From b668c9cf3e58739dac54a1d6f42f2b4bdd980b3e Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sun, 22 Nov 2009 08:53:48 -0800 Subject: rcu: Fix grace-period-stall bug on large systems with CPU hotplug When the last CPU of a given leaf rcu_node structure goes offline, all of the tasks queued on that leaf rcu_node structure (due to having blocked in their current RCU read-side critical sections) are requeued onto the root rcu_node structure. This requeuing is carried out by rcu_preempt_offline_tasks(). However, it is possible that these queued tasks are the only thing preventing the leaf rcu_node structure from reporting a quiescent state up the rcu_node hierarchy. Unfortunately, the old code would fail to do this reporting, resulting in a grace-period stall given the following sequence of events: 1. Kernel built for more than 32 CPUs on 32-bit systems or for more than 64 CPUs on 64-bit systems, so that there is more than one rcu_node structure. (Or CONFIG_RCU_FANOUT is artificially set to a number smaller than CONFIG_NR_CPUS.) 2. The kernel is built with CONFIG_TREE_PREEMPT_RCU. 3. A task running on a CPU associated with a given leaf rcu_node structure blocks while in an RCU read-side critical section -and- that CPU has not yet passed through a quiescent state for the current RCU grace period. This will cause the task to be queued on the leaf rcu_node's blocked_tasks[] array, in particular, on the element of this array corresponding to the current grace period. 4. Each of the remaining CPUs corresponding to this same leaf rcu_node structure pass through a quiescent state. However, the task is still in its RCU read-side critical section, so these quiescent states cannot be reported further up the rcu_node hierarchy. Nevertheless, all bits in the leaf rcu_node structure's ->qsmask field are now zero. 5. Each of the remaining CPUs go offline. (The events in step #4 and #5 can happen in any order as long as each CPU passes through a quiescent state before going offline.) 6. When the last CPU goes offline, __rcu_offline_cpu() will invoke rcu_preempt_offline_tasks(), which will move the task to the root rcu_node structure, but without reporting a quiescent state up the rcu_node hierarchy (and this failure to report a quiescent state is the bug). But because this leaf rcu_node structure's ->qsmask field is already zero and its ->block_tasks[] entries are all empty, force_quiescent_state() will skip this rcu_node structure. Therefore, grace periods are now hung. This patch abstracts some code out of rcu_read_unlock_special(), calling the result task_quiet() by analogy with cpu_quiet(), and invokes task_quiet() from both rcu_read_lock_special() and __rcu_offline_cpu(). Invoking task_quiet() from __rcu_offline_cpu() reports the quiescent state up the rcu_node hierarchy, fixing the bug. This ends up requiring a separate lock_class_key per level of the rcu_node hierarchy, which this patch also provides. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12589088301770-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree_plugin.h | 85 +++++++++++++++++++++++++++++++++++-------------- 1 file changed, 61 insertions(+), 24 deletions(-) (limited to 'kernel/rcutree_plugin.h') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 5ca2d26..0bdb592 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -160,11 +160,51 @@ static int rcu_preempted_readers(struct rcu_node *rnp) return !list_empty(&rnp->blocked_tasks[rnp->gpnum & 0x1]); } +/* + * Record a quiescent state for all tasks that were previously queued + * on the specified rcu_node structure and that were blocking the current + * RCU grace period. The caller must hold the specified rnp->lock with + * irqs disabled, and this lock is released upon return, but irqs remain + * disabled. + */ +static void task_quiet(struct rcu_node *rnp, unsigned long flags) + __releases(rnp->lock) +{ + unsigned long mask; + struct rcu_node *rnp_p; + + if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) { + spin_unlock_irqrestore(&rnp->lock, flags); + return; /* Still need more quiescent states! */ + } + + rnp_p = rnp->parent; + if (rnp_p == NULL) { + /* + * Either there is only one rcu_node in the tree, + * or tasks were kicked up to root rcu_node due to + * CPUs going offline. + */ + cpu_quiet_msk_finish(&rcu_preempt_state, flags); + return; + } + + /* Report up the rest of the hierarchy. */ + mask = rnp->grpmask; + spin_unlock(&rnp->lock); /* irqs remain disabled. */ + spin_lock(&rnp_p->lock); /* irqs already disabled. */ + cpu_quiet_msk(mask, &rcu_preempt_state, rnp_p, flags); +} + +/* + * Handle special cases during rcu_read_unlock(), such as needing to + * notify RCU core processing or task having blocked during the RCU + * read-side critical section. + */ static void rcu_read_unlock_special(struct task_struct *t) { int empty; unsigned long flags; - unsigned long mask; struct rcu_node *rnp; int special; @@ -213,30 +253,15 @@ static void rcu_read_unlock_special(struct task_struct *t) /* * If this was the last task on the current list, and if * we aren't waiting on any CPUs, report the quiescent state. - * Note that both cpu_quiet_msk_finish() and cpu_quiet_msk() - * drop rnp->lock and restore irq. + * Note that task_quiet() releases rnp->lock. */ - if (!empty && rnp->qsmask == 0 && - !rcu_preempted_readers(rnp)) { - struct rcu_node *rnp_p; - - if (rnp->parent == NULL) { - /* Only one rcu_node in the tree. */ - cpu_quiet_msk_finish(&rcu_preempt_state, flags); - return; - } - /* Report up the rest of the hierarchy. */ - mask = rnp->grpmask; + if (empty) spin_unlock_irqrestore(&rnp->lock, flags); - rnp_p = rnp->parent; - spin_lock_irqsave(&rnp_p->lock, flags); - WARN_ON_ONCE(rnp->qsmask); - cpu_quiet_msk(mask, &rcu_preempt_state, rnp_p, flags); - return; - } - spin_unlock(&rnp->lock); + else + task_quiet(rnp, flags); + } else { + local_irq_restore(flags); } - local_irq_restore(flags); } /* @@ -303,6 +328,8 @@ static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp) * rcu_node. The reason for not just moving them to the immediate * parent is to remove the need for rcu_read_unlock_special() to * make more than two attempts to acquire the target rcu_node's lock. + * Returns true if there were tasks blocking the current RCU grace + * period. * * Returns 1 if there was previously a task blocking the current grace * period on the specified rcu_node structure. @@ -316,7 +343,7 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, int i; struct list_head *lp; struct list_head *lp_root; - int retval = rcu_preempted_readers(rnp); + int retval; struct rcu_node *rnp_root = rcu_get_root(rsp); struct task_struct *tp; @@ -334,6 +361,7 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, * rcu_nodes in terms of gp_num value. This fact allows us to * move the blocked_tasks[] array directly, element by element. */ + retval = rcu_preempted_readers(rnp); for (i = 0; i < 2; i++) { lp = &rnp->blocked_tasks[i]; lp_root = &rnp_root->blocked_tasks[i]; @@ -346,7 +374,6 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, spin_unlock(&rnp_root->lock); /* irqs remain disabled */ } } - return retval; } @@ -512,6 +539,16 @@ static int rcu_preempted_readers(struct rcu_node *rnp) return 0; } +#ifdef CONFIG_HOTPLUG_CPU + +/* Because preemptible RCU does not exist, no quieting of tasks. */ +static void task_quiet(struct rcu_node *rnp, unsigned long flags) +{ + spin_unlock_irqrestore(&rnp->lock, flags); +} + +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ + #ifdef CONFIG_RCU_CPU_STALL_DETECTOR /* -- cgit v1.1