> Hi Ian,
> 
> > -----Original Message-----
> > From: Stokes, Ian <[email protected]>
> > Sent: Wednesday, January 16, 2019 12:37 AM
> > To: Kevin Traynor <[email protected]>; Ophir Munk
> > <[email protected]>; [email protected]
> > Cc: Olga Shern <[email protected]>; Ilya Maximets
> > <[email protected]>
> > Subject: RE: [PATCH v5] netdev-dpdk: support port representors
> >
> > > On 01/15/2019 09:47 AM, Ophir Munk wrote:
> > > > Dpdk port representors were introduced in dpdk versions 18.xx.
> > > > Prior to port representors there was a one-to-one relationship
> > > > between an rte device (e.g. PCI bus) and an eth device (referenced
> > > > as dpdk port id in OVS). With port representors the relationship
> > > > becomes one-to-many rte device to eth devices.
> > > > For example in [3] there are two devices (representors) using the
> > > > same PCI physical address 0000:08:00.0:
> > > > "0000:08:00.0,representor=[3]" and "0000:08:00.0,representor=[5]".
> > > > This commit handles the new one-to-many relationship. For example,
> > > > when one of the device port representors in [3] is closed - the
> > > > PCI bus cannot be detached until the other device port representor
> > > > is closed as well. OVS remains backward compatible by supporting
> > > > dpdk legacy PCI ports which do not include port representors.
> > > > Dpdk port representors related commits are listed in [1]. Dpdk
> > > > port representors documentation appears in [2]. A sample
> > > > configuration which uses two representors ports (the output of "ovs-
> vsctl show"
> > > > command) is shown in [3].
> > > >
> > >
> > > Hi Ophir, I had a scan through and there isn't any
> > > documentation/examples for this outside the commit message. I think
> > > a user would need something basic, or at least reference to know
> > > that this
> > exists and how to use it.
> >
> > +1, although I can confirm the backwards compatibility with the legacy
> > +pci
> > port methodology I'm seeing issues around the use of representors
> > which I'm not entirely sure are user configuration related or specific
> > to the patch implementation, will need more time to investigate and
> confirm.
> >
> > Ian
> 
> Can you please inform what exact issues you are seeing with representors?

Hi Ophir, in my case trying to add an i40e port representor was failing with 
invalid port.

I want to confirm this is not an issue in DPDK, or a configuration issue on my 
side, or that it's just not supported (although I would have thought it was).

> What is the configuration?
PF I40e address with 05:00.0 is bound to igb_uio.
VF then created for PF with 'echo 1 > 
/sys/bus/pci/devices/0000\:05\:00.0/max_vfs'.
VF is created with address 05:02.0
VF then bound to igb_uio driver.
OVS DPDK then stated (no new commands are used here when launching).
Add bridge br0
Attempt to add VF with port representor, use the PF address and representor of 
VF

ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk 
options:dpdk-devargs=0000:05:00.0,representor=[0]

Possibly it's above where I'm going wrong. Either on the address passed, or the 
representor value used (though I've tried a range of represntor values besides 
0, no luck, how do you typically identify the value?)

> What is the NIC under test?
The NIC is Intel X710 (Ethernet Controller X710 for 10GbE SFP+ 1572).
The VF created is recognized as Ethernet Virtual Function 700 Series 154c

> What is the expected result?
I was attempting to see if the port representor could be used to add a VF for 
the i40e PF.
There is no issue if I add the VF using the legacy method ' ovs-vsctl add-port 
br0 dpdk0 -- set Interface dpdk0 type=dpdk options:dpdk-devargs=0000:05:02.0'. 

> What is the actual result?

2019-01-16T10:18:38Z|00062|dpdk|INFO|EAL: PCI device 0000:05:00.0 on NUMA 
socket 0
2019-01-16T10:18:38Z|00063|dpdk|ERR|EAL: Failed to attach device on primary 
process
2019-01-16T10:18:38Z|00064|netdev_dpdk|WARN|Error attaching device 
'0000:05:00.0,representor=[0]' to DPDK
2019-01-16T10:18:38Z|00065|netdev|WARN|dpdk0: could not set configuration 
(Invalid argument)
2019-01-16T10:18:38Z|00066|dpdk|ERR|Invalid port_id=32
> 
> Without those details I do not know how the "issues"  can be addressed.

As I said this could be a user configuration issue, I spoke with Awal 
co-authored the i40e port representor code which is how I arrived at the 
commands above but these could be wrong.

Ian
> 
> Regards,
> Ophir
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to