Linux Power Management development
 help / color / mirror / Atom feed
From: Ali Ahmet Memis <ali@iusegentoo.com>
To: Shuah Khan <skhan@linuxfoundation.org>,
	Shuah Khan <shuah@kernel.org>, Thomas Renninger <trenn@suse.com>,
	"John B . Wyatt IV" <jwyatt@redhat.com>,
	John Kacur <jkacur@redhat.com>
Cc: linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH v3 3/3] cpupower: do not count incomplete topology entries as physical cores
Date: Thu,  6 Aug 2026 17:51:40 +0000	[thread overview]
Message-ID: <20260806175140.270935-4-ali@iusegentoo.com> (raw)
In-Reply-To: <20260803175215.117518-1-ali@iusegentoo.com>

The physical core count is derived by sorting core_info by core_cpu_list
and counting how many distinct lists there are. The loop seeds the count
with entry 0 unconditionally:

	last_cpu_list = cpu_top->core_info[0].core_cpu_list;
	cpu_top->cores = 1;

A CPU whose topology could not be read is still a member of the array,
carrying pkg and core of -1 and a core_cpu_list of "-1". That sorts ahead
of any real cpu list, so after the qsort it is entry 0 and it seeds the
count as if it were a core. The pkg check inside the loop only guards the
entries that follow, never the one the count started from.

An offline CPU is enough to reach this. The topology attribute group is
added and removed by a CPU hotplug callback in drivers/base/topology.c, so
physical_package_id and core_id are absent while a CPU is offline and both
reads fail.

Skip entries without complete topology data and count from zero, so only
CPUs with a package and a core contribute.

Fixes: f89cb9cba7a2 ("cpupower: Implement CPU physical core querying")
Signed-off-by: Ali Ahmet Memis <ali@iusegentoo.com>
---
 tools/power/cpupower/lib/cpupower.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/tools/power/cpupower/lib/cpupower.c b/tools/power/cpupower/lib/cpupower.c
index 3d7e22ed1fd7..7784eb499435 100644
--- a/tools/power/cpupower/lib/cpupower.c
+++ b/tools/power/cpupower/lib/cpupower.c
@@ -212,11 +212,16 @@ int get_cpu_topology(struct cpupower_topology *cpu_top)
 	qsort(cpu_top->core_info, cpus, sizeof(struct cpuid_core_info),
 	      __compare_core_cpu_list);
 
-	last_cpu_list = cpu_top->core_info[0].core_cpu_list;
-	cpu_top->cores = 1;
-	for (cpu = 1; cpu < cpus; cpu++) {
-		if (strcmp(cpu_top->core_info[cpu].core_cpu_list, last_cpu_list) != 0 &&
-		    cpu_top->core_info[cpu].pkg != -1) {
+	last_cpu_list = NULL;
+	cpu_top->cores = 0;
+	for (cpu = 0; cpu < cpus; cpu++) {
+		if (cpu_top->core_info[cpu].pkg == -1 ||
+		    cpu_top->core_info[cpu].core == -1)
+			continue;
+
+		if (!last_cpu_list ||
+		    strcmp(cpu_top->core_info[cpu].core_cpu_list,
+			   last_cpu_list) != 0) {
 			last_cpu_list = cpu_top->core_info[cpu].core_cpu_list;
 			cpu_top->cores++;
 		}
-- 
2.55.0


      parent reply	other threads:[~2026-08-06 17:53 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-03 17:52 [PATCH v2 0/2] cpupower: fix topology array handling Ali Ahmet Memis
2026-08-03 17:52 ` [PATCH v2 1/2] cpupower: zero the topology array to avoid uninitialized reads Ali Ahmet Memis
2026-08-03 17:52 ` [PATCH v2 2/2] cpupower: do not count incomplete topology entries as physical cores Ali Ahmet Memis
2026-08-04 20:45 ` [PATCH v2 0/2] cpupower: fix topology array handling Shuah Khan
2026-08-05 11:43   ` Ali Ahmet Memis
2026-08-05 12:10     ` Ali Ahmet Memis
2026-08-06 15:50     ` Shuah Khan
2026-08-06 17:51 ` [PATCH v3 0/3] " Ali Ahmet Memis
2026-08-06 17:51 ` [PATCH v3 1/3] cpupower: zero the topology array to avoid uninitialized reads Ali Ahmet Memis
2026-08-06 17:51 ` [PATCH v3 2/3] cpupower: let the core == -1 check handle failed topology reads Ali Ahmet Memis
2026-08-06 17:51 ` Ali Ahmet Memis [this message]

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=20260806175140.270935-4-ali@iusegentoo.com \
    --to=ali@iusegentoo.com \
    --cc=jkacur@redhat.com \
    --cc=jwyatt@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=shuah@kernel.org \
    --cc=skhan@linuxfoundation.org \
    --cc=trenn@suse.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox