[dpdk-dev] [PATCH v2 00/15] Add support for secondary queue set in nicvf thunderx driver

Bruce Richardson bruce.richardson at intel.com
Wed Oct 12 17:59:10 CEST 2016


On Fri, Sep 30, 2016 at 02:05:39PM +0200, Kamil Rytarowski wrote:
> This series of patches adds support for secondary queue set in nicvf thunderx
> driver
> 
> There are two types of VFs:
>  - Primary VF
>  - Secondary VF
> 
> Each port consist of a primary VF and n secondary VF(s). Each VF provides 8
> Tx/Rx queues to a port. In case port is configured to use more than 8 queues,
> then it requires one (or more) secondary VF. Each secondary VF adds additional
> 8 queues to the queue set.
> 
> During PMD driver initialization, the primary VF's are enumerated by checking the
> specific flag (see READY message). They are at the beginning of  VF list (the remain
> ones are secondary VF's).
> 
> The primary VFs are used as master queue sets. Secondary VFs provide
> additional queue sets for primary ones. If a port is configured for more then
> 8 queues then it will request for additional queues from secondary VFs.
> 
> Secondary VFs cannot be shared between primary VFs.
> 
> Primary VFs are present on the tail of the 'Network devices using kernel
> driver' list, secondary VFs are on the remaining tail of the list.
> 
> The VNIC driver in the multiqueue setup works differently than other drivers
> like `ixgbe`. We need to bind separately each specific queue set device with
> the ``tools/dpdk-devbind.py`` utility.
> 
> Depending on the hardware used, the kernel driver sets a threshold ``vf_id``.
> VFs that try to attach with an id below or equal to this boundary are
> considered primary VFs. VFs that try to attach with an id above this boundary
> are considered secondary VFs.
> 
> This patchset also contains other cleanups and improvements like fixing
> erroneous checksum calculation and preparing the thunderx driver for the multi
> queue set feature support.
> 
> 
> These changes base on the following pending patches:
> 
> [dpdk-dev,1/3] net/thunderx: remove generic passx references from the driver
> http://dpdk.org/dev/patchwork/patch/14963/
> 
> [dpdk-dev,2/3] net/thunderx: introduce cqe_rx2 HW capability flag
> http://dpdk.org/dev/patchwork/patch/14964/
> 
> [dpdk-dev,3/3] net/thunderx: add 81xx SoC support
> http://dpdk.org/dev/patchwork/patch/14965/
> 
> Kamil Rytarowski (15):
>   net/thunderx: cleanup the driver before adding new features
>   net/thunderx: correct transmit checksum handling
>   net/thunderx/base: add family of functions to store qsets
>   net/thunderx/base: add secondary queue set support
>   net/thunderx: add family of functions to store DPDK qsets
>   net/thunderx: add secondary queue set in interrupt functions
>   net/thunderx: remove problematic private_data->eth_dev link
>   net/thunderx: add helper utils for secondary qset support
>   net/thunderx: add secondary qset support in dev stop/close
>   net/thunderx: add secondary qset support in device start
>   net/thunderx: add secondary qset support in device configure
>   net/thunderx: add final bits for secondary queue support
>   net/thunderx: document secondary queue set support
>   ethdev: Support VFs on the different PCI domains
>   net/thunderx: Bump driver version to 2.0
> 
Series applied to dpdk-next-net/rel_16_11, with the exception of patch 14 which
has been resubmitted as a separate ethdev patch: 
	http://dpdk.org/dev/patchwork/patch/16480/

	/Bruce


More information about the dev mailing list