Re: [ANNOUNCE] v4.11.5-rt1

From: Mike Galbraith
Date: Thu Jun 22 2017 - 13:30:25 EST


On Thu, 2017-06-22 at 18:34 +0200, Sebastian Andrzej Siewior wrote:
> On 2017-06-20 09:45:06 [+0200], Mike Galbraith wrote:
> > See ! and ?
>
> See see.
> What about this:

I'll give it a go, likely during the weekend.

I moved 4.11-rt today (also repros nicely) due to ftrace annoying me.
ÂAfter yet more staring at ever more huge traces (opposite of goal;),
then taking a break to stare at source again, I decided that the dual
wake_q business should die.. and the stall died with it.

> diff --git a/include/linux/sched.h b/include/linux/sched.h
> --- a/include/linux/sched.h
> +++ b/include/linux/sched.h
> @@ -1014,8 +1014,20 @@ struct wake_q_head {
> #define WAKE_Q(name) \
> struct wake_q_head name = { WAKE_Q_TAIL, &name.first }
>
> -extern void wake_q_add(struct wake_q_head *head,
> - struct task_struct *task);
> +extern void __wake_q_add(struct wake_q_head *head,
> + struct task_struct *task, bool sleeper);
> +static inline void wake_q_add(struct wake_q_head *head,
> + struct task_struct *task)
> +{
> + __wake_q_add(head, task, false);
> +}
> +
> +static inline void wake_q_add_sleeper(struct wake_q_head *head,
> + struct task_struct *task)
> +{
> + __wake_q_add(head, task, true);
> +}
> +
> extern void __wake_up_q(struct wake_q_head *head, bool sleeper);
>
> static inline void wake_up_q(struct wake_q_head *head)
> @@ -1745,6 +1757,7 @@ struct task_struct {
> raw_spinlock_t pi_lock;
>
> struct wake_q_node wake_q;
> + struct wake_q_node wake_q_sleeper;
>
> #ifdef CONFIG_RT_MUTEXES
> /* PI waiters blocked on a rt_mutex held by this task */
> diff --git a/kernel/fork.c b/kernel/fork.c
> --- a/kernel/fork.c
> +++ b/kernel/fork.c
> @@ -558,6 +558,7 @@ static struct task_struct *dup_task_struct(struct task_struct *orig, int node)
> tsk->splice_pipe = NULL;
> tsk->task_frag.page = NULL;
> tsk->wake_q.next = NULL;
> + tsk->wake_q_sleeper.next = NULL;
>
> account_kernel_stack(tsk, 1);
>
> diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
> --- a/kernel/locking/rtmutex.c
> +++ b/kernel/locking/rtmutex.c
> @@ -1506,7 +1506,7 @@ static void mark_wakeup_next_waiter(struct wake_q_head *wake_q,
> */
> preempt_disable();
> if (waiter->savestate)
> - wake_q_add(wake_sleeper_q, waiter->task);
> + wake_q_add_sleeper(wake_sleeper_q, waiter->task);
> else
> wake_q_add(wake_q, waiter->task);
> raw_spin_unlock(&current->pi_lock);
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -430,9 +430,15 @@ static bool set_nr_if_polling(struct task_struct *p)
> #endif
> #endif
>
> -void wake_q_add(struct wake_q_head *head, struct task_struct *task)
> +void __wake_q_add(struct wake_q_head *head, struct task_struct *task,
> + bool sleeper)
> {
> - struct wake_q_node *node = &task->wake_q;
> + struct wake_q_node *node;
> +
> + if (sleeper)
> + node = &task->wake_q_sleeper;
> + else
> + node = &task->wake_q;
>
> /*
> * Atomically grab the task, if ->wake_q is !nil already it means
> @@ -461,11 +467,17 @@ void __wake_up_q(struct wake_q_head *head, bool sleeper)
> while (node != WAKE_Q_TAIL) {
> struct task_struct *task;
>
> - task = container_of(node, struct task_struct, wake_q);
> + if (sleeper)
> + task = container_of(node, struct task_struct, wake_q_sleeper);
> + else
> + task = container_of(node, struct task_struct, wake_q);
> BUG_ON(!task);
> /* task can safely be re-inserted now */
> node = node->next;
> - task->wake_q.next = NULL;
> + if (sleeper)
> + task->wake_q_sleeper.next = NULL;
> + else
> + task->wake_q.next = NULL;
>
> /*
> * wake_up_process() implies a wmb() to pair with the queueing