Hi Ashish,
You are asking relevant and interesting questions that have been bothering us
for quite some time. Please find some partial answers/thoughts below.
Best regards, Jan
From: Ashish Varma [mailto:ava...@vmware.com]
Sent: Monday, 16 April, 2018 21:28
To: ovs-discuss@openvswitch.org; Jan
cannot connect to
DPDK rings as ports.
BR, Jan
From: Kapil A [mailto:kapil20...@gmail.com]
Sent: Monday, 19 February, 2018 13:41
To: Jan Scheurich
Cc: disc...@openvswitch.org; d...@openvswitch.org
Subject: Re: [ovs-dev] DPDK: How to send packets from one OVS bridge to another
OVS bridge while
vSwitch
in the VM would then use the virtio PMD, I expect this to be supported by all
DPDK vSwitches.
BR, Jan
From: Kapil A [mailto:kapil20...@gmail.com]
Sent: Wednesday, 14 February, 2018 03:47
To: Jan Scheurich
Cc: disc...@openvswitch.org; d...@openvswitch.org
Subject: Re: [ovs-dev] DPDK: How to
Hi Kapil,
This is what patch ports are for. They are only traversed for the first packet
of a flow in the slow path. The resulting datapath flow entry collapses the
processing in both bridges into a single megaflow. So there is no performance
overhead compared to having a single bridge.
Regard
> -Original Message-
> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
> Sent: Monday, 29 January, 2018 09:35
> To: Jan Scheurich ; Venkatesan Pradeep
> ; Stokes, Ian
> ; d...@openvswitch.org
> Cc: Kevin Traynor ; Flavio Leitner ;
> Loftus, Ciara ; Kavanagh,
lt in 2 * [ 4 * 2K + 9 * 2K ] + 8K = 60K mbufs.
BR, Jan
> -Original Message-
> From: Venkatesan Pradeep
> Sent: Friday, 26 January, 2018 18:34
> To: Jan Scheurich ; Stokes, Ian
> ; ovs-discuss@openvswitch.org
> Cc: Kevin Traynor ; Flavio Leitner ;
> Ilya Maximets
> -Original Message-
> From: Stokes, Ian [mailto:ian.sto...@intel.com]
> Sent: Friday, 26 January, 2018 13:01
> To: ovs-discuss@openvswitch.org
> Cc: Kevin Traynor ; Flavio Leitner ;
> Ilya Maximets (i.maxim...@samsung.com)
> ; Loftus, Ciara ; Kavanagh,
> Mark B
>
> I think the mention of pinning was confusing me a little. Let me see if I
> fully understand your use case: You don't 'want' to pin
> anything but you are using it as a way to force the distribution of rxq from
> a single nic across to PMDs on different NUMAs. As without
> pinning all rxqs
Hi Billy,
> You are going to have to take the hit crossing the NUMA boundary at some
> point if your NIC and VM are on different NUMAs.
>
> So are you saying that it is more expensive to cross the NUMA boundary from
> the pmd to the VM that to cross it from the NIC to the
> PMD?
Indeed, that i