cpu/hotplug: Abort disabling secondary CPUs if wakeup is pending
authorPavankumar Kondeti <pkondeti@codeaurora.org>
Mon, 3 Jun 2019 04:31:03 +0000 (10:01 +0530)
committerThomas Gleixner <tglx@linutronix.de>
Wed, 12 Jun 2019 09:03:05 +0000 (11:03 +0200)
When "deep" suspend is enabled, all CPUs except the primary CPU are frozen
via CPU hotplug one by one. After all secondary CPUs are unplugged the
wakeup pending condition is evaluated and if pending the suspend operation
is aborted and the secondary CPUs are brought up again.

CPU hotplug is a slow operation, so it makes sense to check for wakeup
pending in the freezer loop before bringing down the next CPU. This
improves the system suspend abort latency significantly.

[ tglx: Massaged changelog and improved printk message ]

Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Len Brown <len.brown@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: iri Kosina <jkosina@suse.cz>
Cc: Mukesh Ojha <mojha@codeaurora.org>
Cc: linux-pm@vger.kernel.org
Link: https://lkml.kernel.org/r/1559536263-16472-1-git-send-email-pkondeti@codeaurora.org
kernel/cpu.c

index be82cbc..0778249 100644 (file)
@@ -1221,6 +1221,13 @@ int freeze_secondary_cpus(int primary)
        for_each_online_cpu(cpu) {
                if (cpu == primary)
                        continue;
+
+               if (pm_wakeup_pending()) {
+                       pr_info("Wakeup pending. Abort CPU freeze\n");
+                       error = -EBUSY;
+                       break;
+               }
+
                trace_suspend_resume(TPS("CPU_OFF"), cpu, true);
                error = _cpu_down(cpu, 1, CPUHP_OFFLINE);
                trace_suspend_resume(TPS("CPU_OFF"), cpu, false);