On Mon, Dec 07, 2020 at 12:20:54PM +0200, Oz Shlomo wrote:
> On 12/1/2020 11:41 PM, Saeed Mahameed wrote:
> > On Fri, 2020-11-27 at 11:01 -0300, Marcelo Ricardo Leitner wrote:
...
> > > The same is visible on 'ovs-appctl dpctl/dump-conntrack -s' then.
> > > Summing both directions in one like this
Hi Marcelo,
On 12/1/2020 11:41 PM, Saeed Mahameed wrote:
On Fri, 2020-11-27 at 11:01 -0300, Marcelo Ricardo Leitner wrote:
On Wed, Sep 23, 2020 at 03:48:18PM -0700, sa...@kernel.org wrote:
From: Oz Shlomo
Sorry for reviving this one, but seemed better for the context.
A connection is repr
On Fri, 2020-11-27 at 11:01 -0300, Marcelo Ricardo Leitner wrote:
> On Wed, Sep 23, 2020 at 03:48:18PM -0700, sa...@kernel.org wrote:
> > From: Oz Shlomo
>
> Sorry for reviving this one, but seemed better for the context.
>
> > A connection is represented by two 5-tuple entries, one for each
> >
On Wed, Sep 23, 2020 at 03:48:18PM -0700, sa...@kernel.org wrote:
> From: Oz Shlomo
Sorry for reviving this one, but seemed better for the context.
>
> A connection is represented by two 5-tuple entries, one for each direction.
> Currently, each direction allocates its own hw counter, which is
From: Oz Shlomo
A connection is represented by two 5-tuple entries, one for each direction.
Currently, each direction allocates its own hw counter, which is
inefficient as ct aging is managed per connection.
Share the counter that was allocated for the original direction with the
reverse directi
