cpufreq: powernv: Move smp_call_function_any() out of irq safe block
authorAkshay Adiga <akshay.adiga@linux.vnet.ibm.com>
Tue, 3 May 2016 15:19:35 +0000 (20:49 +0530)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Wed, 11 May 2016 00:28:00 +0000 (02:28 +0200)
Fix a WARN_ON caused by smp_call_function_any() when irq is disabled,
because of changes made in the patch ('cpufreq: powernv: Ramp-down
 global pstate slower than local-pstate')
https://patchwork.ozlabs.org/patch/612058/

 WARNING: CPU: 0 PID: 4 at kernel/smp.c:291
smp_call_function_single+0x170/0x180

 Call Trace:
 [c0000007f648f9f0] [c0000007f648fa90] 0xc0000007f648fa90 (unreliable)
 [c0000007f648fa30] [c0000000001430e0] smp_call_function_any+0x170/0x1c0
 [c0000007f648fa90] [c0000000007b4b00]
powernv_cpufreq_target_index+0xe0/0x250
 [c0000007f648fb00] [c0000000007ac9dc]
__cpufreq_driver_target+0x20c/0x3d0
 [c0000007f648fbc0] [c0000000007b1b4c] od_dbs_timer+0xcc/0x260
 [c0000007f648fc10] [c0000000007b3024] dbs_work_handler+0x54/0xa0
 [c0000007f648fc50] [c0000000000c49a8] process_one_work+0x1d8/0x590
 [c0000007f648fce0] [c0000000000c4e08] worker_thread+0xa8/0x660
 [c0000007f648fd80] [c0000000000cca88] kthread+0x108/0x130
 [c0000007f648fe30] [c0000000000095e8] ret_from_kernel_thread+0x5c/0x74

- Calling smp_call_function_any() with interrupt disabled (through
 spin_lock_irqsave) could cause a deadlock, as smp_call_function_any()
 relies on the IPI to complete. This is detected in the
 smp_call_function_any() call and hence the WARN_ON.

- As the spinlock (gpstates->lock) is only used to synchronize access of
 global_pstate_info  between timer irq handler and target_index calls. And
 the timer irq handler just try_locks() hence it would not cause a
 deadlock. Hence could do without making spinlocks irq safe.

- As the smp_call_function_any() is a blocking call and does not access
 global_pstates_info, it could reduce the critcal section by moving
 smp_call_function_any() after giving up the lock.

Reported-by: Abdul Haleem <abdhalee@linux.vnet.linux.com>
Signed-off-by: Akshay Adiga <akshay.adiga@linux.vnet.ibm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
drivers/cpufreq/powernv-cpufreq.c

index 144c73211926393216f8bbf739f8cbf54e0783e6..1f0e20ccc2ff43ae1100e8f6ba1d7253bbf4d69a 100644 (file)
@@ -581,9 +581,10 @@ void gpstate_timer_handler(unsigned long data)
        gpstates->last_gpstate = freq_data.gpstate_id;
        gpstates->last_lpstate = freq_data.pstate_id;
 
+       spin_unlock(&gpstates->gpstate_lock);
+
        /* Timer may get migrated to a different cpu on cpu hot unplug */
        smp_call_function_any(policy->cpus, set_pstate, &freq_data, 1);
-       spin_unlock(&gpstates->gpstate_lock);
 }
 
 /*
@@ -596,7 +597,6 @@ static int powernv_cpufreq_target_index(struct cpufreq_policy *policy,
 {
        struct powernv_smp_call_data freq_data;
        unsigned int cur_msec, gpstate_id;
-       unsigned long flags;
        struct global_pstate_info *gpstates = policy->driver_data;
 
        if (unlikely(rebooting) && new_index != get_nominal_index())
@@ -607,7 +607,7 @@ static int powernv_cpufreq_target_index(struct cpufreq_policy *policy,
 
        cur_msec = jiffies_to_msecs(get_jiffies_64());
 
-       spin_lock_irqsave(&gpstates->gpstate_lock, flags);
+       spin_lock(&gpstates->gpstate_lock);
        freq_data.pstate_id = powernv_freqs[new_index].driver_data;
 
        if (!gpstates->last_sampled_time) {
@@ -654,13 +654,14 @@ gpstates_done:
        gpstates->last_gpstate = freq_data.gpstate_id;
        gpstates->last_lpstate = freq_data.pstate_id;
 
+       spin_unlock(&gpstates->gpstate_lock);
+
        /*
         * Use smp_call_function to send IPI and execute the
         * mtspr on target CPU.  We could do that without IPI
         * if current CPU is within policy->cpus (core)
         */
        smp_call_function_any(policy->cpus, set_pstate, &freq_data, 1);
-       spin_unlock_irqrestore(&gpstates->gpstate_lock, flags);
        return 0;
 }