[tip: sched/core] sched: Improve wake_up_all_idle_cpus() take #2

[tip: sched/core] sched: Improve wake_up_all_idle_cpus() take #2

[tip-bot2 for Peter Zijlstra]

The following commit has been merged into the sched/core branch of tip:

Commit-ID:     96611c26dc351c33f73b48756a9feacc109e5bab
Gitweb:        https://git.kernel.org/tip/96611c26dc351c33f73b48756a9feacc109e5bab
Author:        Peter Zijlstra <peterz@infradead.org>
AuthorDate:    Mon, 18 Oct 2021 16:41:05 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 22 Oct 2021 15:32:46 +02:00

sched: Improve wake_up_all_idle_cpus() take #2

As reported by syzbot and experienced by Pavel, using cpus_read_lock()
in wake_up_all_idle_cpus() generates lock inversion (against mmap_sem
and possibly others).

Instead, shrink the preempt disable region by iterating all CPUs and
checking the online status for each individual CPU while having
preemption disabled.

Fixes: 8850cb663b5c ("sched: Simplify wake_up_*idle*()")
Reported-by: syzbot+d5b23b18d2f4feae8a67@syzkaller.appspotmail.com
Reported-by: Pavel Machek <pavel@ucw.cz>
Reported-by: Qian Cai <quic_qiancai@quicinc.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Qian Cai <quic_qiancai@quicinc.com>
---
 kernel/smp.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/kernel/smp.c b/kernel/smp.c
index ad0b68a..01a7c17 100644
--- a/kernel/smp.c
+++ b/kernel/smp.c
@@ -1170,14 +1170,12 @@ void wake_up_all_idle_cpus(void)
 {
 	int cpu;
 
-	cpus_read_lock();
-	for_each_online_cpu(cpu) {
-		if (cpu == raw_smp_processor_id())
-			continue;
-
-		wake_up_if_idle(cpu);
+	for_each_possible_cpu(cpu) {
+		preempt_disable();
+		if (cpu != smp_processor_id() && cpu_online(cpu))
+			wake_up_if_idle(cpu);
+		preempt_enable();
 	}
-	cpus_read_unlock();
 }
 EXPORT_SYMBOL_GPL(wake_up_all_idle_cpus);