lkml.org 
[lkml]   [2018]   [Apr]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: cpu stopper threads and load balancing leads to deadlock
On Tue, Apr 17, 2018 at 03:21:19PM +0100, Matt Fleming wrote:
> Hi guys,
>
> We've seen a bug in one of our SLE kernels where the cpu stopper
> thread ("migration/15") is entering idle balance. This then triggers
> active load balance.
>
> At the same time, a task on another CPU triggers a page fault and NUMA
> balancing kicks in to try and migrate the task closer to the NUMA node
> for that page (we're inside stop_two_cpus()). This faulting task is
> spinning in try_to_wake_up() (inside smp_cond_load_acquire(&p->on_cpu,
> !VAL)), waiting for "migration/15" to context switch.
>
> Unfortunately, because "migration/15" is doing active load balance
> it's spinning waiting for the NUMA-page-faulting CPU's stopper lock,
> which is already held (since it's inside stop_two_cpus()).
>
> Deadlock ensues.


So if I read that right, something like the following happens:

CPU0 CPU1

schedule(.prev=migrate/0) <fault>
pick_next_task ...
idle_balance migrate_swap()
active_balance stop_two_cpus()
spin_lock(stopper0->lock)
spin_lock(stopper1->lock)
ttwu(migrate/0)
smp_cond_load_acquire() -- waits for schedule()
stop_one_cpu(1)
spin_lock(stopper1->lock) -- waits for stopper lock


Fix _this_ deadlock by taking out the wakeups from under stopper->lock.
I'm not entirely sure there isn't more dragons here, but this particular
one seems fixable by doing that.

Is there any way you can reproduce/test this?

Maybe-signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
kernel/stop_machine.c | 19 ++++++++++++++-----
1 file changed, 14 insertions(+), 5 deletions(-)

diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index b7591261652d..64c0291b579c 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -21,6 +21,7 @@
#include <linux/smpboot.h>
#include <linux/atomic.h>
#include <linux/nmi.h>
+#include <linux/sched/wake_q.h>

/*
* Structure to determine completion condition and record errors. May
@@ -65,27 +66,31 @@ static void cpu_stop_signal_done(struct cpu_stop_done *done)
}

static void __cpu_stop_queue_work(struct cpu_stopper *stopper,
- struct cpu_stop_work *work)
+ struct cpu_stop_work *work,
+ struct wake_q_head *wakeq)
{
list_add_tail(&work->list, &stopper->works);
- wake_up_process(stopper->thread);
+ wake_q_add(wakeq, stopper->thread);
}

/* queue @work to @stopper. if offline, @work is completed immediately */
static bool cpu_stop_queue_work(unsigned int cpu, struct cpu_stop_work *work)
{
struct cpu_stopper *stopper = &per_cpu(cpu_stopper, cpu);
+ DEFINE_WAKE_Q(wakeq);
unsigned long flags;
bool enabled;

spin_lock_irqsave(&stopper->lock, flags);
enabled = stopper->enabled;
if (enabled)
- __cpu_stop_queue_work(stopper, work);
+ __cpu_stop_queue_work(stopper, work, &wakeq);
else if (work->done)
cpu_stop_signal_done(work->done);
spin_unlock_irqrestore(&stopper->lock, flags);

+ wake_up_q(&wakeq);
+
return enabled;
}

@@ -229,6 +234,7 @@ static int cpu_stop_queue_two_works(int cpu1, struct cpu_stop_work *work1,
{
struct cpu_stopper *stopper1 = per_cpu_ptr(&cpu_stopper, cpu1);
struct cpu_stopper *stopper2 = per_cpu_ptr(&cpu_stopper, cpu2);
+ DEFINE_WAKE_Q(wakeq);
int err;
retry:
spin_lock_irq(&stopper1->lock);
@@ -252,8 +258,8 @@ static int cpu_stop_queue_two_works(int cpu1, struct cpu_stop_work *work1,
goto unlock;

err = 0;
- __cpu_stop_queue_work(stopper1, work1);
- __cpu_stop_queue_work(stopper2, work2);
+ __cpu_stop_queue_work(stopper1, work1, &wakeq);
+ __cpu_stop_queue_work(stopper2, work2, &wakeq);
unlock:
spin_unlock(&stopper2->lock);
spin_unlock_irq(&stopper1->lock);
@@ -263,6 +269,9 @@ static int cpu_stop_queue_two_works(int cpu1, struct cpu_stop_work *work1,
cpu_relax();
goto retry;
}
+
+ wake_up_q(&wakeq);
+
return err;
}
/**
\
 
 \ /
  Last update: 2018-04-20 11:51    [W:0.112 / U:0.236 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site