mbox series

[net-next,v9,0/6] Improve neigh_flush_dev performance

Message ID 20241107160444.2913124-1-gnaaman@drivenets.com (mailing list archive)
Headers show
Series Improve neigh_flush_dev performance | expand

Message

Gilad Naaman Nov. 7, 2024, 4:04 p.m. UTC
This patchsets improves the performance of neigh_flush_dev.

Currently, the only way to implement it requires traversing
all neighbours known to the kernel, across all network-namespaces.

This means that some flows are slowed down as a function of neigh-scale,
even if the specific link they're handling has little to no neighbours.

In order to solve this, this patchset adds a netdev->neighbours list,
as well as making the original linked-list doubly-, so that it is
possible to unlink neighbours without traversing the hash-bucket to
obtain the previous neighbour.

The original use-case we encountered was mass-deletion of links (12K
VLANs) while there are 50K ARPs and 50K NDPs in the system; though the
slowdowns would also appear when the links are set down.

Changes in v9:
 - Use RCU iteration in ___neigh_lookup_noref and neigh_dump_table

I kept the reviewed-by tags, as instructed, but note that there are
minor changes in commits 2/6 and 4/6 (just adding the `_rcu`).

Gilad Naaman (6):
  neighbour: Add hlist_node to struct neighbour
  neighbour: Define neigh_for_each_in_bucket
  neighbour: Convert seq_file functions to use hlist
  neighbour: Convert iteration to use hlist+macro
  neighbour: Remove bare neighbour::next pointer
  neighbour: Create netdev->neighbour association

 .../networking/net_cachelines/net_device.rst  |   1 +
 include/linux/netdevice.h                     |   7 +
 include/net/neighbour.h                       |  26 +-
 include/net/neighbour_tables.h                |  12 +
 net/core/neighbour.c                          | 325 ++++++++----------
 net/ipv4/arp.c                                |   2 +-
 6 files changed, 170 insertions(+), 203 deletions(-)
 create mode 100644 include/net/neighbour_tables.h

Comments

patchwork-bot+netdevbpf@kernel.org Nov. 9, 2024, 9:50 p.m. UTC | #1
Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Thu,  7 Nov 2024 16:04:37 +0000 you wrote:
> This patchsets improves the performance of neigh_flush_dev.
> 
> Currently, the only way to implement it requires traversing
> all neighbours known to the kernel, across all network-namespaces.
> 
> This means that some flows are slowed down as a function of neigh-scale,
> even if the specific link they're handling has little to no neighbours.
> 
> [...]

Here is the summary with links:
  - [net-next,v9,1/6] neighbour: Add hlist_node to struct neighbour
    https://git.kernel.org/netdev/net-next/c/41b3caa7c076
  - [net-next,v9,2/6] neighbour: Define neigh_for_each_in_bucket
    https://git.kernel.org/netdev/net-next/c/d7ddee1a522d
  - [net-next,v9,3/6] neighbour: Convert seq_file functions to use hlist
    https://git.kernel.org/netdev/net-next/c/00df5e1a3fdf
  - [net-next,v9,4/6] neighbour: Convert iteration to use hlist+macro
    https://git.kernel.org/netdev/net-next/c/0e3bcb0f78a0
  - [net-next,v9,5/6] neighbour: Remove bare neighbour::next pointer
    https://git.kernel.org/netdev/net-next/c/a01a67ab2fff
  - [net-next,v9,6/6] neighbour: Create netdev->neighbour association
    https://git.kernel.org/netdev/net-next/c/f7f52738637f

You are awesome, thank you!