#include <linux/notifier.h>
#include <linux/cpu.h>
#include <linux/mutex.h>
+#include <linux/time.h>
#ifdef CONFIG_DEBUG_LOCK_ALLOC
static struct lock_class_key rcu_lock_key;
{
int cpu;
cpumask_t cpumask;
+ unsigned long flags;
+
set_need_resched();
+ spin_lock_irqsave(&rcp->lock, flags);
if (unlikely(!rcp->signaled)) {
rcp->signaled = 1;
/*
for_each_cpu_mask_nr(cpu, cpumask)
smp_send_reschedule(cpu);
}
+ spin_unlock_irqrestore(&rcp->lock, flags);
}
#else
static inline void force_quiescent_state(struct rcu_data *rdp,
struct rcu_data *rdp)
{
long batch;
- smp_mb(); /* reads the most recently updated value of rcu->cur. */
+
+ head->next = NULL;
+ smp_mb(); /* Read of rcu->cur must happen after any change by caller. */
/*
* Determine the batch number of this callback.
}
}
+#ifdef CONFIG_RCU_CPU_STALL_DETECTOR
+
+static void record_gp_stall_check_time(struct rcu_ctrlblk *rcp)
+{
+ rcp->gp_start = jiffies;
+ rcp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_CHECK;
+}
+
+static void print_other_cpu_stall(struct rcu_ctrlblk *rcp)
+{
+ int cpu;
+ long delta;
+ unsigned long flags;
+
+ /* Only let one CPU complain about others per time interval. */
+
+ spin_lock_irqsave(&rcp->lock, flags);
+ delta = jiffies - rcp->jiffies_stall;
+ if (delta < 2 || rcp->cur != rcp->completed) {
+ spin_unlock_irqrestore(&rcp->lock, flags);
+ return;
+ }
+ rcp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK;
+ spin_unlock_irqrestore(&rcp->lock, flags);
+
+ /* OK, time to rat on our buddy... */
+
+ printk(KERN_ERR "RCU detected CPU stalls:");
+ for_each_possible_cpu(cpu) {
+ if (cpu_isset(cpu, rcp->cpumask))
+ printk(" %d", cpu);
+ }
+ printk(" (detected by %d, t=%ld jiffies)\n",
+ smp_processor_id(), (long)(jiffies - rcp->gp_start));
+}
+
+static void print_cpu_stall(struct rcu_ctrlblk *rcp)
+{
+ unsigned long flags;
+
+ printk(KERN_ERR "RCU detected CPU %d stall (t=%lu/%lu jiffies)\n",
+ smp_processor_id(), jiffies,
+ jiffies - rcp->gp_start);
+ dump_stack();
+ spin_lock_irqsave(&rcp->lock, flags);
+ if ((long)(jiffies - rcp->jiffies_stall) >= 0)
+ rcp->jiffies_stall =
+ jiffies + RCU_SECONDS_TILL_STALL_RECHECK;
+ spin_unlock_irqrestore(&rcp->lock, flags);
+ set_need_resched(); /* kick ourselves to get things going. */
+}
+
+static void check_cpu_stall(struct rcu_ctrlblk *rcp)
+{
+ long delta;
+
+ delta = jiffies - rcp->jiffies_stall;
+ if (cpu_isset(smp_processor_id(), rcp->cpumask) && delta >= 0) {
+
+ /* We haven't checked in, so go dump stack. */
+ print_cpu_stall(rcp);
+
+ } else if (rcp->cur != rcp->completed && delta >= 2) {
+
+ /* They had two seconds to dump stack, so complain. */
+ print_other_cpu_stall(rcp);
+ }
+}
+
+#else /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
+
+static void record_gp_stall_check_time(struct rcu_ctrlblk *rcp)
+{
+}
+
+static inline void check_cpu_stall(struct rcu_ctrlblk *rcp)
+{
+}
+
+#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
+
/**
* call_rcu - Queue an RCU callback for invocation after a grace period.
* @head: structure to be used for queueing the RCU updates.
unsigned long flags;
head->func = func;
- head->next = NULL;
local_irq_save(flags);
__call_rcu(head, &rcu_ctrlblk, &__get_cpu_var(rcu_data));
local_irq_restore(flags);
unsigned long flags;
head->func = func;
- head->next = NULL;
local_irq_save(flags);
__call_rcu(head, &rcu_bh_ctrlblk, &__get_cpu_var(rcu_bh_data));
local_irq_restore(flags);
*/
static void rcu_do_batch(struct rcu_data *rdp)
{
+ unsigned long flags;
struct rcu_head *next, *list;
int count = 0;
}
rdp->donelist = list;
- local_irq_disable();
+ local_irq_save(flags);
rdp->qlen -= count;
- local_irq_enable();
+ local_irq_restore(flags);
if (rdp->blimit == INT_MAX && rdp->qlen <= qlowmark)
rdp->blimit = blimit;
* rcu_check_quiescent_state calls rcu_start_batch(0) to start the next grace
* period (if necessary).
*/
+
/*
* Register a new batch of callbacks, and start it up if there is currently no
* active batch and the batch to be registered has not already occurred.
if (rcp->cur != rcp->pending &&
rcp->completed == rcp->cur) {
rcp->cur++;
+ record_gp_stall_check_time(rcp);
/*
* Accessing nohz_cpu_mask before incrementing rcp->cur needs a
static void rcu_check_quiescent_state(struct rcu_ctrlblk *rcp,
struct rcu_data *rdp)
{
+ unsigned long flags;
+
if (rdp->quiescbatch != rcp->cur) {
/* start new grace period: */
rdp->qs_pending = 1;
return;
rdp->qs_pending = 0;
- spin_lock(&rcp->lock);
+ spin_lock_irqsave(&rcp->lock, flags);
/*
* rdp->quiescbatch/rcp->cur and the cpu bitmap can come out of sync
* during cpu startup. Ignore the quiescent state.
if (likely(rdp->quiescbatch == rcp->cur))
cpu_quiet(rdp->cpu, rcp);
- spin_unlock(&rcp->lock);
+ spin_unlock_irqrestore(&rcp->lock, flags);
}
static void rcu_move_batch(struct rcu_data *this_rdp, struct rcu_head *list,
struct rcu_head **tail, long batch)
{
+ unsigned long flags;
+
if (list) {
- local_irq_disable();
+ local_irq_save(flags);
this_rdp->batch = batch;
*this_rdp->nxttail[2] = list;
this_rdp->nxttail[2] = tail;
- local_irq_enable();
+ local_irq_restore(flags);
}
}
static void __rcu_offline_cpu(struct rcu_data *this_rdp,
struct rcu_ctrlblk *rcp, struct rcu_data *rdp)
{
- /* if the cpu going offline owns the grace period
+ unsigned long flags;
+
+ /*
+ * if the cpu going offline owns the grace period
* we can block indefinitely waiting for it, so flush
* it here
*/
- spin_lock_bh(&rcp->lock);
+ spin_lock_irqsave(&rcp->lock, flags);
if (rcp->cur != rcp->completed)
cpu_quiet(rdp->cpu, rcp);
- spin_unlock_bh(&rcp->lock);
- /* spin_lock implies smp_mb() */
rcu_move_batch(this_rdp, rdp->donelist, rdp->donetail, rcp->cur + 1);
rcu_move_batch(this_rdp, rdp->nxtlist, rdp->nxttail[2], rcp->cur + 1);
+ spin_unlock(&rcp->lock);
- local_irq_disable();
this_rdp->qlen += rdp->qlen;
- local_irq_enable();
+ local_irq_restore(flags);
}
static void rcu_offline_cpu(int cpu)
static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp,
struct rcu_data *rdp)
{
+ unsigned long flags;
+ long completed_snap;
+
if (rdp->nxtlist) {
- local_irq_disable();
+ local_irq_save(flags);
+ completed_snap = ACCESS_ONCE(rcp->completed);
/*
* move the other grace-period-completed entries to
* [rdp->nxtlist, *rdp->nxttail[0]) temporarily
*/
- if (!rcu_batch_before(rcp->completed, rdp->batch))
+ if (!rcu_batch_before(completed_snap, rdp->batch))
rdp->nxttail[0] = rdp->nxttail[1] = rdp->nxttail[2];
- else if (!rcu_batch_before(rcp->completed, rdp->batch - 1))
+ else if (!rcu_batch_before(completed_snap, rdp->batch - 1))
rdp->nxttail[0] = rdp->nxttail[1];
/*
rdp->nxttail[0] = &rdp->nxtlist;
}
- local_irq_enable();
+ local_irq_restore(flags);
if (rcu_batch_after(rdp->batch, rcp->pending)) {
+ unsigned long flags2;
+
/* and start it/schedule start if it's a new batch */
- spin_lock(&rcp->lock);
+ spin_lock_irqsave(&rcp->lock, flags2);
if (rcu_batch_after(rdp->batch, rcp->pending)) {
rcp->pending = rdp->batch;
rcu_start_batch(rcp);
}
- spin_unlock(&rcp->lock);
+ spin_unlock_irqrestore(&rcp->lock, flags2);
}
}
static void rcu_process_callbacks(struct softirq_action *unused)
{
+ /*
+ * Memory references from any prior RCU read-side critical sections
+ * executed by the interrupted code must be see before any RCU
+ * grace-period manupulations below.
+ */
+
+ smp_mb(); /* See above block comment. */
+
__rcu_process_callbacks(&rcu_ctrlblk, &__get_cpu_var(rcu_data));
__rcu_process_callbacks(&rcu_bh_ctrlblk, &__get_cpu_var(rcu_bh_data));
+
+ /*
+ * Memory references from any later RCU read-side critical sections
+ * executed by the interrupted code must be see after any RCU
+ * grace-period manupulations above.
+ */
+
+ smp_mb(); /* See above block comment. */
}
static int __rcu_pending(struct rcu_ctrlblk *rcp, struct rcu_data *rdp)
{
+ /* Check for CPU stalls, if enabled. */
+ check_cpu_stall(rcp);
+
if (rdp->nxtlist) {
+ long completed_snap = ACCESS_ONCE(rcp->completed);
+
/*
* This cpu has pending rcu entries and the grace period
* for them has completed.
*/
- if (!rcu_batch_before(rcp->completed, rdp->batch))
+ if (!rcu_batch_before(completed_snap, rdp->batch))
return 1;
- if (!rcu_batch_before(rcp->completed, rdp->batch - 1) &&
+ if (!rcu_batch_before(completed_snap, rdp->batch - 1) &&
rdp->nxttail[0] != rdp->nxttail[1])
return 1;
if (rdp->nxttail[0] != &rdp->nxtlist)
return !!rdp->nxtlist || !!rdp_bh->nxtlist || rcu_pending(cpu);
}
+/*
+ * Top-level function driving RCU grace-period detection, normally
+ * invoked from the scheduler-clock interrupt. This function simply
+ * increments counters that are read only from softirq by this same
+ * CPU, so there are no memory barriers required.
+ */
void rcu_check_callbacks(int cpu, int user)
{
if (user ||
static void rcu_init_percpu_data(int cpu, struct rcu_ctrlblk *rcp,
struct rcu_data *rdp)
{
+ unsigned long flags;
+
+ spin_lock_irqsave(&rcp->lock, flags);
memset(rdp, 0, sizeof(*rdp));
rdp->nxttail[0] = rdp->nxttail[1] = rdp->nxttail[2] = &rdp->nxtlist;
rdp->donetail = &rdp->donelist;
rdp->qs_pending = 0;
rdp->cpu = cpu;
rdp->blimit = blimit;
+ spin_unlock_irqrestore(&rcp->lock, flags);
}
static void __cpuinit rcu_online_cpu(int cpu)
*/
void __init __rcu_init(void)
{
+#ifdef CONFIG_RCU_CPU_STALL_DETECTOR
+ printk(KERN_INFO "RCU-based detection of stalled CPUs is enabled.\n");
+#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE,
(void *)(long)smp_processor_id());
/* Register notifier for non-boot CPUs */