1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
|
Subject: sched: Move mmdrop to RCU on RT
From: Thomas Gleixner <tglx@linutronix.de>
Date: Mon, 06 Jun 2011 12:20:33 +0200
Origin: https://www.kernel.org/pub/linux/kernel/projects/rt/5.2/older/patches-5.2.17-rt9.tar.xz
Takes sleeping locks and calls into the memory allocator, so nothing
we want to do in task switch and oder atomic contexts.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
include/linux/mm_types.h | 4 ++++
include/linux/sched/mm.h | 11 +++++++++++
kernel/fork.c | 13 +++++++++++++
kernel/sched/core.c | 18 ++++++++++++++++--
4 files changed, 44 insertions(+), 2 deletions(-)
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -12,6 +12,7 @@
#include <linux/completion.h>
#include <linux/cpumask.h>
#include <linux/uprobes.h>
+#include <linux/rcupdate.h>
#include <linux/page-flags-layout.h>
#include <linux/workqueue.h>
@@ -496,6 +497,9 @@ struct mm_struct {
bool tlb_flush_batched;
#endif
struct uprobes_state uprobes_state;
+#ifdef CONFIG_PREEMPT_RT_BASE
+ struct rcu_head delayed_drop;
+#endif
#ifdef CONFIG_HUGETLB_PAGE
atomic_long_t hugetlb_usage;
#endif
--- a/include/linux/sched/mm.h
+++ b/include/linux/sched/mm.h
@@ -49,6 +49,17 @@ static inline void mmdrop(struct mm_stru
__mmdrop(mm);
}
+#ifdef CONFIG_PREEMPT_RT_BASE
+extern void __mmdrop_delayed(struct rcu_head *rhp);
+static inline void mmdrop_delayed(struct mm_struct *mm)
+{
+ if (atomic_dec_and_test(&mm->mm_count))
+ call_rcu(&mm->delayed_drop, __mmdrop_delayed);
+}
+#else
+# define mmdrop_delayed(mm) mmdrop(mm)
+#endif
+
/*
* This has to be called after a get_task_mm()/mmget_not_zero()
* followed by taking the mmap_sem for writing before modifying the
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -685,6 +685,19 @@ void __mmdrop(struct mm_struct *mm)
}
EXPORT_SYMBOL_GPL(__mmdrop);
+#ifdef CONFIG_PREEMPT_RT_BASE
+/*
+ * RCU callback for delayed mm drop. Not strictly rcu, but we don't
+ * want another facility to make this work.
+ */
+void __mmdrop_delayed(struct rcu_head *rhp)
+{
+ struct mm_struct *mm = container_of(rhp, struct mm_struct, delayed_drop);
+
+ __mmdrop(mm);
+}
+#endif
+
static void mmdrop_async_fn(struct work_struct *work)
{
struct mm_struct *mm;
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2709,9 +2709,13 @@ static struct rq *finish_task_switch(str
* provided by mmdrop(),
* - a sync_core for SYNC_CORE.
*/
+ /*
+ * We use mmdrop_delayed() here so we don't have to do the
+ * full __mmdrop() when we are the last user.
+ */
if (mm) {
membarrier_mm_sync_core_before_usermode(mm);
- mmdrop(mm);
+ mmdrop_delayed(mm);
}
if (unlikely(prev_state == TASK_DEAD)) {
if (prev->sched_class->task_dead)
@@ -5539,6 +5543,8 @@ void sched_setnuma(struct task_struct *p
#endif /* CONFIG_NUMA_BALANCING */
#ifdef CONFIG_HOTPLUG_CPU
+static DEFINE_PER_CPU(struct mm_struct *, idle_last_mm);
+
/*
* Ensure that the idle task is using init_mm right before its CPU goes
* offline.
@@ -5554,7 +5560,11 @@ void idle_task_exit(void)
current->active_mm = &init_mm;
finish_arch_post_lock_switch();
}
- mmdrop(mm);
+ /*
+ * Defer the cleanup to an alive cpu. On RT we can neither
+ * call mmdrop() nor mmdrop_delayed() from here.
+ */
+ per_cpu(idle_last_mm, smp_processor_id()) = mm;
}
/*
@@ -5866,6 +5876,10 @@ int sched_cpu_dying(unsigned int cpu)
update_max_interval();
nohz_balance_exit_idle(rq);
hrtick_clear(rq);
+ if (per_cpu(idle_last_mm, cpu)) {
+ mmdrop_delayed(per_cpu(idle_last_mm, cpu));
+ per_cpu(idle_last_mm, cpu) = NULL;
+ }
return 0;
}
#endif
|
