Netlink sockets are maintained in a hash table to allow efficient lookup
via the port ID for unicast messages. However, lookups currently require
a read lock to be taken. This series adds a new generic, resizable,
scalable, concurrent hash table based on the paper referenced in the first
patch. It
Heavy Netlink users such as Open vSwitch spend a considerable amount of
time in netlink_lookup() due to the read-lock on nl_table_lock. Use of
RCU relieves the lock contention.
Makes use of the new resizable hash table to avoid locking on the
lookup.
The hash table will grow if entries exceeds 75
Generic implementation of a resizable, scalable, concurrent hash table
based on [0]. The implementation supports both, fixed size keys specified
via an offset and length, or arbitrary keys via own hash and compare
functions.
Lookups are lockless and protected as RCU read side critical sections.
Au
Hello guys,
While working on the integration between our kernel netlink component and your
driver, I had my attention drawn a bit more towards your implementation of the
buffer management.
I expect that there was a discussion upon it, but I'd like to highlight a few
things myself, hoping that t
On 08/02/2014 11:47 AM, Thomas Graf wrote:
> Generic implementation of a resizable, scalable, concurrent hash table
> based on [0]. The implementation supports both, fixed size keys specified
> via an offset and length, or arbitrary keys via own hash and compare
> functions.
>
> Lookups are lockle
On 08/02/2014 11:47 AM, Thomas Graf wrote:
> Heavy Netlink users such as Open vSwitch spend a considerable amount of
> time in netlink_lookup() due to the read-lock on nl_table_lock. Use of
> RCU relieves the lock contention.
>
> Makes use of the new resizable hash table to avoid locking on the
>
On 08/02/2014 11:47 AM, Thomas Graf wrote:
> The sizing of the hash table and the practice of requiring a lookup
> to retrieve the pprev to be stored in the element cookie before the
> deletion of an entry is left intact.
>
> Signed-off-by: Thomas Graf
> Acked-by: Patrick McHardy
> ---
> net/ne
Hello guys,
I wanted to ask you: do you have buffer management functionality to duplicate a
packet?
I have seen that the function OvsOutputBeforeSetAction CLONES instead of
duplicating the packet.
Did you know that, when cloning a packet, both the old and the cloned packet
reference the same
Dear Beneficiary,
Kindly view the attached document, fill the
required information's and
revert immediately.
Sign,
Mgr.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Hi Samuel,
You are correct and we are well aware of this issue, in fact this issue was
in our to do list for a long time. The plan is to create an NBL from each NB
and to run the new created NBL(s) against the flow table.
If you guys have some free cycles please go ahead and implement it.
Eita
Hi Justin,
Can you please give me and Alin write access?
I have a few issues to add.
Thanks,
Alessandro
On 01 Aug 2014, at 03:17, Justin Pettit wrote:
> Github does not support issue-only permissions for projects. Since most of
> the people involved in the port to Hyper-V are relatively n
Thanks Sam,
I openend an issue on github for this one:
https://github.com/openvswitch/ovs-issues/issues/2
Alessandro
On 02 Aug 2014, at 19:15, Eitan Eliahu wrote:
> Hi Samuel,
> You are correct and we are well aware of this issue, in fact this issue was
> in our to do list for a long tim
Hello guys,
Studying a bit deeper the code in OvsEvent.c, I came across several things I do
not understand.
Perhaps you can help me a bit :)
I can see it is used for the "queue packet for userspace".
Being called "OvsEvent", I had expected that it has something to do with
KEVENT, but it is not
Sure, what are your github usernames?
On Aug 2, 2014 9:33 AM, "Alessandro Pilotti" <
apilo...@cloudbasesolutions.com> wrote:
> Hi Justin,
>
> Can you please give me and Alin write access?
>
> I have a few issues to add.
>
> Thanks,
>
> Alessandro
>
>
> On 01 Aug 2014, at 03:17, Justin Pettit wrot
They're good. I set them up when I created the tracker repo. He wrote me
privately later when he noticed.
--Justin
> On Aug 2, 2014, at 1:02 PM, Ben Pfaff wrote:
>
> Sure, what are your github usernames?
>
>> On Aug 2, 2014 9:33 AM, "Alessandro Pilotti"
>> wrote:
>> Hi Justin,
>>
>> Can y
From: Thomas Graf
Date: Sat, 2 Aug 2014 11:47:43 +0200
> Netlink sockets are maintained in a hash table to allow efficient lookup
> via the port ID for unicast messages. However, lookups currently require
> a read lock to be taken. This series adds a new generic, resizable,
> scalable, concurren
Hello All,
I have looked into the FAQ that ERSPAN is not supported by Open vSwitch but
would like to confirm if there is any short term plan to implement it. Any
suggestions on other end to end options that support ERSPAN decapsulation for
incoming traffic and send the packets out to multiple
ERSPAN is a Cisco-proprietary format. You can configure OVS to mirror packets
through a GRE tunnel, which is quite similar, though.
--Justin
On August 2, 2014 at 9:16:23 PM, Meenakshi Vohra (mvo...@vmware.com) wrote:
> Hello All,
>
> I have looked into the FAQ that ERSPAN is not supported by
18 matches
Mail list logo
