DPDK patches and discussions
 help / color / mirror / Atom feed
From: Ferruh Yigit <ferruh.yigit@intel.com>
To: Andrew Rybchenko <arybchenko@solarflare.com>, dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH v2 00/13] Improve Solarflare PMD performance
Date: Mon, 20 Mar 2017 15:37:30 +0000	[thread overview]
Message-ID: <9f2cd334-cde4-ca24-0db8-90cc27062f57@intel.com> (raw)
In-Reply-To: <1490004919-2177-1-git-send-email-arybchenko@solarflare.com>

On 3/20/2017 10:15 AM, Andrew Rybchenko wrote:
> Implement EF10 (SFN7xxx and SFN8xxx) native datapaths which may be
> chosen per device using PCI whitelist device arguments.
> 
> libefx-based datapath implementation is bound to API and structure
> imposed by the libefx. It has many indirect function calls to
> provide HW abstraction (bad for CPU pipeline) and uses many data
> structures: driver Rx/Tx queue, driver event queue, libefx Rx/Tx
> queue, libefx event queue, libefx NIC (bad for cache).
> 
> Native datapath implementation is fully separated from control
> path to be able to use alternative control path if required
> (e.g. kernel-aware).
> 
> Native datapaths show better performance than libefx-based.
> 
> v2:
>  - fix spelling, reword commit messages as requested
>  - exclude packed stream support since it shows worse performance yet
>    because of indirect mbufs usage
>  - use uint16_t for port_id to avoid changes when corresponding
>    mbuf patches are applied
>  - add header with functions shared by EF10 Rx and Tx
>  - clear event queue entries by cache-lines
>  - remove unnecessary checks in refill code
>  - add missing BSD LICENSE line to new files
>  - avoid usage of function pointers in state structures, make EF10
>    native datapath multi-process support friendly to avoid code
>    shuffling in the future
>  - remove unnecessary memory barriers, add corresponding comments
>  - do not use libefx macros for Rx/Tx queue limit, define own which
>    take event queue clear by cache-line into account
> 
> Andrew Rybchenko (13):
>   net/sfc: use different callbacks for event queues
>   net/sfc: emphasis that RSS hash flag is an Rx queue flag
>   net/sfc: do not use Rx queue control state on datapath
>   net/sfc: factor out libefx-based Rx datapath
>   net/sfc: make Rx scatter a datapath-dependent feature
>   net/sfc: remove few conditions in Rx queue refill
>   net/sfc: implement EF10 native Rx datapath
>   net/sfc: factor out libefx-based Tx datapath
>   net/sfc: make VLAN insertion a datapath-dependent feature
>   net/sfc: make TSO a datapath-dependent feature
>   net/sfc: implement EF10 native Tx datapath
>   net/sfc: make multi-segment support a Tx datapath feature
>   net/sfc: implement simple EF10 native Tx datapath

Series applied to dpdk-next-net/master, thanks.

      parent reply	other threads:[~2017-03-20 15:37 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-02  7:07 [dpdk-dev] [PATCH " Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 01/13] net/sfc: callbacks should depend on EvQ usage Andrew Rybchenko
2017-03-04 21:04   ` Ferruh Yigit
2017-03-02  7:07 ` [dpdk-dev] [PATCH 02/13] net/sfc: emphasis that RSS hash flag is an Rx queue flag Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 03/13] net/sfc: do not use Rx queue control state on datapath Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 04/13] net/sfc: factor out libefx-based Rx datapath Andrew Rybchenko
2017-03-04 21:05   ` Ferruh Yigit
2017-03-13 13:12     ` Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 05/13] net/sfc: Rx scatter is a datapath-dependent feature Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 06/13] net/sfc: implement EF10 native Rx datapath Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 07/13] net/sfc: factory out libefx-based Tx datapath Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 08/13] net/sfc: VLAN insertion is a datapath dependent feature Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 09/13] net/sfc: TSO " Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 10/13] net/sfc: implement EF10 native Tx datapath Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 11/13] net/sfc: multi-segment support as is Tx datapath features Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 12/13] net/sfc: implement simple EF10 native Tx datapath Andrew Rybchenko
2017-03-02  7:07 ` [dpdk-dev] [PATCH 13/13] net/sfc: support Rx packed stream EF10-specific datapath Andrew Rybchenko
2017-03-04 21:07 ` [dpdk-dev] [PATCH 00/13] Improve Solarflare PMD performance Ferruh Yigit
2017-03-20 10:15 ` [dpdk-dev] [PATCH v2 " Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 01/13] net/sfc: use different callbacks for event queues Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 02/13] net/sfc: emphasis that RSS hash flag is an Rx queue flag Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 03/13] net/sfc: do not use Rx queue control state on datapath Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 04/13] net/sfc: factor out libefx-based Rx datapath Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 05/13] net/sfc: make Rx scatter a datapath-dependent feature Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 06/13] net/sfc: remove few conditions in Rx queue refill Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 07/13] net/sfc: implement EF10 native Rx datapath Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 08/13] net/sfc: factor out libefx-based Tx datapath Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 09/13] net/sfc: make VLAN insertion a datapath-dependent feature Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 10/13] net/sfc: make TSO " Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 11/13] net/sfc: implement EF10 native Tx datapath Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 12/13] net/sfc: make multi-segment support a Tx datapath feature Andrew Rybchenko
2017-03-20 10:15   ` [dpdk-dev] [PATCH v2 13/13] net/sfc: implement simple EF10 native Tx datapath Andrew Rybchenko
2017-03-20 15:37   ` Ferruh Yigit [this message]

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=9f2cd334-cde4-ca24-0db8-90cc27062f57@intel.com \
    --to=ferruh.yigit@intel.com \
    --cc=arybchenko@solarflare.com \
    --cc=dev@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).