[dpdk-dev] doc: more clarification for i40e 16 bytes desc

  The description of i40e 16 bytes descriptor is not clear
enough. Add more details.

Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
---
 doc/guides/nics/i40e.rst | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
  

On Wed, Oct 25, 2017 at 10:54:53AM +0800, Wenzhuo Lu wrote:
> The description of i40e 16 bytes descriptor is not clear
> enough. Add more details.
> 
> Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
> ---

Good idea to document this better.

>  doc/guides/nics/i40e.rst | 11 +++++++++--
>  1 file changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/doc/guides/nics/i40e.rst b/doc/guides/nics/i40e.rst
> index a0946e6..e0237a8 100644
> --- a/doc/guides/nics/i40e.rst
> +++ b/doc/guides/nics/i40e.rst
> @@ -398,13 +398,20 @@ used to classify MPLS packet by using a command in testpmd like:
>     testpmd> ethertype_filter 0 add mac_ignr 00:00:00:00:00:00 ethertype \
>              0x8847 fwd queue <M>
>  
> -16 Byte Descriptor cannot be used on DPDK VF
> -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> +16 Bytes Descriptor setting on DPDK VF
> +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>  
> +Until 2.1.26, the Linux i40e driver doesn't support 16 bytes RX descriptor.

I assume that 2.1.26 is referring to the i40e driver version, not the
linux kernel version. This should be made clearer, as it also could be
read as a very old kernel version.

>  If the Linux i40e kernel driver is used as host driver, while DPDK i40e PMD
>  is used as the VF driver, DPDK cannot choose 16 byte receive descriptor. That
>  is to say, user should keep ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n`` in
>  config file.

Can you explain - either on-list here, or actually in the docs - why
there is this restriction? Is it HW or SW? A reference to the section in
the datasheet for it would also be good. For example Table 8-10 in the
datasheet lists no restriction on the Dsize bit in the queue context.

> +In the future, if the Linux i40e driver supports 16 bytes RX descriptor, user
> +should make sure the DPDK VF uses the same RX descriptor mode, 16 bytes or 32
> +bytes, as the PF driver.
> +
> +The same rule for DPDK PF + DPDK VF. The PF and VF should use the same RX
> +descriptor mode. Or the VF RX will not work.
>  
>  Receive packets with Ethertype 0x88A8
>  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> -- 
> 1.9.3
> 


Regards,
/Bruce
  

Hi Bruce,

> -----Original Message-----
> From: Richardson, Bruce
> Sent: Wednesday, October 25, 2017 4:37 PM
> To: Lu, Wenzhuo <wenzhuo.lu@intel.com>
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] doc: more clarification for i40e 16 bytes desc
> 
> On Wed, Oct 25, 2017 at 10:54:53AM +0800, Wenzhuo Lu wrote:
> > The description of i40e 16 bytes descriptor is not clear enough. Add
> > more details.
> >
> > Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
> > ---
> 
> Good idea to document this better.
> 
> >  doc/guides/nics/i40e.rst | 11 +++++++++--
> >  1 file changed, 9 insertions(+), 2 deletions(-)
> >
> > diff --git a/doc/guides/nics/i40e.rst b/doc/guides/nics/i40e.rst index
> > a0946e6..e0237a8 100644
> > --- a/doc/guides/nics/i40e.rst
> > +++ b/doc/guides/nics/i40e.rst
> > @@ -398,13 +398,20 @@ used to classify MPLS packet by using a command
> in testpmd like:
> >     testpmd> ethertype_filter 0 add mac_ignr 00:00:00:00:00:00 ethertype \
> >              0x8847 fwd queue <M>
> >
> > -16 Byte Descriptor cannot be used on DPDK VF
> > -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > +16 Bytes Descriptor setting on DPDK VF
> > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> > +Until 2.1.26, the Linux i40e driver doesn't support 16 bytes RX descriptor.
> 
> I assume that 2.1.26 is referring to the i40e driver version, not the linux
> kernel version. This should be made clearer, as it also could be read as a very
> old kernel version.
OK, I'll change it to i40e 2.1.26.

> 
> >  If the Linux i40e kernel driver is used as host driver, while DPDK
> > i40e PMD  is used as the VF driver, DPDK cannot choose 16 byte receive
> > descriptor. That  is to say, user should keep
> > ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n`` in  config file.
> 
> Can you explain - either on-list here, or actually in the docs - why there is
> this restriction? Is it HW or SW? A reference to the section in the datasheet
> for it would also be good. For example Table 8-10 in the datasheet lists no
> restriction on the Dsize bit in the queue context.
It's not a HW limitation. To my opinion, it's SW design and implementation limitation.
I'll add more details.

> 
> > +In the future, if the Linux i40e driver supports 16 bytes RX
> > +descriptor, user should make sure the DPDK VF uses the same RX
> > +descriptor mode, 16 bytes or 32 bytes, as the PF driver.
> > +
> > +The same rule for DPDK PF + DPDK VF. The PF and VF should use the
> > +same RX descriptor mode. Or the VF RX will not work.
> >
> >  Receive packets with Ethertype 0x88A8
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > --
> > 1.9.3
> >
> 
> 
> Regards,
> /Bruce