An instance of 'struct dbs_data' can support multiple 'struct policy_dbs_info' instances. To traverse all policy_dbs supported by a dbs_data, create a list of policy_dbs within dbs_data.
We can traverse this list now, instead of traversing the loop for all online CPUs in update_sampling_rate(), to solve the circular dependency lockdep reported by Juri (and verified by Shilpa) earlier:
====================================================== [ INFO: possible circular locking dependency detected ] 4.4.0+ #445 Not tainted ------------------------------------------------------- trace.sh/1723 is trying to acquire lock: (s_active#48){++++.+}, at: [<c01f78c8>] kernfs_remove_by_name_ns+0x4c/0x94
but task is already holding lock: (od_dbs_cdata.mutex){+.+.+.}, at: [<c05824a0>] cpufreq_governor_dbs+0x34/0x5d4
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #2 (od_dbs_cdata.mutex){+.+.+.}: [<c075b040>] mutex_lock_nested+0x7c/0x434 [<c05824a0>] cpufreq_governor_dbs+0x34/0x5d4 [<c0017c10>] return_to_handler+0x0/0x18
-> #1 (&policy->rwsem){+++++.}: [<c075ca8c>] down_read+0x58/0x94 [<c057c244>] show+0x30/0x60 [<c01f934c>] sysfs_kf_seq_show+0x90/0xfc [<c01f7ad8>] kernfs_seq_show+0x34/0x38 [<c01a22ec>] seq_read+0x1e4/0x4e4 [<c01f8694>] kernfs_fop_read+0x120/0x1a0 [<c01794b4>] __vfs_read+0x3c/0xe0 [<c017a378>] vfs_read+0x98/0x104 [<c017a434>] SyS_read+0x50/0x90 [<c000fd40>] ret_fast_syscall+0x0/0x1c
-> #0 (s_active#48){++++.+}: [<c008238c>] lock_acquire+0xd4/0x20c [<c01f6ae4>] __kernfs_remove+0x288/0x328 [<c01f78c8>] kernfs_remove_by_name_ns+0x4c/0x94 [<c01fa024>] remove_files+0x44/0x88 [<c01fa5a4>] sysfs_remove_group+0x50/0xa4 [<c058285c>] cpufreq_governor_dbs+0x3f0/0x5d4 [<c0017c10>] return_to_handler+0x0/0x18
other info that might help us debug this:
Chain exists of: s_active#48 --> &policy->rwsem --> od_dbs_cdata.mutex
Possible unsafe locking scenario:
CPU0 CPU1 ---- ---- lock(od_dbs_cdata.mutex); lock(&policy->rwsem); lock(od_dbs_cdata.mutex); lock(s_active#48);
*** DEADLOCK ***
5 locks held by trace.sh/1723: #0: (sb_writers#6){.+.+.+}, at: [<c017beb8>] __sb_start_write+0xb4/0xc0 #1: (&of->mutex){+.+.+.}, at: [<c01f8418>] kernfs_fop_write+0x6c/0x1c8 #2: (s_active#35){.+.+.+}, at: [<c01f8420>] kernfs_fop_write+0x74/0x1c8 #3: (cpu_hotplug.lock){++++++}, at: [<c0029e6c>] get_online_cpus+0x48/0xb8 #4: (od_dbs_cdata.mutex){+.+.+.}, at: [<c05824a0>] cpufreq_governor_dbs+0x34/0x5d4
stack backtrace: CPU: 2 PID: 1723 Comm: trace.sh Not tainted 4.4.0+ #445 Hardware name: ARM-Versatile Express [<c001883c>] (unwind_backtrace) from [<c0013f50>] (show_stack+0x20/0x24) [<c0013f50>] (show_stack) from [<c044ad90>] (dump_stack+0x80/0xb4) [<c044ad90>] (dump_stack) from [<c0128edc>] (print_circular_bug+0x29c/0x2f0) [<c0128edc>] (print_circular_bug) from [<c0081708>] (__lock_acquire+0x163c/0x1d74) [<c0081708>] (__lock_acquire) from [<c008238c>] (lock_acquire+0xd4/0x20c) [<c008238c>] (lock_acquire) from [<c01f6ae4>] (__kernfs_remove+0x288/0x328) [<c01f6ae4>] (__kernfs_remove) from [<c01f78c8>] (kernfs_remove_by_name_ns+0x4c/0x94) [<c01f78c8>] (kernfs_remove_by_name_ns) from [<c01fa024>] (remove_files+0x44/0x88) [<c01fa024>] (remove_files) from [<c01fa5a4>] (sysfs_remove_group+0x50/0xa4) [<c01fa5a4>] (sysfs_remove_group) from [<c058285c>] (cpufreq_governor_dbs+0x3f0/0x5d4) [<c058285c>] (cpufreq_governor_dbs) from [<c0017c10>] (return_to_handler+0x0/0x18)
This also updates the comment above update_sampling_rate() to make it more relevant to the current state of code.
Signed-off-by: Viresh Kumar viresh.kumar@linaro.org Reported-by: Juri Lelli juri.lelli@arm.com Tested-by: Juri Lelli juri.lelli@arm.com Tested-by: Shilpasri G Bhat shilpa.bhat@linux.vnet.ibm.com --- drivers/cpufreq/cpufreq_governor.c | 22 ++++++++-- drivers/cpufreq/cpufreq_governor.h | 7 ++- drivers/cpufreq/cpufreq_ondemand.c | 89 +++++++++++++------------------------- 3 files changed, 54 insertions(+), 64 deletions(-)
diff --git a/drivers/cpufreq/cpufreq_governor.c b/drivers/cpufreq/cpufreq_governor.c index bba9d3fb8103..8e53f804a5af 100644 --- a/drivers/cpufreq/cpufreq_governor.c +++ b/drivers/cpufreq/cpufreq_governor.c @@ -384,9 +384,14 @@ static int cpufreq_governor_init(struct cpufreq_policy *policy) ret = -EINVAL; goto free_policy_dbs_info; } - dbs_data->usage_count++; policy_dbs->dbs_data = dbs_data; policy->governor_data = policy_dbs; + + mutex_lock(&dbs_data->mutex); + dbs_data->usage_count++; + list_add(&policy_dbs->list, &dbs_data->policy_dbs_list); + mutex_unlock(&dbs_data->mutex); + return 0; }
@@ -396,7 +401,7 @@ static int cpufreq_governor_init(struct cpufreq_policy *policy) goto free_policy_dbs_info; }
- dbs_data->usage_count = 1; + INIT_LIST_HEAD(&dbs_data->policy_dbs_list); mutex_init(&dbs_data->mutex);
ret = gov->init(dbs_data, !policy->governor->initialized); @@ -417,9 +422,12 @@ static int cpufreq_governor_init(struct cpufreq_policy *policy) if (!have_governor_per_policy()) gov->gdbs_data = dbs_data;
- policy_dbs->dbs_data = dbs_data; policy->governor_data = policy_dbs;
+ policy_dbs->dbs_data = dbs_data; + dbs_data->usage_count = 1; + list_add(&policy_dbs->list, &dbs_data->policy_dbs_list); + gov->kobj_type.sysfs_ops = &governor_sysfs_ops; ret = kobject_init_and_add(&dbs_data->kobj, &gov->kobj_type, get_governor_parent_kobj(policy), @@ -450,12 +458,18 @@ static int cpufreq_governor_exit(struct cpufreq_policy *policy) struct dbs_governor *gov = dbs_governor_of(policy); struct policy_dbs_info *policy_dbs = policy->governor_data; struct dbs_data *dbs_data = policy_dbs->dbs_data; + int count;
/* State should be equivalent to INIT */ if (policy_dbs->policy) return -EBUSY;
- if (!--dbs_data->usage_count) { + mutex_lock(&dbs_data->mutex); + list_del(&policy_dbs->list); + count = dbs_data->usage_count--; + mutex_unlock(&dbs_data->mutex); + + if (!count) { kobject_put(&dbs_data->kobj);
policy->governor_data = NULL; diff --git a/drivers/cpufreq/cpufreq_governor.h b/drivers/cpufreq/cpufreq_governor.h index d46ebcb4f16d..4e77efb7db67 100644 --- a/drivers/cpufreq/cpufreq_governor.h +++ b/drivers/cpufreq/cpufreq_governor.h @@ -73,7 +73,11 @@ struct dbs_data { unsigned int up_threshold;
struct kobject kobj; - /* Protect concurrent updates to governor tunables from sysfs */ + struct list_head policy_dbs_list; + /* + * Protect concurrent updates to governor tunables from sysfs, + * policy_dbs_list and usage_count. + */ struct mutex mutex; };
@@ -125,6 +129,7 @@ struct policy_dbs_info { struct work_struct work; /* dbs_data may be shared between multiple policy objects */ struct dbs_data *dbs_data; + struct list_head list; };
static inline void gov_update_sample_delay(struct policy_dbs_info *policy_dbs, diff --git a/drivers/cpufreq/cpufreq_ondemand.c b/drivers/cpufreq/cpufreq_ondemand.c index e36792f60348..38301c6b31c7 100644 --- a/drivers/cpufreq/cpufreq_ondemand.c +++ b/drivers/cpufreq/cpufreq_ondemand.c @@ -226,84 +226,55 @@ static struct dbs_governor od_dbs_gov; * @new_rate: new sampling rate * * If new rate is smaller than the old, simply updating - * dbs_tuners_int.sampling_rate might not be appropriate. For example, if the + * dbs.sampling_rate might not be appropriate. For example, if the * original sampling_rate was 1 second and the requested new sampling rate is 10 * ms because the user needs immediate reaction from ondemand governor, but not * sure if higher frequency will be required or not, then, the governor may * change the sampling rate too late; up to 1 second later. Thus, if we are * reducing the sampling rate, we need to make the new value effective * immediately. + * + * On the other hand, if new rate is larger than the old, then we may evaluate + * the load too soon, and it might we worth updating sample_delay_ns then as + * well. + * + * This must be called with dbs_data->mutex held, otherwise traversing + * policy_dbs_list isn't safe. */ static void update_sampling_rate(struct dbs_data *dbs_data, unsigned int new_rate) { - struct cpumask cpumask; - int cpu; + struct policy_dbs_info *policy_dbs;
dbs_data->sampling_rate = new_rate = max(new_rate, dbs_data->min_sampling_rate);
/* - * Lock governor so that governor start/stop can't execute in parallel. + * We are operating under dbs_data->mutex and so the list and its + * entries can't be freed concurrently. */ - mutex_lock(&dbs_data_mutex); - - cpumask_copy(&cpumask, cpu_online_mask); - - for_each_cpu(cpu, &cpumask) { - struct cpufreq_policy *policy; - struct od_cpu_dbs_info_s *dbs_info; - struct cpu_dbs_info *cdbs; - struct policy_dbs_info *policy_dbs; - - dbs_info = &per_cpu(od_cpu_dbs_info, cpu); - cdbs = &dbs_info->cdbs; - policy_dbs = cdbs->policy_dbs; - + list_for_each_entry(policy_dbs, &dbs_data->policy_dbs_list, list) { + mutex_lock(&policy_dbs->timer_mutex); /* - * A valid policy_dbs and policy_dbs->policy means governor - * hasn't stopped or exited yet. + * On 32-bit architectures this may race with the + * sample_delay_ns read in dbs_update_util_handler(), but that + * really doesn't matter. If the read returns a value that's + * too big, the sample will be skipped, but the next invocation + * of dbs_update_util_handler() (when the update has been + * completed) will take a sample. If the returned value is too + * small, the sample will be taken immediately, but that isn't a + * problem, as we want the new rate to take effect immediately + * anyway. + * + * If this runs in parallel with dbs_work_handler(), we may end + * up overwriting the sample_delay_ns value that it has just + * written, but the difference should not be too big and it will + * be corrected next time a sample is taken, so it shouldn't be + * significant. */ - if (!policy_dbs || !policy_dbs->policy) - continue; - - policy = policy_dbs->policy; - - /* clear all CPUs of this policy */ - cpumask_andnot(&cpumask, &cpumask, policy->cpus); - - /* - * Update sampling rate for CPUs whose policy is governed by - * dbs_data. In case of governor_per_policy, only a single - * policy will be governed by dbs_data, otherwise there can be - * multiple policies that are governed by the same dbs_data. - */ - if (dbs_data == policy_dbs->dbs_data) { - mutex_lock(&policy_dbs->timer_mutex); - /* - * On 32-bit architectures this may race with the - * sample_delay_ns read in dbs_update_util_handler(), - * but that really doesn't matter. If the read returns - * a value that's too big, the sample will be skipped, - * but the next invocation of dbs_update_util_handler() - * (when the update has been completed) will take a - * sample. If the returned value is too small, the - * sample will be taken immediately, but that isn't a - * problem, as we want the new rate to take effect - * immediately anyway. - * - * If this runs in parallel with dbs_work_handler(), we - * may end up overwriting the sample_delay_ns value that - * it has just written, but the difference should not be - * too big and it will be corrected next time a sample - * is taken, so it shouldn't be significant. - */ - gov_update_sample_delay(policy_dbs, new_rate); - mutex_unlock(&policy_dbs->timer_mutex); - } + gov_update_sample_delay(policy_dbs, new_rate); + mutex_unlock(&policy_dbs->timer_mutex); } - - mutex_unlock(&dbs_data_mutex); }
static ssize_t store_sampling_rate(struct dbs_data *dbs_data, const char *buf,