All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] hwmon: (coretemp) Fix handling of isolated SMT siblings
@ 2026-10-16  8:36 ` Ximing Zhang
  0 siblings, 0 replies; 3+ messages in thread
From: Ximing Zhang via B4 Relay @ 2026-10-16  8:36 UTC (permalink / raw)
  To: Guenter Roeck, Marcelo Tosatti
  Cc: linux-hwmon, linux-kernel, stable, Ximing Zhang

From: Ximing Zhang <xzhangjr@gmail.com>

Commit 0f8b916bc5b5 ("hwmon: (coretemp) avoid RDMSR interrupts to
isolated CPUs") prevents create_core_data() from creating temperature
data for an isolated CPU. coretemp_cpu_online(), however, still adds
that CPU to pdata->cpumask.

When an isolated SMT thread is processed before a housekeeping sibling,
the mask makes coretemp_cpu_online() believe that the core is already
represented. The housekeeping sibling then gets no temperature
attributes. Package attributes are similarly not retried when the first
CPU processed in a die is isolated.

The mask is also used to select a replacement CPU when a sensor target
goes offline. Including isolated CPUs permits the driver to retarget a
sensor to one, causing subsequent sysfs reads to send an RDMSR IPI to an
isolated CPU.

Keep isolated CPUs out of pdata->cpumask by moving the housekeeping check
to coretemp_cpu_online(). Sensor creation is then deferred to an eligible
CPU, and hotplug replacement targets remain on housekeeping CPUs.

Tested on an Intel Core i7-10510U running 6.18.50-3-cachyos-lts, with
nohz_full=1 on CPU 1 and its housekeeping SMT sibling on CPU 5. The stock
driver omitted Core 1 after module load and retained its readable sensor
after CPU 5 was offlined. The patched driver exposed Core 1, removed it
when CPU 5 went offline, and restored it when CPU 5 returned.

Fixes: 0f8b916bc5b5 ("hwmon: (coretemp) avoid RDMSR interrupts to isolated CPUs")
Cc: stable@vger.kernel.org
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Ximing Zhang <xzhangjr@gmail.com>
---
The patch was tested on an Intel Core i7-10510U running
6.18.50-3-cachyos-lts with CPU 1 isolated using nohz_full=1 and
housekeeping SMT sibling CPU 5.

The stock driver omitted Core 1 after module load and retained its
readable sensor after CPU 5 was offlined. The patched driver exposed
Core 1, removed it when CPU 5 went offline, and restored it when CPU 5
returned.
---
 drivers/hwmon/coretemp.c | 26 ++++++++++++++++----------
 1 file changed, 16 insertions(+), 10 deletions(-)

diff --git a/drivers/hwmon/coretemp.c b/drivers/hwmon/coretemp.c
index 0ab6bbff5637..ad9cadd99dc8 100644
--- a/drivers/hwmon/coretemp.c
+++ b/drivers/hwmon/coretemp.c
@@ -559,9 +559,6 @@ static int create_core_data(struct platform_device *pdev, unsigned int cpu,
 	u64 val;
 	int err;
 
-	if (!housekeeping_cpu(cpu, HK_TYPE_MISC))
-		return 0;
-
 	tdata = init_temp_data(pdata, cpu, pkg_flag);
 	if (!tdata)
 		return -ENOMEM;
@@ -681,6 +678,14 @@ static int coretemp_cpu_online(unsigned int cpu)
 	if (!cpu_has(c, X86_FEATURE_DTHERM))
 		return -ENODEV;
 
+	/*
+	 * Keep isolated CPUs out of pdata->cpumask. Besides tracking the CPUs
+	 * available to this driver, the mask is used to select the CPU on which
+	 * to read the temperature MSRs.
+	 */
+	if (!housekeeping_cpu(cpu, HK_TYPE_MISC))
+		return 0;
+
 	pdata = platform_get_drvdata(pdev);
 	if (!pdata->hwmon_dev) {
 		struct device *hwmon;
@@ -710,8 +715,8 @@ static int coretemp_cpu_online(unsigned int cpu)
 	}
 
 	/*
-	 * Check whether a thread sibling is already online. If not add the
-	 * interface for this CPU core.
+	 * Check whether a housekeeping thread sibling is already online. If
+	 * not, add the interface for this CPU core.
 	 */
 	if (!cpumask_intersects(&pdata->cpumask, topology_sibling_cpumask(cpu)))
 		coretemp_add_core(pdev, cpu, 0);
@@ -741,9 +746,9 @@ static int coretemp_cpu_offline(unsigned int cpu)
 	cpumask_clear_cpu(cpu, &pd->cpumask);
 
 	/*
-	 * If this is the last thread sibling, remove the CPU core
-	 * interface, If there is still a sibling online, transfer the
-	 * target cpu of that core interface to it.
+	 * If this is the last housekeeping thread sibling, remove the CPU core
+	 * interface. If there is still a housekeeping sibling online, transfer
+	 * the target CPU of that core interface to it.
 	 */
 	target = cpumask_any_and(&pd->cpumask, topology_sibling_cpumask(cpu));
 	if (target >= nr_cpu_ids) {
@@ -755,7 +760,8 @@ static int coretemp_cpu_offline(unsigned int cpu)
 	}
 
 	/*
-	 * If all cores in this pkg are offline, remove the interface.
+	 * If all housekeeping cores in this pkg are offline, remove the
+	 * interface.
 	 */
 	tdata = get_temp_data(pd, -1);
 	if (cpumask_empty(&pd->cpumask)) {
@@ -768,7 +774,7 @@ static int coretemp_cpu_offline(unsigned int cpu)
 
 	/*
 	 * Check whether this core is the target for the package
-	 * interface. We need to assign it to some other cpu.
+	 * interface. We need to assign it to some other housekeeping CPU.
 	 */
 	if (tdata && tdata->cpu == cpu) {
 		target = cpumask_first(&pd->cpumask);

---
base-commit: 4781ca52761e666cf18b591e6bb0478396c90320
change-id: 20261015-codex-coretemp-housekeeping-e7a16253adea

Best regards,
--  
Ximing Zhang <xzhangjr@gmail.com>



^ permalink raw reply related	[flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-10-16  8:46 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-16  8:36 [PATCH] hwmon: (coretemp) Fix handling of isolated SMT siblings Ximing Zhang via B4 Relay
2026-10-16  8:36 ` Ximing Zhang
2026-10-16  8:46 ` sashiko-bot

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.