mbox series

[v3,0/2] add mbuf fast free offload support

Message ID 1611335529-26503-1-git-send-email-viacheslavo@nvidia.com (mailing list archive)
Headers
Series add mbuf fast free offload support |

Message

Slava Ovsiienko Jan. 22, 2021, 5:12 p.m. UTC
  This patch adds support of the mbuf fast free offload to the
transmit datapath. This offload allows to free the mbufs on
transmit completion in the most efficient way. It requires
the all mbufs were allocated from the same pool, have
the reference counter value as 1, and have no any externally
attached buffers.

The patchset is split in two parts, the first one is overall
send loop optimization and can be ported back to stable release,
and also is the preparation step before introducing the fast
free offload. The second part provides the code for the feature.

Signed-off-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>

---
v1: http://patches.dpdk.org/patch/85482/
v2: release notes, PMD features and documentation update
v3: typo fixes, documentation update

Viacheslav Ovsiienko (2):
  net/mlx5: optimize inline mbuf freeing
  net/mlx5: add mbuf fast free offload support

 doc/guides/nics/features/mlx5.ini      |  1 +
 doc/guides/nics/mlx5.rst               | 12 ++++++
 doc/guides/rel_notes/release_21_02.rst |  2 +-
 drivers/net/mlx5/mlx5_rxtx.c           | 67 +++++++++++++++++++++++++++++-----
 drivers/net/mlx5/mlx5_rxtx.h           |  2 +
 drivers/net/mlx5/mlx5_txq.c            |  6 +++
 6 files changed, 80 insertions(+), 10 deletions(-)
  

Comments

Matan Azrad Jan. 26, 2021, 10:34 a.m. UTC | #1
From: Viacheslav Ovsiienko
> This patch adds support of the mbuf fast free offload to the transmit datapath.
> This offload allows to free the mbufs on transmit completion in the most
> efficient way. It requires the all mbufs were allocated from the same pool,
> have the reference counter value as 1, and have no any externally attached
> buffers.
> 
> The patchset is split in two parts, the first one is overall send loop optimization
> and can be ported back to stable release, and also is the preparation step
> before introducing the fast free offload. The second part provides the code for
> the feature.
> 
> Signed-off-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
> 
> ---
> v1:
> https://nam11.safelinks.protection.outlook.com/?url=http%3A%2F%2Fpatches.d
> pdk.org%2Fpatch%2F85482%2F&amp;data=04%7C01%7Cmatan%40nvidia.com
> %7Cd873758f2c354a1af0aa08d8bef8ddd9%7C43083d15727340c1b7db39efd9c
> cc17a%7C0%7C0%7C637469323804842896%7CUnknown%7CTWFpbGZsb3d8eyJ
> WIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C
> 1000&amp;sdata=BqgJn08GepCw1gpCfG7rKmqZcp1ht02HG%2Bi%2FlEXKjjc%3
> D&amp;reserved=0
> v2: release notes, PMD features and documentation update
> v3: typo fixes, documentation update
> 
> Viacheslav Ovsiienko (2):
>   net/mlx5: optimize inline mbuf freeing
>   net/mlx5: add mbuf fast free offload support
> 
>  doc/guides/nics/features/mlx5.ini      |  1 +
>  doc/guides/nics/mlx5.rst               | 12 ++++++
>  doc/guides/rel_notes/release_21_02.rst |  2 +-
>  drivers/net/mlx5/mlx5_rxtx.c           | 67 +++++++++++++++++++++++++++++---
> --
>  drivers/net/mlx5/mlx5_rxtx.h           |  2 +
>  drivers/net/mlx5/mlx5_txq.c            |  6 +++
>  6 files changed, 80 insertions(+), 10 deletions(-)
> 
> --
> 1.8.3.1
Series-acked-by: Matan Azrad <matan@nvidia.com>
  
Raslan Darawsheh Jan. 27, 2021, 9:31 a.m. UTC | #2
Hi,

> -----Original Message-----
> From: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
> Sent: Friday, January 22, 2021 7:12 PM
> To: dev@dpdk.org
> Cc: Raslan Darawsheh <rasland@nvidia.com>; Matan Azrad
> <matan@nvidia.com>; Ori Kam <orika@nvidia.com>; NBU-Contact-Thomas
> Monjalon <thomas@monjalon.net>; Alexander Kozyrev
> <akozyrev@nvidia.com>
> Subject: [PATCH v3 0/2] add mbuf fast free offload support
> 
> This patch adds support of the mbuf fast free offload to the
> transmit datapath. This offload allows to free the mbufs on
> transmit completion in the most efficient way. It requires
> the all mbufs were allocated from the same pool, have
> the reference counter value as 1, and have no any externally
> attached buffers.
> 
> The patchset is split in two parts, the first one is overall
> send loop optimization and can be ported back to stable release,
> and also is the preparation step before introducing the fast
> free offload. The second part provides the code for the feature.
> 
> Signed-off-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
> 
> ---
> v1:
> https://nam11.safelinks.protection.outlook.com/?url=http%3A%2F%2Fpatch
> es.dpdk.org%2Fpatch%2F85482%2F&amp;data=04%7C01%7Crasland%40nvid
> ia.com%7Cb616de17445740f7ea7908d8bef8ddea%7C43083d15727340c1b7db
> 39efd9ccc17a%7C0%7C0%7C637469323379495159%7CUnknown%7CTWFpbG
> Zsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6
> Mn0%3D%7C1000&amp;sdata=5GosR%2BxQfWziqF3bWmOUfa%2FuTKuf%2
> BOkjdklt%2BJrnK%2FA%3D&amp;reserved=0
> v2: release notes, PMD features and documentation update
> v3: typo fixes, documentation update
> 
> Viacheslav Ovsiienko (2):
>   net/mlx5: optimize inline mbuf freeing
>   net/mlx5: add mbuf fast free offload support
> 
>  doc/guides/nics/features/mlx5.ini      |  1 +
>  doc/guides/nics/mlx5.rst               | 12 ++++++
>  doc/guides/rel_notes/release_21_02.rst |  2 +-
>  drivers/net/mlx5/mlx5_rxtx.c           | 67
> +++++++++++++++++++++++++++++-----
>  drivers/net/mlx5/mlx5_rxtx.h           |  2 +
>  drivers/net/mlx5/mlx5_txq.c            |  6 +++
>  6 files changed, 80 insertions(+), 10 deletions(-)
> 
> --
> 1.8.3.1

Series applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh