|
|
Log in / Subscribe / Register

sched, net: NUMA-aware CPU spreading interface

From:  Valentin Schneider <vschneid-AT-redhat.com>
To:  netdev-AT-vger.kernel.org, linux-rdma-AT-vger.kernel.org, linux-kernel-AT-vger.kernel.org
Subject:  [PATCH v5 0/3] sched, net: NUMA-aware CPU spreading interface
Date:  Fri, 21 Oct 2022 13:19:24 +0100
Message-ID:  <20221021121927.2893692-1-vschneid@redhat.com>
Cc:  Saeed Mahameed <saeedm-AT-nvidia.com>, Leon Romanovsky <leon-AT-kernel.org>, "David S. Miller" <davem-AT-davemloft.net>, Eric Dumazet <edumazet-AT-google.com>, Jakub Kicinski <kuba-AT-kernel.org>, Paolo Abeni <pabeni-AT-redhat.com>, Yury Norov <yury.norov-AT-gmail.com>, Andy Shevchenko <andriy.shevchenko-AT-linux.intel.com>, Rasmus Villemoes <linux-AT-rasmusvillemoes.dk>, Ingo Molnar <mingo-AT-kernel.org>, Peter Zijlstra <peterz-AT-infradead.org>, Vincent Guittot <vincent.guittot-AT-linaro.org>, Dietmar Eggemann <dietmar.eggemann-AT-arm.com>, Steven Rostedt <rostedt-AT-goodmis.org>, Mel Gorman <mgorman-AT-suse.de>, Greg Kroah-Hartman <gregkh-AT-linuxfoundation.org>, Heiko Carstens <hca-AT-linux.ibm.com>, Tony Luck <tony.luck-AT-intel.com>, Jonathan Cameron <Jonathan.Cameron-AT-huawei.com>, Gal Pressman <gal-AT-nvidia.com>, Tariq Toukan <tariqt-AT-nvidia.com>, Jesse Brandeburg <jesse.brandeburg-AT-intel.com>
Archive-link:  Article

Hi folks,

Tariq pointed out in [1] that drivers allocating IRQ vectors would benefit
from having smarter NUMA-awareness (cpumask_local_spread() doesn't quite cut
it).

The proposed interface involved an array of CPUs and a temporary cpumask, and
being my difficult self what I'm proposing here is an interface that doesn't
require any temporary storage other than some stack variables (at the cost of
one wild macro).

[1]: https://lore.kernel.org/all/20220728191203.4055-1-tariqt@...

Revisions
=========

v4 -> v5
++++++++
o Rebased onto 6.1-rc1
o Ditched the CPU iterator, moved to a cpumask iterator (Yury)

v3 -> v4
++++++++

o Rebased on top of Yury's bitmap-for-next
o Added Tariq's mlx5e patch
o Made sched_numa_hop_mask() return cpu_online_mask for the NUMA_NO_NODE &&
  hops=0 case

v2 -> v3
++++++++

o Added for_each_cpu_and() and for_each_cpu_andnot() tests (Yury)
o New patches to fix issues raised by running the above

o New patch to use for_each_cpu_andnot() in sched/core.c (Yury)

v1 -> v2
++++++++

o Split _find_next_bit() @invert into @invert1 and @invert2 (Yury)
o Rebase onto v6.0-rc1

Cheers,
Valentin

Tariq Toukan (1):
  net/mlx5e: Improve remote NUMA preferences used for the IRQ affinity
    hints

Valentin Schneider (2):
  sched/topology: Introduce sched_numa_hop_mask()
  sched/topology: Introduce for_each_numa_hop_mask()

 drivers/net/ethernet/mellanox/mlx5/core/eq.c | 18 +++++++++--
 include/linux/topology.h                     | 32 ++++++++++++++++++++
 kernel/sched/topology.c                      | 31 +++++++++++++++++++
 3 files changed, 79 insertions(+), 2 deletions(-)

--
2.31.1



Copyright © 2022, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds