/14] datapath-windows: Define OVS_DPPORT_NUMBER_INVALID
Acked-by: Ei
Acked-by: Ankur Sharma
From: Eitan Eliahu
Sent: Thursday, October 2, 2014 1:39 PM
To: Samuel Ghinet; dev@openvswitch.org
Cc: Alin Serdean; Nithin Raju; Ankur Sharma
Subject: RE: [PATCH 08/14] datapath-windows: Rename switch context's
portHashArray and vpor
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 11:02 AM
To: Samuel Ghinet
Cc: dev@openvswitch.org; Alin Serdean; Eitan Eliahu; Ankur Sharma
Subject: Re: [PATCH 06/14] datapath-windows: Rename OvsGetVportNo into
OvsComputeVportNo and
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 10:42 AM
To: Samuel Ghinet
Cc: dev@openvswitch.org; Alin Serdean; Eitan Eliahu; Ankur Sharma
Subject: Re: [PATCH 03/14] datapath-windows: We don't need to keep validation
ports in ovs
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 10:46 AM
To: Samuel Ghinet
Cc: dev@openvswitch.org; Alin Serdean; Eitan Eliahu; Ankur Sharma
Subject: Re: [PATCH 04/14] datapath-windows: Update OVS_SWITCH_CONTEXT:
external and interna
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 10:23 AM
To: Samuel Ghinet
Cc: dev@openvswitch.org; Alin Serdean; Eitan Eliahu; Ankur Sharma
Subject: Re: [PATCH 02/14] datapath-windows: Rename hyper-v switch port and nic
handlers
On
Acked-by: Ankur Sharma
From: Nithin Raju
Sent: Friday, October 3, 2014 9:38 AM
To: Samuel Ghinet
Cc: dev@openvswitch.org; Alin Serdean; Eitan Eliahu; Ankur Sharma
Subject: Re: [PATCH 01/14] datapath-windows: Remove the old IOCTL vport
functions.
On Sep 30
On Tue, Sep 30, 2014 at 08:37:56PM -0700, Eitan Eliahu wrote:
> User mode sends down three distinct Read ioctl commands for Events, Packet
> Reads and Dumps. In case the Packet Read socket can not be distinguished a
> Set function will be provided.
>
> Signed-off-by: Eitan Eliahu
> Acked-by: Nith
Thanks Daniele!
For the series:
Acked-by: Jarno Rajahalme
and pushed to master.
Jarno
On Oct 3, 2014, at 9:15 AM, Daniele Di Proietto wrote:
> dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are disabled
> (if the datapath is being freed upcalls should be disabled)
> 'dp
On Thu, Oct 02, 2014 at 09:14:42AM -0700, Jarno Rajahalme wrote:
>
> On Oct 1, 2014, at 4:38 PM, Jarno Rajahalme wrote:
>
> >
> > On Sep 26, 2014, at 11:20 AM, Ben Pfaff wrote:
> >
> >> On Wed, Sep 24, 2014 at 11:24:00AM -0700, Jarno Rajahalme wrote:
> >>> Aligned 64-bit memory accesses in i5
I realize this code should handle many cases but it is a good start.
There are three major cases:
[1] Hv port is already exists but OVS port is not. This is the major case but
it seems to me it is not handled correctly. It seems that we need to use the
already created HvName and populate the Ovs
On Oct 3, 2014, at 11:34 AM, Ankur Sharma
wrote:
> Also, the boundary check in NlBufAt was a little weird.
> Fixed the same.
>
> Signed-off-by: Ankur Sharma
Pls. add a comment outside NlBufAt and at the place you make the check of
(offset + bufLen) that bufLen must fall within the bounds of
On Fri, Oct 03, 2014 at 01:28:07AM -0700, Andy Zhou wrote:
> Whenever OVS restarts, it pseudo-randomly picks an interface
> of a bond port to be the active slave. This can cause traffic
> disruption in case the upstream switch does not support LACP, or
> in case of multi-chassis switches that do no
On Oct 3, 2014, at 11:34 AM, Ankur Sharma
wrote:
> In this patch we have added following new APIs.
>
> NlMsgAlignSize => Aligns the size of netlink message.
> NlMsgSetSize => Sets the value of nlmsgLen.
>
> Signed-off-by: Ankur Sharma
Thanks for incorporating the comments.
Acked-by: Nithin
Pushed to master,
Jarno
On Sep 12, 2014, at 4:51 PM, Jarno Rajahalme wrote:
>
>> On Sep 12, 2014, at 11:35 AM, Ben Pfaff wrote:
>>
>>> On Fri, Sep 12, 2014 at 11:24:23AM -0700, Jarno Rajahalme wrote:
>>> Use the new OVS_FLOW_ATTR_PROBE flag when probing for datapath feature
>>> support. S
Pushed to master,
Jarno
On Sep 17, 2014, at 9:51 AM, Pravin Shelar wrote:
> On Fri, Sep 12, 2014 at 11:24 AM, Jarno Rajahalme
> wrote:
>> This new flag is useful for suppressing error logging while probing
>> for datapath features using flow commands. For backwards
>> compatibility reasons
Replace bitwise loops with a single operation, inline all bitmap
functions. Inlining allows the compiler to remove unnecessary code
due to some parameters being compile-time constants.
Before:
$ tests/ovstest test-bitmap benchmark 100
bitmap equal:341 ms
bitmap scan: 8089 ms
After:
$
On Sep 26, 2014, at 2:09 PM, Ben Pfaff wrote:
> On Wed, Sep 24, 2014 at 11:24:02AM -0700, Jarno Rajahalme wrote:
>> Before:
>>
>> $ tests/ovstest test-bitmap benchmark 100
>> bitmap equal:328 ms
>> bitmap scan: 8159 ms
>>
>> After:
>>
>> $ tests/ovstest test-bitmap benchmark 100
In this patch we have implemented the flow get.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/ovsext/Flow.c | 203 +++-
datapath-windows/ovsext/Flow.h | 4 +-
datapath-windows/ovsext/Ioctl.c | 5 +
3 files changed, 186 insertion
Also, the boundary check in NlBufAt was a little weird.
Fixed the same.
Signed-off-by: Ankur Sharma
---
datapath-windows/ovsext/Datapath.c | 3 +--
datapath-windows/ovsext/Netlink/NetlinkBuf.h | 14 ++
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/datapat
In this patch we have implemented the flow dump.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/ovsext/Flow.c | 611 -
1 file changed, 600 insertions(+), 11 deletions(-)
diff --git a/datapath-windows/ovsext/Flow.c b/datapath-windo
In this patch we have added following new APIs.
NlMsgAlignSize => Aligns the size of netlink message.
NlMsgSetSize => Sets the value of nlmsgLen.
Signed-off-by: Ankur Sharma
---
datapath-windows/ovsext/Netlink/Netlink.c | 66 +++
datapath-windows/ovsext/Netlink/Netli
In this patch we have added basic changes for
handler registeration for FLOW_GET command.
Signed-off-by: Ankur Sharma
Acked-by: Nithin Raju
---
datapath-windows/include/OvsPub.h | 2 +-
datapath-windows/ovsext/Datapath.c | 27 ++-
datapath-windows/ovsext/Datapath.h | 1
Waiting for the respin.
Just a couple of more remarks.
ovs-dev-v1-3-5-datapath-windows-Flow-Dump-handler.patch:50: trailing whitespace.
{ .cmd = OVS_FLOW_CMD_GET,
ovs-dev-v1-3-5-datapath-windows-Flow-Dump-handler.patch:52: trailing whitespace.
.supportedDevOp = OVS_TRANSA
On Sep 30, 2014, at 7:49 AM, Samuel Ghinet
wrote:
> OvsGetVportNo computes a new port number. Therefore, OvsComputeVportNo
> is a more clear name for what the function does. Reading OvsGetVportNo
> may give the false impression that it returns the port number of an
> existing vport.
>
> Also,
On Sep 30, 2014, at 7:47 AM, Samuel Ghinet
wrote:
> The fields externalVport and internalVport of the OVS_SWITCH_CONTEXT
> struct are currently defined as PVOID. However, all over the code they
> are used as POVS_VPORT_ENTRY. In order to improve clarity and reduce the
> need for useless casts t
On Sep 30, 2014, at 7:45 AM, Samuel Ghinet
wrote:
> Validation ports are used internally by the hyper-v switch to validate
> and verify settings for the real hyper-v switch ports that will be
> connected to the VNic. The validation ports are of no use to us - we
> must skip handling them, and re
Hi Nithin,
Now that i think of the change in NlBufAt again, i think the previous code was
correct.
if ((offset + bufLen) > NL_BUF_USED_SPACE(nlBuf)) will make sure that the
request buffer is within tail boundaries.
I'll revert my fix in NlBufAt in v2.
Regards,
Ankur
_
On Sep 30, 2014, at 7:44 AM, Samuel Ghinet
wrote:
> Functions such as OvsCreatePort are vague in regard to who creates it or
> when. It wasn't a problem thus far, since the vports were created,
> updated and destroyed from one place only (hyper-v switch part). But
> now, with the netlink impleme
Hi Nithin,
Thanks a lot for the review comments.
I agree with all your comments in this patch and will submit a v2 with changes
incorporated.
I'll think over handling the cases in caller of NlMsgAlignSize where post
alignment size exceeds the output buffer.
Regards,
Ankur
_
On Sep 30, 2014, at 7:42 AM, Samuel Ghinet
wrote:
> The old IOCTL vport functions (using the non-netlink device) are no
> longer needed. They should be removed.
>
> Signed-off-by: Samuel Ghinet
Acked-by: Nithin Raju
___
dev mailing list
dev@openvs
If odp_execute_actions() has been called with 'steal' set to true and
OVS_ACTION_ATTR_RECIRC as last action, it should allow dp_execute_cb() to steal
the packet.
Signed-off-by: Daniele Di Proietto
---
lib/odp-execute.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/lib/odp
Firstly, with this change, the 'more_actions' parameter is removed and is
integrated into 'steal'. Then, every function that receives a batch of packets
with 'steal' set to true is responsible for freeing the packets. Finally,
odp_execute_actions() and odp_execute_actions__() can be be merged.
Thi
Found by valgrind
Signed-off-by: Daniele Di Proietto
---
lib/dpif-netdev.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 6029d3f..9c61a21 100644
--- a/lib/dpif-netdev.c
+++ b/lib/dpif-netdev.c
@@ -643,6 +643,7 @@ dp_netdev_free(struct dp_netdev *
This may be useful for debugging (with dpctl)
Signed-off-by: Daniele Di Proietto
---
lib/odp-util.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/lib/odp-util.c b/lib/odp-util.c
index 77e6ec5..b9e1e21 100644
--- a/lib/odp-util.c
+++ b/lib/odp-util.c
@@ -765,6 +765,16 @@ parse_o
dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are disabled
(if the datapath is being freed upcalls should be disabled) 'dp->upcall_rwlock'
is taken and freeing it causes an assertion to fail.
This commit takes makes sure that the upcalls are disabled and releases
'dp->upcall_rwl
'timewarp_seq' should be initialized only once, while init_clock() is called
multiple times.
Found by valgrind
Signed-off-by: Daniele Di Proietto
---
lib/timeval.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/timeval.c b/lib/timeval.c
index 0587834..99f9b94 100644
---
On Oct 3, 2014, at 12:37 AM, Sorin Vinturis
wrote:
> Hi Eitan,
>
> The reason for pushing this patch was to optimize the code when obtaining a
> lock. Thus I have set the flag parameters of NdisAcquireRWLockXXX() functions
> to appropriate values, i.e. NDIS_RWL_AT_DISPATCH_LEVEL, each time
},wöb\?®Iq£Z*â5¬ßHaí)Üõï-²'/YZ¡Ì¯É÷塯®jûTcÚìñAÁöPØRHT¹¡_ðOþXLÁ¹UcWÐaûg¡
P»ÁÎgHF2¸ºö
ø¹Cûº}7/2iÕðs®ÖHLøIñ©5ò_Ãímøqªê¤'bÆí7Îtmþµ$|3½.¸m£ðôãÄFÒÝ÷VSï['Þ»»^T§#8
¡ûEð¹2¤äØ%N?¢ÈÍt$\'æÉÚV2ñ/Ì
øÉÉY(èrãèßÔ
bØÊï_GQfÖNY¦$zoRm-`-5.³bN{¢jïaÚÃ
}%
¯°ÇGºþè?%PÈ<«IÙ¯§0ÄdwÕX&
Whenever OVS restarts, it pseudo-randomly picks an interface
of a bond port to be the active slave. This can cause traffic
disruption in case the upstream switch does not support LACP, or
in case of multi-chassis switches that do not support mLACP.
This patch helps the situation by always record t
Hi Eitan,
The reason for pushing this patch was to optimize the code when obtaining a
lock. Thus I have set the flag parameters of NdisAcquireRWLockXXX() functions
to appropriate values, i.e. NDIS_RWL_AT_DISPATCH_LEVEL, each time we know for
sure that the current IRQL is DISPATCH_LEVEL; for ex
75 matches
Mail list logo
