DPDK patches and discussions
 help / color / mirror / Atom feed
From: Olivier Matz <olivier.matz@6wind.com>
To: Somnath Kotur <somnath.kotur@broadcom.com>
Cc: dev <dev@dpdk.org>, Ferruh Yigit <ferruh.yigit@intel.com>
Subject: Re: [dpdk-dev] [PATCH] mbuf: fix to update documentation of QinQ stripped bit interpretation
Date: Fri, 7 Feb 2020 15:29:59 +0100
Message-ID: <20200207142959.GR22738@platinum> (raw)
In-Reply-To: <CAOBf=mtRbtkc5BJxkBRok8BYOm3ZrNCvY4ncEd5uM5W6KmQ9uA@mail.gmail.com>

Hi Somnath,

On Fri, Feb 07, 2020 at 07:13:04PM +0530, Somnath Kotur wrote:
> Olivier,
> 
> On Thu, Feb 6, 2020 at 10:55 PM Olivier Matz <olivier.matz@6wind.com> wrote:
> >
> > Hi Somnath,
> >
> > Sorry for the delay, please find some comments below.
> >
> > I suggest the following title instead:
> >
> >   mbuf: extend meaning of QinQ stripped bit
> >
> > On Mon, Jan 06, 2020 at 02:04:23PM +0530, Somnath Kotur wrote:
> > > Certain hardware may be able to strip and/or save only the outermost
> > > VLAN instead of both the VLANs in the mbuf in a QinQ scenario.
> > > To handle such cases, we could re-interpret setting of just
> > > PKT_RX_QINQ_STRIPPED to indicate that only the outermost VLAN has
> > > been stripped and saved in mbuf->vlan_tci_outer.
> >
> > To be sure we're on the same page: we are talking about case 7 of this
> > link: http://inbox.dpdk.org/dev/20191227145041.GQ22738@platinum/
> 
> I'm not sure we are on the same page then, please see my response inline below
> >
> > So, even if the inner vlan_tci is not stripped from packet data, it has
> > to be saved in m->vlan_tci, because it is implied by PKT_RX_VLAN.
> >
> > From the same link, the case where the driver only strips+saves the
> > outer vlan without saving or stripping the inner one is case 3.
> >
> While this is how it works currently, I'm wondering how will the
> application know if this was
> a double VLAN pkt, correct?

If the hardware supports it and configured for it, the flag PKT_RX_QINQ
is set when there are 2 vlans.

FYI, the m->packet_type field can also be useful. It describes what is
really present in the packet data, as described in rte_mbuf_core.h:

	/*
	 * The packet type, which is the combination of outer/inner L2, L3, L4
	 * and tunnel types. The packet_type is about data really present in the
	 * mbuf. Example: if vlan stripping is enabled, a received vlan packet
	 * would have RTE_PTYPE_L2_ETHER and not RTE_PTYPE_L2_VLAN because the
	 * vlan is stripped from the data.
	 */

> Also when i look at options 5 and 7 I don't really see the difference
> in semantics between them ?
> Both seem to store the outer-vlan and inner-vlan in m->vlan_tci_outer
> and m->vlan_tci_inner respectively
> 
> 7/ PKT_RX_VLAN | PKT_RX_QINQ | PKT_RX_QINQ_STRIPPED
>    outer-vlan is removed from packet data
>    m->vlan_tci_outer=outer-vlan
>    m->vlan_tci=inner-vlan

The difference is about what was stripped or not from mbuf data.

> I was hoping that with the new interpretation of PKT_RX_QINQ_STRIPPED,
> it only meant that outer-vlan is removed from packet data
> and m->vlan_tci_outer = outer_vlan, while PKT_RX_QINQ implies it is a
> double-vlan pkt and PKT_RX_VLAN implies that pkt has VLAN
> associated with it?   Not m->vlan_tci = inner-vlan

The meaning of each flag should be as simple as possible, I think we can
summarize them like this:

- PKT_RX_VLAN: the vlan is saved in vlan tci.
- PKT_RX_VLAN_STRIPPED: the vlan hdr is removed from packet data.
- PKT_RX_QINQ: the outer vlan is saved in vlan tci.
- PKT_RX_QINQ_STRIPPED: the inner vlan is stripped from packet data.
- When PKT_RX_QINQ is set, PKT_RX_VLAN* refer to the inner vlan
  of initial packet, else it refers to the first vlan of the packet.

There is a link between vlan flag and vlan_tci field, and qinq flag and
vlan_tci_outer field.

I'm still not sure to understand what you expect. Can you give an
example with flags (which are set), and the expected content of m->vlan_tci
and m->vlan_tci_outer?

By the way, the case 5/ is not very well described too, maybe we should
add something about it.

Thanks,
Olivier


> 
> Thanks
> Som
> 
> 
> > > Only When both PKT_RX_QINQ_STRIPPED and PKT_RX_VLAN_STRIPPED are set, the 2
> > > VLANs have been stripped by the hardware and their TCI are saved in
> > > mbuf->vlan_tci (inner) and mbuf->vlan_tci_outer (outer).
> > >
> > > Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
> > > ---
> > >  lib/librte_mbuf/rte_mbuf_core.h | 15 +++++++++++----
> > >  1 file changed, 11 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/lib/librte_mbuf/rte_mbuf_core.h b/lib/librte_mbuf/rte_mbuf_core.h
> > > index 9a8557d..db1070b 100644
> > > --- a/lib/librte_mbuf/rte_mbuf_core.h
> > > +++ b/lib/librte_mbuf/rte_mbuf_core.h
> > > @@ -124,12 +124,19 @@
> > >  #define PKT_RX_FDIR_FLX      (1ULL << 14)
> > >
> > >  /**
> > > - * The 2 vlans have been stripped by the hardware and their tci are
> > > - * saved in mbuf->vlan_tci (inner) and mbuf->vlan_tci_outer (outer).
> > > + * The outer vlan has been stripped by the hardware and their tci are
> > > + * saved in mbuf->vlan_tci_outer (outer).
> >
> > their tci are -> its tci is
> >
> > >   * This can only happen if vlan stripping is enabled in the RX
> > >   * configuration of the PMD.
> > > - * When PKT_RX_QINQ_STRIPPED is set, the flags (PKT_RX_VLAN |
> > > - * PKT_RX_VLAN_STRIPPED | PKT_RX_QINQ) must also be set.
> > > + * When PKT_RX_QINQ_STRIPPED is set, the flags (PKT_RX_VLAN |  PKT_RX_QINQ)
> > > + * must also be set.
> >
> > ok
> >
> > > + * When both PKT_RX_QINQ_STRIPPED and PKT_RX_VLAN_STRIPPED are set, the 2 vlans
> > > + * have been stripped by the hardware and their tci are saved in
> > > + * mbuf->vlan_tci (inner) and mbuf->vlan_tci_outer (outer).
> >
> > This is correct, but I'd use a bullet list to add another sentence:
> >
> >  * - If both PKT_RX_QINQ_STRIPPED and PKT_RX_VLAN_STRIPPED are set, the 2 vlans
> >  *   have been stripped by the hardware and their tci are saved in
> >  *   mbuf->vlan_tci (inner) and mbuf->vlan_tci_outer (outer).
> >  * - If PKT_RX_QINQ_STRIPPED is set and PKT_RX_VLAN_STRIPPED is unset, only the
> >  *   outer vlan is removed from packet data, but both tci are saved in
> >  *   mbuf->vlan_tci (inner) and mbuf->vlan_tci_outer (outer).
> >
> > > + * This can only happen if vlan stripping is enabled in the RX configuration
> > > + * of the PMD.
> >
> > The same exact sentence is above, this one can be removed.
> >
> > > + * When PKT_RX_QINQ_STRIPPED and PKT_RX_VLAN_STRIPPED are set,
> > > + * (PKT_RX_VLAN | PKT_RX_QINQ) must also be set.
> >
> > This can be removed too as it is redundant with above sentence:
> >
> >  * When PKT_RX_QINQ_STRIPPED is set, the flags (PKT_RX_VLAN |  PKT_RX_QINQ)
> >  * must also be set.
> >
> >
> > Thanks,
> > Olivier

  reply	other threads:[~2020-02-07 14:30 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-06  8:34 Somnath Kotur
2020-02-06 16:27 ` Somnath Kotur
2020-02-06 17:25 ` Olivier Matz
2020-02-07 13:43   ` Somnath Kotur
2020-02-07 14:29     ` Olivier Matz [this message]
2020-02-26  0:55       ` Stephen Hemminger
2020-04-24 18:24         ` Thomas Monjalon
2020-05-24 15:34           ` Thomas Monjalon
2020-06-11  6:11             ` Thomas Monjalon
2020-04-25 13:08 ` Andrew Rybchenko
2020-10-06  7:22 ` [dpdk-dev] [PATCH v2] mbuf: extend meaning of QinQ stripped bit Olivier Matz
2020-10-08 13:13   ` David Marchand
2020-10-12  8:20   ` Andrew Rybchenko
2020-10-15 20:58   ` David Marchand

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=20200207142959.GR22738@platinum \
    --to=olivier.matz@6wind.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=somnath.kotur@broadcom.com \
    /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

DPDK patches and discussions

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/dev/0 dev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dev dev/ https://inbox.dpdk.org/dev \
		dev@dpdk.org
	public-inbox-index dev

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.dev


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git