DPDK usage discussions
 help / color / mirror / Atom feed
From: Laurent Dumont <laurentfdumont@gmail.com>
To: "David Harton (dharton)" <dharton@cisco.com>
Cc: "users@dpdk.org" <users@dpdk.org>
Subject: Re: [dpdk-users] net/i40e VF state does not track PF state
Date: Wed, 2 Oct 2019 21:10:04 -0400	[thread overview]
Message-ID: <CAOAKi8wPFheFU1n5mVP0HHbujcbk_vh8O_WWtY5WteAcUyZDeg@mail.gmail.com> (raw)
In-Reply-To: <CY4PR11MB143024CC318FA990489DC348AA820@CY4PR11MB1430.namprd11.prod.outlook.com>

Not sure for the DPDK behavior. If a non-dpdk VF is not shutdown, the other
way around with DPDK would be surpring. I'll see if it's possible for us to
test.

Thanks

On Mon, Sep 30, 2019 at 12:32 PM David Harton (dharton) <dharton@cisco.com>
wrote:

>
>
>
>
> *From:* Laurent Dumont <laurentfdumont@gmail.com>
> *Sent:* Monday, September 30, 2019 12:25 PM
> *To:* David Harton (dharton) <dharton@cisco.com>
> *Cc:* users@dpdk.org
> *Subject:* Re: [dpdk-users] net/i40e VF state does not track PF state
>
>
>
> Hey David,
>
>
>
> Another possibility is that you are shutting down the PF through the
> compute with " ip link"?
>
> *<dharton> Yes, but shouldn’t matter the VF state should follow the PF
> state.  For example, if I shut a ixgbe PF the VFs go down.*
>
>
>
> It seems similar to a behavior we saw with sriov port where the host OS
> seems to lose a portion of control over the nic.
>
>
>
> We could shutdown the port from the compute but we still saw the traffic
> being forwarded from the VF towards the TOR.
>
>
>
> You can try to physically shutdown the port from the TOR or unplugging the
> cable.
>
>
>
> We are running 7.6 hosts with the 2.3.2 driver.
>
> *<dharton> I assume DPDK is packaged with that release...which version is
> it?*
>
>
>
> *Thanks,*
>
> *Dave*
>
>
>
> Thanks!
>
>
>
>
>
>
>
> On Mon, Sep 30, 2019, 12:14 PM David Harton (dharton) <dharton@cisco.com>
> wrote:
>
> Yes, the VFs are set to the default ‘link-state auto’:
>
> ip link show enp9s0f1
> 15: enp9s0f1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP
> mode DEFAULT group default qlen 1000
>     link/ether 90:e2:ba:ac:16:19 brd ff:ff:ff:ff:ff:ff
>     vf 0 MAC 52:54:00:63:12:97, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 1 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 2 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 3 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 4 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 5 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 6 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 7 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 8 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>     vf 9 MAC 00:00:00:00:00:00, spoof checking on, link-state auto, trust
> off, query_rss off
>
>
> > From: Laurent Dumont <laurentfdumont@gmail.com>
> > Sent: Monday, September 30, 2019 11:53 AM
> > To: David Harton (dharton) <dharton@cisco.com>
> > Cc: users@dpdk.org
> > Subject: Re: [dpdk-users] net/i40e VF state does not track PF state
> >
> > Hey David,
> >
> > From the compute, can you check if the VF "link-state" status is set to
> "auto"? That parameter is required for the physical > state of the PF to be
> reflected by the VF.
> >
> > Thanks
> >
> >
> >
> > > On Mon, Sep 30, 2019, 11:12 AM David Harton (dharton) <mailto:
> dharton@cisco.com> wrote:
> > > I'm using i40evf based interfaces using DPDK 18.11.
> > >
> > > The host is rhel 7.3 and the PF driver details are:
> > > ethtool -i enp9s0f1
> > > driver: i40e
> > > version: 1.5.10-k
> > > firmware-version: 6.01 0x800036b8 0.385.33
> > > expansion-rom-version:
> > > bus-info: 0000:09:00.1
> > > supports-statistics: yes
> > > supports-test: yes
> > > supports-eeprom-access: yes
> > > supports-register-dump: yes
> > > supports-priv-flags: yes
> > >
> > > When the PF is shut the VFs are not showing a 'down' state.
> > >
> > > Is this a known issue?  Any tips on how to debug?
> > >
> > > Thanks,
> > > Dave
>
>

  reply	other threads:[~2019-10-03  1:10 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-30 15:12 David Harton (dharton)
2019-09-30 15:52 ` Laurent Dumont
2019-09-30 16:14   ` David Harton (dharton)
2019-09-30 16:24     ` Laurent Dumont
2019-09-30 16:32       ` David Harton (dharton)
2019-10-03  1:10         ` Laurent Dumont [this message]
2019-10-03 13:24           ` David Harton (dharton)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAOAKi8wPFheFU1n5mVP0HHbujcbk_vh8O_WWtY5WteAcUyZDeg@mail.gmail.com \
    --to=laurentfdumont@gmail.com \
    --cc=dharton@cisco.com \
    --cc=users@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).