All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Yury Norov (NVIDIA)" <yury.norov@gmail.com>
To: Andrew Morton <akpm@linux-foundation.org>,
	Thomas Gleixner <tglx@linutronix.de>
Cc: "Yury Norov (NVIDIA)" <yury.norov@gmail.com>,
	Rasmus Villemoes <linux@rasmusvillemoes.dk>,
	linux-kernel@vger.kernel.org
Subject: [PATCH 2/3] group_cpus: don't call cpumask_weight() prematurely
Date: Tue, 18 Nov 2025 22:13:04 -0500	[thread overview]
Message-ID: <20251119031306.644129-3-yury.norov@gmail.com> (raw)
In-Reply-To: <20251119031306.644129-1-yury.norov@gmail.com>

alloc_nodes_groups() and __group_cpus_evenly() call cpumask_weight()
unconditionally in the for_each() loops. cpumask_weight() is O(N), so
the complexity of the functions become O(MAX_NUMNODES * nr_cpu_ids).

This call may be avoided if the nmsk is empty.

Signed-off-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
---
 lib/group_cpus.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)

diff --git a/lib/group_cpus.c b/lib/group_cpus.c
index 6d08ac05f371..6aae1560b796 100644
--- a/lib/group_cpus.c
+++ b/lib/group_cpus.c
@@ -142,15 +142,11 @@ static void alloc_nodes_groups(unsigned int numgrps,
 	}
 
 	for_each_node_mask(n, nodemsk) {
-		unsigned ncpus;
-
-		cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
-		ncpus = cpumask_weight(nmsk);
-
-		if (!ncpus)
+		if (!cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]))
 			continue;
-		remaining_ncpus += ncpus;
-		node_groups[n].ncpus = ncpus;
+
+		node_groups[n].ncpus = cpumask_weight(nmsk);
+		remaining_ncpus += node_groups[n].ncpus;
 	}
 
 	numgrps = min_t(unsigned, remaining_ncpus, numgrps);
@@ -294,11 +290,10 @@ static int __group_cpus_evenly(unsigned int startgrp, unsigned int numgrps,
 			continue;
 
 		/* Get the cpus on this node which are in the mask */
-		cpumask_and(nmsk, cpu_mask, node_to_cpumask[nv->id]);
-		ncpus = cpumask_weight(nmsk);
-		if (!ncpus)
+		if (!cpumask_and(nmsk, cpu_mask, node_to_cpumask[nv->id]))
 			continue;
 
+		ncpus = cpumask_weight(nmsk);
 		WARN_ON_ONCE(nv->ngroups > ncpus);
 
 		/* Account for rounding errors */
-- 
2.43.0


  parent reply	other threads:[~2025-11-19  3:13 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-19  3:13 [PATCH 0/3] improve group_cpus initialization routines Yury Norov (NVIDIA)
2025-11-19  3:13 ` [PATCH 1/3] bitmap: cpumask: introduce and_andnot search helper and iterator Yury Norov (NVIDIA)
2025-11-19  3:13 ` Yury Norov (NVIDIA) [this message]
2025-11-19  3:13 ` [PATCH 3/3] group_cpus: simplify inner loop in grp_spread_init_one() Yury Norov (NVIDIA)
2025-11-28  2:25 ` [PATCH 0/3] improve group_cpus initialization routines Yury Norov
  -- strict thread matches above, loose matches on Subject: below --
2025-09-10 21:08 Yury Norov
2025-09-10 21:08 ` [PATCH 2/3] group_cpus: don't call cpumask_weight() prematurely Yury Norov

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=20251119031306.644129-3-yury.norov@gmail.com \
    --to=yury.norov@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@rasmusvillemoes.dk \
    --cc=tglx@linutronix.de \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.