Linux Documentation
 help / color / mirror / Atom feed
From: Waiman Long <longman@redhat.com>
To: "Tejun Heo" <tj@kernel.org>,
	"Johannes Weiner" <hannes@cmpxchg.org>,
	"Michal Koutný" <mkoutny@suse.com>,
	"Jonathan Corbet" <corbet@lwn.net>,
	"Shuah Khan" <skhan@linuxfoundation.org>,
	"Randy Dunlap" <rdunlap@infradead.org>,
	"Ridong Chen" <ridong.chen@linux.dev>,
	"Ingo Molnar" <mingo@redhat.com>,
	"Peter Zijlstra" <peterz@infradead.org>,
	"Juri Lelli" <juri.lelli@redhat.com>,
	"Vincent Guittot" <vincent.guittot@linaro.org>,
	"Dietmar Eggemann" <dietmar.eggemann@arm.com>,
	"Steven Rostedt" <rostedt@goodmis.org>,
	"Ben Segall" <bsegall@google.com>, "Mel Gorman" <mgorman@suse.de>,
	"Valentin Schneider" <vschneid@redhat.com>,
	"K Prateek Nayak" <kprateek.nayak@amd.com>,
	"Anna-Maria Behnsen" <anna-maria@linutronix.de>,
	"Frederic Weisbecker" <frederic@kernel.org>,
	"Thomas Gleixner" <tglx@kernel.org>
Cc: cgroups@vger.kernel.org, linux-doc@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
	Guopeng Zhang <guopeng.zhang@linux.dev>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	Waiman Long <longman@redhat.com>
Subject: [PATCH-next 2/3] sched/isolation: Forbid the boot CPU from being isolated
Date: Fri, 20 Nov 2026 22:13:28 -0500	[thread overview]
Message-ID: <20261121031329.240177-3-longman@redhat.com> (raw)
In-Reply-To: <20261121031329.240177-1-longman@redhat.com>

It is reported that with CONFIG_CPUSETS=n, a kernel booted with
isolcpus=domain,0 hits the BUG() in select_fallback_rq() during boot
[1]. At boot time, cpu 0 is the only CPU online and an unbound kworker
cannot be run on a domain isolated CPU causing kernel panic.

Sashiko also reports that domain isolating the boot CPU can break system
suspend and trigger a BUG_ON() [2]. So it is problematic if the boot CPU
is domain isolated.

There may be other kernel subsystems that have problem if the boot CPU
is domain isolated as most of them won't take into account the fact
that boot CPU is currently allowed to be domain isolated. The current
housekeeping code enforces the rule that if both isolcpus and nohz_full
boot parameters are set, they should match each other. Given the fact
that the boot CPU is likely to be disallowed from the nohz_full list,
it makes sense that we should not allow the boot CPU from being set
in the isolcpus boot parameter as well. This will also avoid the above
problems due to boot CPU being domain isolated.

Do that by clearing the boot CPU from the isolcpus range when it is
specified in the boot command line. A warning will also be printed when
that happens.

This is indeed a change in behavior which is customer visible. So the
kernel-paremeters.txt is updated to mention the fact that boot CPU
cannot be isolated.

[1] https://lore.kernel.org/lkml/20261117144817.265109-1-mathieu.desnoyers@efficios.com
[2] https://sashiko.dev/#/patchset/20261108180849.1268263-1-longman%40redhat.com

Signed-off-by: Waiman Long <longman@redhat.com>
---
 Documentation/admin-guide/kernel-parameters.txt |  7 ++++---
 kernel/sched/isolation.c                        | 16 ++++++++++++----
 2 files changed, 16 insertions(+), 7 deletions(-)

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 930cee4de79a..1c8bf3583b0b 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -2954,11 +2954,12 @@ Kernel parameters
 
 			  You can move a process onto or off an "isolated" CPU via
 			  the CPU affinity syscalls or cpuset.
-			  <cpu number> begins at 0 and the maximum value is
-			  "number of CPUs in system - 1".
+			  <cpu number> begins at 0 and the maximum value
+			  is "number of CPUs in system - 1". However boot
+			  CPU (cpu 0) cannot be domain isolated as it
+			  may cause problem with some kernel subsystems.
 
 			managed_irq
-
 			  Isolate from being targeted by managed interrupts
 			  which have an interrupt mask containing isolated
 			  CPUs. The affinity of managed interrupts is
diff --git a/kernel/sched/isolation.c b/kernel/sched/isolation.c
index 18b50c0d807a..ef4aa5126dc5 100644
--- a/kernel/sched/isolation.c
+++ b/kernel/sched/isolation.c
@@ -232,6 +232,7 @@ static int __init housekeeping_setup(char *str, unsigned long flags)
 {
 	cpumask_var_t non_housekeeping_mask, housekeeping_staging;
 	unsigned int first_cpu;
+	int boot_cpu = smp_processor_id();
 	int err = 0;
 
 	if ((flags & HK_FLAG_KERNEL_NOISE) && !(housekeeping.flags & HK_FLAG_KERNEL_NOISE)) {
@@ -254,14 +255,21 @@ static int __init housekeeping_setup(char *str, unsigned long flags)
 
 	first_cpu = cpumask_first_and(cpu_present_mask, housekeeping_staging);
 	if (first_cpu >= nr_cpu_ids || first_cpu >= setup_max_cpus) {
-		__cpumask_set_cpu(smp_processor_id(), housekeeping_staging);
-		__cpumask_clear_cpu(smp_processor_id(), non_housekeeping_mask);
+		__cpumask_set_cpu(boot_cpu, housekeeping_staging);
+		__cpumask_clear_cpu(boot_cpu, non_housekeeping_mask);
 		if (!housekeeping.flags) {
-			pr_warn("Housekeeping: must include one present CPU, "
-				"using boot CPU:%d\n", smp_processor_id());
+			pr_warn("Housekeeping: must include one present CPU, using boot CPU:%d\n",
+				boot_cpu);
 		}
 	}
 
+	if ((flags & HK_FLAG_DOMAIN) && cpumask_test_cpu(boot_cpu, non_housekeeping_mask)) {
+		pr_warn("Housekeeping: Boot CPU must not be isolated, clearing %d from isolcpus range\n",
+			 boot_cpu);
+		__cpumask_set_cpu(boot_cpu, housekeeping_staging);
+		__cpumask_clear_cpu(boot_cpu, non_housekeeping_mask);
+	}
+
 	if (cpumask_empty(non_housekeeping_mask))
 		goto free_housekeeping_staging;
 
-- 
2.56.1


  parent reply	other threads:[~2026-11-21  3:14 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-21  3:13 [PATCH-next 0/3] Skip boot CPU or don't allow it from being used in isolcpus and isolated partitions Waiman Long
2026-11-21  3:13 ` [PATCH-next 1/3] tick/sched: Make tick_nohz_full_mask and HK_TYPE_KERNEL_NOISE cpumask consistent Waiman Long
2026-11-21  3:13 ` Waiman Long [this message]
2026-11-21  3:13 ` [PATCH-next 3/3] cgroup/cpuset: Disallow CPU 0 from being used in isolated partition Waiman Long

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=20261121031329.240177-3-longman@redhat.com \
    --to=longman@redhat.com \
    --cc=anna-maria@linutronix.de \
    --cc=bsegall@google.com \
    --cc=cgroups@vger.kernel.org \
    --cc=corbet@lwn.net \
    --cc=dietmar.eggemann@arm.com \
    --cc=frederic@kernel.org \
    --cc=guopeng.zhang@linux.dev \
    --cc=hannes@cmpxchg.org \
    --cc=juri.lelli@redhat.com \
    --cc=kprateek.nayak@amd.com \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mgorman@suse.de \
    --cc=mingo@redhat.com \
    --cc=mkoutny@suse.com \
    --cc=peterz@infradead.org \
    --cc=rdunlap@infradead.org \
    --cc=ridong.chen@linux.dev \
    --cc=rostedt@goodmis.org \
    --cc=skhan@linuxfoundation.org \
    --cc=tglx@kernel.org \
    --cc=tj@kernel.org \
    --cc=vincent.guittot@linaro.org \
    --cc=vschneid@redhat.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