Re: [patch] x86: avoid unnecessary smp alternatives switch duringsuspend/resume

From: Nigel Cunningham
Date: Sun Nov 21 2010 - 04:37:05 EST


Hi.

On 20/11/10 11:09, Suresh Siddha wrote:
During suspend, we disable all the non boot cpus. And during resume we bring
them all back again. So no need to do alternatives_smp_switch() in between.

This speeds up both suspend and resume paths.

Signed-off-by: Suresh Siddha<suresh.b.siddha@xxxxxxxxx>
---

arch/x86/kernel/smpboot.c | 7 ++++++-
include/linux/cpu.h | 2 ++
kernel/cpu.c | 3 +++
3 files changed, 11 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index f0a0624..0b04ca3 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1349,7 +1349,12 @@ void native_cpu_die(unsigned int cpu)
if (system_state == SYSTEM_RUNNING)
pr_info("CPU %u is now offline\n", cpu);

- if (1 == num_online_cpus())
+ /*
+ * Don't do the smp alternatives switch during
+ * suspend. We will be back in the SMP mode after
+ * resume.
+ */
+ if (1 == num_online_cpus()&& !pm_sleep_smp)
alternatives_smp_switch(0);
return;
}
diff --git a/include/linux/cpu.h b/include/linux/cpu.h
index 4823af6..8cab04c 100644
--- a/include/linux/cpu.h
+++ b/include/linux/cpu.h
@@ -169,11 +169,13 @@ static inline void cpu_hotplug_driver_unlock(void)

#ifdef CONFIG_PM_SLEEP_SMP
extern int suspend_cpu_hotplug;
+extern int pm_sleep_smp;

extern int disable_nonboot_cpus(void);
extern void enable_nonboot_cpus(void);
#else /* !CONFIG_PM_SLEEP_SMP */
#define suspend_cpu_hotplug 0
+#define pm_sleep_smp 0

static inline int disable_nonboot_cpus(void) { return 0; }
static inline void enable_nonboot_cpus(void) {}
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 8615aa6..2eed810 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -381,6 +381,7 @@ out:

#ifdef CONFIG_PM_SLEEP_SMP
static cpumask_var_t frozen_cpus;
+int pm_sleep_smp;

int disable_nonboot_cpus(void)
{
@@ -393,6 +394,7 @@ int disable_nonboot_cpus(void)
* with the userspace trying to use the CPU hotplug at the same time
*/
cpumask_clear(frozen_cpus);
+ pm_sleep_smp = 1;

printk("Disabling non-boot CPUs ...\n");
for_each_online_cpu(cpu) {
@@ -454,6 +456,7 @@ void __ref enable_nonboot_cpus(void)

cpumask_clear(frozen_cpus);
out:
+ pm_sleep_smp = 0;
cpu_maps_update_done();
}
We have a few others things that want to modify their behaviour according to whether we're doing the atomic copy/restore. Perhaps it would be an idea to just use a single flag, perhaps a value for system_state?

Nigel
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/