Re: [PATCH 3/3] rcu: use simple wait queues where possible in rcutree

From: Paul E. McKenney
Date: Wed Dec 11 2013 - 22:42:54 EST

Next message: Joe Perches: "[PATCH -next 0/2] seq_file/netfilter: Start removing returns from seq_<foo>"
Previous message: Sasha Levin: "Re: kernel BUG in munlock_vma_pages_range"
In reply to: Paul Gortmaker: "[PATCH 3/3] rcu: use simple wait queues where possible in rcutree"
Next in thread: Paul Gortmaker: "[PATCH 2/3] sched/core: convert completions to use simple wait queues"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, Dec 11, 2013 at 08:06:39PM -0500, Paul Gortmaker wrote:
> From: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
>
> As of commit dae6e64d2bcfd4b06304ab864c7e3a4f6b5fedf4 ("rcu: Introduce
> proper blocking to no-CBs kthreads GP waits") the rcu subsystem started
> making use of wait queues.
>
> Here we convert all additions of rcu wait queues to use simple wait queues,
> since they don't need the extra overhead of the full wait queue features.
>
> Originally this was done for RT kernels, since we would get things like...
>
> BUG: sleeping function called from invalid context at kernel/rtmutex.c:659
> in_atomic(): 1, irqs_disabled(): 1, pid: 8, name: rcu_preempt
> Pid: 8, comm: rcu_preempt Not tainted
> Call Trace:
> [<ffffffff8106c8d0>] __might_sleep+0xd0/0xf0
> [<ffffffff817d77b4>] rt_spin_lock+0x24/0x50
> [<ffffffff8106fcf6>] __wake_up+0x36/0x70
> [<ffffffff810c4542>] rcu_gp_kthread+0x4d2/0x680
> [<ffffffff8105f910>] ? __init_waitqueue_head+0x50/0x50
> [<ffffffff810c4070>] ? rcu_gp_fqs+0x80/0x80
> [<ffffffff8105eabb>] kthread+0xdb/0xe0
> [<ffffffff8106b912>] ? finish_task_switch+0x52/0x100
> [<ffffffff817e0754>] kernel_thread_helper+0x4/0x10
> [<ffffffff8105e9e0>] ? __init_kthread_worker+0x60/0x60
> [<ffffffff817e0750>] ? gs_change+0xb/0xb
>
> ...and hence simple wait queues were deployed on RT out of necessity
> (as simple wait uses a raw lock), but mainline might as well take
> advantage of the more streamline support as well.
>
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> Signed-off-by: Steven Rostedt <rostedt@xxxxxxxxxxx>
> [PG: adapt from multiple v3.10-rt patches and add a commit log.]
> Signed-off-by: Paul Gortmaker <paul.gortmaker@xxxxxxxxxxxxx>

You got the swake_up_all() this time, so:

Reviewed-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>

;-)

> ---
> kernel/rcu/tree.c | 16 ++++++++--------
> kernel/rcu/tree.h | 7 ++++---
> kernel/rcu/tree_plugin.h | 14 +++++++-------
> 3 files changed, 19 insertions(+), 18 deletions(-)
>
> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index dd08198..b35babb 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -1550,9 +1550,9 @@ static int __noreturn rcu_gp_kthread(void *arg)
> trace_rcu_grace_period(rsp->name,
> ACCESS_ONCE(rsp->gpnum),
> TPS("reqwait"));
> - wait_event_interruptible(rsp->gp_wq,
> - ACCESS_ONCE(rsp->gp_flags) &
> - RCU_GP_FLAG_INIT);
> + swait_event_interruptible(rsp->gp_wq,
> + ACCESS_ONCE(rsp->gp_flags) &
> + RCU_GP_FLAG_INIT);
> if (rcu_gp_init(rsp))
> break;
> cond_resched();
> @@ -1576,7 +1576,7 @@ static int __noreturn rcu_gp_kthread(void *arg)
> trace_rcu_grace_period(rsp->name,
> ACCESS_ONCE(rsp->gpnum),
> TPS("fqswait"));
> - ret = wait_event_interruptible_timeout(rsp->gp_wq,
> + ret = swait_event_interruptible_timeout(rsp->gp_wq,
> ((gf = ACCESS_ONCE(rsp->gp_flags)) &
> RCU_GP_FLAG_FQS) ||
> (!ACCESS_ONCE(rnp->qsmask) &&
> @@ -1625,7 +1625,7 @@ static void rsp_wakeup(struct irq_work *work)
> struct rcu_state *rsp = container_of(work, struct rcu_state, wakeup_work);
>
> /* Wake up rcu_gp_kthread() to start the grace period. */
> - wake_up(&rsp->gp_wq);
> + swake_up(&rsp->gp_wq);
> }
>
> /*
> @@ -1701,7 +1701,7 @@ static void rcu_report_qs_rsp(struct rcu_state *rsp, unsigned long flags)
> {
> WARN_ON_ONCE(!rcu_gp_in_progress(rsp));
> raw_spin_unlock_irqrestore(&rcu_get_root(rsp)->lock, flags);
> - wake_up(&rsp->gp_wq); /* Memory barrier implied by wake_up() path. */
> + swake_up(&rsp->gp_wq); /* Memory barrier implied by swake_up() path. */
> }
>
> /*
> @@ -2271,7 +2271,7 @@ static void force_quiescent_state(struct rcu_state *rsp)
> }
> rsp->gp_flags |= RCU_GP_FLAG_FQS;
> raw_spin_unlock_irqrestore(&rnp_old->lock, flags);
> - wake_up(&rsp->gp_wq); /* Memory barrier implied by wake_up() path. */
> + swake_up(&rsp->gp_wq); /* Memory barrier implied by swake_up() path. */
> }
>
> /*
> @@ -3304,7 +3304,7 @@ static void __init rcu_init_one(struct rcu_state *rsp,
> }
>
> rsp->rda = rda;
> - init_waitqueue_head(&rsp->gp_wq);
> + init_swaitqueue_head(&rsp->gp_wq);
> init_irq_work(&rsp->wakeup_work, rsp_wakeup);
> rnp = rsp->level[rcu_num_lvls - 1];
> for_each_possible_cpu(i) {
> diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h
> index 52be957..01476e1 100644
> --- a/kernel/rcu/tree.h
> +++ b/kernel/rcu/tree.h
> @@ -28,6 +28,7 @@
> #include <linux/cpumask.h>
> #include <linux/seqlock.h>
> #include <linux/irq_work.h>
> +#include <linux/swait.h>
>
> /*
> * Define shape of hierarchy based on NR_CPUS, CONFIG_RCU_FANOUT, and
> @@ -200,7 +201,7 @@ struct rcu_node {
> /* This can happen due to race conditions. */
> #endif /* #ifdef CONFIG_RCU_BOOST */
> #ifdef CONFIG_RCU_NOCB_CPU
> - wait_queue_head_t nocb_gp_wq[2];
> + swait_queue_head_t nocb_gp_wq[2];
> /* Place for rcu_nocb_kthread() to wait GP. */
> #endif /* #ifdef CONFIG_RCU_NOCB_CPU */
> int need_future_gp[2];
> @@ -333,7 +334,7 @@ struct rcu_data {
> atomic_long_t nocb_q_count_lazy; /* (approximate). */
> int nocb_p_count; /* # CBs being invoked by kthread */
> int nocb_p_count_lazy; /* (approximate). */
> - wait_queue_head_t nocb_wq; /* For nocb kthreads to sleep on. */
> + swait_queue_head_t nocb_wq; /* For nocb kthreads to sleep on. */
> struct task_struct *nocb_kthread;
> #endif /* #ifdef CONFIG_RCU_NOCB_CPU */
>
> @@ -403,7 +404,7 @@ struct rcu_state {
> unsigned long gpnum; /* Current gp number. */
> unsigned long completed; /* # of last completed gp. */
> struct task_struct *gp_kthread; /* Task for grace periods. */
> - wait_queue_head_t gp_wq; /* Where GP task waits. */
> + swait_queue_head_t gp_wq; /* Where GP task waits. */
> int gp_flags; /* Commands for GP task. */
>
> /* End of fields guarded by root rcu_node's lock. */
> diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> index 08a7652..b565ebd 100644
> --- a/kernel/rcu/tree_plugin.h
> +++ b/kernel/rcu/tree_plugin.h
> @@ -2060,7 +2060,7 @@ static int rcu_nocb_needs_gp(struct rcu_state *rsp)
> */
> static void rcu_nocb_gp_cleanup(struct rcu_state *rsp, struct rcu_node *rnp)
> {
> - wake_up_all(&rnp->nocb_gp_wq[rnp->completed & 0x1]);
> + swake_up_all(&rnp->nocb_gp_wq[rnp->completed & 0x1]);
> }
>
> /*
> @@ -2078,8 +2078,8 @@ static void rcu_nocb_gp_set(struct rcu_node *rnp, int nrq)
>
> static void rcu_init_one_nocb(struct rcu_node *rnp)
> {
> - init_waitqueue_head(&rnp->nocb_gp_wq[0]);
> - init_waitqueue_head(&rnp->nocb_gp_wq[1]);
> + init_swaitqueue_head(&rnp->nocb_gp_wq[0]);
> + init_swaitqueue_head(&rnp->nocb_gp_wq[1]);
> }
>
> /* Is the specified CPU a no-CPUs CPU? */
> @@ -2122,7 +2122,7 @@ static void __call_rcu_nocb_enqueue(struct rcu_data *rdp,
> }
> len = atomic_long_read(&rdp->nocb_q_count);
> if (old_rhpp == &rdp->nocb_head) {
> - wake_up(&rdp->nocb_wq); /* ... only if queue was empty ... */
> + swake_up(&rdp->nocb_wq); /* ... only if queue was empty ... */
> rdp->qlen_last_fqs_check = 0;
> trace_rcu_nocb_wake(rdp->rsp->name, rdp->cpu, TPS("WakeEmpty"));
> } else if (len > rdp->qlen_last_fqs_check + qhimark) {
> @@ -2218,7 +2218,7 @@ static void rcu_nocb_wait_gp(struct rcu_data *rdp)
> */
> trace_rcu_future_gp(rnp, rdp, c, TPS("StartWait"));
> for (;;) {
> - wait_event_interruptible(
> + swait_event_interruptible(
> rnp->nocb_gp_wq[c & 0x1],
> (d = ULONG_CMP_GE(ACCESS_ONCE(rnp->completed), c)));
> if (likely(d))
> @@ -2249,7 +2249,7 @@ static int rcu_nocb_kthread(void *arg)
> if (!rcu_nocb_poll) {
> trace_rcu_nocb_wake(rdp->rsp->name, rdp->cpu,
> TPS("Sleep"));
> - wait_event_interruptible(rdp->nocb_wq, rdp->nocb_head);
> + swait_event_interruptible(rdp->nocb_wq, rdp->nocb_head);
> } else if (firsttime) {
> firsttime = 0;
> trace_rcu_nocb_wake(rdp->rsp->name, rdp->cpu,
> @@ -2314,7 +2314,7 @@ static int rcu_nocb_kthread(void *arg)
> static void __init rcu_boot_init_nocb_percpu_data(struct rcu_data *rdp)
> {
> rdp->nocb_tail = &rdp->nocb_head;
> - init_waitqueue_head(&rdp->nocb_wq);
> + init_swaitqueue_head(&rdp->nocb_wq);
> }
>
> /* Create a kthread for each RCU flavor for each no-CBs CPU. */
> --
> 1.8.5.1
>

--
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/

Next message: Joe Perches: "[PATCH -next 0/2] seq_file/netfilter: Start removing returns from seq_<foo>"
Previous message: Sasha Levin: "Re: kernel BUG in munlock_vma_pages_range"
In reply to: Paul Gortmaker: "[PATCH 3/3] rcu: use simple wait queues where possible in rcutree"
Next in thread: Paul Gortmaker: "[PATCH 2/3] sched/core: convert completions to use simple wait queues"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]