lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Message-Id: <20241205162336.1675428-1-visitorckw@gmail.com>
Date: Fri,  6 Dec 2024 00:23:36 +0800
From: Kuan-Wei Chiu <visitorckw@...il.com>
To: mingo@...hat.com,
	peterz@...radead.org,
	juri.lelli@...hat.com,
	vincent.guittot@...aro.org
Cc: dietmar.eggemann@....com,
	rostedt@...dmis.org,
	bsegall@...gle.com,
	mgorman@...e.de,
	vschneid@...hat.com,
	jserv@...s.ncku.edu.tw,
	linux-kernel@...r.kernel.org,
	Kuan-Wei Chiu <visitorckw@...il.com>
Subject: [PATCH] sched/topology: Optimize sched_numa_find_nth_cpu() by inlining bsearch()

When CONFIG_MITIGATION_RETPOLINE is enabled, indirect function calls
become costly. Replacing bsearch() with an inline version of the binary
search reduces the overhead of indirect function calls, improving
efficiency. This change also results in a reduction of the code size by
128 bytes on x86-64 systems.

Before the patch:
$ size ./kernel/sched/build_utility.o
   text    data     bss     dec     hex filename
  40113   12379    2176   54668    d58c ./kernel/sched/build_utility.o

After the patch:
$ size ./kernel/sched/build_utility.o
   text    data     bss     dec     hex filename
  39993   12371    2176   54540    d50c ./kernel/sched/build_utility.o

Signed-off-by: Kuan-Wei Chiu <visitorckw@...il.com>
---
 kernel/sched/topology.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index 9748a4c8d668..7790060d12ca 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -2173,7 +2173,8 @@ int sched_numa_find_nth_cpu(const struct cpumask *cpus, int cpu, int node)
 	if (!k.masks)
 		goto unlock;
 
-	hop_masks = bsearch(&k, k.masks, sched_domains_numa_levels, sizeof(k.masks[0]), hop_cmp);
+	hop_masks = __inline_bsearch(&k, k.masks, sched_domains_numa_levels, sizeof(k.masks[0]),
+				     hop_cmp);
 	hop = hop_masks	- k.masks;
 
 	ret = hop ?
-- 
2.34.1


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ