aboutsummaryrefslogtreecommitdiffstats
path: root/include
diff options
context:
space:
mode:
Diffstat (limited to 'include')
-rw-r--r--include/asm-generic/barrier.h4
-rw-r--r--include/asm-generic/cmpxchg.h3
-rw-r--r--include/asm-generic/futex.h7
-rw-r--r--include/asm-generic/io.h17
-rw-r--r--include/asm-generic/iomap.h4
-rw-r--r--include/asm-generic/pgtable.h4
-rw-r--r--include/asm-generic/preempt.h7
-rw-r--r--include/asm-generic/qspinlock.h139
-rw-r--r--include/asm-generic/qspinlock_types.h79
-rw-r--r--include/linux/alarmtimer.h4
-rw-r--r--include/linux/backing-dev.h1
-rw-r--r--include/linux/bottom_half.h1
-rw-r--r--include/linux/brcmphy.h2
-rw-r--r--include/linux/clockchips.h37
-rw-r--r--include/linux/clocksource.h1
-rw-r--r--include/linux/compiler.h20
-rw-r--r--include/linux/context_tracking.h10
-rw-r--r--include/linux/context_tracking_state.h1
-rw-r--r--include/linux/cpumask.h6
-rw-r--r--include/linux/debugfs.h1
-rw-r--r--include/linux/dmar.h85
-rw-r--r--include/linux/efi.h12
-rw-r--r--include/linux/fs.h16
-rw-r--r--include/linux/hardirq.h2
-rw-r--r--include/linux/highmem.h2
-rw-r--r--include/linux/hrtimer.h167
-rw-r--r--include/linux/htirq.h22
-rw-r--r--include/linux/init_task.h5
-rw-r--r--include/linux/intel-iommu.h8
-rw-r--r--include/linux/interrupt.h9
-rw-r--r--include/linux/io-mapping.h2
-rw-r--r--include/linux/io.h8
-rw-r--r--include/linux/jiffies.h130
-rw-r--r--include/linux/kernel.h3
-rw-r--r--include/linux/ktime.h27
-rw-r--r--include/linux/lglock.h5
-rw-r--r--include/linux/lockdep.h4
-rw-r--r--include/linux/namei.h41
-rw-r--r--include/linux/of.h6
-rw-r--r--include/linux/osq_lock.h5
-rw-r--r--include/linux/percpu_counter.h13
-rw-r--r--include/linux/perf_event.h50
-rw-r--r--include/linux/preempt.h159
-rw-r--r--include/linux/preempt_mask.h117
-rw-r--r--include/linux/rculist.h10
-rw-r--r--include/linux/rcupdate.h76
-rw-r--r--include/linux/rcutiny.h16
-rw-r--r--include/linux/rcutree.h9
-rw-r--r--include/linux/sched.h145
-rw-r--r--include/linux/sched/sysctl.h12
-rw-r--r--include/linux/security.h13
-rw-r--r--include/linux/seqlock.h47
-rw-r--r--include/linux/spinlock.h2
-rw-r--r--include/linux/tick.h7
-rw-r--r--include/linux/time64.h2
-rw-r--r--include/linux/timekeeper_internal.h19
-rw-r--r--include/linux/timekeeping.h2
-rw-r--r--include/linux/timer.h63
-rw-r--r--include/linux/timerqueue.h8
-rw-r--r--include/linux/topology.h6
-rw-r--r--include/linux/uaccess.h48
-rw-r--r--include/linux/wait.h17
-rw-r--r--include/net/inet_connection_sock.h3
-rw-r--r--include/net/mac80211.h7
-rw-r--r--include/net/sctp/sctp.h7
-rw-r--r--include/sound/hda_regmap.h2
-rw-r--r--include/target/target_core_backend.h8
-rw-r--r--include/target/target_core_configfs.h2
-rw-r--r--include/target/target_core_fabric.h4
-rw-r--r--include/trace/events/kmem.h54
-rw-r--r--include/trace/events/sched.h3
-rw-r--r--include/trace/events/timer.h12
-rw-r--r--include/trace/events/writeback.h1
-rw-r--r--include/uapi/drm/radeon_drm.h1
-rw-r--r--include/uapi/linux/perf_event.h18
-rw-r--r--include/uapi/linux/virtio_balloon.h1
76 files changed, 1331 insertions, 540 deletions
diff --git a/include/asm-generic/barrier.h b/include/asm-generic/barrier.h
index f5c40b0fadc2..e6a83d712ef6 100644
--- a/include/asm-generic/barrier.h
+++ b/include/asm-generic/barrier.h
@@ -66,8 +66,8 @@
66#define smp_read_barrier_depends() do { } while (0) 66#define smp_read_barrier_depends() do { } while (0)
67#endif 67#endif
68 68
69#ifndef set_mb 69#ifndef smp_store_mb
70#define set_mb(var, value) do { (var) = (value); mb(); } while (0) 70#define smp_store_mb(var, value) do { WRITE_ONCE(var, value); mb(); } while (0)
71#endif 71#endif
72 72
73#ifndef smp_mb__before_atomic 73#ifndef smp_mb__before_atomic
diff --git a/include/asm-generic/cmpxchg.h b/include/asm-generic/cmpxchg.h
index 811fb1e9b061..3766ab34aa45 100644
--- a/include/asm-generic/cmpxchg.h
+++ b/include/asm-generic/cmpxchg.h
@@ -86,9 +86,6 @@ unsigned long __xchg(unsigned long x, volatile void *ptr, int size)
86 86
87/* 87/*
88 * Atomic compare and exchange. 88 * Atomic compare and exchange.
89 *
90 * Do not define __HAVE_ARCH_CMPXCHG because we want to use it to check whether
91 * a cmpxchg primitive faster than repeated local irq save/restore exists.
92 */ 89 */
93#include <asm-generic/cmpxchg-local.h> 90#include <asm-generic/cmpxchg-local.h>
94 91
diff --git a/include/asm-generic/futex.h b/include/asm-generic/futex.h
index b59b5a52637e..e56272c919b5 100644
--- a/include/asm-generic/futex.h
+++ b/include/asm-generic/futex.h
@@ -8,8 +8,7 @@
8#ifndef CONFIG_SMP 8#ifndef CONFIG_SMP
9/* 9/*
10 * The following implementation only for uniprocessor machines. 10 * The following implementation only for uniprocessor machines.
11 * For UP, it's relies on the fact that pagefault_disable() also disables 11 * It relies on preempt_disable() ensuring mutual exclusion.
12 * preemption to ensure mutual exclusion.
13 * 12 *
14 */ 13 */
15 14
@@ -38,6 +37,7 @@ futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
38 if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28)) 37 if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
39 oparg = 1 << oparg; 38 oparg = 1 << oparg;
40 39
40 preempt_disable();
41 pagefault_disable(); 41 pagefault_disable();
42 42
43 ret = -EFAULT; 43 ret = -EFAULT;
@@ -72,6 +72,7 @@ futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
72 72
73out_pagefault_enable: 73out_pagefault_enable:
74 pagefault_enable(); 74 pagefault_enable();
75 preempt_enable();
75 76
76 if (ret == 0) { 77 if (ret == 0) {
77 switch (cmp) { 78 switch (cmp) {
@@ -106,6 +107,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
106{ 107{
107 u32 val; 108 u32 val;
108 109
110 preempt_disable();
109 if (unlikely(get_user(val, uaddr) != 0)) 111 if (unlikely(get_user(val, uaddr) != 0))
110 return -EFAULT; 112 return -EFAULT;
111 113
@@ -113,6 +115,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
113 return -EFAULT; 115 return -EFAULT;
114 116
115 *uval = val; 117 *uval = val;
118 preempt_enable();
116 119
117 return 0; 120 return 0;
118} 121}
diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
index 9db042304df3..f56094cfdeff 100644
--- a/include/asm-generic/io.h
+++ b/include/asm-generic/io.h
@@ -769,6 +769,14 @@ static inline void __iomem *ioremap_nocache(phys_addr_t offset, size_t size)
769} 769}
770#endif 770#endif
771 771
772#ifndef ioremap_uc
773#define ioremap_uc ioremap_uc
774static inline void __iomem *ioremap_uc(phys_addr_t offset, size_t size)
775{
776 return ioremap_nocache(offset, size);
777}
778#endif
779
772#ifndef ioremap_wc 780#ifndef ioremap_wc
773#define ioremap_wc ioremap_wc 781#define ioremap_wc ioremap_wc
774static inline void __iomem *ioremap_wc(phys_addr_t offset, size_t size) 782static inline void __iomem *ioremap_wc(phys_addr_t offset, size_t size)
@@ -777,8 +785,17 @@ static inline void __iomem *ioremap_wc(phys_addr_t offset, size_t size)
777} 785}
778#endif 786#endif
779 787
788#ifndef ioremap_wt
789#define ioremap_wt ioremap_wt
790static inline void __iomem *ioremap_wt(phys_addr_t offset, size_t size)
791{
792 return ioremap_nocache(offset, size);
793}
794#endif
795
780#ifndef iounmap 796#ifndef iounmap
781#define iounmap iounmap 797#define iounmap iounmap
798
782static inline void iounmap(void __iomem *addr) 799static inline void iounmap(void __iomem *addr)
783{ 800{
784} 801}
diff --git a/include/asm-generic/iomap.h b/include/asm-generic/iomap.h
index 1b41011643a5..d8f8622fa044 100644
--- a/include/asm-generic/iomap.h
+++ b/include/asm-generic/iomap.h
@@ -66,6 +66,10 @@ extern void ioport_unmap(void __iomem *);
66#define ioremap_wc ioremap_nocache 66#define ioremap_wc ioremap_nocache
67#endif 67#endif
68 68
69#ifndef ARCH_HAS_IOREMAP_WT
70#define ioremap_wt ioremap_nocache
71#endif
72
69#ifdef CONFIG_PCI 73#ifdef CONFIG_PCI
70/* Destroy a virtual mapping cookie for a PCI BAR (memory or IO) */ 74/* Destroy a virtual mapping cookie for a PCI BAR (memory or IO) */
71struct pci_dev; 75struct pci_dev;
diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
index 39f1d6a2b04d..bd910ceaccfa 100644
--- a/include/asm-generic/pgtable.h
+++ b/include/asm-generic/pgtable.h
@@ -262,6 +262,10 @@ static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b)
262#define pgprot_writecombine pgprot_noncached 262#define pgprot_writecombine pgprot_noncached
263#endif 263#endif
264 264
265#ifndef pgprot_writethrough
266#define pgprot_writethrough pgprot_noncached
267#endif
268
265#ifndef pgprot_device 269#ifndef pgprot_device
266#define pgprot_device pgprot_noncached 270#define pgprot_device pgprot_noncached
267#endif 271#endif
diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h
index eb6f9e6c3075..d0a7a4753db2 100644
--- a/include/asm-generic/preempt.h
+++ b/include/asm-generic/preempt.h
@@ -79,11 +79,8 @@ static __always_inline bool should_resched(void)
79#ifdef CONFIG_PREEMPT 79#ifdef CONFIG_PREEMPT
80extern asmlinkage void preempt_schedule(void); 80extern asmlinkage void preempt_schedule(void);
81#define __preempt_schedule() preempt_schedule() 81#define __preempt_schedule() preempt_schedule()
82 82extern asmlinkage void preempt_schedule_notrace(void);
83#ifdef CONFIG_CONTEXT_TRACKING 83#define __preempt_schedule_notrace() preempt_schedule_notrace()
84extern asmlinkage void preempt_schedule_context(void);
85#define __preempt_schedule_context() preempt_schedule_context()
86#endif
87#endif /* CONFIG_PREEMPT */ 84#endif /* CONFIG_PREEMPT */
88 85
89#endif /* __ASM_PREEMPT_H */ 86#endif /* __ASM_PREEMPT_H */
diff --git a/include/asm-generic/qspinlock.h b/include/asm-generic/qspinlock.h
new file mode 100644
index 000000000000..83bfb87f5bf1
--- /dev/null
+++ b/include/asm-generic/qspinlock.h
@@ -0,0 +1,139 @@
1/*
2 * Queued spinlock
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; either version 2 of the License, or
7 * (at your option) any later version.
8 *
9 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
13 *
14 * (C) Copyright 2013-2015 Hewlett-Packard Development Company, L.P.
15 *
16 * Authors: Waiman Long <waiman.long@hp.com>
17 */
18#ifndef __ASM_GENERIC_QSPINLOCK_H
19#define __ASM_GENERIC_QSPINLOCK_H
20
21#include <asm-generic/qspinlock_types.h>
22
23/**
24 * queued_spin_is_locked - is the spinlock locked?
25 * @lock: Pointer to queued spinlock structure
26 * Return: 1 if it is locked, 0 otherwise
27 */
28static __always_inline int queued_spin_is_locked(struct qspinlock *lock)
29{
30 return atomic_read(&lock->val);
31}
32
33/**
34 * queued_spin_value_unlocked - is the spinlock structure unlocked?
35 * @lock: queued spinlock structure
36 * Return: 1 if it is unlocked, 0 otherwise
37 *
38 * N.B. Whenever there are tasks waiting for the lock, it is considered
39 * locked wrt the lockref code to avoid lock stealing by the lockref
40 * code and change things underneath the lock. This also allows some
41 * optimizations to be applied without conflict with lockref.
42 */
43static __always_inline int queued_spin_value_unlocked(struct qspinlock lock)
44{
45 return !atomic_read(&lock.val);
46}
47
48/**
49 * queued_spin_is_contended - check if the lock is contended
50 * @lock : Pointer to queued spinlock structure
51 * Return: 1 if lock contended, 0 otherwise
52 */
53static __always_inline int queued_spin_is_contended(struct qspinlock *lock)
54{
55 return atomic_read(&lock->val) & ~_Q_LOCKED_MASK;
56}
57/**
58 * queued_spin_trylock - try to acquire the queued spinlock
59 * @lock : Pointer to queued spinlock structure
60 * Return: 1 if lock acquired, 0 if failed
61 */
62static __always_inline int queued_spin_trylock(struct qspinlock *lock)
63{
64 if (!atomic_read(&lock->val) &&
65 (atomic_cmpxchg(&lock->val, 0, _Q_LOCKED_VAL) == 0))
66 return 1;
67 return 0;
68}
69
70extern void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val);
71
72/**
73 * queued_spin_lock - acquire a queued spinlock
74 * @lock: Pointer to queued spinlock structure
75 */
76static __always_inline void queued_spin_lock(struct qspinlock *lock)
77{
78 u32 val;
79
80 val = atomic_cmpxchg(&lock->val, 0, _Q_LOCKED_VAL);
81 if (likely(val == 0))
82 return;
83 queued_spin_lock_slowpath(lock, val);
84}
85
86#ifndef queued_spin_unlock
87/**
88 * queued_spin_unlock - release a queued spinlock
89 * @lock : Pointer to queued spinlock structure
90 */
91static __always_inline void queued_spin_unlock(struct qspinlock *lock)
92{
93 /*
94 * smp_mb__before_atomic() in order to guarantee release semantics
95 */
96 smp_mb__before_atomic_dec();
97 atomic_sub(_Q_LOCKED_VAL, &lock->val);
98}
99#endif
100
101/**
102 * queued_spin_unlock_wait - wait until current lock holder releases the lock
103 * @lock : Pointer to queued spinlock structure
104 *
105 * There is a very slight possibility of live-lock if the lockers keep coming
106 * and the waiter is just unfortunate enough to not see any unlock state.
107 */
108static inline void queued_spin_unlock_wait(struct qspinlock *lock)
109{
110 while (atomic_read(&lock->val) & _Q_LOCKED_MASK)
111 cpu_relax();
112}
113
114#ifndef virt_queued_spin_lock
115static __always_inline bool virt_queued_spin_lock(struct qspinlock *lock)
116{
117 return false;
118}
119#endif
120
121/*
122 * Initializier
123 */
124#define __ARCH_SPIN_LOCK_UNLOCKED { ATOMIC_INIT(0) }
125
126/*
127 * Remapping spinlock architecture specific functions to the corresponding
128 * queued spinlock functions.
129 */
130#define arch_spin_is_locked(l) queued_spin_is_locked(l)
131#define arch_spin_is_contended(l) queued_spin_is_contended(l)
132#define arch_spin_value_unlocked(l) queued_spin_value_unlocked(l)
133#define arch_spin_lock(l) queued_spin_lock(l)
134#define arch_spin_trylock(l) queued_spin_trylock(l)
135#define arch_spin_unlock(l) queued_spin_unlock(l)
136#define arch_spin_lock_flags(l, f) queued_spin_lock(l)
137#define arch_spin_unlock_wait(l) queued_spin_unlock_wait(l)
138
139#endif /* __ASM_GENERIC_QSPINLOCK_H */
diff --git a/include/asm-generic/qspinlock_types.h b/include/asm-generic/qspinlock_types.h
new file mode 100644
index 000000000000..85f888e86761
--- /dev/null
+++ b/include/asm-generic/qspinlock_types.h
@@ -0,0 +1,79 @@
1/*
2 * Queued spinlock
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; either version 2 of the License, or
7 * (at your option) any later version.
8 *
9 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
13 *
14 * (C) Copyright 2013-2015 Hewlett-Packard Development Company, L.P.
15 *
16 * Authors: Waiman Long <waiman.long@hp.com>
17 */
18#ifndef __ASM_GENERIC_QSPINLOCK_TYPES_H
19#define __ASM_GENERIC_QSPINLOCK_TYPES_H
20
21/*
22 * Including atomic.h with PARAVIRT on will cause compilation errors because
23 * of recursive header file incluson via paravirt_types.h. So don't include
24 * it if PARAVIRT is on.
25 */
26#ifndef CONFIG_PARAVIRT
27#include <linux/types.h>
28#include <linux/atomic.h>
29#endif
30
31typedef struct qspinlock {
32 atomic_t val;
33} arch_spinlock_t;
34
35/*
36 * Bitfields in the atomic value:
37 *
38 * When NR_CPUS < 16K
39 * 0- 7: locked byte
40 * 8: pending
41 * 9-15: not used
42 * 16-17: tail index
43 * 18-31: tail cpu (+1)
44 *
45 * When NR_CPUS >= 16K
46 * 0- 7: locked byte
47 * 8: pending
48 * 9-10: tail index
49 * 11-31: tail cpu (+1)
50 */
51#define _Q_SET_MASK(type) (((1U << _Q_ ## type ## _BITS) - 1)\
52 << _Q_ ## type ## _OFFSET)
53#define _Q_LOCKED_OFFSET 0
54#define _Q_LOCKED_BITS 8
55#define _Q_LOCKED_MASK _Q_SET_MASK(LOCKED)
56
57#define _Q_PENDING_OFFSET (_Q_LOCKED_OFFSET + _Q_LOCKED_BITS)
58#if CONFIG_NR_CPUS < (1U << 14)
59#define _Q_PENDING_BITS 8
60#else
61#define _Q_PENDING_BITS 1
62#endif
63#define _Q_PENDING_MASK _Q_SET_MASK(PENDING)
64
65#define _Q_TAIL_IDX_OFFSET (_Q_PENDING_OFFSET + _Q_PENDING_BITS)
66#define _Q_TAIL_IDX_BITS 2
67#define _Q_TAIL_IDX_MASK _Q_SET_MASK(TAIL_IDX)
68
69#define _Q_TAIL_CPU_OFFSET (_Q_TAIL_IDX_OFFSET + _Q_TAIL_IDX_BITS)
70#define _Q_TAIL_CPU_BITS (32 - _Q_TAIL_CPU_OFFSET)
71#define _Q_TAIL_CPU_MASK _Q_SET_MASK(TAIL_CPU)
72
73#define _Q_TAIL_OFFSET _Q_TAIL_IDX_OFFSET
74#define _Q_TAIL_MASK (_Q_TAIL_IDX_MASK | _Q_TAIL_CPU_MASK)
75
76#define _Q_LOCKED_VAL (1U << _Q_LOCKED_OFFSET)
77#define _Q_PENDING_VAL (1U << _Q_PENDING_OFFSET)
78
79#endif /* __ASM_GENERIC_QSPINLOCK_TYPES_H */
diff --git a/include/linux/alarmtimer.h b/include/linux/alarmtimer.h
index a899402a5a0e..52f3b7da4f2d 100644
--- a/include/linux/alarmtimer.h
+++ b/include/linux/alarmtimer.h
@@ -43,8 +43,8 @@ struct alarm {
43 43
44void alarm_init(struct alarm *alarm, enum alarmtimer_type type, 44void alarm_init(struct alarm *alarm, enum alarmtimer_type type,
45 enum alarmtimer_restart (*function)(struct alarm *, ktime_t)); 45 enum alarmtimer_restart (*function)(struct alarm *, ktime_t));
46int alarm_start(struct alarm *alarm, ktime_t start); 46void alarm_start(struct alarm *alarm, ktime_t start);
47int alarm_start_relative(struct alarm *alarm, ktime_t start); 47void alarm_start_relative(struct alarm *alarm, ktime_t start);
48void alarm_restart(struct alarm *alarm); 48void alarm_restart(struct alarm *alarm);
49int alarm_try_to_cancel(struct alarm *alarm); 49int alarm_try_to_cancel(struct alarm *alarm);
50int alarm_cancel(struct alarm *alarm); 50int alarm_cancel(struct alarm *alarm);
diff --git a/include/linux/backing-dev.h b/include/linux/backing-dev.h
index aff923ae8c4b..d87d8eced064 100644
--- a/include/linux/backing-dev.h
+++ b/include/linux/backing-dev.h
@@ -116,7 +116,6 @@ __printf(3, 4)
116int bdi_register(struct backing_dev_info *bdi, struct device *parent, 116int bdi_register(struct backing_dev_info *bdi, struct device *parent,
117 const char *fmt, ...); 117 const char *fmt, ...);
118int bdi_register_dev(struct backing_dev_info *bdi, dev_t dev); 118int bdi_register_dev(struct backing_dev_info *bdi, dev_t dev);
119void bdi_unregister(struct backing_dev_info *bdi);
120int __must_check bdi_setup_and_register(struct backing_dev_info *, char *); 119int __must_check bdi_setup_and_register(struct backing_dev_info *, char *);
121void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages, 120void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages,
122 enum wb_reason reason); 121 enum wb_reason reason);
diff --git a/include/linux/bottom_half.h b/include/linux/bottom_half.h
index 86c12c93e3cf..8fdcb783197d 100644
--- a/include/linux/bottom_half.h
+++ b/include/linux/bottom_half.h
@@ -2,7 +2,6 @@
2#define _LINUX_BH_H 2#define _LINUX_BH_H
3 3
4#include <linux/preempt.h> 4#include <linux/preempt.h>
5#include <linux/preempt_mask.h>
6 5
7#ifdef CONFIG_TRACE_IRQFLAGS 6#ifdef CONFIG_TRACE_IRQFLAGS
8extern void __local_bh_disable_ip(unsigned long ip, unsigned int cnt); 7extern void __local_bh_disable_ip(unsigned long ip, unsigned int cnt);
diff --git a/include/linux/brcmphy.h b/include/linux/brcmphy.h
index ae2982c0f7a6..656da2a12ffe 100644
--- a/include/linux/brcmphy.h
+++ b/include/linux/brcmphy.h
@@ -17,7 +17,7 @@
17#define PHY_ID_BCM7250 0xae025280 17#define PHY_ID_BCM7250 0xae025280
18#define PHY_ID_BCM7364 0xae025260 18#define PHY_ID_BCM7364 0xae025260
19#define PHY_ID_BCM7366 0x600d8490 19#define PHY_ID_BCM7366 0x600d8490
20#define PHY_ID_BCM7425 0x03625e60 20#define PHY_ID_BCM7425 0x600d86b0
21#define PHY_ID_BCM7429 0x600d8730 21#define PHY_ID_BCM7429 0x600d8730
22#define PHY_ID_BCM7439 0x600d8480 22#define PHY_ID_BCM7439 0x600d8480
23#define PHY_ID_BCM7439_2 0xae025080 23#define PHY_ID_BCM7439_2 0xae025080
diff --git a/include/linux/clockchips.h b/include/linux/clockchips.h
index 96c280b2c263..597a1e836f22 100644
--- a/include/linux/clockchips.h
+++ b/include/linux/clockchips.h
@@ -37,12 +37,15 @@ enum clock_event_mode {
37 * reached from DETACHED or SHUTDOWN. 37 * reached from DETACHED or SHUTDOWN.
38 * ONESHOT: Device is programmed to generate event only once. Can be reached 38 * ONESHOT: Device is programmed to generate event only once. Can be reached
39 * from DETACHED or SHUTDOWN. 39 * from DETACHED or SHUTDOWN.
40 * ONESHOT_STOPPED: Device was programmed in ONESHOT mode and is temporarily
41 * stopped.
40 */ 42 */
41enum clock_event_state { 43enum clock_event_state {
42 CLOCK_EVT_STATE_DETACHED, 44 CLOCK_EVT_STATE_DETACHED,
43 CLOCK_EVT_STATE_SHUTDOWN, 45 CLOCK_EVT_STATE_SHUTDOWN,
44 CLOCK_EVT_STATE_PERIODIC, 46 CLOCK_EVT_STATE_PERIODIC,
45 CLOCK_EVT_STATE_ONESHOT, 47 CLOCK_EVT_STATE_ONESHOT,
48 CLOCK_EVT_STATE_ONESHOT_STOPPED,
46}; 49};
47 50
48/* 51/*
@@ -84,12 +87,13 @@ enum clock_event_state {
84 * @mult: nanosecond to cycles multiplier 87 * @mult: nanosecond to cycles multiplier
85 * @shift: nanoseconds to cycles divisor (power of two) 88 * @shift: nanoseconds to cycles divisor (power of two)
86 * @mode: operating mode, relevant only to ->set_mode(), OBSOLETE 89 * @mode: operating mode, relevant only to ->set_mode(), OBSOLETE
87 * @state: current state of the device, assigned by the core code 90 * @state_use_accessors:current state of the device, assigned by the core code
88 * @features: features 91 * @features: features
89 * @retries: number of forced programming retries 92 * @retries: number of forced programming retries
90 * @set_mode: legacy set mode function, only for modes <= CLOCK_EVT_MODE_RESUME. 93 * @set_mode: legacy set mode function, only for modes <= CLOCK_EVT_MODE_RESUME.
91 * @set_state_periodic: switch state to periodic, if !set_mode 94 * @set_state_periodic: switch state to periodic, if !set_mode
92 * @set_state_oneshot: switch state to oneshot, if !set_mode 95 * @set_state_oneshot: switch state to oneshot, if !set_mode
96 * @set_state_oneshot_stopped: switch state to oneshot_stopped, if !set_mode
93 * @set_state_shutdown: switch state to shutdown, if !set_mode 97 * @set_state_shutdown: switch state to shutdown, if !set_mode
94 * @tick_resume: resume clkevt device, if !set_mode 98 * @tick_resume: resume clkevt device, if !set_mode
95 * @broadcast: function to broadcast events 99 * @broadcast: function to broadcast events
@@ -113,7 +117,7 @@ struct clock_event_device {
113 u32 mult; 117 u32 mult;
114 u32 shift; 118 u32 shift;
115 enum clock_event_mode mode; 119 enum clock_event_mode mode;
116 enum clock_event_state state; 120 enum clock_event_state state_use_accessors;
117 unsigned int features; 121 unsigned int features;
118 unsigned long retries; 122 unsigned long retries;
119 123
@@ -121,11 +125,12 @@ struct clock_event_device {
121 * State transition callback(s): Only one of the two groups should be 125 * State transition callback(s): Only one of the two groups should be
122 * defined: 126 * defined:
123 * - set_mode(), only for modes <= CLOCK_EVT_MODE_RESUME. 127 * - set_mode(), only for modes <= CLOCK_EVT_MODE_RESUME.
124 * - set_state_{shutdown|periodic|oneshot}(), tick_resume(). 128 * - set_state_{shutdown|periodic|oneshot|oneshot_stopped}(), tick_resume().
125 */ 129 */
126 void (*set_mode)(enum clock_event_mode mode, struct clock_event_device *); 130 void (*set_mode)(enum clock_event_mode mode, struct clock_event_device *);
127 int (*set_state_periodic)(struct clock_event_device *); 131 int (*set_state_periodic)(struct clock_event_device *);
128 int (*set_state_oneshot)(struct clock_event_device *); 132 int (*set_state_oneshot)(struct clock_event_device *);
133 int (*set_state_oneshot_stopped)(struct clock_event_device *);
129 int (*set_state_shutdown)(struct clock_event_device *); 134 int (*set_state_shutdown)(struct clock_event_device *);
130 int (*tick_resume)(struct clock_event_device *); 135 int (*tick_resume)(struct clock_event_device *);
131 136
@@ -144,6 +149,32 @@ struct clock_event_device {
144 struct module *owner; 149 struct module *owner;
145} ____cacheline_aligned; 150} ____cacheline_aligned;
146 151
152/* Helpers to verify state of a clockevent device */
153static inline bool clockevent_state_detached(struct clock_event_device *dev)
154{
155 return dev->state_use_accessors == CLOCK_EVT_STATE_DETACHED;
156}
157
158static inline bool clockevent_state_shutdown(struct clock_event_device *dev)
159{
160 return dev->state_use_accessors == CLOCK_EVT_STATE_SHUTDOWN;
161}
162
163static inline bool clockevent_state_periodic(struct clock_event_device *dev)
164{
165 return dev->state_use_accessors == CLOCK_EVT_STATE_PERIODIC;
166}
167
168static inline bool clockevent_state_oneshot(struct clock_event_device *dev)
169{
170 return dev->state_use_accessors == CLOCK_EVT_STATE_ONESHOT;
171}
172
173static inline bool clockevent_state_oneshot_stopped(struct clock_event_device *dev)
174{
175 return dev->state_use_accessors == CLOCK_EVT_STATE_ONESHOT_STOPPED;
176}
177
147/* 178/*
148 * Calculate a multiplication factor for scaled math, which is used to convert 179 * Calculate a multiplication factor for scaled math, which is used to convert
149 * nanoseconds based values to clock ticks: 180 * nanoseconds based values to clock ticks:
diff --git a/include/linux/clocksource.h b/include/linux/clocksource.h
index d27d0152271f..278dd279a7a8 100644
--- a/include/linux/clocksource.h
+++ b/include/linux/clocksource.h
@@ -181,7 +181,6 @@ static inline s64 clocksource_cyc2ns(cycle_t cycles, u32 mult, u32 shift)
181 181
182extern int clocksource_unregister(struct clocksource*); 182extern int clocksource_unregister(struct clocksource*);
183extern void clocksource_touch_watchdog(void); 183extern void clocksource_touch_watchdog(void);
184extern struct clocksource* clocksource_get_next(void);
185extern void clocksource_change_rating(struct clocksource *cs, int rating); 184extern void clocksource_change_rating(struct clocksource *cs, int rating);
186extern void clocksource_suspend(void); 185extern void clocksource_suspend(void);
187extern void clocksource_resume(void); 186extern void clocksource_resume(void);
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index 867722591be2..05be2352fef8 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -250,7 +250,23 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
250 ({ union { typeof(x) __val; char __c[1]; } __u; __read_once_size(&(x), __u.__c, sizeof(x)); __u.__val; }) 250 ({ union { typeof(x) __val; char __c[1]; } __u; __read_once_size(&(x), __u.__c, sizeof(x)); __u.__val; })
251 251
252#define WRITE_ONCE(x, val) \ 252#define WRITE_ONCE(x, val) \
253 ({ typeof(x) __val = (val); __write_once_size(&(x), &__val, sizeof(__val)); __val; }) 253 ({ union { typeof(x) __val; char __c[1]; } __u = { .__val = (val) }; __write_once_size(&(x), __u.__c, sizeof(x)); __u.__val; })
254
255/**
256 * READ_ONCE_CTRL - Read a value heading a control dependency
257 * @x: The value to be read, heading the control dependency
258 *
259 * Control dependencies are tricky. See Documentation/memory-barriers.txt
260 * for important information on how to use them. Note that in many cases,
261 * use of smp_load_acquire() will be much simpler. Control dependencies
262 * should be avoided except on the hottest of hotpaths.
263 */
264#define READ_ONCE_CTRL(x) \
265({ \
266 typeof(x) __val = READ_ONCE(x); \
267 smp_read_barrier_depends(); /* Enforce control dependency. */ \
268 __val; \
269})
254 270
255#endif /* __KERNEL__ */ 271#endif /* __KERNEL__ */
256 272
@@ -450,7 +466,7 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
450 * with an explicit memory barrier or atomic instruction that provides the 466 * with an explicit memory barrier or atomic instruction that provides the
451 * required ordering. 467 * required ordering.
452 * 468 *
453 * If possible use READ_ONCE/ASSIGN_ONCE instead. 469 * If possible use READ_ONCE()/WRITE_ONCE() instead.
454 */ 470 */
455#define __ACCESS_ONCE(x) ({ \ 471#define __ACCESS_ONCE(x) ({ \
456 __maybe_unused typeof(x) __var = (__force typeof(x)) 0; \ 472 __maybe_unused typeof(x) __var = (__force typeof(x)) 0; \
diff --git a/include/linux/context_tracking.h b/include/linux/context_tracking.h
index 2821838256b4..b96bd299966f 100644
--- a/include/linux/context_tracking.h
+++ b/include/linux/context_tracking.h
@@ -14,8 +14,6 @@ extern void context_tracking_enter(enum ctx_state state);
14extern void context_tracking_exit(enum ctx_state state); 14extern void context_tracking_exit(enum ctx_state state);
15extern void context_tracking_user_enter(void); 15extern void context_tracking_user_enter(void);
16extern void context_tracking_user_exit(void); 16extern void context_tracking_user_exit(void);
17extern void __context_tracking_task_switch(struct task_struct *prev,
18 struct task_struct *next);
19 17
20static inline void user_enter(void) 18static inline void user_enter(void)
21{ 19{
@@ -51,19 +49,11 @@ static inline void exception_exit(enum ctx_state prev_ctx)
51 } 49 }
52} 50}
53 51
54static inline void context_tracking_task_switch(struct task_struct *prev,
55 struct task_struct *next)
56{
57 if (context_tracking_is_enabled())
58 __context_tracking_task_switch(prev, next);
59}
60#else 52#else
61static inline void user_enter(void) { } 53static inline void user_enter(void) { }
62static inline void user_exit(void) { } 54static inline void user_exit(void) { }
63static inline enum ctx_state exception_enter(void) { return 0; } 55static inline enum ctx_state exception_enter(void) { return 0; }
64static inline void exception_exit(enum ctx_state prev_ctx) { } 56static inline void exception_exit(enum ctx_state prev_ctx) { }
65static inline void context_tracking_task_switch(struct task_struct *prev,
66 struct task_struct *next) { }
67#endif /* !CONFIG_CONTEXT_TRACKING */ 57#endif /* !CONFIG_CONTEXT_TRACKING */
68 58
69 59
diff --git a/include/linux/context_tracking_state.h b/include/linux/context_tracking_state.h
index 6b7b96a32b75..678ecdf90cf6 100644
--- a/include/linux/context_tracking_state.h
+++ b/include/linux/context_tracking_state.h
@@ -12,6 +12,7 @@ struct context_tracking {
12 * may be further optimized using static keys. 12 * may be further optimized using static keys.
13 */ 13 */
14 bool active; 14 bool active;
15 int recursion;
15 enum ctx_state { 16 enum ctx_state {
16 CONTEXT_KERNEL = 0, 17 CONTEXT_KERNEL = 0,
17 CONTEXT_USER, 18 CONTEXT_USER,
diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
index 27e285b92b5f..59915ea5373c 100644
--- a/include/linux/cpumask.h
+++ b/include/linux/cpumask.h
@@ -151,10 +151,8 @@ static inline unsigned int cpumask_any_but(const struct cpumask *mask,
151 return 1; 151 return 1;
152} 152}
153 153
154static inline int cpumask_set_cpu_local_first(int i, int numa_node, cpumask_t *dstp) 154static inline unsigned int cpumask_local_spread(unsigned int i, int node)
155{ 155{
156 set_bit(0, cpumask_bits(dstp));
157
158 return 0; 156 return 0;
159} 157}
160 158
@@ -208,7 +206,7 @@ static inline unsigned int cpumask_next_zero(int n, const struct cpumask *srcp)
208 206
209int cpumask_next_and(int n, const struct cpumask *, const struct cpumask *); 207int cpumask_next_and(int n, const struct cpumask *, const struct cpumask *);
210int cpumask_any_but(const struct cpumask *mask, unsigned int cpu); 208int cpumask_any_but(const struct cpumask *mask, unsigned int cpu);
211int cpumask_set_cpu_local_first(int i, int numa_node, cpumask_t *dstp); 209unsigned int cpumask_local_spread(unsigned int i, int node);
212 210
213/** 211/**
214 * for_each_cpu - iterate over every cpu in a mask 212 * for_each_cpu - iterate over every cpu in a mask
diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h
index cb25af461054..420311bcee38 100644
--- a/include/linux/debugfs.h
+++ b/include/linux/debugfs.h
@@ -45,7 +45,6 @@ extern struct dentry *arch_debugfs_dir;
45 45
46/* declared over in file.c */ 46/* declared over in file.c */
47extern const struct file_operations debugfs_file_operations; 47extern const struct file_operations debugfs_file_operations;
48extern const struct inode_operations debugfs_link_operations;
49 48
50struct dentry *debugfs_create_file(const char *name, umode_t mode, 49struct dentry *debugfs_create_file(const char *name, umode_t mode,
51 struct dentry *parent, void *data, 50 struct dentry *parent, void *data,
diff --git a/include/linux/dmar.h b/include/linux/dmar.h
index 30624954dec5..e9bc9292bd3a 100644
--- a/include/linux/dmar.h
+++ b/include/linux/dmar.h
@@ -185,33 +185,85 @@ static inline int dmar_device_remove(void *handle)
185 185
186struct irte { 186struct irte {
187 union { 187 union {
188 /* Shared between remapped and posted mode*/
188 struct { 189 struct {
189 __u64 present : 1, 190 __u64 present : 1, /* 0 */
190 fpd : 1, 191 fpd : 1, /* 1 */
191 dst_mode : 1, 192 __res0 : 6, /* 2 - 6 */
192 redir_hint : 1, 193 avail : 4, /* 8 - 11 */
193 trigger_mode : 1, 194 __res1 : 3, /* 12 - 14 */
194 dlvry_mode : 3, 195 pst : 1, /* 15 */
195 avail : 4, 196 vector : 8, /* 16 - 23 */
196 __reserved_1 : 4, 197 __res2 : 40; /* 24 - 63 */
197 vector : 8, 198 };
198 __reserved_2 : 8, 199
199 dest_id : 32; 200 /* Remapped mode */
201 struct {
202 __u64 r_present : 1, /* 0 */
203 r_fpd : 1, /* 1 */
204 dst_mode : 1, /* 2 */
205 redir_hint : 1, /* 3 */
206 trigger_mode : 1, /* 4 */
207 dlvry_mode : 3, /* 5 - 7 */
208 r_avail : 4, /* 8 - 11 */
209 r_res0 : 4, /* 12 - 15 */
210 r_vector : 8, /* 16 - 23 */
211 r_res1 : 8, /* 24 - 31 */
212 dest_id : 32; /* 32 - 63 */
213 };
214
215 /* Posted mode */
216 struct {
217 __u64 p_present : 1, /* 0 */
218 p_fpd : 1, /* 1 */
219 p_res0 : 6, /* 2 - 7 */
220 p_avail : 4, /* 8 - 11 */
221 p_res1 : 2, /* 12 - 13 */
222 p_urgent : 1, /* 14 */
223 p_pst : 1, /* 15 */
224 p_vector : 8, /* 16 - 23 */
225 p_res2 : 14, /* 24 - 37 */
226 pda_l : 26; /* 38 - 63 */
200 }; 227 };
201 __u64 low; 228 __u64 low;
202 }; 229 };
203 230
204 union { 231 union {
232 /* Shared between remapped and posted mode*/
205 struct { 233 struct {
206 __u64 sid : 16, 234 __u64 sid : 16, /* 64 - 79 */
207 sq : 2, 235 sq : 2, /* 80 - 81 */
208 svt : 2, 236 svt : 2, /* 82 - 83 */
209 __reserved_3 : 44; 237 __res3 : 44; /* 84 - 127 */
238 };
239
240 /* Posted mode*/
241 struct {
242 __u64 p_sid : 16, /* 64 - 79 */
243 p_sq : 2, /* 80 - 81 */
244 p_svt : 2, /* 82 - 83 */
245 p_res3 : 12, /* 84 - 95 */
246 pda_h : 32; /* 96 - 127 */
210 }; 247 };
211 __u64 high; 248 __u64 high;
212 }; 249 };
213}; 250};
214 251
252static inline void dmar_copy_shared_irte(struct irte *dst, struct irte *src)
253{
254 dst->present = src->present;
255 dst->fpd = src->fpd;
256 dst->avail = src->avail;
257 dst->pst = src->pst;
258 dst->vector = src->vector;
259 dst->sid = src->sid;
260 dst->sq = src->sq;
261 dst->svt = src->svt;
262}
263
264#define PDA_LOW_BIT 26
265#define PDA_HIGH_BIT 32
266
215enum { 267enum {
216 IRQ_REMAP_XAPIC_MODE, 268 IRQ_REMAP_XAPIC_MODE,
217 IRQ_REMAP_X2APIC_MODE, 269 IRQ_REMAP_X2APIC_MODE,
@@ -227,6 +279,7 @@ extern void dmar_msi_read(int irq, struct msi_msg *msg);
227extern void dmar_msi_write(int irq, struct msi_msg *msg); 279extern void dmar_msi_write(int irq, struct msi_msg *msg);
228extern int dmar_set_interrupt(struct intel_iommu *iommu); 280extern int dmar_set_interrupt(struct intel_iommu *iommu);
229extern irqreturn_t dmar_fault(int irq, void *dev_id); 281extern irqreturn_t dmar_fault(int irq, void *dev_id);
230extern int arch_setup_dmar_msi(unsigned int irq); 282extern int dmar_alloc_hwirq(int id, int node, void *arg);
283extern void dmar_free_hwirq(int irq);
231 284
232#endif /* __DMAR_H__ */ 285#endif /* __DMAR_H__ */
diff --git a/include/linux/efi.h b/include/linux/efi.h
index af5be0368dec..2092965afca3 100644
--- a/include/linux/efi.h
+++ b/include/linux/efi.h
@@ -583,6 +583,9 @@ void efi_native_runtime_setup(void);
583#define EFI_FILE_INFO_ID \ 583#define EFI_FILE_INFO_ID \
584 EFI_GUID( 0x9576e92, 0x6d3f, 0x11d2, 0x8e, 0x39, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b ) 584 EFI_GUID( 0x9576e92, 0x6d3f, 0x11d2, 0x8e, 0x39, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b )
585 585
586#define EFI_SYSTEM_RESOURCE_TABLE_GUID \
587 EFI_GUID( 0xb122a263, 0x3661, 0x4f68, 0x99, 0x29, 0x78, 0xf8, 0xb0, 0xd6, 0x21, 0x80 )
588
586#define EFI_FILE_SYSTEM_GUID \ 589#define EFI_FILE_SYSTEM_GUID \
587 EFI_GUID( 0x964e5b22, 0x6459, 0x11d2, 0x8e, 0x39, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b ) 590 EFI_GUID( 0x964e5b22, 0x6459, 0x11d2, 0x8e, 0x39, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b )
588 591
@@ -823,6 +826,7 @@ extern struct efi {
823 unsigned long fw_vendor; /* fw_vendor */ 826 unsigned long fw_vendor; /* fw_vendor */
824 unsigned long runtime; /* runtime table */ 827 unsigned long runtime; /* runtime table */
825 unsigned long config_table; /* config tables */ 828 unsigned long config_table; /* config tables */
829 unsigned long esrt; /* ESRT table */
826 efi_get_time_t *get_time; 830 efi_get_time_t *get_time;
827 efi_set_time_t *set_time; 831 efi_set_time_t *set_time;
828 efi_get_wakeup_time_t *get_wakeup_time; 832 efi_get_wakeup_time_t *get_wakeup_time;
@@ -875,6 +879,11 @@ static inline efi_status_t efi_query_variable_store(u32 attributes, unsigned lon
875#endif 879#endif
876extern void __iomem *efi_lookup_mapped_addr(u64 phys_addr); 880extern void __iomem *efi_lookup_mapped_addr(u64 phys_addr);
877extern int efi_config_init(efi_config_table_type_t *arch_tables); 881extern int efi_config_init(efi_config_table_type_t *arch_tables);
882#ifdef CONFIG_EFI_ESRT
883extern void __init efi_esrt_init(void);
884#else
885static inline void efi_esrt_init(void) { }
886#endif
878extern int efi_config_parse_tables(void *config_tables, int count, int sz, 887extern int efi_config_parse_tables(void *config_tables, int count, int sz,
879 efi_config_table_type_t *arch_tables); 888 efi_config_table_type_t *arch_tables);
880extern u64 efi_get_iobase (void); 889extern u64 efi_get_iobase (void);
@@ -882,12 +891,15 @@ extern u32 efi_mem_type (unsigned long phys_addr);
882extern u64 efi_mem_attributes (unsigned long phys_addr); 891extern u64 efi_mem_attributes (unsigned long phys_addr);
883extern u64 efi_mem_attribute (unsigned long phys_addr, unsigned long size); 892extern u64 efi_mem_attribute (unsigned long phys_addr, unsigned long size);
884extern int __init efi_uart_console_only (void); 893extern int __init efi_uart_console_only (void);
894extern u64 efi_mem_desc_end(efi_memory_desc_t *md);
895extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md);
885extern void efi_initialize_iomem_resources(struct resource *code_resource, 896extern void efi_initialize_iomem_resources(struct resource *code_resource,
886 struct resource *data_resource, struct resource *bss_resource); 897 struct resource *data_resource, struct resource *bss_resource);
887extern void efi_get_time(struct timespec *now); 898extern void efi_get_time(struct timespec *now);
888extern void efi_reserve_boot_services(void); 899extern void efi_reserve_boot_services(void);
889extern int efi_get_fdt_params(struct efi_fdt_params *params, int verbose); 900extern int efi_get_fdt_params(struct efi_fdt_params *params, int verbose);
890extern struct efi_memory_map memmap; 901extern struct efi_memory_map memmap;
902extern struct kobject *efi_kobj;
891 903
892extern int efi_reboot_quirk_mode; 904extern int efi_reboot_quirk_mode;
893extern bool efi_poweroff_required(void); 905extern bool efi_poweroff_required(void);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 35ec87e490b1..b577e801b4af 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -38,7 +38,6 @@ struct backing_dev_info;
38struct export_operations; 38struct export_operations;
39struct hd_geometry; 39struct hd_geometry;
40struct iovec; 40struct iovec;
41struct nameidata;
42struct kiocb; 41struct kiocb;
43struct kobject; 42struct kobject;
44struct pipe_inode_info; 43struct pipe_inode_info;
@@ -656,6 +655,7 @@ struct inode {
656 struct pipe_inode_info *i_pipe; 655 struct pipe_inode_info *i_pipe;
657 struct block_device *i_bdev; 656 struct block_device *i_bdev;
658 struct cdev *i_cdev; 657 struct cdev *i_cdev;
658 char *i_link;
659 }; 659 };
660 660
661 __u32 i_generation; 661 __u32 i_generation;
@@ -1607,12 +1607,12 @@ struct file_operations {
1607 1607
1608struct inode_operations { 1608struct inode_operations {
1609 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); 1609 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1610 void * (*follow_link) (struct dentry *, struct nameidata *); 1610 const char * (*follow_link) (struct dentry *, void **);
1611 int (*permission) (struct inode *, int); 1611 int (*permission) (struct inode *, int);
1612 struct posix_acl * (*get_acl)(struct inode *, int); 1612 struct posix_acl * (*get_acl)(struct inode *, int);
1613 1613
1614 int (*readlink) (struct dentry *, char __user *,int); 1614 int (*readlink) (struct dentry *, char __user *,int);
1615 void (*put_link) (struct dentry *, struct nameidata *, void *); 1615 void (*put_link) (struct inode *, void *);
1616 1616
1617 int (*create) (struct inode *,struct dentry *, umode_t, bool); 1617 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1618 int (*link) (struct dentry *,struct inode *,struct dentry *); 1618 int (*link) (struct dentry *,struct inode *,struct dentry *);
@@ -1879,6 +1879,7 @@ enum file_time_flags {
1879 S_VERSION = 8, 1879 S_VERSION = 8,
1880}; 1880};
1881 1881
1882extern bool atime_needs_update(const struct path *, struct inode *);
1882extern void touch_atime(const struct path *); 1883extern void touch_atime(const struct path *);
1883static inline void file_accessed(struct file *file) 1884static inline void file_accessed(struct file *file)
1884{ 1885{
@@ -2704,13 +2705,14 @@ extern const struct file_operations generic_ro_fops;
2704 2705
2705extern int readlink_copy(char __user *, int, const char *); 2706extern int readlink_copy(char __user *, int, const char *);
2706extern int page_readlink(struct dentry *, char __user *, int); 2707extern int page_readlink(struct dentry *, char __user *, int);
2707extern void *page_follow_link_light(struct dentry *, struct nameidata *); 2708extern const char *page_follow_link_light(struct dentry *, void **);
2708extern void page_put_link(struct dentry *, struct nameidata *, void *); 2709extern void page_put_link(struct inode *, void *);
2709extern int __page_symlink(struct inode *inode, const char *symname, int len, 2710extern int __page_symlink(struct inode *inode, const char *symname, int len,
2710 int nofs); 2711 int nofs);
2711extern int page_symlink(struct inode *inode, const char *symname, int len); 2712extern int page_symlink(struct inode *inode, const char *symname, int len);
2712extern const struct inode_operations page_symlink_inode_operations; 2713extern const struct inode_operations page_symlink_inode_operations;
2713extern void kfree_put_link(struct dentry *, struct nameidata *, void *); 2714extern void kfree_put_link(struct inode *, void *);
2715extern void free_page_put_link(struct inode *, void *);
2714extern int generic_readlink(struct dentry *, char __user *, int); 2716extern int generic_readlink(struct dentry *, char __user *, int);
2715extern void generic_fillattr(struct inode *, struct kstat *); 2717extern void generic_fillattr(struct inode *, struct kstat *);
2716int vfs_getattr_nosec(struct path *path, struct kstat *stat); 2718int vfs_getattr_nosec(struct path *path, struct kstat *stat);
@@ -2721,6 +2723,8 @@ void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2721void inode_sub_bytes(struct inode *inode, loff_t bytes); 2723void inode_sub_bytes(struct inode *inode, loff_t bytes);
2722loff_t inode_get_bytes(struct inode *inode); 2724loff_t inode_get_bytes(struct inode *inode);
2723void inode_set_bytes(struct inode *inode, loff_t bytes); 2725void inode_set_bytes(struct inode *inode, loff_t bytes);
2726const char *simple_follow_link(struct dentry *, void **);
2727extern const struct inode_operations simple_symlink_inode_operations;
2724 2728
2725extern int iterate_dir(struct file *, struct dir_context *); 2729extern int iterate_dir(struct file *, struct dir_context *);
2726 2730
diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h
index f4af03404b97..dfd59d6bc6f0 100644
--- a/include/linux/hardirq.h
+++ b/include/linux/hardirq.h
@@ -1,7 +1,7 @@
1#ifndef LINUX_HARDIRQ_H 1#ifndef LINUX_HARDIRQ_H
2#define LINUX_HARDIRQ_H 2#define LINUX_HARDIRQ_H
3 3
4#include <linux/preempt_mask.h> 4#include <linux/preempt.h>
5#include <linux/lockdep.h> 5#include <linux/lockdep.h>
6#include <linux/ftrace_irq.h> 6#include <linux/ftrace_irq.h>
7#include <linux/vtime.h> 7#include <linux/vtime.h>
diff --git a/include/linux/highmem.h b/include/linux/highmem.h
index 9286a46b7d69..6aefcd0031a6 100644
--- a/include/linux/highmem.h
+++ b/include/linux/highmem.h
@@ -65,6 +65,7 @@ static inline void kunmap(struct page *page)
65 65
66static inline void *kmap_atomic(struct page *page) 66static inline void *kmap_atomic(struct page *page)
67{ 67{
68 preempt_disable();
68 pagefault_disable(); 69 pagefault_disable();
69 return page_address(page); 70 return page_address(page);
70} 71}
@@ -73,6 +74,7 @@ static inline void *kmap_atomic(struct page *page)
73static inline void __kunmap_atomic(void *addr) 74static inline void __kunmap_atomic(void *addr)
74{ 75{
75 pagefault_enable(); 76 pagefault_enable();
77 preempt_enable();
76} 78}
77 79
78#define kmap_atomic_pfn(pfn) kmap_atomic(pfn_to_page(pfn)) 80#define kmap_atomic_pfn(pfn) kmap_atomic(pfn_to_page(pfn))
diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h
index 05f6df1fdf5b..76dd4f0da5ca 100644
--- a/include/linux/hrtimer.h
+++ b/include/linux/hrtimer.h
@@ -53,34 +53,25 @@ enum hrtimer_restart {
53 * 53 *
54 * 0x00 inactive 54 * 0x00 inactive
55 * 0x01 enqueued into rbtree 55 * 0x01 enqueued into rbtree
56 * 0x02 callback function running
57 * 0x04 timer is migrated to another cpu
58 * 56 *
59 * Special cases: 57 * The callback state is not part of the timer->state because clearing it would
60 * 0x03 callback function running and enqueued 58 * mean touching the timer after the callback, this makes it impossible to free
61 * (was requeued on another CPU) 59 * the timer from the callback function.
62 * 0x05 timer was migrated on CPU hotunplug
63 * 60 *
64 * The "callback function running and enqueued" status is only possible on 61 * Therefore we track the callback state in:
65 * SMP. It happens for example when a posix timer expired and the callback 62 *
63 * timer->base->cpu_base->running == timer
64 *
65 * On SMP it is possible to have a "callback function running and enqueued"
66 * status. It happens for example when a posix timer expired and the callback
66 * queued a signal. Between dropping the lock which protects the posix timer 67 * queued a signal. Between dropping the lock which protects the posix timer
67 * and reacquiring the base lock of the hrtimer, another CPU can deliver the 68 * and reacquiring the base lock of the hrtimer, another CPU can deliver the
68 * signal and rearm the timer. We have to preserve the callback running state, 69 * signal and rearm the timer.
69 * as otherwise the timer could be removed before the softirq code finishes the
70 * the handling of the timer.
71 *
72 * The HRTIMER_STATE_ENQUEUED bit is always or'ed to the current state
73 * to preserve the HRTIMER_STATE_CALLBACK in the above scenario. This
74 * also affects HRTIMER_STATE_MIGRATE where the preservation is not
75 * necessary. HRTIMER_STATE_MIGRATE is cleared after the timer is
76 * enqueued on the new cpu.
77 * 70 *
78 * All state transitions are protected by cpu_base->lock. 71 * All state transitions are protected by cpu_base->lock.
79 */ 72 */
80#define HRTIMER_STATE_INACTIVE 0x00 73#define HRTIMER_STATE_INACTIVE 0x00
81#define HRTIMER_STATE_ENQUEUED 0x01 74#define HRTIMER_STATE_ENQUEUED 0x01
82#define HRTIMER_STATE_CALLBACK 0x02
83#define HRTIMER_STATE_MIGRATE 0x04
84 75
85/** 76/**
86 * struct hrtimer - the basic hrtimer structure 77 * struct hrtimer - the basic hrtimer structure
@@ -130,6 +121,12 @@ struct hrtimer_sleeper {
130 struct task_struct *task; 121 struct task_struct *task;
131}; 122};
132 123
124#ifdef CONFIG_64BIT
125# define HRTIMER_CLOCK_BASE_ALIGN 64
126#else
127# define HRTIMER_CLOCK_BASE_ALIGN 32
128#endif
129
133/** 130/**
134 * struct hrtimer_clock_base - the timer base for a specific clock 131 * struct hrtimer_clock_base - the timer base for a specific clock
135 * @cpu_base: per cpu clock base 132 * @cpu_base: per cpu clock base
@@ -137,9 +134,7 @@ struct hrtimer_sleeper {
137 * timer to a base on another cpu. 134 * timer to a base on another cpu.
138 * @clockid: clock id for per_cpu support 135 * @clockid: clock id for per_cpu support
139 * @active: red black tree root node for the active timers 136 * @active: red black tree root node for the active timers
140 * @resolution: the resolution of the clock, in nanoseconds
141 * @get_time: function to retrieve the current time of the clock 137 * @get_time: function to retrieve the current time of the clock
142 * @softirq_time: the time when running the hrtimer queue in the softirq
143 * @offset: offset of this clock to the monotonic base 138 * @offset: offset of this clock to the monotonic base
144 */ 139 */
145struct hrtimer_clock_base { 140struct hrtimer_clock_base {
@@ -147,11 +142,9 @@ struct hrtimer_clock_base {
147 int index; 142 int index;
148 clockid_t clockid; 143 clockid_t clockid;
149 struct timerqueue_head active; 144 struct timerqueue_head active;
150 ktime_t resolution;
151 ktime_t (*get_time)(void); 145 ktime_t (*get_time)(void);
152 ktime_t softirq_time;
153 ktime_t offset; 146 ktime_t offset;
154}; 147} __attribute__((__aligned__(HRTIMER_CLOCK_BASE_ALIGN)));
155 148
156enum hrtimer_base_type { 149enum hrtimer_base_type {
157 HRTIMER_BASE_MONOTONIC, 150 HRTIMER_BASE_MONOTONIC,
@@ -165,11 +158,16 @@ enum hrtimer_base_type {
165 * struct hrtimer_cpu_base - the per cpu clock bases 158 * struct hrtimer_cpu_base - the per cpu clock bases
166 * @lock: lock protecting the base and associated clock bases 159 * @lock: lock protecting the base and associated clock bases
167 * and timers 160 * and timers
161 * @seq: seqcount around __run_hrtimer
162 * @running: pointer to the currently running hrtimer
168 * @cpu: cpu number 163 * @cpu: cpu number
169 * @active_bases: Bitfield to mark bases with active timers 164 * @active_bases: Bitfield to mark bases with active timers
170 * @clock_was_set: Indicates that clock was set from irq context. 165 * @clock_was_set_seq: Sequence counter of clock was set events
166 * @migration_enabled: The migration of hrtimers to other cpus is enabled
167 * @nohz_active: The nohz functionality is enabled
171 * @expires_next: absolute time of the next event which was scheduled 168 * @expires_next: absolute time of the next event which was scheduled
172 * via clock_set_next_event() 169 * via clock_set_next_event()
170 * @next_timer: Pointer to the first expiring timer
173 * @in_hrtirq: hrtimer_interrupt() is currently executing 171 * @in_hrtirq: hrtimer_interrupt() is currently executing
174 * @hres_active: State of high resolution mode 172 * @hres_active: State of high resolution mode
175 * @hang_detected: The last hrtimer interrupt detected a hang 173 * @hang_detected: The last hrtimer interrupt detected a hang
@@ -178,27 +176,38 @@ enum hrtimer_base_type {
178 * @nr_hangs: Total number of hrtimer interrupt hangs 176 * @nr_hangs: Total number of hrtimer interrupt hangs
179 * @max_hang_time: Maximum time spent in hrtimer_interrupt 177 * @max_hang_time: Maximum time spent in hrtimer_interrupt
180 * @clock_base: array of clock bases for this cpu 178 * @clock_base: array of clock bases for this cpu
179 *
180 * Note: next_timer is just an optimization for __remove_hrtimer().
181 * Do not dereference the pointer because it is not reliable on
182 * cross cpu removals.
181 */ 183 */
182struct hrtimer_cpu_base { 184struct hrtimer_cpu_base {
183 raw_spinlock_t lock; 185 raw_spinlock_t lock;
186 seqcount_t seq;
187 struct hrtimer *running;
184 unsigned int cpu; 188 unsigned int cpu;
185 unsigned int active_bases; 189 unsigned int active_bases;
186 unsigned int clock_was_set; 190 unsigned int clock_was_set_seq;
191 bool migration_enabled;
192 bool nohz_active;
187#ifdef CONFIG_HIGH_RES_TIMERS 193#ifdef CONFIG_HIGH_RES_TIMERS
194 unsigned int in_hrtirq : 1,
195 hres_active : 1,
196 hang_detected : 1;
188 ktime_t expires_next; 197 ktime_t expires_next;
189 int in_hrtirq; 198 struct hrtimer *next_timer;
190 int hres_active; 199 unsigned int nr_events;
191 int hang_detected; 200 unsigned int nr_retries;
192 unsigned long nr_events; 201 unsigned int nr_hangs;
193 unsigned long nr_retries; 202 unsigned int max_hang_time;
194 unsigned long nr_hangs;
195 ktime_t max_hang_time;
196#endif 203#endif
197 struct hrtimer_clock_base clock_base[HRTIMER_MAX_CLOCK_BASES]; 204 struct hrtimer_clock_base clock_base[HRTIMER_MAX_CLOCK_BASES];
198}; 205} ____cacheline_aligned;
199 206
200static inline void hrtimer_set_expires(struct hrtimer *timer, ktime_t time) 207static inline void hrtimer_set_expires(struct hrtimer *timer, ktime_t time)
201{ 208{
209 BUILD_BUG_ON(sizeof(struct hrtimer_clock_base) > HRTIMER_CLOCK_BASE_ALIGN);
210
202 timer->node.expires = time; 211 timer->node.expires = time;
203 timer->_softexpires = time; 212 timer->_softexpires = time;
204} 213}
@@ -262,19 +271,16 @@ static inline ktime_t hrtimer_expires_remaining(const struct hrtimer *timer)
262 return ktime_sub(timer->node.expires, timer->base->get_time()); 271 return ktime_sub(timer->node.expires, timer->base->get_time());
263} 272}
264 273
265#ifdef CONFIG_HIGH_RES_TIMERS
266struct clock_event_device;
267
268extern void hrtimer_interrupt(struct clock_event_device *dev);
269
270/*
271 * In high resolution mode the time reference must be read accurate
272 */
273static inline ktime_t hrtimer_cb_get_time(struct hrtimer *timer) 274static inline ktime_t hrtimer_cb_get_time(struct hrtimer *timer)
274{ 275{
275 return timer->base->get_time(); 276 return timer->base->get_time();
276} 277}
277 278
279#ifdef CONFIG_HIGH_RES_TIMERS
280struct clock_event_device;
281
282extern void hrtimer_interrupt(struct clock_event_device *dev);
283
278static inline int hrtimer_is_hres_active(struct hrtimer *timer) 284static inline int hrtimer_is_hres_active(struct hrtimer *timer)
279{ 285{
280 return timer->base->cpu_base->hres_active; 286 return timer->base->cpu_base->hres_active;
@@ -295,21 +301,16 @@ extern void hrtimer_peek_ahead_timers(void);
295 301
296extern void clock_was_set_delayed(void); 302extern void clock_was_set_delayed(void);
297 303
304extern unsigned int hrtimer_resolution;
305
298#else 306#else
299 307
300# define MONOTONIC_RES_NSEC LOW_RES_NSEC 308# define MONOTONIC_RES_NSEC LOW_RES_NSEC
301# define KTIME_MONOTONIC_RES KTIME_LOW_RES 309# define KTIME_MONOTONIC_RES KTIME_LOW_RES
302 310
303static inline void hrtimer_peek_ahead_timers(void) { } 311#define hrtimer_resolution (unsigned int)LOW_RES_NSEC
304 312
305/* 313static inline void hrtimer_peek_ahead_timers(void) { }
306 * In non high resolution mode the time reference is taken from
307 * the base softirq time variable.
308 */
309static inline ktime_t hrtimer_cb_get_time(struct hrtimer *timer)
310{
311 return timer->base->softirq_time;
312}
313 314
314static inline int hrtimer_is_hres_active(struct hrtimer *timer) 315static inline int hrtimer_is_hres_active(struct hrtimer *timer)
315{ 316{
@@ -353,49 +354,47 @@ static inline void destroy_hrtimer_on_stack(struct hrtimer *timer) { }
353#endif 354#endif
354 355
355/* Basic timer operations: */ 356/* Basic timer operations: */
356extern int hrtimer_start(struct hrtimer *timer, ktime_t tim, 357extern void hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim,
357 const enum hrtimer_mode mode);
358extern int hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim,
359 unsigned long range_ns, const enum hrtimer_mode mode); 358 unsigned long range_ns, const enum hrtimer_mode mode);
360extern int 359
361__hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, 360/**
362 unsigned long delta_ns, 361 * hrtimer_start - (re)start an hrtimer on the current CPU
363 const enum hrtimer_mode mode, int wakeup); 362 * @timer: the timer to be added
363 * @tim: expiry time
364 * @mode: expiry mode: absolute (HRTIMER_MODE_ABS) or
365 * relative (HRTIMER_MODE_REL)
366 */
367static inline void hrtimer_start(struct hrtimer *timer, ktime_t tim,
368 const enum hrtimer_mode mode)
369{
370 hrtimer_start_range_ns(timer, tim, 0, mode);
371}
364 372
365extern int hrtimer_cancel(struct hrtimer *timer); 373extern int hrtimer_cancel(struct hrtimer *timer);
366extern int hrtimer_try_to_cancel(struct hrtimer *timer); 374extern int hrtimer_try_to_cancel(struct hrtimer *timer);
367 375
368static inline int hrtimer_start_expires(struct hrtimer *timer, 376static inline void hrtimer_start_expires(struct hrtimer *timer,
369 enum hrtimer_mode mode) 377 enum hrtimer_mode mode)
370{ 378{
371 unsigned long delta; 379 unsigned long delta;
372 ktime_t soft, hard; 380 ktime_t soft, hard;
373 soft = hrtimer_get_softexpires(timer); 381 soft = hrtimer_get_softexpires(timer);
374 hard = hrtimer_get_expires(timer); 382 hard = hrtimer_get_expires(timer);
375 delta = ktime_to_ns(ktime_sub(hard, soft)); 383 delta = ktime_to_ns(ktime_sub(hard, soft));
376 return hrtimer_start_range_ns(timer, soft, delta, mode); 384 hrtimer_start_range_ns(timer, soft, delta, mode);
377} 385}
378 386
379static inline int hrtimer_restart(struct hrtimer *timer) 387static inline void hrtimer_restart(struct hrtimer *timer)
380{ 388{
381 return hrtimer_start_expires(timer, HRTIMER_MODE_ABS); 389 hrtimer_start_expires(timer, HRTIMER_MODE_ABS);
382} 390}
383 391
384/* Query timers: */ 392/* Query timers: */
385extern ktime_t hrtimer_get_remaining(const struct hrtimer *timer); 393extern ktime_t hrtimer_get_remaining(const struct hrtimer *timer);
386extern int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp);
387 394
388extern ktime_t hrtimer_get_next_event(void); 395extern u64 hrtimer_get_next_event(void);
389 396
390/* 397extern bool hrtimer_active(const struct hrtimer *timer);
391 * A timer is active, when it is enqueued into the rbtree or the
392 * callback function is running or it's in the state of being migrated
393 * to another cpu.
394 */
395static inline int hrtimer_active(const struct hrtimer *timer)
396{
397 return timer->state != HRTIMER_STATE_INACTIVE;
398}
399 398
400/* 399/*
401 * Helper function to check, whether the timer is on one of the queues 400 * Helper function to check, whether the timer is on one of the queues
@@ -411,14 +410,29 @@ static inline int hrtimer_is_queued(struct hrtimer *timer)
411 */ 410 */
412static inline int hrtimer_callback_running(struct hrtimer *timer) 411static inline int hrtimer_callback_running(struct hrtimer *timer)
413{ 412{
414 return timer->state & HRTIMER_STATE_CALLBACK; 413 return timer->base->cpu_base->running == timer;
415} 414}
416 415
417/* Forward a hrtimer so it expires after now: */ 416/* Forward a hrtimer so it expires after now: */
418extern u64 417extern u64
419hrtimer_forward(struct hrtimer *timer, ktime_t now, ktime_t interval); 418hrtimer_forward(struct hrtimer *timer, ktime_t now, ktime_t interval);
420 419
421/* Forward a hrtimer so it expires after the hrtimer's current now */ 420/**
421 * hrtimer_forward_now - forward the timer expiry so it expires after now
422 * @timer: hrtimer to forward
423 * @interval: the interval to forward
424 *
425 * Forward the timer expiry so it will expire after the current time
426 * of the hrtimer clock base. Returns the number of overruns.
427 *
428 * Can be safely called from the callback function of @timer. If
429 * called from other contexts @timer must neither be enqueued nor
430 * running the callback and the caller needs to take care of
431 * serialization.
432 *
433 * Note: This only updates the timer expiry value and does not requeue
434 * the timer.
435 */
422static inline u64 hrtimer_forward_now(struct hrtimer *timer, 436static inline u64 hrtimer_forward_now(struct hrtimer *timer,
423 ktime_t interval) 437 ktime_t interval)
424{ 438{
@@ -443,7 +457,6 @@ extern int schedule_hrtimeout(ktime_t *expires, const enum hrtimer_mode mode);
443 457
444/* Soft interrupt function to run the hrtimer queues: */ 458/* Soft interrupt function to run the hrtimer queues: */
445extern void hrtimer_run_queues(void); 459extern void hrtimer_run_queues(void);
446extern void hrtimer_run_pending(void);
447 460
448/* Bootup initialization: */ 461/* Bootup initialization: */
449extern void __init hrtimers_init(void); 462extern void __init hrtimers_init(void);
diff --git a/include/linux/htirq.h b/include/linux/htirq.h
index 70a1dbbf2093..d4a527e58434 100644
--- a/include/linux/htirq.h
+++ b/include/linux/htirq.h
@@ -1,24 +1,38 @@
1#ifndef LINUX_HTIRQ_H 1#ifndef LINUX_HTIRQ_H
2#define LINUX_HTIRQ_H 2#define LINUX_HTIRQ_H
3 3
4struct pci_dev;
5struct irq_data;
6
4struct ht_irq_msg { 7struct ht_irq_msg {
5 u32 address_lo; /* low 32 bits of the ht irq message */ 8 u32 address_lo; /* low 32 bits of the ht irq message */
6 u32 address_hi; /* high 32 bits of the it irq message */ 9 u32 address_hi; /* high 32 bits of the it irq message */
7}; 10};
8 11
12typedef void (ht_irq_update_t)(struct pci_dev *dev, int irq,
13 struct ht_irq_msg *msg);
14
15struct ht_irq_cfg {
16 struct pci_dev *dev;
17 /* Update callback used to cope with buggy hardware */
18 ht_irq_update_t *update;
19 unsigned pos;
20 unsigned idx;
21 struct ht_irq_msg msg;
22};
23
9/* Helper functions.. */ 24/* Helper functions.. */
10void fetch_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg); 25void fetch_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg);
11void write_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg); 26void write_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg);
12struct irq_data;
13void mask_ht_irq(struct irq_data *data); 27void mask_ht_irq(struct irq_data *data);
14void unmask_ht_irq(struct irq_data *data); 28void unmask_ht_irq(struct irq_data *data);
15 29
16/* The arch hook for getting things started */ 30/* The arch hook for getting things started */
17int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev); 31int arch_setup_ht_irq(int idx, int pos, struct pci_dev *dev,
32 ht_irq_update_t *update);
33void arch_teardown_ht_irq(unsigned int irq);
18 34
19/* For drivers of buggy hardware */ 35/* For drivers of buggy hardware */
20typedef void (ht_irq_update_t)(struct pci_dev *dev, int irq,
21 struct ht_irq_msg *msg);
22int __ht_create_irq(struct pci_dev *dev, int idx, ht_irq_update_t *update); 36int __ht_create_irq(struct pci_dev *dev, int idx, ht_irq_update_t *update);
23 37
24#endif /* LINUX_HTIRQ_H */ 38#endif /* LINUX_HTIRQ_H */
diff --git a/include/linux/init_task.h b/include/linux/init_task.h
index 696d22312b31..bb9b075f0eb0 100644
--- a/include/linux/init_task.h
+++ b/include/linux/init_task.h
@@ -50,9 +50,8 @@ extern struct fs_struct init_fs;
50 .cpu_timers = INIT_CPU_TIMERS(sig.cpu_timers), \ 50 .cpu_timers = INIT_CPU_TIMERS(sig.cpu_timers), \
51 .rlim = INIT_RLIMITS, \ 51 .rlim = INIT_RLIMITS, \
52 .cputimer = { \ 52 .cputimer = { \
53 .cputime = INIT_CPUTIME, \ 53 .cputime_atomic = INIT_CPUTIME_ATOMIC, \
54 .running = 0, \ 54 .running = 0, \
55 .lock = __RAW_SPIN_LOCK_UNLOCKED(sig.cputimer.lock), \
56 }, \ 55 }, \
57 .cred_guard_mutex = \ 56 .cred_guard_mutex = \
58 __MUTEX_INITIALIZER(sig.cred_guard_mutex), \ 57 __MUTEX_INITIALIZER(sig.cred_guard_mutex), \
diff --git a/include/linux/intel-iommu.h b/include/linux/intel-iommu.h
index 796ef9645827..3665cb331ca1 100644
--- a/include/linux/intel-iommu.h
+++ b/include/linux/intel-iommu.h
@@ -87,6 +87,7 @@ static inline void dmar_writeq(void __iomem *addr, u64 val)
87/* 87/*
88 * Decoding Capability Register 88 * Decoding Capability Register
89 */ 89 */
90#define cap_pi_support(c) (((c) >> 59) & 1)
90#define cap_read_drain(c) (((c) >> 55) & 1) 91#define cap_read_drain(c) (((c) >> 55) & 1)
91#define cap_write_drain(c) (((c) >> 54) & 1) 92#define cap_write_drain(c) (((c) >> 54) & 1)
92#define cap_max_amask_val(c) (((c) >> 48) & 0x3f) 93#define cap_max_amask_val(c) (((c) >> 48) & 0x3f)
@@ -115,13 +116,14 @@ static inline void dmar_writeq(void __iomem *addr, u64 val)
115 * Extended Capability Register 116 * Extended Capability Register
116 */ 117 */
117 118
119#define ecap_pasid(e) ((e >> 40) & 0x1)
118#define ecap_pss(e) ((e >> 35) & 0x1f) 120#define ecap_pss(e) ((e >> 35) & 0x1f)
119#define ecap_eafs(e) ((e >> 34) & 0x1) 121#define ecap_eafs(e) ((e >> 34) & 0x1)
120#define ecap_nwfs(e) ((e >> 33) & 0x1) 122#define ecap_nwfs(e) ((e >> 33) & 0x1)
121#define ecap_srs(e) ((e >> 31) & 0x1) 123#define ecap_srs(e) ((e >> 31) & 0x1)
122#define ecap_ers(e) ((e >> 30) & 0x1) 124#define ecap_ers(e) ((e >> 30) & 0x1)
123#define ecap_prs(e) ((e >> 29) & 0x1) 125#define ecap_prs(e) ((e >> 29) & 0x1)
124#define ecap_pasid(e) ((e >> 28) & 0x1) 126/* PASID support used to be on bit 28 */
125#define ecap_dis(e) ((e >> 27) & 0x1) 127#define ecap_dis(e) ((e >> 27) & 0x1)
126#define ecap_nest(e) ((e >> 26) & 0x1) 128#define ecap_nest(e) ((e >> 26) & 0x1)
127#define ecap_mts(e) ((e >> 25) & 0x1) 129#define ecap_mts(e) ((e >> 25) & 0x1)
@@ -298,6 +300,8 @@ struct q_inval {
298 300
299#define INTR_REMAP_TABLE_ENTRIES 65536 301#define INTR_REMAP_TABLE_ENTRIES 65536
300 302
303struct irq_domain;
304
301struct ir_table { 305struct ir_table {
302 struct irte *base; 306 struct irte *base;
303 unsigned long *bitmap; 307 unsigned long *bitmap;
@@ -347,6 +351,8 @@ struct intel_iommu {
347 351
348#ifdef CONFIG_IRQ_REMAP 352#ifdef CONFIG_IRQ_REMAP
349 struct ir_table *ir_table; /* Interrupt remapping info */ 353 struct ir_table *ir_table; /* Interrupt remapping info */
354 struct irq_domain *ir_domain;
355 struct irq_domain *ir_msi_domain;
350#endif 356#endif
351 struct device *iommu_dev; /* IOMMU-sysfs device */ 357 struct device *iommu_dev; /* IOMMU-sysfs device */
352 int node; 358 int node;
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 950ae4501826..be7e75c945e9 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -413,7 +413,8 @@ enum
413 BLOCK_IOPOLL_SOFTIRQ, 413 BLOCK_IOPOLL_SOFTIRQ,
414 TASKLET_SOFTIRQ, 414 TASKLET_SOFTIRQ,
415 SCHED_SOFTIRQ, 415 SCHED_SOFTIRQ,
416 HRTIMER_SOFTIRQ, 416 HRTIMER_SOFTIRQ, /* Unused, but kept as tools rely on the
417 numbering. Sigh! */
417 RCU_SOFTIRQ, /* Preferable RCU should always be the last softirq */ 418 RCU_SOFTIRQ, /* Preferable RCU should always be the last softirq */
418 419
419 NR_SOFTIRQS 420 NR_SOFTIRQS
@@ -592,10 +593,10 @@ tasklet_hrtimer_init(struct tasklet_hrtimer *ttimer,
592 clockid_t which_clock, enum hrtimer_mode mode); 593 clockid_t which_clock, enum hrtimer_mode mode);
593 594
594static inline 595static inline
595int tasklet_hrtimer_start(struct tasklet_hrtimer *ttimer, ktime_t time, 596void tasklet_hrtimer_start(struct tasklet_hrtimer *ttimer, ktime_t time,
596 const enum hrtimer_mode mode) 597 const enum hrtimer_mode mode)
597{ 598{
598 return hrtimer_start(&ttimer->timer, time, mode); 599 hrtimer_start(&ttimer->timer, time, mode);
599} 600}
600 601
601static inline 602static inline
diff --git a/include/linux/io-mapping.h b/include/linux/io-mapping.h
index 657fab4efab3..c27dde7215b5 100644
--- a/include/linux/io-mapping.h
+++ b/include/linux/io-mapping.h
@@ -141,6 +141,7 @@ static inline void __iomem *
141io_mapping_map_atomic_wc(struct io_mapping *mapping, 141io_mapping_map_atomic_wc(struct io_mapping *mapping,
142 unsigned long offset) 142 unsigned long offset)
143{ 143{
144 preempt_disable();
144 pagefault_disable(); 145 pagefault_disable();
145 return ((char __force __iomem *) mapping) + offset; 146 return ((char __force __iomem *) mapping) + offset;
146} 147}
@@ -149,6 +150,7 @@ static inline void
149io_mapping_unmap_atomic(void __iomem *vaddr) 150io_mapping_unmap_atomic(void __iomem *vaddr)
150{ 151{
151 pagefault_enable(); 152 pagefault_enable();
153 preempt_enable();
152} 154}
153 155
154/* Non-atomic map/unmap */ 156/* Non-atomic map/unmap */
diff --git a/include/linux/io.h b/include/linux/io.h
index 986f2bffea1e..fb5a99800e77 100644
--- a/include/linux/io.h
+++ b/include/linux/io.h
@@ -19,6 +19,7 @@
19#define _LINUX_IO_H 19#define _LINUX_IO_H
20 20
21#include <linux/types.h> 21#include <linux/types.h>
22#include <linux/init.h>
22#include <asm/io.h> 23#include <asm/io.h>
23#include <asm/page.h> 24#include <asm/page.h>
24 25
@@ -111,6 +112,13 @@ static inline void arch_phys_wc_del(int handle)
111} 112}
112 113
113#define arch_phys_wc_add arch_phys_wc_add 114#define arch_phys_wc_add arch_phys_wc_add
115#ifndef arch_phys_wc_index
116static inline int arch_phys_wc_index(int handle)
117{
118 return -1;
119}
120#define arch_phys_wc_index arch_phys_wc_index
121#endif
114#endif 122#endif
115 123
116#endif /* _LINUX_IO_H */ 124#endif /* _LINUX_IO_H */
diff --git a/include/linux/jiffies.h b/include/linux/jiffies.h
index c367cbdf73ab..535fd3bb1ba8 100644
--- a/include/linux/jiffies.h
+++ b/include/linux/jiffies.h
@@ -7,6 +7,7 @@
7#include <linux/time.h> 7#include <linux/time.h>
8#include <linux/timex.h> 8#include <linux/timex.h>
9#include <asm/param.h> /* for HZ */ 9#include <asm/param.h> /* for HZ */
10#include <generated/timeconst.h>
10 11
11/* 12/*
12 * The following defines establish the engineering parameters of the PLL 13 * The following defines establish the engineering parameters of the PLL
@@ -288,8 +289,133 @@ static inline u64 jiffies_to_nsecs(const unsigned long j)
288 return (u64)jiffies_to_usecs(j) * NSEC_PER_USEC; 289 return (u64)jiffies_to_usecs(j) * NSEC_PER_USEC;
289} 290}
290 291
291extern unsigned long msecs_to_jiffies(const unsigned int m); 292extern unsigned long __msecs_to_jiffies(const unsigned int m);
292extern unsigned long usecs_to_jiffies(const unsigned int u); 293#if HZ <= MSEC_PER_SEC && !(MSEC_PER_SEC % HZ)
294/*
295 * HZ is equal to or smaller than 1000, and 1000 is a nice round
296 * multiple of HZ, divide with the factor between them, but round
297 * upwards:
298 */
299static inline unsigned long _msecs_to_jiffies(const unsigned int m)
300{
301 return (m + (MSEC_PER_SEC / HZ) - 1) / (MSEC_PER_SEC / HZ);
302}
303#elif HZ > MSEC_PER_SEC && !(HZ % MSEC_PER_SEC)
304/*
305 * HZ is larger than 1000, and HZ is a nice round multiple of 1000 -
306 * simply multiply with the factor between them.
307 *
308 * But first make sure the multiplication result cannot overflow:
309 */
310static inline unsigned long _msecs_to_jiffies(const unsigned int m)
311{
312 if (m > jiffies_to_msecs(MAX_JIFFY_OFFSET))
313 return MAX_JIFFY_OFFSET;
314 return m * (HZ / MSEC_PER_SEC);
315}
316#else
317/*
318 * Generic case - multiply, round and divide. But first check that if
319 * we are doing a net multiplication, that we wouldn't overflow:
320 */
321static inline unsigned long _msecs_to_jiffies(const unsigned int m)
322{
323 if (HZ > MSEC_PER_SEC && m > jiffies_to_msecs(MAX_JIFFY_OFFSET))
324 return MAX_JIFFY_OFFSET;
325
326 return (MSEC_TO_HZ_MUL32 * m + MSEC_TO_HZ_ADJ32) >> MSEC_TO_HZ_SHR32;
327}
328#endif
329/**
330 * msecs_to_jiffies: - convert milliseconds to jiffies
331 * @m: time in milliseconds
332 *
333 * conversion is done as follows:
334 *
335 * - negative values mean 'infinite timeout' (MAX_JIFFY_OFFSET)
336 *
337 * - 'too large' values [that would result in larger than
338 * MAX_JIFFY_OFFSET values] mean 'infinite timeout' too.
339 *
340 * - all other values are converted to jiffies by either multiplying
341 * the input value by a factor or dividing it with a factor and
342 * handling any 32-bit overflows.
343 * for the details see __msecs_to_jiffies()
344 *
345 * msecs_to_jiffies() checks for the passed in value being a constant
346 * via __builtin_constant_p() allowing gcc to eliminate most of the
347 * code, __msecs_to_jiffies() is called if the value passed does not
348 * allow constant folding and the actual conversion must be done at
349 * runtime.
350 * the HZ range specific helpers _msecs_to_jiffies() are called both
351 * directly here and from __msecs_to_jiffies() in the case where
352 * constant folding is not possible.
353 */
354static inline unsigned long msecs_to_jiffies(const unsigned int m)
355{
356 if (__builtin_constant_p(m)) {
357 if ((int)m < 0)
358 return MAX_JIFFY_OFFSET;
359 return _msecs_to_jiffies(m);
360 } else {
361 return __msecs_to_jiffies(m);
362 }
363}
364
365extern unsigned long __usecs_to_jiffies(const unsigned int u);
366#if HZ <= USEC_PER_SEC && !(USEC_PER_SEC % HZ)
367static inline unsigned long _usecs_to_jiffies(const unsigned int u)
368{
369 return (u + (USEC_PER_SEC / HZ) - 1) / (USEC_PER_SEC / HZ);
370}
371#elif HZ > USEC_PER_SEC && !(HZ % USEC_PER_SEC)
372static inline unsigned long _usecs_to_jiffies(const unsigned int u)
373{
374 return u * (HZ / USEC_PER_SEC);
375}
376static inline unsigned long _usecs_to_jiffies(const unsigned int u)
377{
378#else
379static inline unsigned long _usecs_to_jiffies(const unsigned int u)
380{
381 return (USEC_TO_HZ_MUL32 * u + USEC_TO_HZ_ADJ32)
382 >> USEC_TO_HZ_SHR32;
383}
384#endif
385
386/**
387 * usecs_to_jiffies: - convert microseconds to jiffies
388 * @u: time in microseconds
389 *
390 * conversion is done as follows:
391 *
392 * - 'too large' values [that would result in larger than
393 * MAX_JIFFY_OFFSET values] mean 'infinite timeout' too.
394 *
395 * - all other values are converted to jiffies by either multiplying
396 * the input value by a factor or dividing it with a factor and
397 * handling any 32-bit overflows as for msecs_to_jiffies.
398 *
399 * usecs_to_jiffies() checks for the passed in value being a constant
400 * via __builtin_constant_p() allowing gcc to eliminate most of the
401 * code, __usecs_to_jiffies() is called if the value passed does not
402 * allow constant folding and the actual conversion must be done at
403 * runtime.
404 * the HZ range specific helpers _usecs_to_jiffies() are called both
405 * directly here and from __msecs_to_jiffies() in the case where
406 * constant folding is not possible.
407 */
408static inline unsigned long usecs_to_jiffies(const unsigned int u)
409{
410 if (__builtin_constant_p(u)) {
411 if (u > jiffies_to_usecs(MAX_JIFFY_OFFSET))
412 return MAX_JIFFY_OFFSET;
413 return _usecs_to_jiffies(u);
414 } else {
415 return __usecs_to_jiffies(u);
416 }
417}
418
293extern unsigned long timespec_to_jiffies(const struct timespec *value); 419extern unsigned long timespec_to_jiffies(const struct timespec *value);
294extern void jiffies_to_timespec(const unsigned long jiffies, 420extern void jiffies_to_timespec(const unsigned long jiffies,
295 struct timespec *value); 421 struct timespec *value);
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 3a5b48e52a9e..060dd7b61c6d 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -244,7 +244,8 @@ static inline u32 reciprocal_scale(u32 val, u32 ep_ro)
244 244
245#if defined(CONFIG_MMU) && \ 245#if defined(CONFIG_MMU) && \
246 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP)) 246 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP))
247void might_fault(void); 247#define might_fault() __might_fault(__FILE__, __LINE__)
248void __might_fault(const char *file, int line);
248#else 249#else
249static inline void might_fault(void) { } 250static inline void might_fault(void) { }
250#endif 251#endif
diff --git a/include/linux/ktime.h b/include/linux/ktime.h
index 5fc3d1083071..2b6a204bd8d4 100644
--- a/include/linux/ktime.h
+++ b/include/linux/ktime.h
@@ -166,19 +166,34 @@ static inline bool ktime_before(const ktime_t cmp1, const ktime_t cmp2)
166} 166}
167 167
168#if BITS_PER_LONG < 64 168#if BITS_PER_LONG < 64
169extern u64 __ktime_divns(const ktime_t kt, s64 div); 169extern s64 __ktime_divns(const ktime_t kt, s64 div);
170static inline u64 ktime_divns(const ktime_t kt, s64 div) 170static inline s64 ktime_divns(const ktime_t kt, s64 div)
171{ 171{
172 /*
173 * Negative divisors could cause an inf loop,
174 * so bug out here.
175 */
176 BUG_ON(div < 0);
172 if (__builtin_constant_p(div) && !(div >> 32)) { 177 if (__builtin_constant_p(div) && !(div >> 32)) {
173 u64 ns = kt.tv64; 178 s64 ns = kt.tv64;
174 do_div(ns, div); 179 u64 tmp = ns < 0 ? -ns : ns;
175 return ns; 180
181 do_div(tmp, div);
182 return ns < 0 ? -tmp : tmp;
176 } else { 183 } else {
177 return __ktime_divns(kt, div); 184 return __ktime_divns(kt, div);
178 } 185 }
179} 186}
180#else /* BITS_PER_LONG < 64 */ 187#else /* BITS_PER_LONG < 64 */
181# define ktime_divns(kt, div) (u64)((kt).tv64 / (div)) 188static inline s64 ktime_divns(const ktime_t kt, s64 div)
189{
190 /*
191 * 32-bit implementation cannot handle negative divisors,
192 * so catch them on 64bit as well.
193 */
194 WARN_ON(div < 0);
195 return kt.tv64 / div;
196}
182#endif 197#endif
183 198
184static inline s64 ktime_to_us(const ktime_t kt) 199static inline s64 ktime_to_us(const ktime_t kt)
diff --git a/include/linux/lglock.h b/include/linux/lglock.h
index 0081f000e34b..c92ebd100d9b 100644
--- a/include/linux/lglock.h
+++ b/include/linux/lglock.h
@@ -52,10 +52,15 @@ struct lglock {
52 static struct lglock name = { .lock = &name ## _lock } 52 static struct lglock name = { .lock = &name ## _lock }
53 53
54void lg_lock_init(struct lglock *lg, char *name); 54void lg_lock_init(struct lglock *lg, char *name);
55
55void lg_local_lock(struct lglock *lg); 56void lg_local_lock(struct lglock *lg);
56void lg_local_unlock(struct lglock *lg); 57void lg_local_unlock(struct lglock *lg);
57void lg_local_lock_cpu(struct lglock *lg, int cpu); 58void lg_local_lock_cpu(struct lglock *lg, int cpu);
58void lg_local_unlock_cpu(struct lglock *lg, int cpu); 59void lg_local_unlock_cpu(struct lglock *lg, int cpu);
60
61void lg_double_lock(struct lglock *lg, int cpu1, int cpu2);
62void lg_double_unlock(struct lglock *lg, int cpu1, int cpu2);
63
59void lg_global_lock(struct lglock *lg); 64void lg_global_lock(struct lglock *lg);
60void lg_global_unlock(struct lglock *lg); 65void lg_global_unlock(struct lglock *lg);
61 66
diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
index 066ba4157541..2722111591a3 100644
--- a/include/linux/lockdep.h
+++ b/include/linux/lockdep.h
@@ -130,8 +130,8 @@ enum bounce_type {
130}; 130};
131 131
132struct lock_class_stats { 132struct lock_class_stats {
133 unsigned long contention_point[4]; 133 unsigned long contention_point[LOCKSTAT_POINTS];
134 unsigned long contending_point[4]; 134 unsigned long contending_point[LOCKSTAT_POINTS];
135 struct lock_time read_waittime; 135 struct lock_time read_waittime;
136 struct lock_time write_waittime; 136 struct lock_time write_waittime;
137 struct lock_time read_holdtime; 137 struct lock_time read_holdtime;
diff --git a/include/linux/namei.h b/include/linux/namei.h
index c8990779f0c3..d8c6334cd150 100644
--- a/include/linux/namei.h
+++ b/include/linux/namei.h
@@ -1,16 +1,15 @@
1#ifndef _LINUX_NAMEI_H 1#ifndef _LINUX_NAMEI_H
2#define _LINUX_NAMEI_H 2#define _LINUX_NAMEI_H
3 3
4#include <linux/dcache.h> 4#include <linux/kernel.h>
5#include <linux/errno.h>
6#include <linux/linkage.h>
7#include <linux/path.h> 5#include <linux/path.h>
8 6#include <linux/fcntl.h>
9struct vfsmount; 7#include <linux/errno.h>
10struct nameidata;
11 8
12enum { MAX_NESTED_LINKS = 8 }; 9enum { MAX_NESTED_LINKS = 8 };
13 10
11#define MAXSYMLINKS 40
12
14/* 13/*
15 * Type of the last component on LOOKUP_PARENT 14 * Type of the last component on LOOKUP_PARENT
16 */ 15 */
@@ -45,13 +44,29 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND};
45#define LOOKUP_ROOT 0x2000 44#define LOOKUP_ROOT 0x2000
46#define LOOKUP_EMPTY 0x4000 45#define LOOKUP_EMPTY 0x4000
47 46
48extern int user_path_at(int, const char __user *, unsigned, struct path *);
49extern int user_path_at_empty(int, const char __user *, unsigned, struct path *, int *empty); 47extern int user_path_at_empty(int, const char __user *, unsigned, struct path *, int *empty);
50 48
51#define user_path(name, path) user_path_at(AT_FDCWD, name, LOOKUP_FOLLOW, path) 49static inline int user_path_at(int dfd, const char __user *name, unsigned flags,
52#define user_lpath(name, path) user_path_at(AT_FDCWD, name, 0, path) 50 struct path *path)
53#define user_path_dir(name, path) \ 51{
54 user_path_at(AT_FDCWD, name, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, path) 52 return user_path_at_empty(dfd, name, flags, path, NULL);
53}
54
55static inline int user_path(const char __user *name, struct path *path)
56{
57 return user_path_at_empty(AT_FDCWD, name, LOOKUP_FOLLOW, path, NULL);
58}
59
60static inline int user_lpath(const char __user *name, struct path *path)
61{
62 return user_path_at_empty(AT_FDCWD, name, 0, path, NULL);
63}
64
65static inline int user_path_dir(const char __user *name, struct path *path)
66{
67 return user_path_at_empty(AT_FDCWD, name,
68 LOOKUP_FOLLOW | LOOKUP_DIRECTORY, path, NULL);
69}
55 70
56extern int kern_path(const char *, unsigned, struct path *); 71extern int kern_path(const char *, unsigned, struct path *);
57 72
@@ -70,9 +85,7 @@ extern int follow_up(struct path *);
70extern struct dentry *lock_rename(struct dentry *, struct dentry *); 85extern struct dentry *lock_rename(struct dentry *, struct dentry *);
71extern void unlock_rename(struct dentry *, struct dentry *); 86extern void unlock_rename(struct dentry *, struct dentry *);
72 87
73extern void nd_jump_link(struct nameidata *nd, struct path *path); 88extern void nd_jump_link(struct path *path);
74extern void nd_set_link(struct nameidata *nd, char *path);
75extern char *nd_get_link(struct nameidata *nd);
76 89
77static inline void nd_terminate_link(void *name, size_t len, size_t maxlen) 90static inline void nd_terminate_link(void *name, size_t len, size_t maxlen)
78{ 91{
diff --git a/include/linux/of.h b/include/linux/of.h
index ddeaae6d2083..b871ff9d81d7 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -121,6 +121,8 @@ extern struct device_node *of_stdout;
121extern raw_spinlock_t devtree_lock; 121extern raw_spinlock_t devtree_lock;
122 122
123#ifdef CONFIG_OF 123#ifdef CONFIG_OF
124void of_core_init(void);
125
124static inline bool is_of_node(struct fwnode_handle *fwnode) 126static inline bool is_of_node(struct fwnode_handle *fwnode)
125{ 127{
126 return fwnode && fwnode->type == FWNODE_OF; 128 return fwnode && fwnode->type == FWNODE_OF;
@@ -376,6 +378,10 @@ bool of_console_check(struct device_node *dn, char *name, int index);
376 378
377#else /* CONFIG_OF */ 379#else /* CONFIG_OF */
378 380
381static inline void of_core_init(void)
382{
383}
384
379static inline bool is_of_node(struct fwnode_handle *fwnode) 385static inline bool is_of_node(struct fwnode_handle *fwnode)
380{ 386{
381 return false; 387 return false;
diff --git a/include/linux/osq_lock.h b/include/linux/osq_lock.h
index 3a6490e81b28..703ea5c30a33 100644
--- a/include/linux/osq_lock.h
+++ b/include/linux/osq_lock.h
@@ -32,4 +32,9 @@ static inline void osq_lock_init(struct optimistic_spin_queue *lock)
32extern bool osq_lock(struct optimistic_spin_queue *lock); 32extern bool osq_lock(struct optimistic_spin_queue *lock);
33extern void osq_unlock(struct optimistic_spin_queue *lock); 33extern void osq_unlock(struct optimistic_spin_queue *lock);
34 34
35static inline bool osq_is_locked(struct optimistic_spin_queue *lock)
36{
37 return atomic_read(&lock->tail) != OSQ_UNLOCKED_VAL;
38}
39
35#endif 40#endif
diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h
index 50e50095c8d1..84a109449610 100644
--- a/include/linux/percpu_counter.h
+++ b/include/linux/percpu_counter.h
@@ -41,7 +41,12 @@ void percpu_counter_destroy(struct percpu_counter *fbc);
41void percpu_counter_set(struct percpu_counter *fbc, s64 amount); 41void percpu_counter_set(struct percpu_counter *fbc, s64 amount);
42void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch); 42void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch);
43s64 __percpu_counter_sum(struct percpu_counter *fbc); 43s64 __percpu_counter_sum(struct percpu_counter *fbc);
44int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs); 44int __percpu_counter_compare(struct percpu_counter *fbc, s64 rhs, s32 batch);
45
46static inline int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs)
47{
48 return __percpu_counter_compare(fbc, rhs, percpu_counter_batch);
49}
45 50
46static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount) 51static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount)
47{ 52{
@@ -116,6 +121,12 @@ static inline int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs)
116 return 0; 121 return 0;
117} 122}
118 123
124static inline int
125__percpu_counter_compare(struct percpu_counter *fbc, s64 rhs, s32 batch)
126{
127 return percpu_counter_compare(fbc, rhs);
128}
129
119static inline void 130static inline void
120percpu_counter_add(struct percpu_counter *fbc, s64 amount) 131percpu_counter_add(struct percpu_counter *fbc, s64 amount)
121{ 132{
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 61992cf2e977..1b82d44b0a02 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -92,8 +92,6 @@ struct hw_perf_event_extra {
92 int idx; /* index in shared_regs->regs[] */ 92 int idx; /* index in shared_regs->regs[] */
93}; 93};
94 94
95struct event_constraint;
96
97/** 95/**
98 * struct hw_perf_event - performance event hardware details: 96 * struct hw_perf_event - performance event hardware details:
99 */ 97 */
@@ -112,8 +110,6 @@ struct hw_perf_event {
112 110
113 struct hw_perf_event_extra extra_reg; 111 struct hw_perf_event_extra extra_reg;
114 struct hw_perf_event_extra branch_reg; 112 struct hw_perf_event_extra branch_reg;
115
116 struct event_constraint *constraint;
117 }; 113 };
118 struct { /* software */ 114 struct { /* software */
119 struct hrtimer hrtimer; 115 struct hrtimer hrtimer;
@@ -124,7 +120,7 @@ struct hw_perf_event {
124 }; 120 };
125 struct { /* intel_cqm */ 121 struct { /* intel_cqm */
126 int cqm_state; 122 int cqm_state;
127 int cqm_rmid; 123 u32 cqm_rmid;
128 struct list_head cqm_events_entry; 124 struct list_head cqm_events_entry;
129 struct list_head cqm_groups_entry; 125 struct list_head cqm_groups_entry;
130 struct list_head cqm_group_entry; 126 struct list_head cqm_group_entry;
@@ -566,8 +562,12 @@ struct perf_cpu_context {
566 struct perf_event_context *task_ctx; 562 struct perf_event_context *task_ctx;
567 int active_oncpu; 563 int active_oncpu;
568 int exclusive; 564 int exclusive;
565
566 raw_spinlock_t hrtimer_lock;
569 struct hrtimer hrtimer; 567 struct hrtimer hrtimer;
570 ktime_t hrtimer_interval; 568 ktime_t hrtimer_interval;
569 unsigned int hrtimer_active;
570
571 struct pmu *unique_pmu; 571 struct pmu *unique_pmu;
572 struct perf_cgroup *cgrp; 572 struct perf_cgroup *cgrp;
573}; 573};
@@ -734,6 +734,22 @@ extern int perf_event_overflow(struct perf_event *event,
734 struct perf_sample_data *data, 734 struct perf_sample_data *data,
735 struct pt_regs *regs); 735 struct pt_regs *regs);
736 736
737extern void perf_event_output(struct perf_event *event,
738 struct perf_sample_data *data,
739 struct pt_regs *regs);
740
741extern void
742perf_event_header__init_id(struct perf_event_header *header,
743 struct perf_sample_data *data,
744 struct perf_event *event);
745extern void
746perf_event__output_id_sample(struct perf_event *event,
747 struct perf_output_handle *handle,
748 struct perf_sample_data *sample);
749
750extern void
751perf_log_lost_samples(struct perf_event *event, u64 lost);
752
737static inline bool is_sampling_event(struct perf_event *event) 753static inline bool is_sampling_event(struct perf_event *event)
738{ 754{
739 return event->attr.sample_period != 0; 755 return event->attr.sample_period != 0;
@@ -798,11 +814,33 @@ perf_sw_event_sched(u32 event_id, u64 nr, u64 addr)
798 814
799extern struct static_key_deferred perf_sched_events; 815extern struct static_key_deferred perf_sched_events;
800 816
817static __always_inline bool
818perf_sw_migrate_enabled(void)
819{
820 if (static_key_false(&perf_swevent_enabled[PERF_COUNT_SW_CPU_MIGRATIONS]))
821 return true;
822 return false;
823}
824
825static inline void perf_event_task_migrate(struct task_struct *task)
826{
827 if (perf_sw_migrate_enabled())
828 task->sched_migrated = 1;
829}
830
801static inline void perf_event_task_sched_in(struct task_struct *prev, 831static inline void perf_event_task_sched_in(struct task_struct *prev,
802 struct task_struct *task) 832 struct task_struct *task)
803{ 833{
804 if (static_key_false(&perf_sched_events.key)) 834 if (static_key_false(&perf_sched_events.key))
805 __perf_event_task_sched_in(prev, task); 835 __perf_event_task_sched_in(prev, task);
836
837 if (perf_sw_migrate_enabled() && task->sched_migrated) {
838 struct pt_regs *regs = this_cpu_ptr(&__perf_regs[0]);
839
840 perf_fetch_caller_regs(regs);
841 ___perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, regs, 0);
842 task->sched_migrated = 0;
843 }
806} 844}
807 845
808static inline void perf_event_task_sched_out(struct task_struct *prev, 846static inline void perf_event_task_sched_out(struct task_struct *prev,
@@ -925,6 +963,8 @@ perf_aux_output_skip(struct perf_output_handle *handle,
925static inline void * 963static inline void *
926perf_get_aux(struct perf_output_handle *handle) { return NULL; } 964perf_get_aux(struct perf_output_handle *handle) { return NULL; }
927static inline void 965static inline void
966perf_event_task_migrate(struct task_struct *task) { }
967static inline void
928perf_event_task_sched_in(struct task_struct *prev, 968perf_event_task_sched_in(struct task_struct *prev,
929 struct task_struct *task) { } 969 struct task_struct *task) { }
930static inline void 970static inline void
diff --git a/include/linux/preempt.h b/include/linux/preempt.h
index de83b4eb1642..0f1534acaf60 100644
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -10,13 +10,117 @@
10#include <linux/list.h> 10#include <linux/list.h>
11 11
12/* 12/*
13 * We use the MSB mostly because its available; see <linux/preempt_mask.h> for 13 * We put the hardirq and softirq counter into the preemption
14 * the other bits -- can't include that header due to inclusion hell. 14 * counter. The bitmask has the following meaning:
15 *
16 * - bits 0-7 are the preemption count (max preemption depth: 256)
17 * - bits 8-15 are the softirq count (max # of softirqs: 256)
18 *
19 * The hardirq count could in theory be the same as the number of
20 * interrupts in the system, but we run all interrupt handlers with
21 * interrupts disabled, so we cannot have nesting interrupts. Though
22 * there are a few palaeontologic drivers which reenable interrupts in
23 * the handler, so we need more than one bit here.
24 *
25 * PREEMPT_MASK: 0x000000ff
26 * SOFTIRQ_MASK: 0x0000ff00
27 * HARDIRQ_MASK: 0x000f0000
28 * NMI_MASK: 0x00100000
29 * PREEMPT_ACTIVE: 0x00200000
30 * PREEMPT_NEED_RESCHED: 0x80000000
15 */ 31 */
32#define PREEMPT_BITS 8
33#define SOFTIRQ_BITS 8
34#define HARDIRQ_BITS 4
35#define NMI_BITS 1
36
37#define PREEMPT_SHIFT 0
38#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS)
39#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS)
40#define NMI_SHIFT (HARDIRQ_SHIFT + HARDIRQ_BITS)
41
42#define __IRQ_MASK(x) ((1UL << (x))-1)
43
44#define PREEMPT_MASK (__IRQ_MASK(PREEMPT_BITS) << PREEMPT_SHIFT)
45#define SOFTIRQ_MASK (__IRQ_MASK(SOFTIRQ_BITS) << SOFTIRQ_SHIFT)
46#define HARDIRQ_MASK (__IRQ_MASK(HARDIRQ_BITS) << HARDIRQ_SHIFT)
47#define NMI_MASK (__IRQ_MASK(NMI_BITS) << NMI_SHIFT)
48
49#define PREEMPT_OFFSET (1UL << PREEMPT_SHIFT)
50#define SOFTIRQ_OFFSET (1UL << SOFTIRQ_SHIFT)
51#define HARDIRQ_OFFSET (1UL << HARDIRQ_SHIFT)
52#define NMI_OFFSET (1UL << NMI_SHIFT)
53
54#define SOFTIRQ_DISABLE_OFFSET (2 * SOFTIRQ_OFFSET)
55
56#define PREEMPT_ACTIVE_BITS 1
57#define PREEMPT_ACTIVE_SHIFT (NMI_SHIFT + NMI_BITS)
58#define PREEMPT_ACTIVE (__IRQ_MASK(PREEMPT_ACTIVE_BITS) << PREEMPT_ACTIVE_SHIFT)
59
60/* We use the MSB mostly because its available */
16#define PREEMPT_NEED_RESCHED 0x80000000 61#define PREEMPT_NEED_RESCHED 0x80000000
17 62
63/* preempt_count() and related functions, depends on PREEMPT_NEED_RESCHED */
18#include <asm/preempt.h> 64#include <asm/preempt.h>
19 65
66#define hardirq_count() (preempt_count() & HARDIRQ_MASK)
67#define softirq_count() (preempt_count() & SOFTIRQ_MASK)
68#define irq_count() (preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK \
69 | NMI_MASK))
70
71/*
72 * Are we doing bottom half or hardware interrupt processing?
73 * Are we in a softirq context? Interrupt context?
74 * in_softirq - Are we currently processing softirq or have bh disabled?
75 * in_serving_softirq - Are we currently processing softirq?
76 */
77#define in_irq() (hardirq_count())
78#define in_softirq() (softirq_count())
79#define in_interrupt() (irq_count())
80#define in_serving_softirq() (softirq_count() & SOFTIRQ_OFFSET)
81
82/*
83 * Are we in NMI context?
84 */
85#define in_nmi() (preempt_count() & NMI_MASK)
86
87#if defined(CONFIG_PREEMPT_COUNT)
88# define PREEMPT_DISABLE_OFFSET 1
89#else
90# define PREEMPT_DISABLE_OFFSET 0
91#endif
92
93/*
94 * The preempt_count offset needed for things like:
95 *
96 * spin_lock_bh()
97 *
98 * Which need to disable both preemption (CONFIG_PREEMPT_COUNT) and
99 * softirqs, such that unlock sequences of:
100 *
101 * spin_unlock();
102 * local_bh_enable();
103 *
104 * Work as expected.
105 */
106#define SOFTIRQ_LOCK_OFFSET (SOFTIRQ_DISABLE_OFFSET + PREEMPT_DISABLE_OFFSET)
107
108/*
109 * Are we running in atomic context? WARNING: this macro cannot
110 * always detect atomic context; in particular, it cannot know about
111 * held spinlocks in non-preemptible kernels. Thus it should not be
112 * used in the general case to determine whether sleeping is possible.
113 * Do not use in_atomic() in driver code.
114 */
115#define in_atomic() (preempt_count() != 0)
116
117/*
118 * Check whether we were atomic before we did preempt_disable():
119 * (used by the scheduler)
120 */
121#define in_atomic_preempt_off() \
122 ((preempt_count() & ~PREEMPT_ACTIVE) != PREEMPT_DISABLE_OFFSET)
123
20#if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_PREEMPT_TRACER) 124#if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_PREEMPT_TRACER)
21extern void preempt_count_add(int val); 125extern void preempt_count_add(int val);
22extern void preempt_count_sub(int val); 126extern void preempt_count_sub(int val);
@@ -33,6 +137,18 @@ extern void preempt_count_sub(int val);
33#define preempt_count_inc() preempt_count_add(1) 137#define preempt_count_inc() preempt_count_add(1)
34#define preempt_count_dec() preempt_count_sub(1) 138#define preempt_count_dec() preempt_count_sub(1)
35 139
140#define preempt_active_enter() \
141do { \
142 preempt_count_add(PREEMPT_ACTIVE + PREEMPT_DISABLE_OFFSET); \
143 barrier(); \
144} while (0)
145
146#define preempt_active_exit() \
147do { \
148 barrier(); \
149 preempt_count_sub(PREEMPT_ACTIVE + PREEMPT_DISABLE_OFFSET); \
150} while (0)
151
36#ifdef CONFIG_PREEMPT_COUNT 152#ifdef CONFIG_PREEMPT_COUNT
37 153
38#define preempt_disable() \ 154#define preempt_disable() \
@@ -49,6 +165,8 @@ do { \
49 165
50#define preempt_enable_no_resched() sched_preempt_enable_no_resched() 166#define preempt_enable_no_resched() sched_preempt_enable_no_resched()
51 167
168#define preemptible() (preempt_count() == 0 && !irqs_disabled())
169
52#ifdef CONFIG_PREEMPT 170#ifdef CONFIG_PREEMPT
53#define preempt_enable() \ 171#define preempt_enable() \
54do { \ 172do { \
@@ -57,52 +175,46 @@ do { \
57 __preempt_schedule(); \ 175 __preempt_schedule(); \
58} while (0) 176} while (0)
59 177
178#define preempt_enable_notrace() \
179do { \
180 barrier(); \
181 if (unlikely(__preempt_count_dec_and_test())) \
182 __preempt_schedule_notrace(); \
183} while (0)
184
60#define preempt_check_resched() \ 185#define preempt_check_resched() \
61do { \ 186do { \
62 if (should_resched()) \ 187 if (should_resched()) \
63 __preempt_schedule(); \ 188 __preempt_schedule(); \
64} while (0) 189} while (0)
65 190
66#else 191#else /* !CONFIG_PREEMPT */
67#define preempt_enable() \ 192#define preempt_enable() \
68do { \ 193do { \
69 barrier(); \ 194 barrier(); \
70 preempt_count_dec(); \ 195 preempt_count_dec(); \
71} while (0) 196} while (0)
72#define preempt_check_resched() do { } while (0)
73#endif
74
75#define preempt_disable_notrace() \
76do { \
77 __preempt_count_inc(); \
78 barrier(); \
79} while (0)
80 197
81#define preempt_enable_no_resched_notrace() \ 198#define preempt_enable_notrace() \
82do { \ 199do { \
83 barrier(); \ 200 barrier(); \
84 __preempt_count_dec(); \ 201 __preempt_count_dec(); \
85} while (0) 202} while (0)
86 203
87#ifdef CONFIG_PREEMPT 204#define preempt_check_resched() do { } while (0)
88 205#endif /* CONFIG_PREEMPT */
89#ifndef CONFIG_CONTEXT_TRACKING
90#define __preempt_schedule_context() __preempt_schedule()
91#endif
92 206
93#define preempt_enable_notrace() \ 207#define preempt_disable_notrace() \
94do { \ 208do { \
209 __preempt_count_inc(); \
95 barrier(); \ 210 barrier(); \
96 if (unlikely(__preempt_count_dec_and_test())) \
97 __preempt_schedule_context(); \
98} while (0) 211} while (0)
99#else 212
100#define preempt_enable_notrace() \ 213#define preempt_enable_no_resched_notrace() \
101do { \ 214do { \
102 barrier(); \ 215 barrier(); \
103 __preempt_count_dec(); \ 216 __preempt_count_dec(); \
104} while (0) 217} while (0)
105#endif
106 218
107#else /* !CONFIG_PREEMPT_COUNT */ 219#else /* !CONFIG_PREEMPT_COUNT */
108 220
@@ -121,6 +233,7 @@ do { \
121#define preempt_disable_notrace() barrier() 233#define preempt_disable_notrace() barrier()
122#define preempt_enable_no_resched_notrace() barrier() 234#define preempt_enable_no_resched_notrace() barrier()
123#define preempt_enable_notrace() barrier() 235#define preempt_enable_notrace() barrier()
236#define preemptible() 0
124 237
125#endif /* CONFIG_PREEMPT_COUNT */ 238#endif /* CONFIG_PREEMPT_COUNT */
126 239
diff --git a/include/linux/preempt_mask.h b/include/linux/preempt_mask.h
deleted file mode 100644
index dbeec4d4a3be..000000000000
--- a/include/linux/preempt_mask.h
+++ /dev/null
@@ -1,117 +0,0 @@
1#ifndef LINUX_PREEMPT_MASK_H
2#define LINUX_PREEMPT_MASK_H
3
4#include <linux/preempt.h>
5
6/*
7 * We put the hardirq and softirq counter into the preemption
8 * counter. The bitmask has the following meaning:
9 *
10 * - bits 0-7 are the preemption count (max preemption depth: 256)
11 * - bits 8-15 are the softirq count (max # of softirqs: 256)
12 *
13 * The hardirq count could in theory be the same as the number of
14 * interrupts in the system, but we run all interrupt handlers with
15 * interrupts disabled, so we cannot have nesting interrupts. Though
16 * there are a few palaeontologic drivers which reenable interrupts in
17 * the handler, so we need more than one bit here.
18 *
19 * PREEMPT_MASK: 0x000000ff
20 * SOFTIRQ_MASK: 0x0000ff00
21 * HARDIRQ_MASK: 0x000f0000
22 * NMI_MASK: 0x00100000
23 * PREEMPT_ACTIVE: 0x00200000
24 */
25#define PREEMPT_BITS 8
26#define SOFTIRQ_BITS 8
27#define HARDIRQ_BITS 4
28#define NMI_BITS 1
29
30#define PREEMPT_SHIFT 0
31#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS)
32#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS)
33#define NMI_SHIFT (HARDIRQ_SHIFT + HARDIRQ_BITS)
34
35#define __IRQ_MASK(x) ((1UL << (x))-1)
36
37#define PREEMPT_MASK (__IRQ_MASK(PREEMPT_BITS) << PREEMPT_SHIFT)
38#define SOFTIRQ_MASK (__IRQ_MASK(SOFTIRQ_BITS) << SOFTIRQ_SHIFT)
39#define HARDIRQ_MASK (__IRQ_MASK(HARDIRQ_BITS) << HARDIRQ_SHIFT)
40#define NMI_MASK (__IRQ_MASK(NMI_BITS) << NMI_SHIFT)
41
42#define PREEMPT_OFFSET (1UL << PREEMPT_SHIFT)
43#define SOFTIRQ_OFFSET (1UL << SOFTIRQ_SHIFT)
44#define HARDIRQ_OFFSET (1UL << HARDIRQ_SHIFT)
45#define NMI_OFFSET (1UL << NMI_SHIFT)
46
47#define SOFTIRQ_DISABLE_OFFSET (2 * SOFTIRQ_OFFSET)
48
49#define PREEMPT_ACTIVE_BITS 1
50#define PREEMPT_ACTIVE_SHIFT (NMI_SHIFT + NMI_BITS)
51#define PREEMPT_ACTIVE (__IRQ_MASK(PREEMPT_ACTIVE_BITS) << PREEMPT_ACTIVE_SHIFT)
52
53#define hardirq_count() (preempt_count() & HARDIRQ_MASK)
54#define softirq_count() (preempt_count() & SOFTIRQ_MASK)
55#define irq_count() (preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK \
56 | NMI_MASK))
57
58/*
59 * Are we doing bottom half or hardware interrupt processing?
60 * Are we in a softirq context? Interrupt context?
61 * in_softirq - Are we currently processing softirq or have bh disabled?
62 * in_serving_softirq - Are we currently processing softirq?
63 */
64#define in_irq() (hardirq_count())
65#define in_softirq() (softirq_count())
66#define in_interrupt() (irq_count())
67#define in_serving_softirq() (softirq_count() & SOFTIRQ_OFFSET)
68
69/*
70 * Are we in NMI context?
71 */
72#define in_nmi() (preempt_count() & NMI_MASK)
73
74#if defined(CONFIG_PREEMPT_COUNT)
75# define PREEMPT_CHECK_OFFSET 1
76#else
77# define PREEMPT_CHECK_OFFSET 0
78#endif
79
80/*
81 * The preempt_count offset needed for things like:
82 *
83 * spin_lock_bh()
84 *
85 * Which need to disable both preemption (CONFIG_PREEMPT_COUNT) and
86 * softirqs, such that unlock sequences of:
87 *
88 * spin_unlock();
89 * local_bh_enable();
90 *
91 * Work as expected.
92 */
93#define SOFTIRQ_LOCK_OFFSET (SOFTIRQ_DISABLE_OFFSET + PREEMPT_CHECK_OFFSET)
94
95/*
96 * Are we running in atomic context? WARNING: this macro cannot
97 * always detect atomic context; in particular, it cannot know about
98 * held spinlocks in non-preemptible kernels. Thus it should not be
99 * used in the general case to determine whether sleeping is possible.
100 * Do not use in_atomic() in driver code.
101 */
102#define in_atomic() ((preempt_count() & ~PREEMPT_ACTIVE) != 0)
103
104/*
105 * Check whether we were atomic before we did preempt_disable():
106 * (used by the scheduler, *after* releasing the kernel lock)
107 */
108#define in_atomic_preempt_off() \
109 ((preempt_count() & ~PREEMPT_ACTIVE) != PREEMPT_CHECK_OFFSET)
110
111#ifdef CONFIG_PREEMPT_COUNT
112# define preemptible() (preempt_count() == 0 && !irqs_disabled())
113#else
114# define preemptible() 0
115#endif
116
117#endif /* LINUX_PREEMPT_MASK_H */
diff --git a/include/linux/rculist.h b/include/linux/rculist.h
index a18b16f1dc0e..17c6b1f84a77 100644
--- a/include/linux/rculist.h
+++ b/include/linux/rculist.h
@@ -29,8 +29,8 @@
29 */ 29 */
30static inline void INIT_LIST_HEAD_RCU(struct list_head *list) 30static inline void INIT_LIST_HEAD_RCU(struct list_head *list)
31{ 31{
32 ACCESS_ONCE(list->next) = list; 32 WRITE_ONCE(list->next, list);
33 ACCESS_ONCE(list->prev) = list; 33 WRITE_ONCE(list->prev, list);
34} 34}
35 35
36/* 36/*
@@ -288,7 +288,7 @@ static inline void list_splice_init_rcu(struct list_head *list,
288#define list_first_or_null_rcu(ptr, type, member) \ 288#define list_first_or_null_rcu(ptr, type, member) \
289({ \ 289({ \
290 struct list_head *__ptr = (ptr); \ 290 struct list_head *__ptr = (ptr); \
291 struct list_head *__next = ACCESS_ONCE(__ptr->next); \ 291 struct list_head *__next = READ_ONCE(__ptr->next); \
292 likely(__ptr != __next) ? list_entry_rcu(__next, type, member) : NULL; \ 292 likely(__ptr != __next) ? list_entry_rcu(__next, type, member) : NULL; \
293}) 293})
294 294
@@ -549,8 +549,8 @@ static inline void hlist_add_behind_rcu(struct hlist_node *n,
549 */ 549 */
550#define hlist_for_each_entry_from_rcu(pos, member) \ 550#define hlist_for_each_entry_from_rcu(pos, member) \
551 for (; pos; \ 551 for (; pos; \
552 pos = hlist_entry_safe(rcu_dereference((pos)->member.next),\ 552 pos = hlist_entry_safe(rcu_dereference_raw(hlist_next_rcu( \
553 typeof(*(pos)), member)) 553 &(pos)->member)), typeof(*(pos)), member))
554 554
555#endif /* __KERNEL__ */ 555#endif /* __KERNEL__ */
556#endif 556#endif
diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index 573a5afd5ed8..33a056bb886f 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -44,6 +44,8 @@
44#include <linux/debugobjects.h> 44#include <linux/debugobjects.h>
45#include <linux/bug.h> 45#include <linux/bug.h>
46#include <linux/compiler.h> 46#include <linux/compiler.h>
47#include <linux/ktime.h>
48
47#include <asm/barrier.h> 49#include <asm/barrier.h>
48 50
49extern int rcu_expedited; /* for sysctl */ 51extern int rcu_expedited; /* for sysctl */
@@ -292,10 +294,6 @@ void rcu_sched_qs(void);
292void rcu_bh_qs(void); 294void rcu_bh_qs(void);
293void rcu_check_callbacks(int user); 295void rcu_check_callbacks(int user);
294struct notifier_block; 296struct notifier_block;
295void rcu_idle_enter(void);
296void rcu_idle_exit(void);
297void rcu_irq_enter(void);
298void rcu_irq_exit(void);
299int rcu_cpu_notify(struct notifier_block *self, 297int rcu_cpu_notify(struct notifier_block *self,
300 unsigned long action, void *hcpu); 298 unsigned long action, void *hcpu);
301 299
@@ -364,8 +362,8 @@ extern struct srcu_struct tasks_rcu_exit_srcu;
364#define rcu_note_voluntary_context_switch(t) \ 362#define rcu_note_voluntary_context_switch(t) \
365 do { \ 363 do { \
366 rcu_all_qs(); \ 364 rcu_all_qs(); \
367 if (ACCESS_ONCE((t)->rcu_tasks_holdout)) \ 365 if (READ_ONCE((t)->rcu_tasks_holdout)) \
368 ACCESS_ONCE((t)->rcu_tasks_holdout) = false; \ 366 WRITE_ONCE((t)->rcu_tasks_holdout, false); \
369 } while (0) 367 } while (0)
370#else /* #ifdef CONFIG_TASKS_RCU */ 368#else /* #ifdef CONFIG_TASKS_RCU */
371#define TASKS_RCU(x) do { } while (0) 369#define TASKS_RCU(x) do { } while (0)
@@ -609,7 +607,7 @@ static inline void rcu_preempt_sleep_check(void)
609 607
610#define __rcu_access_pointer(p, space) \ 608#define __rcu_access_pointer(p, space) \
611({ \ 609({ \
612 typeof(*p) *_________p1 = (typeof(*p) *__force)ACCESS_ONCE(p); \ 610 typeof(*p) *_________p1 = (typeof(*p) *__force)READ_ONCE(p); \
613 rcu_dereference_sparse(p, space); \ 611 rcu_dereference_sparse(p, space); \
614 ((typeof(*p) __force __kernel *)(_________p1)); \ 612 ((typeof(*p) __force __kernel *)(_________p1)); \
615}) 613})
@@ -628,21 +626,6 @@ static inline void rcu_preempt_sleep_check(void)
628 ((typeof(*p) __force __kernel *)(p)); \ 626 ((typeof(*p) __force __kernel *)(p)); \
629}) 627})
630 628
631#define __rcu_access_index(p, space) \
632({ \
633 typeof(p) _________p1 = ACCESS_ONCE(p); \
634 rcu_dereference_sparse(p, space); \
635 (_________p1); \
636})
637#define __rcu_dereference_index_check(p, c) \
638({ \
639 /* Dependency order vs. p above. */ \
640 typeof(p) _________p1 = lockless_dereference(p); \
641 rcu_lockdep_assert(c, \
642 "suspicious rcu_dereference_index_check() usage"); \
643 (_________p1); \
644})
645
646/** 629/**
647 * RCU_INITIALIZER() - statically initialize an RCU-protected global variable 630 * RCU_INITIALIZER() - statically initialize an RCU-protected global variable
648 * @v: The value to statically initialize with. 631 * @v: The value to statically initialize with.
@@ -659,7 +642,7 @@ static inline void rcu_preempt_sleep_check(void)
659 */ 642 */
660#define lockless_dereference(p) \ 643#define lockless_dereference(p) \
661({ \ 644({ \
662 typeof(p) _________p1 = ACCESS_ONCE(p); \ 645 typeof(p) _________p1 = READ_ONCE(p); \
663 smp_read_barrier_depends(); /* Dependency order vs. p above. */ \ 646 smp_read_barrier_depends(); /* Dependency order vs. p above. */ \
664 (_________p1); \ 647 (_________p1); \
665}) 648})
@@ -702,7 +685,7 @@ static inline void rcu_preempt_sleep_check(void)
702 * @p: The pointer to read 685 * @p: The pointer to read
703 * 686 *
704 * Return the value of the specified RCU-protected pointer, but omit the 687 * Return the value of the specified RCU-protected pointer, but omit the
705 * smp_read_barrier_depends() and keep the ACCESS_ONCE(). This is useful 688 * smp_read_barrier_depends() and keep the READ_ONCE(). This is useful
706 * when the value of this pointer is accessed, but the pointer is not 689 * when the value of this pointer is accessed, but the pointer is not
707 * dereferenced, for example, when testing an RCU-protected pointer against 690 * dereferenced, for example, when testing an RCU-protected pointer against
708 * NULL. Although rcu_access_pointer() may also be used in cases where 691 * NULL. Although rcu_access_pointer() may also be used in cases where
@@ -787,47 +770,12 @@ static inline void rcu_preempt_sleep_check(void)
787#define rcu_dereference_raw_notrace(p) __rcu_dereference_check((p), 1, __rcu) 770#define rcu_dereference_raw_notrace(p) __rcu_dereference_check((p), 1, __rcu)
788 771
789/** 772/**
790 * rcu_access_index() - fetch RCU index with no dereferencing
791 * @p: The index to read
792 *
793 * Return the value of the specified RCU-protected index, but omit the
794 * smp_read_barrier_depends() and keep the ACCESS_ONCE(). This is useful
795 * when the value of this index is accessed, but the index is not
796 * dereferenced, for example, when testing an RCU-protected index against
797 * -1. Although rcu_access_index() may also be used in cases where
798 * update-side locks prevent the value of the index from changing, you
799 * should instead use rcu_dereference_index_protected() for this use case.
800 */
801#define rcu_access_index(p) __rcu_access_index((p), __rcu)
802
803/**
804 * rcu_dereference_index_check() - rcu_dereference for indices with debug checking
805 * @p: The pointer to read, prior to dereferencing
806 * @c: The conditions under which the dereference will take place
807 *
808 * Similar to rcu_dereference_check(), but omits the sparse checking.
809 * This allows rcu_dereference_index_check() to be used on integers,
810 * which can then be used as array indices. Attempting to use
811 * rcu_dereference_check() on an integer will give compiler warnings
812 * because the sparse address-space mechanism relies on dereferencing
813 * the RCU-protected pointer. Dereferencing integers is not something
814 * that even gcc will put up with.
815 *
816 * Note that this function does not implicitly check for RCU read-side
817 * critical sections. If this function gains lots of uses, it might
818 * make sense to provide versions for each flavor of RCU, but it does
819 * not make sense as of early 2010.
820 */
821#define rcu_dereference_index_check(p, c) \
822 __rcu_dereference_index_check((p), (c))
823
824/**
825 * rcu_dereference_protected() - fetch RCU pointer when updates prevented 773 * rcu_dereference_protected() - fetch RCU pointer when updates prevented
826 * @p: The pointer to read, prior to dereferencing 774 * @p: The pointer to read, prior to dereferencing
827 * @c: The conditions under which the dereference will take place 775 * @c: The conditions under which the dereference will take place
828 * 776 *
829 * Return the value of the specified RCU-protected pointer, but omit 777 * Return the value of the specified RCU-protected pointer, but omit
830 * both the smp_read_barrier_depends() and the ACCESS_ONCE(). This 778 * both the smp_read_barrier_depends() and the READ_ONCE(). This
831 * is useful in cases where update-side locks prevent the value of the 779 * is useful in cases where update-side locks prevent the value of the
832 * pointer from changing. Please note that this primitive does -not- 780 * pointer from changing. Please note that this primitive does -not-
833 * prevent the compiler from repeating this reference or combining it 781 * prevent the compiler from repeating this reference or combining it
@@ -1153,13 +1101,13 @@ static inline notrace void rcu_read_unlock_sched_notrace(void)
1153#define kfree_rcu(ptr, rcu_head) \ 1101#define kfree_rcu(ptr, rcu_head) \
1154 __kfree_rcu(&((ptr)->rcu_head), offsetof(typeof(*(ptr)), rcu_head)) 1102 __kfree_rcu(&((ptr)->rcu_head), offsetof(typeof(*(ptr)), rcu_head))
1155 1103
1156#if defined(CONFIG_TINY_RCU) || defined(CONFIG_RCU_NOCB_CPU_ALL) 1104#ifdef CONFIG_TINY_RCU
1157static inline int rcu_needs_cpu(unsigned long *delta_jiffies) 1105static inline int rcu_needs_cpu(u64 basemono, u64 *nextevt)
1158{ 1106{
1159 *delta_jiffies = ULONG_MAX; 1107 *nextevt = KTIME_MAX;
1160 return 0; 1108 return 0;
1161} 1109}
1162#endif /* #if defined(CONFIG_TINY_RCU) || defined(CONFIG_RCU_NOCB_CPU_ALL) */ 1110#endif /* #ifdef CONFIG_TINY_RCU */
1163 1111
1164#if defined(CONFIG_RCU_NOCB_CPU_ALL) 1112#if defined(CONFIG_RCU_NOCB_CPU_ALL)
1165static inline bool rcu_is_nocb_cpu(int cpu) { return true; } 1113static inline bool rcu_is_nocb_cpu(int cpu) { return true; }
diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h
index 937edaeb150d..3df6c1ec4e25 100644
--- a/include/linux/rcutiny.h
+++ b/include/linux/rcutiny.h
@@ -159,6 +159,22 @@ static inline void rcu_cpu_stall_reset(void)
159{ 159{
160} 160}
161 161
162static inline void rcu_idle_enter(void)
163{
164}
165
166static inline void rcu_idle_exit(void)
167{
168}
169
170static inline void rcu_irq_enter(void)
171{
172}
173
174static inline void rcu_irq_exit(void)
175{
176}
177
162static inline void exit_rcu(void) 178static inline void exit_rcu(void)
163{ 179{
164} 180}
diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h
index d2e583a6aaca..456879143f89 100644
--- a/include/linux/rcutree.h
+++ b/include/linux/rcutree.h
@@ -31,9 +31,7 @@
31#define __LINUX_RCUTREE_H 31#define __LINUX_RCUTREE_H
32 32
33void rcu_note_context_switch(void); 33void rcu_note_context_switch(void);
34#ifndef CONFIG_RCU_NOCB_CPU_ALL 34int rcu_needs_cpu(u64 basem, u64 *nextevt);
35int rcu_needs_cpu(unsigned long *delta_jiffies);
36#endif /* #ifndef CONFIG_RCU_NOCB_CPU_ALL */
37void rcu_cpu_stall_reset(void); 35void rcu_cpu_stall_reset(void);
38 36
39/* 37/*
@@ -93,6 +91,11 @@ void rcu_force_quiescent_state(void);
93void rcu_bh_force_quiescent_state(void); 91void rcu_bh_force_quiescent_state(void);
94void rcu_sched_force_quiescent_state(void); 92void rcu_sched_force_quiescent_state(void);
95 93
94void rcu_idle_enter(void);
95void rcu_idle_exit(void);
96void rcu_irq_enter(void);
97void rcu_irq_exit(void);
98
96void exit_rcu(void); 99void exit_rcu(void);
97 100
98void rcu_scheduler_starting(void); 101void rcu_scheduler_starting(void);
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 26a2e6122734..6633e83e608a 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -25,7 +25,7 @@ struct sched_param {
25#include <linux/errno.h> 25#include <linux/errno.h>
26#include <linux/nodemask.h> 26#include <linux/nodemask.h>
27#include <linux/mm_types.h> 27#include <linux/mm_types.h>
28#include <linux/preempt_mask.h> 28#include <linux/preempt.h>
29 29
30#include <asm/page.h> 30#include <asm/page.h>
31#include <asm/ptrace.h> 31#include <asm/ptrace.h>
@@ -132,6 +132,7 @@ struct fs_struct;
132struct perf_event_context; 132struct perf_event_context;
133struct blk_plug; 133struct blk_plug;
134struct filename; 134struct filename;
135struct nameidata;
135 136
136#define VMACACHE_BITS 2 137#define VMACACHE_BITS 2
137#define VMACACHE_SIZE (1U << VMACACHE_BITS) 138#define VMACACHE_SIZE (1U << VMACACHE_BITS)
@@ -173,7 +174,12 @@ extern unsigned long nr_iowait_cpu(int cpu);
173extern void get_iowait_load(unsigned long *nr_waiters, unsigned long *load); 174extern void get_iowait_load(unsigned long *nr_waiters, unsigned long *load);
174 175
175extern void calc_global_load(unsigned long ticks); 176extern void calc_global_load(unsigned long ticks);
177
178#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
176extern void update_cpu_load_nohz(void); 179extern void update_cpu_load_nohz(void);
180#else
181static inline void update_cpu_load_nohz(void) { }
182#endif
177 183
178extern unsigned long get_parent_ip(unsigned long addr); 184extern unsigned long get_parent_ip(unsigned long addr);
179 185
@@ -213,9 +219,10 @@ print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
213#define TASK_WAKEKILL 128 219#define TASK_WAKEKILL 128
214#define TASK_WAKING 256 220#define TASK_WAKING 256
215#define TASK_PARKED 512 221#define TASK_PARKED 512
216#define TASK_STATE_MAX 1024 222#define TASK_NOLOAD 1024
223#define TASK_STATE_MAX 2048
217 224
218#define TASK_STATE_TO_CHAR_STR "RSDTtXZxKWP" 225#define TASK_STATE_TO_CHAR_STR "RSDTtXZxKWPN"
219 226
220extern char ___assert_task_state[1 - 2*!!( 227extern char ___assert_task_state[1 - 2*!!(
221 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)]; 228 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
@@ -225,6 +232,8 @@ extern char ___assert_task_state[1 - 2*!!(
225#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED) 232#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
226#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED) 233#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
227 234
235#define TASK_IDLE (TASK_UNINTERRUPTIBLE | TASK_NOLOAD)
236
228/* Convenience macros for the sake of wake_up */ 237/* Convenience macros for the sake of wake_up */
229#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE) 238#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
230#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED) 239#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
@@ -240,7 +249,8 @@ extern char ___assert_task_state[1 - 2*!!(
240 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0) 249 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
241#define task_contributes_to_load(task) \ 250#define task_contributes_to_load(task) \
242 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \ 251 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
243 (task->flags & PF_FROZEN) == 0) 252 (task->flags & PF_FROZEN) == 0 && \
253 (task->state & TASK_NOLOAD) == 0)
244 254
245#ifdef CONFIG_DEBUG_ATOMIC_SLEEP 255#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
246 256
@@ -252,7 +262,7 @@ extern char ___assert_task_state[1 - 2*!!(
252#define set_task_state(tsk, state_value) \ 262#define set_task_state(tsk, state_value) \
253 do { \ 263 do { \
254 (tsk)->task_state_change = _THIS_IP_; \ 264 (tsk)->task_state_change = _THIS_IP_; \
255 set_mb((tsk)->state, (state_value)); \ 265 smp_store_mb((tsk)->state, (state_value)); \
256 } while (0) 266 } while (0)
257 267
258/* 268/*
@@ -274,7 +284,7 @@ extern char ___assert_task_state[1 - 2*!!(
274#define set_current_state(state_value) \ 284#define set_current_state(state_value) \
275 do { \ 285 do { \
276 current->task_state_change = _THIS_IP_; \ 286 current->task_state_change = _THIS_IP_; \
277 set_mb(current->state, (state_value)); \ 287 smp_store_mb(current->state, (state_value)); \
278 } while (0) 288 } while (0)
279 289
280#else 290#else
@@ -282,7 +292,7 @@ extern char ___assert_task_state[1 - 2*!!(
282#define __set_task_state(tsk, state_value) \ 292#define __set_task_state(tsk, state_value) \
283 do { (tsk)->state = (state_value); } while (0) 293 do { (tsk)->state = (state_value); } while (0)
284#define set_task_state(tsk, state_value) \ 294#define set_task_state(tsk, state_value) \
285 set_mb((tsk)->state, (state_value)) 295 smp_store_mb((tsk)->state, (state_value))
286 296
287/* 297/*
288 * set_current_state() includes a barrier so that the write of current->state 298 * set_current_state() includes a barrier so that the write of current->state
@@ -298,7 +308,7 @@ extern char ___assert_task_state[1 - 2*!!(
298#define __set_current_state(state_value) \ 308#define __set_current_state(state_value) \
299 do { current->state = (state_value); } while (0) 309 do { current->state = (state_value); } while (0)
300#define set_current_state(state_value) \ 310#define set_current_state(state_value) \
301 set_mb(current->state, (state_value)) 311 smp_store_mb(current->state, (state_value))
302 312
303#endif 313#endif
304 314
@@ -335,14 +345,10 @@ extern int runqueue_is_locked(int cpu);
335#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON) 345#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
336extern void nohz_balance_enter_idle(int cpu); 346extern void nohz_balance_enter_idle(int cpu);
337extern void set_cpu_sd_state_idle(void); 347extern void set_cpu_sd_state_idle(void);
338extern int get_nohz_timer_target(int pinned); 348extern int get_nohz_timer_target(void);
339#else 349#else
340static inline void nohz_balance_enter_idle(int cpu) { } 350static inline void nohz_balance_enter_idle(int cpu) { }
341static inline void set_cpu_sd_state_idle(void) { } 351static inline void set_cpu_sd_state_idle(void) { }
342static inline int get_nohz_timer_target(int pinned)
343{
344 return smp_processor_id();
345}
346#endif 352#endif
347 353
348/* 354/*
@@ -567,6 +573,23 @@ struct task_cputime {
567 .sum_exec_runtime = 0, \ 573 .sum_exec_runtime = 0, \
568 } 574 }
569 575
576/*
577 * This is the atomic variant of task_cputime, which can be used for
578 * storing and updating task_cputime statistics without locking.
579 */
580struct task_cputime_atomic {
581 atomic64_t utime;
582 atomic64_t stime;
583 atomic64_t sum_exec_runtime;
584};
585
586#define INIT_CPUTIME_ATOMIC \
587 (struct task_cputime_atomic) { \
588 .utime = ATOMIC64_INIT(0), \
589 .stime = ATOMIC64_INIT(0), \
590 .sum_exec_runtime = ATOMIC64_INIT(0), \
591 }
592
570#ifdef CONFIG_PREEMPT_COUNT 593#ifdef CONFIG_PREEMPT_COUNT
571#define PREEMPT_DISABLED (1 + PREEMPT_ENABLED) 594#define PREEMPT_DISABLED (1 + PREEMPT_ENABLED)
572#else 595#else
@@ -584,18 +607,16 @@ struct task_cputime {
584 607
585/** 608/**
586 * struct thread_group_cputimer - thread group interval timer counts 609 * struct thread_group_cputimer - thread group interval timer counts
587 * @cputime: thread group interval timers. 610 * @cputime_atomic: atomic thread group interval timers.
588 * @running: non-zero when there are timers running and 611 * @running: non-zero when there are timers running and
589 * @cputime receives updates. 612 * @cputime receives updates.
590 * @lock: lock for fields in this struct.
591 * 613 *
592 * This structure contains the version of task_cputime, above, that is 614 * This structure contains the version of task_cputime, above, that is
593 * used for thread group CPU timer calculations. 615 * used for thread group CPU timer calculations.
594 */ 616 */
595struct thread_group_cputimer { 617struct thread_group_cputimer {
596 struct task_cputime cputime; 618 struct task_cputime_atomic cputime_atomic;
597 int running; 619 int running;
598 raw_spinlock_t lock;
599}; 620};
600 621
601#include <linux/rwsem.h> 622#include <linux/rwsem.h>
@@ -900,6 +921,50 @@ enum cpu_idle_type {
900#define SCHED_CAPACITY_SCALE (1L << SCHED_CAPACITY_SHIFT) 921#define SCHED_CAPACITY_SCALE (1L << SCHED_CAPACITY_SHIFT)
901 922
902/* 923/*
924 * Wake-queues are lists of tasks with a pending wakeup, whose
925 * callers have already marked the task as woken internally,
926 * and can thus carry on. A common use case is being able to
927 * do the wakeups once the corresponding user lock as been
928 * released.
929 *
930 * We hold reference to each task in the list across the wakeup,
931 * thus guaranteeing that the memory is still valid by the time
932 * the actual wakeups are performed in wake_up_q().
933 *
934 * One per task suffices, because there's never a need for a task to be
935 * in two wake queues simultaneously; it is forbidden to abandon a task
936 * in a wake queue (a call to wake_up_q() _must_ follow), so if a task is
937 * already in a wake queue, the wakeup will happen soon and the second
938 * waker can just skip it.
939 *
940 * The WAKE_Q macro declares and initializes the list head.
941 * wake_up_q() does NOT reinitialize the list; it's expected to be
942 * called near the end of a function, where the fact that the queue is
943 * not used again will be easy to see by inspection.
944 *
945 * Note that this can cause spurious wakeups. schedule() callers
946 * must ensure the call is done inside a loop, confirming that the
947 * wakeup condition has in fact occurred.
948 */
949struct wake_q_node {
950 struct wake_q_node *next;
951};
952
953struct wake_q_head {
954 struct wake_q_node *first;
955 struct wake_q_node **lastp;
956};
957
958#define WAKE_Q_TAIL ((struct wake_q_node *) 0x01)
959
960#define WAKE_Q(name) \
961 struct wake_q_head name = { WAKE_Q_TAIL, &name.first }
962
963extern void wake_q_add(struct wake_q_head *head,
964 struct task_struct *task);
965extern void wake_up_q(struct wake_q_head *head);
966
967/*
903 * sched-domains (multiprocessor balancing) declarations: 968 * sched-domains (multiprocessor balancing) declarations:
904 */ 969 */
905#ifdef CONFIG_SMP 970#ifdef CONFIG_SMP
@@ -1334,8 +1399,6 @@ struct task_struct {
1334 int rcu_read_lock_nesting; 1399 int rcu_read_lock_nesting;
1335 union rcu_special rcu_read_unlock_special; 1400 union rcu_special rcu_read_unlock_special;
1336 struct list_head rcu_node_entry; 1401 struct list_head rcu_node_entry;
1337#endif /* #ifdef CONFIG_PREEMPT_RCU */
1338#ifdef CONFIG_PREEMPT_RCU
1339 struct rcu_node *rcu_blocked_node; 1402 struct rcu_node *rcu_blocked_node;
1340#endif /* #ifdef CONFIG_PREEMPT_RCU */ 1403#endif /* #ifdef CONFIG_PREEMPT_RCU */
1341#ifdef CONFIG_TASKS_RCU 1404#ifdef CONFIG_TASKS_RCU
@@ -1356,9 +1419,6 @@ struct task_struct {
1356#endif 1419#endif
1357 1420
1358 struct mm_struct *mm, *active_mm; 1421 struct mm_struct *mm, *active_mm;
1359#ifdef CONFIG_COMPAT_BRK
1360 unsigned brk_randomized:1;
1361#endif
1362 /* per-thread vma caching */ 1422 /* per-thread vma caching */
1363 u32 vmacache_seqnum; 1423 u32 vmacache_seqnum;
1364 struct vm_area_struct *vmacache[VMACACHE_SIZE]; 1424 struct vm_area_struct *vmacache[VMACACHE_SIZE];
@@ -1369,7 +1429,7 @@ struct task_struct {
1369 int exit_state; 1429 int exit_state;
1370 int exit_code, exit_signal; 1430 int exit_code, exit_signal;
1371 int pdeath_signal; /* The signal sent when the parent dies */ 1431 int pdeath_signal; /* The signal sent when the parent dies */
1372 unsigned int jobctl; /* JOBCTL_*, siglock protected */ 1432 unsigned long jobctl; /* JOBCTL_*, siglock protected */
1373 1433
1374 /* Used for emulating ABI behavior of previous Linux versions */ 1434 /* Used for emulating ABI behavior of previous Linux versions */
1375 unsigned int personality; 1435 unsigned int personality;
@@ -1381,10 +1441,14 @@ struct task_struct {
1381 /* Revert to default priority/policy when forking */ 1441 /* Revert to default priority/policy when forking */
1382 unsigned sched_reset_on_fork:1; 1442 unsigned sched_reset_on_fork:1;
1383 unsigned sched_contributes_to_load:1; 1443 unsigned sched_contributes_to_load:1;
1444 unsigned sched_migrated:1;
1384 1445
1385#ifdef CONFIG_MEMCG_KMEM 1446#ifdef CONFIG_MEMCG_KMEM
1386 unsigned memcg_kmem_skip_account:1; 1447 unsigned memcg_kmem_skip_account:1;
1387#endif 1448#endif
1449#ifdef CONFIG_COMPAT_BRK
1450 unsigned brk_randomized:1;
1451#endif
1388 1452
1389 unsigned long atomic_flags; /* Flags needing atomic access. */ 1453 unsigned long atomic_flags; /* Flags needing atomic access. */
1390 1454
@@ -1461,7 +1525,7 @@ struct task_struct {
1461 it with task_lock()) 1525 it with task_lock())
1462 - initialized normally by setup_new_exec */ 1526 - initialized normally by setup_new_exec */
1463/* file system info */ 1527/* file system info */
1464 int link_count, total_link_count; 1528 struct nameidata *nameidata;
1465#ifdef CONFIG_SYSVIPC 1529#ifdef CONFIG_SYSVIPC
1466/* ipc stuff */ 1530/* ipc stuff */
1467 struct sysv_sem sysvsem; 1531 struct sysv_sem sysvsem;
@@ -1511,6 +1575,8 @@ struct task_struct {
1511 /* Protection of the PI data structures: */ 1575 /* Protection of the PI data structures: */
1512 raw_spinlock_t pi_lock; 1576 raw_spinlock_t pi_lock;
1513 1577
1578 struct wake_q_node wake_q;
1579
1514#ifdef CONFIG_RT_MUTEXES 1580#ifdef CONFIG_RT_MUTEXES
1515 /* PI waiters blocked on a rt_mutex held by this task */ 1581 /* PI waiters blocked on a rt_mutex held by this task */
1516 struct rb_root pi_waiters; 1582 struct rb_root pi_waiters;
@@ -1724,6 +1790,7 @@ struct task_struct {
1724#ifdef CONFIG_DEBUG_ATOMIC_SLEEP 1790#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
1725 unsigned long task_state_change; 1791 unsigned long task_state_change;
1726#endif 1792#endif
1793 int pagefault_disabled;
1727}; 1794};
1728 1795
1729/* Future-safe accessor for struct task_struct's cpus_allowed. */ 1796/* Future-safe accessor for struct task_struct's cpus_allowed. */
@@ -2077,22 +2144,22 @@ TASK_PFA_CLEAR(SPREAD_SLAB, spread_slab)
2077#define JOBCTL_TRAPPING_BIT 21 /* switching to TRACED */ 2144#define JOBCTL_TRAPPING_BIT 21 /* switching to TRACED */
2078#define JOBCTL_LISTENING_BIT 22 /* ptracer is listening for events */ 2145#define JOBCTL_LISTENING_BIT 22 /* ptracer is listening for events */
2079 2146
2080#define JOBCTL_STOP_DEQUEUED (1 << JOBCTL_STOP_DEQUEUED_BIT) 2147#define JOBCTL_STOP_DEQUEUED (1UL << JOBCTL_STOP_DEQUEUED_BIT)
2081#define JOBCTL_STOP_PENDING (1 << JOBCTL_STOP_PENDING_BIT) 2148#define JOBCTL_STOP_PENDING (1UL << JOBCTL_STOP_PENDING_BIT)
2082#define JOBCTL_STOP_CONSUME (1 << JOBCTL_STOP_CONSUME_BIT) 2149#define JOBCTL_STOP_CONSUME (1UL << JOBCTL_STOP_CONSUME_BIT)
2083#define JOBCTL_TRAP_STOP (1 << JOBCTL_TRAP_STOP_BIT) 2150#define JOBCTL_TRAP_STOP (1UL << JOBCTL_TRAP_STOP_BIT)
2084#define JOBCTL_TRAP_NOTIFY (1 << JOBCTL_TRAP_NOTIFY_BIT) 2151#define JOBCTL_TRAP_NOTIFY (1UL << JOBCTL_TRAP_NOTIFY_BIT)
2085#define JOBCTL_TRAPPING (1 << JOBCTL_TRAPPING_BIT) 2152#define JOBCTL_TRAPPING (1UL << JOBCTL_TRAPPING_BIT)
2086#define JOBCTL_LISTENING (1 << JOBCTL_LISTENING_BIT) 2153#define JOBCTL_LISTENING (1UL << JOBCTL_LISTENING_BIT)
2087 2154
2088#define JOBCTL_TRAP_MASK (JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY) 2155#define JOBCTL_TRAP_MASK (JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY)
2089#define JOBCTL_PENDING_MASK (JOBCTL_STOP_PENDING | JOBCTL_TRAP_MASK) 2156#define JOBCTL_PENDING_MASK (JOBCTL_STOP_PENDING | JOBCTL_TRAP_MASK)
2090 2157
2091extern bool task_set_jobctl_pending(struct task_struct *task, 2158extern bool task_set_jobctl_pending(struct task_struct *task,
2092 unsigned int mask); 2159 unsigned long mask);
2093extern void task_clear_jobctl_trapping(struct task_struct *task); 2160extern void task_clear_jobctl_trapping(struct task_struct *task);
2094extern void task_clear_jobctl_pending(struct task_struct *task, 2161extern void task_clear_jobctl_pending(struct task_struct *task,
2095 unsigned int mask); 2162 unsigned long mask);
2096 2163
2097static inline void rcu_copy_process(struct task_struct *p) 2164static inline void rcu_copy_process(struct task_struct *p)
2098{ 2165{
@@ -2532,6 +2599,9 @@ static inline unsigned long wait_task_inactive(struct task_struct *p,
2532} 2599}
2533#endif 2600#endif
2534 2601
2602#define tasklist_empty() \
2603 list_empty(&init_task.tasks)
2604
2535#define next_task(p) \ 2605#define next_task(p) \
2536 list_entry_rcu((p)->tasks.next, struct task_struct, tasks) 2606 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2537 2607
@@ -2962,11 +3032,6 @@ static __always_inline bool need_resched(void)
2962void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times); 3032void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
2963void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times); 3033void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
2964 3034
2965static inline void thread_group_cputime_init(struct signal_struct *sig)
2966{
2967 raw_spin_lock_init(&sig->cputimer.lock);
2968}
2969
2970/* 3035/*
2971 * Reevaluate whether the task has signals pending delivery. 3036 * Reevaluate whether the task has signals pending delivery.
2972 * Wake the task if so. 3037 * Wake the task if so.
@@ -3080,13 +3145,13 @@ static inline void mm_update_next_owner(struct mm_struct *mm)
3080static inline unsigned long task_rlimit(const struct task_struct *tsk, 3145static inline unsigned long task_rlimit(const struct task_struct *tsk,
3081 unsigned int limit) 3146 unsigned int limit)
3082{ 3147{
3083 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur); 3148 return READ_ONCE(tsk->signal->rlim[limit].rlim_cur);
3084} 3149}
3085 3150
3086static inline unsigned long task_rlimit_max(const struct task_struct *tsk, 3151static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
3087 unsigned int limit) 3152 unsigned int limit)
3088{ 3153{
3089 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max); 3154 return READ_ONCE(tsk->signal->rlim[limit].rlim_max);
3090} 3155}
3091 3156
3092static inline unsigned long rlimit(unsigned int limit) 3157static inline unsigned long rlimit(unsigned int limit)
diff --git a/include/linux/sched/sysctl.h b/include/linux/sched/sysctl.h
index 596a0e007c62..c9e4731cf10b 100644
--- a/include/linux/sched/sysctl.h
+++ b/include/linux/sched/sysctl.h
@@ -57,24 +57,12 @@ extern unsigned int sysctl_numa_balancing_scan_size;
57extern unsigned int sysctl_sched_migration_cost; 57extern unsigned int sysctl_sched_migration_cost;
58extern unsigned int sysctl_sched_nr_migrate; 58extern unsigned int sysctl_sched_nr_migrate;
59extern unsigned int sysctl_sched_time_avg; 59extern unsigned int sysctl_sched_time_avg;
60extern unsigned int sysctl_timer_migration;
61extern unsigned int sysctl_sched_shares_window; 60extern unsigned int sysctl_sched_shares_window;
62 61
63int sched_proc_update_handler(struct ctl_table *table, int write, 62int sched_proc_update_handler(struct ctl_table *table, int write,
64 void __user *buffer, size_t *length, 63 void __user *buffer, size_t *length,
65 loff_t *ppos); 64 loff_t *ppos);
66#endif 65#endif
67#ifdef CONFIG_SCHED_DEBUG
68static inline unsigned int get_sysctl_timer_migration(void)
69{
70 return sysctl_timer_migration;
71}
72#else
73static inline unsigned int get_sysctl_timer_migration(void)
74{
75 return 1;
76}
77#endif
78 66
79/* 67/*
80 * control realtime throttling: 68 * control realtime throttling:
diff --git a/include/linux/security.h b/include/linux/security.h
index 18264ea9e314..52febde52479 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -43,7 +43,6 @@ struct file;
43struct vfsmount; 43struct vfsmount;
44struct path; 44struct path;
45struct qstr; 45struct qstr;
46struct nameidata;
47struct iattr; 46struct iattr;
48struct fown_struct; 47struct fown_struct;
49struct file_operations; 48struct file_operations;
@@ -477,7 +476,8 @@ static inline void security_free_mnt_opts(struct security_mnt_opts *opts)
477 * @inode_follow_link: 476 * @inode_follow_link:
478 * Check permission to follow a symbolic link when looking up a pathname. 477 * Check permission to follow a symbolic link when looking up a pathname.
479 * @dentry contains the dentry structure for the link. 478 * @dentry contains the dentry structure for the link.
480 * @nd contains the nameidata structure for the parent directory. 479 * @inode contains the inode, which itself is not stable in RCU-walk
480 * @rcu indicates whether we are in RCU-walk mode.
481 * Return 0 if permission is granted. 481 * Return 0 if permission is granted.
482 * @inode_permission: 482 * @inode_permission:
483 * Check permission before accessing an inode. This hook is called by the 483 * Check permission before accessing an inode. This hook is called by the
@@ -1553,7 +1553,8 @@ struct security_operations {
1553 int (*inode_rename) (struct inode *old_dir, struct dentry *old_dentry, 1553 int (*inode_rename) (struct inode *old_dir, struct dentry *old_dentry,
1554 struct inode *new_dir, struct dentry *new_dentry); 1554 struct inode *new_dir, struct dentry *new_dentry);
1555 int (*inode_readlink) (struct dentry *dentry); 1555 int (*inode_readlink) (struct dentry *dentry);
1556 int (*inode_follow_link) (struct dentry *dentry, struct nameidata *nd); 1556 int (*inode_follow_link) (struct dentry *dentry, struct inode *inode,
1557 bool rcu);
1557 int (*inode_permission) (struct inode *inode, int mask); 1558 int (*inode_permission) (struct inode *inode, int mask);
1558 int (*inode_setattr) (struct dentry *dentry, struct iattr *attr); 1559 int (*inode_setattr) (struct dentry *dentry, struct iattr *attr);
1559 int (*inode_getattr) (const struct path *path); 1560 int (*inode_getattr) (const struct path *path);
@@ -1839,7 +1840,8 @@ int security_inode_rename(struct inode *old_dir, struct dentry *old_dentry,
1839 struct inode *new_dir, struct dentry *new_dentry, 1840 struct inode *new_dir, struct dentry *new_dentry,
1840 unsigned int flags); 1841 unsigned int flags);
1841int security_inode_readlink(struct dentry *dentry); 1842int security_inode_readlink(struct dentry *dentry);
1842int security_inode_follow_link(struct dentry *dentry, struct nameidata *nd); 1843int security_inode_follow_link(struct dentry *dentry, struct inode *inode,
1844 bool rcu);
1843int security_inode_permission(struct inode *inode, int mask); 1845int security_inode_permission(struct inode *inode, int mask);
1844int security_inode_setattr(struct dentry *dentry, struct iattr *attr); 1846int security_inode_setattr(struct dentry *dentry, struct iattr *attr);
1845int security_inode_getattr(const struct path *path); 1847int security_inode_getattr(const struct path *path);
@@ -2242,7 +2244,8 @@ static inline int security_inode_readlink(struct dentry *dentry)
2242} 2244}
2243 2245
2244static inline int security_inode_follow_link(struct dentry *dentry, 2246static inline int security_inode_follow_link(struct dentry *dentry,
2245 struct nameidata *nd) 2247 struct inode *inode,
2248 bool rcu)
2246{ 2249{
2247 return 0; 2250 return 0;
2248} 2251}
diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h
index 5f68d0a391ce..486e685a226a 100644
--- a/include/linux/seqlock.h
+++ b/include/linux/seqlock.h
@@ -233,6 +233,47 @@ static inline void raw_write_seqcount_end(seqcount_t *s)
233 s->sequence++; 233 s->sequence++;
234} 234}
235 235
236/**
237 * raw_write_seqcount_barrier - do a seq write barrier
238 * @s: pointer to seqcount_t
239 *
240 * This can be used to provide an ordering guarantee instead of the
241 * usual consistency guarantee. It is one wmb cheaper, because we can
242 * collapse the two back-to-back wmb()s.
243 *
244 * seqcount_t seq;
245 * bool X = true, Y = false;
246 *
247 * void read(void)
248 * {
249 * bool x, y;
250 *
251 * do {
252 * int s = read_seqcount_begin(&seq);
253 *
254 * x = X; y = Y;
255 *
256 * } while (read_seqcount_retry(&seq, s));
257 *
258 * BUG_ON(!x && !y);
259 * }
260 *
261 * void write(void)
262 * {
263 * Y = true;
264 *
265 * raw_write_seqcount_barrier(seq);
266 *
267 * X = false;
268 * }
269 */
270static inline void raw_write_seqcount_barrier(seqcount_t *s)
271{
272 s->sequence++;
273 smp_wmb();
274 s->sequence++;
275}
276
236/* 277/*
237 * raw_write_seqcount_latch - redirect readers to even/odd copy 278 * raw_write_seqcount_latch - redirect readers to even/odd copy
238 * @s: pointer to seqcount_t 279 * @s: pointer to seqcount_t
@@ -266,13 +307,13 @@ static inline void write_seqcount_end(seqcount_t *s)
266} 307}
267 308
268/** 309/**
269 * write_seqcount_barrier - invalidate in-progress read-side seq operations 310 * write_seqcount_invalidate - invalidate in-progress read-side seq operations
270 * @s: pointer to seqcount_t 311 * @s: pointer to seqcount_t
271 * 312 *
272 * After write_seqcount_barrier, no read-side seq operations will complete 313 * After write_seqcount_invalidate, no read-side seq operations will complete
273 * successfully and see data older than this. 314 * successfully and see data older than this.
274 */ 315 */
275static inline void write_seqcount_barrier(seqcount_t *s) 316static inline void write_seqcount_invalidate(seqcount_t *s)
276{ 317{
277 smp_wmb(); 318 smp_wmb();
278 s->sequence+=2; 319 s->sequence+=2;
diff --git a/include/linux/spinlock.h b/include/linux/spinlock.h
index 3e18379dfa6f..0063b24b4f36 100644
--- a/include/linux/spinlock.h
+++ b/include/linux/spinlock.h
@@ -120,7 +120,7 @@ do { \
120/* 120/*
121 * Despite its name it doesn't necessarily has to be a full barrier. 121 * Despite its name it doesn't necessarily has to be a full barrier.
122 * It should only guarantee that a STORE before the critical section 122 * It should only guarantee that a STORE before the critical section
123 * can not be reordered with a LOAD inside this section. 123 * can not be reordered with LOADs and STOREs inside this section.
124 * spin_lock() is the one-way barrier, this LOAD can not escape out 124 * spin_lock() is the one-way barrier, this LOAD can not escape out
125 * of the region. So the default implementation simply ensures that 125 * of the region. So the default implementation simply ensures that
126 * a STORE can not move into the critical section, smp_wmb() should 126 * a STORE can not move into the critical section, smp_wmb() should
diff --git a/include/linux/tick.h b/include/linux/tick.h
index f8492da57ad3..4191b5623a28 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -134,6 +134,12 @@ static inline bool tick_nohz_full_cpu(int cpu)
134 return cpumask_test_cpu(cpu, tick_nohz_full_mask); 134 return cpumask_test_cpu(cpu, tick_nohz_full_mask);
135} 135}
136 136
137static inline void tick_nohz_full_add_cpus_to(struct cpumask *mask)
138{
139 if (tick_nohz_full_enabled())
140 cpumask_or(mask, mask, tick_nohz_full_mask);
141}
142
137extern void __tick_nohz_full_check(void); 143extern void __tick_nohz_full_check(void);
138extern void tick_nohz_full_kick(void); 144extern void tick_nohz_full_kick(void);
139extern void tick_nohz_full_kick_cpu(int cpu); 145extern void tick_nohz_full_kick_cpu(int cpu);
@@ -142,6 +148,7 @@ extern void __tick_nohz_task_switch(struct task_struct *tsk);
142#else 148#else
143static inline bool tick_nohz_full_enabled(void) { return false; } 149static inline bool tick_nohz_full_enabled(void) { return false; }
144static inline bool tick_nohz_full_cpu(int cpu) { return false; } 150static inline bool tick_nohz_full_cpu(int cpu) { return false; }
151static inline void tick_nohz_full_add_cpus_to(struct cpumask *mask) { }
145static inline void __tick_nohz_full_check(void) { } 152static inline void __tick_nohz_full_check(void) { }
146static inline void tick_nohz_full_kick_cpu(int cpu) { } 153static inline void tick_nohz_full_kick_cpu(int cpu) { }
147static inline void tick_nohz_full_kick(void) { } 154static inline void tick_nohz_full_kick(void) { }
diff --git a/include/linux/time64.h b/include/linux/time64.h
index a3831478d9cf..77b5df2acd2a 100644
--- a/include/linux/time64.h
+++ b/include/linux/time64.h
@@ -2,6 +2,7 @@
2#define _LINUX_TIME64_H 2#define _LINUX_TIME64_H
3 3
4#include <uapi/linux/time.h> 4#include <uapi/linux/time.h>
5#include <linux/math64.h>
5 6
6typedef __s64 time64_t; 7typedef __s64 time64_t;
7 8
@@ -28,6 +29,7 @@ struct timespec64 {
28#define FSEC_PER_SEC 1000000000000000LL 29#define FSEC_PER_SEC 1000000000000000LL
29 30
30/* Located here for timespec[64]_valid_strict */ 31/* Located here for timespec[64]_valid_strict */
32#define TIME64_MAX ((s64)~((u64)1 << 63))
31#define KTIME_MAX ((s64)~((u64)1 << 63)) 33#define KTIME_MAX ((s64)~((u64)1 << 63))
32#define KTIME_SEC_MAX (KTIME_MAX / NSEC_PER_SEC) 34#define KTIME_SEC_MAX (KTIME_MAX / NSEC_PER_SEC)
33 35
diff --git a/include/linux/timekeeper_internal.h b/include/linux/timekeeper_internal.h
index fb86963859c7..25247220b4b7 100644
--- a/include/linux/timekeeper_internal.h
+++ b/include/linux/timekeeper_internal.h
@@ -49,6 +49,8 @@ struct tk_read_base {
49 * @offs_boot: Offset clock monotonic -> clock boottime 49 * @offs_boot: Offset clock monotonic -> clock boottime
50 * @offs_tai: Offset clock monotonic -> clock tai 50 * @offs_tai: Offset clock monotonic -> clock tai
51 * @tai_offset: The current UTC to TAI offset in seconds 51 * @tai_offset: The current UTC to TAI offset in seconds
52 * @clock_was_set_seq: The sequence number of clock was set events
53 * @next_leap_ktime: CLOCK_MONOTONIC time value of a pending leap-second
52 * @raw_time: Monotonic raw base time in timespec64 format 54 * @raw_time: Monotonic raw base time in timespec64 format
53 * @cycle_interval: Number of clock cycles in one NTP interval 55 * @cycle_interval: Number of clock cycles in one NTP interval
54 * @xtime_interval: Number of clock shifted nano seconds in one NTP 56 * @xtime_interval: Number of clock shifted nano seconds in one NTP
@@ -60,6 +62,9 @@ struct tk_read_base {
60 * shifted nano seconds. 62 * shifted nano seconds.
61 * @ntp_error_shift: Shift conversion between clock shifted nano seconds and 63 * @ntp_error_shift: Shift conversion between clock shifted nano seconds and
62 * ntp shifted nano seconds. 64 * ntp shifted nano seconds.
65 * @last_warning: Warning ratelimiter (DEBUG_TIMEKEEPING)
66 * @underflow_seen: Underflow warning flag (DEBUG_TIMEKEEPING)
67 * @overflow_seen: Overflow warning flag (DEBUG_TIMEKEEPING)
63 * 68 *
64 * Note: For timespec(64) based interfaces wall_to_monotonic is what 69 * Note: For timespec(64) based interfaces wall_to_monotonic is what
65 * we need to add to xtime (or xtime corrected for sub jiffie times) 70 * we need to add to xtime (or xtime corrected for sub jiffie times)
@@ -85,6 +90,8 @@ struct timekeeper {
85 ktime_t offs_boot; 90 ktime_t offs_boot;
86 ktime_t offs_tai; 91 ktime_t offs_tai;
87 s32 tai_offset; 92 s32 tai_offset;
93 unsigned int clock_was_set_seq;
94 ktime_t next_leap_ktime;
88 struct timespec64 raw_time; 95 struct timespec64 raw_time;
89 96
90 /* The following members are for timekeeping internal use */ 97 /* The following members are for timekeeping internal use */
@@ -104,6 +111,18 @@ struct timekeeper {
104 s64 ntp_error; 111 s64 ntp_error;
105 u32 ntp_error_shift; 112 u32 ntp_error_shift;
106 u32 ntp_err_mult; 113 u32 ntp_err_mult;
114#ifdef CONFIG_DEBUG_TIMEKEEPING
115 long last_warning;
116 /*
117 * These simple flag variables are managed
118 * without locks, which is racy, but they are
119 * ok since we don't really care about being
120 * super precise about how many events were
121 * seen, just that a problem was observed.
122 */
123 int underflow_seen;
124 int overflow_seen;
125#endif
107}; 126};
108 127
109#ifdef CONFIG_GENERIC_TIME_VSYSCALL 128#ifdef CONFIG_GENERIC_TIME_VSYSCALL
diff --git a/include/linux/timekeeping.h b/include/linux/timekeeping.h
index 99176af216af..3aa72e648650 100644
--- a/include/linux/timekeeping.h
+++ b/include/linux/timekeeping.h
@@ -163,6 +163,7 @@ extern ktime_t ktime_get(void);
163extern ktime_t ktime_get_with_offset(enum tk_offsets offs); 163extern ktime_t ktime_get_with_offset(enum tk_offsets offs);
164extern ktime_t ktime_mono_to_any(ktime_t tmono, enum tk_offsets offs); 164extern ktime_t ktime_mono_to_any(ktime_t tmono, enum tk_offsets offs);
165extern ktime_t ktime_get_raw(void); 165extern ktime_t ktime_get_raw(void);
166extern u32 ktime_get_resolution_ns(void);
166 167
167/** 168/**
168 * ktime_get_real - get the real (wall-) time in ktime_t format 169 * ktime_get_real - get the real (wall-) time in ktime_t format
@@ -266,7 +267,6 @@ extern int persistent_clock_is_local;
266 267
267extern void read_persistent_clock(struct timespec *ts); 268extern void read_persistent_clock(struct timespec *ts);
268extern void read_persistent_clock64(struct timespec64 *ts); 269extern void read_persistent_clock64(struct timespec64 *ts);
269extern void read_boot_clock(struct timespec *ts);
270extern void read_boot_clock64(struct timespec64 *ts); 270extern void read_boot_clock64(struct timespec64 *ts);
271extern int update_persistent_clock(struct timespec now); 271extern int update_persistent_clock(struct timespec now);
272extern int update_persistent_clock64(struct timespec64 now); 272extern int update_persistent_clock64(struct timespec64 now);
diff --git a/include/linux/timer.h b/include/linux/timer.h
index 8c5a197e1587..61aa61dc410c 100644
--- a/include/linux/timer.h
+++ b/include/linux/timer.h
@@ -14,27 +14,23 @@ struct timer_list {
14 * All fields that change during normal runtime grouped to the 14 * All fields that change during normal runtime grouped to the
15 * same cacheline 15 * same cacheline
16 */ 16 */
17 struct list_head entry; 17 struct hlist_node entry;
18 unsigned long expires; 18 unsigned long expires;
19 struct tvec_base *base; 19 void (*function)(unsigned long);
20 20 unsigned long data;
21 void (*function)(unsigned long); 21 u32 flags;
22 unsigned long data; 22 int slack;
23
24 int slack;
25 23
26#ifdef CONFIG_TIMER_STATS 24#ifdef CONFIG_TIMER_STATS
27 int start_pid; 25 int start_pid;
28 void *start_site; 26 void *start_site;
29 char start_comm[16]; 27 char start_comm[16];
30#endif 28#endif
31#ifdef CONFIG_LOCKDEP 29#ifdef CONFIG_LOCKDEP
32 struct lockdep_map lockdep_map; 30 struct lockdep_map lockdep_map;
33#endif 31#endif
34}; 32};
35 33
36extern struct tvec_base boot_tvec_bases;
37
38#ifdef CONFIG_LOCKDEP 34#ifdef CONFIG_LOCKDEP
39/* 35/*
40 * NB: because we have to copy the lockdep_map, setting the lockdep_map key 36 * NB: because we have to copy the lockdep_map, setting the lockdep_map key
@@ -49,9 +45,6 @@ extern struct tvec_base boot_tvec_bases;
49#endif 45#endif
50 46
51/* 47/*
52 * Note that all tvec_bases are at least 4 byte aligned and lower two bits
53 * of base in timer_list is guaranteed to be zero. Use them for flags.
54 *
55 * A deferrable timer will work normally when the system is busy, but 48 * A deferrable timer will work normally when the system is busy, but
56 * will not cause a CPU to come out of idle just to service it; instead, 49 * will not cause a CPU to come out of idle just to service it; instead,
57 * the timer will be serviced when the CPU eventually wakes up with a 50 * the timer will be serviced when the CPU eventually wakes up with a
@@ -65,17 +58,18 @@ extern struct tvec_base boot_tvec_bases;
65 * workqueue locking issues. It's not meant for executing random crap 58 * workqueue locking issues. It's not meant for executing random crap
66 * with interrupts disabled. Abuse is monitored! 59 * with interrupts disabled. Abuse is monitored!
67 */ 60 */
68#define TIMER_DEFERRABLE 0x1LU 61#define TIMER_CPUMASK 0x0007FFFF
69#define TIMER_IRQSAFE 0x2LU 62#define TIMER_MIGRATING 0x00080000
70 63#define TIMER_BASEMASK (TIMER_CPUMASK | TIMER_MIGRATING)
71#define TIMER_FLAG_MASK 0x3LU 64#define TIMER_DEFERRABLE 0x00100000
65#define TIMER_IRQSAFE 0x00200000
72 66
73#define __TIMER_INITIALIZER(_function, _expires, _data, _flags) { \ 67#define __TIMER_INITIALIZER(_function, _expires, _data, _flags) { \
74 .entry = { .prev = TIMER_ENTRY_STATIC }, \ 68 .entry = { .next = TIMER_ENTRY_STATIC }, \
75 .function = (_function), \ 69 .function = (_function), \
76 .expires = (_expires), \ 70 .expires = (_expires), \
77 .data = (_data), \ 71 .data = (_data), \
78 .base = (void *)((unsigned long)&boot_tvec_bases + (_flags)), \ 72 .flags = (_flags), \
79 .slack = -1, \ 73 .slack = -1, \
80 __TIMER_LOCKDEP_MAP_INITIALIZER( \ 74 __TIMER_LOCKDEP_MAP_INITIALIZER( \
81 __FILE__ ":" __stringify(__LINE__)) \ 75 __FILE__ ":" __stringify(__LINE__)) \
@@ -168,7 +162,7 @@ static inline void init_timer_on_stack_key(struct timer_list *timer,
168 */ 162 */
169static inline int timer_pending(const struct timer_list * timer) 163static inline int timer_pending(const struct timer_list * timer)
170{ 164{
171 return timer->entry.next != NULL; 165 return timer->entry.pprev != NULL;
172} 166}
173 167
174extern void add_timer_on(struct timer_list *timer, int cpu); 168extern void add_timer_on(struct timer_list *timer, int cpu);
@@ -188,26 +182,16 @@ extern void set_timer_slack(struct timer_list *time, int slack_hz);
188#define NEXT_TIMER_MAX_DELTA ((1UL << 30) - 1) 182#define NEXT_TIMER_MAX_DELTA ((1UL << 30) - 1)
189 183
190/* 184/*
191 * Return when the next timer-wheel timeout occurs (in absolute jiffies),
192 * locks the timer base and does the comparison against the given
193 * jiffie.
194 */
195extern unsigned long get_next_timer_interrupt(unsigned long now);
196
197/*
198 * Timer-statistics info: 185 * Timer-statistics info:
199 */ 186 */
200#ifdef CONFIG_TIMER_STATS 187#ifdef CONFIG_TIMER_STATS
201 188
202extern int timer_stats_active; 189extern int timer_stats_active;
203 190
204#define TIMER_STATS_FLAG_DEFERRABLE 0x1
205
206extern void init_timer_stats(void); 191extern void init_timer_stats(void);
207 192
208extern void timer_stats_update_stats(void *timer, pid_t pid, void *startf, 193extern void timer_stats_update_stats(void *timer, pid_t pid, void *startf,
209 void *timerf, char *comm, 194 void *timerf, char *comm, u32 flags);
210 unsigned int timer_flag);
211 195
212extern void __timer_stats_timer_set_start_info(struct timer_list *timer, 196extern void __timer_stats_timer_set_start_info(struct timer_list *timer,
213 void *addr); 197 void *addr);
@@ -254,6 +238,15 @@ extern void run_local_timers(void);
254struct hrtimer; 238struct hrtimer;
255extern enum hrtimer_restart it_real_fn(struct hrtimer *); 239extern enum hrtimer_restart it_real_fn(struct hrtimer *);
256 240
241#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
242#include <linux/sysctl.h>
243
244extern unsigned int sysctl_timer_migration;
245int timer_migration_handler(struct ctl_table *table, int write,
246 void __user *buffer, size_t *lenp,
247 loff_t *ppos);
248#endif
249
257unsigned long __round_jiffies(unsigned long j, int cpu); 250unsigned long __round_jiffies(unsigned long j, int cpu);
258unsigned long __round_jiffies_relative(unsigned long j, int cpu); 251unsigned long __round_jiffies_relative(unsigned long j, int cpu);
259unsigned long round_jiffies(unsigned long j); 252unsigned long round_jiffies(unsigned long j);
diff --git a/include/linux/timerqueue.h b/include/linux/timerqueue.h
index a520fd70a59f..7eec17ad7fa1 100644
--- a/include/linux/timerqueue.h
+++ b/include/linux/timerqueue.h
@@ -16,10 +16,10 @@ struct timerqueue_head {
16}; 16};
17 17
18 18
19extern void timerqueue_add(struct timerqueue_head *head, 19extern bool timerqueue_add(struct timerqueue_head *head,
20 struct timerqueue_node *node); 20 struct timerqueue_node *node);
21extern void timerqueue_del(struct timerqueue_head *head, 21extern bool timerqueue_del(struct timerqueue_head *head,
22 struct timerqueue_node *node); 22 struct timerqueue_node *node);
23extern struct timerqueue_node *timerqueue_iterate_next( 23extern struct timerqueue_node *timerqueue_iterate_next(
24 struct timerqueue_node *node); 24 struct timerqueue_node *node);
25 25
diff --git a/include/linux/topology.h b/include/linux/topology.h
index 909b6e43b694..73ddad1e0fa3 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -191,8 +191,8 @@ static inline int cpu_to_mem(int cpu)
191#ifndef topology_core_id 191#ifndef topology_core_id
192#define topology_core_id(cpu) ((void)(cpu), 0) 192#define topology_core_id(cpu) ((void)(cpu), 0)
193#endif 193#endif
194#ifndef topology_thread_cpumask 194#ifndef topology_sibling_cpumask
195#define topology_thread_cpumask(cpu) cpumask_of(cpu) 195#define topology_sibling_cpumask(cpu) cpumask_of(cpu)
196#endif 196#endif
197#ifndef topology_core_cpumask 197#ifndef topology_core_cpumask
198#define topology_core_cpumask(cpu) cpumask_of(cpu) 198#define topology_core_cpumask(cpu) cpumask_of(cpu)
@@ -201,7 +201,7 @@ static inline int cpu_to_mem(int cpu)
201#ifdef CONFIG_SCHED_SMT 201#ifdef CONFIG_SCHED_SMT
202static inline const struct cpumask *cpu_smt_mask(int cpu) 202static inline const struct cpumask *cpu_smt_mask(int cpu)
203{ 203{
204 return topology_thread_cpumask(cpu); 204 return topology_sibling_cpumask(cpu);
205} 205}
206#endif 206#endif
207 207
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ecd3319dac33..ae572c138607 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -1,21 +1,30 @@
1#ifndef __LINUX_UACCESS_H__ 1#ifndef __LINUX_UACCESS_H__
2#define __LINUX_UACCESS_H__ 2#define __LINUX_UACCESS_H__
3 3
4#include <linux/preempt.h> 4#include <linux/sched.h>
5#include <asm/uaccess.h> 5#include <asm/uaccess.h>
6 6
7static __always_inline void pagefault_disabled_inc(void)
8{
9 current->pagefault_disabled++;
10}
11
12static __always_inline void pagefault_disabled_dec(void)
13{
14 current->pagefault_disabled--;
15 WARN_ON(current->pagefault_disabled < 0);
16}
17
7/* 18/*
8 * These routines enable/disable the pagefault handler in that 19 * These routines enable/disable the pagefault handler. If disabled, it will
9 * it will not take any locks and go straight to the fixup table. 20 * not take any locks and go straight to the fixup table.
10 * 21 *
11 * They have great resemblance to the preempt_disable/enable calls 22 * User access methods will not sleep when called from a pagefault_disabled()
12 * and in fact they are identical; this is because currently there is 23 * environment.
13 * no other way to make the pagefault handlers do this. So we do
14 * disable preemption but we don't necessarily care about that.
15 */ 24 */
16static inline void pagefault_disable(void) 25static inline void pagefault_disable(void)
17{ 26{
18 preempt_count_inc(); 27 pagefault_disabled_inc();
19 /* 28 /*
20 * make sure to have issued the store before a pagefault 29 * make sure to have issued the store before a pagefault
21 * can hit. 30 * can hit.
@@ -25,18 +34,31 @@ static inline void pagefault_disable(void)
25 34
26static inline void pagefault_enable(void) 35static inline void pagefault_enable(void)
27{ 36{
28#ifndef CONFIG_PREEMPT
29 /* 37 /*
30 * make sure to issue those last loads/stores before enabling 38 * make sure to issue those last loads/stores before enabling
31 * the pagefault handler again. 39 * the pagefault handler again.
32 */ 40 */
33 barrier(); 41 barrier();
34 preempt_count_dec(); 42 pagefault_disabled_dec();
35#else
36 preempt_enable();
37#endif
38} 43}
39 44
45/*
46 * Is the pagefault handler disabled? If so, user access methods will not sleep.
47 */
48#define pagefault_disabled() (current->pagefault_disabled != 0)
49
50/*
51 * The pagefault handler is in general disabled by pagefault_disable() or
52 * when in irq context (via in_atomic()).
53 *
54 * This function should only be used by the fault handlers. Other users should
55 * stick to pagefault_disabled().
56 * Please NEVER use preempt_disable() to disable the fault handler. With
57 * !CONFIG_PREEMPT_COUNT, this is like a NOP. So the handler won't be disabled.
58 * in_atomic() will report different values based on !CONFIG_PREEMPT_COUNT.
59 */
60#define faulthandler_disabled() (pagefault_disabled() || in_atomic())
61
40#ifndef ARCH_HAS_NOCACHE_UACCESS 62#ifndef ARCH_HAS_NOCACHE_UACCESS
41 63
42static inline unsigned long __copy_from_user_inatomic_nocache(void *to, 64static inline unsigned long __copy_from_user_inatomic_nocache(void *to,
diff --git a/include/linux/wait.h b/include/linux/wait.h
index 2db83349865b..d69ac4ecc88b 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -969,7 +969,7 @@ extern int bit_wait_io_timeout(struct wait_bit_key *);
969 * on that signal. 969 * on that signal.
970 */ 970 */
971static inline int 971static inline int
972wait_on_bit(void *word, int bit, unsigned mode) 972wait_on_bit(unsigned long *word, int bit, unsigned mode)
973{ 973{
974 might_sleep(); 974 might_sleep();
975 if (!test_bit(bit, word)) 975 if (!test_bit(bit, word))
@@ -994,7 +994,7 @@ wait_on_bit(void *word, int bit, unsigned mode)
994 * on that signal. 994 * on that signal.
995 */ 995 */
996static inline int 996static inline int
997wait_on_bit_io(void *word, int bit, unsigned mode) 997wait_on_bit_io(unsigned long *word, int bit, unsigned mode)
998{ 998{
999 might_sleep(); 999 might_sleep();
1000 if (!test_bit(bit, word)) 1000 if (!test_bit(bit, word))
@@ -1020,7 +1020,8 @@ wait_on_bit_io(void *word, int bit, unsigned mode)
1020 * received a signal and the mode permitted wakeup on that signal. 1020 * received a signal and the mode permitted wakeup on that signal.
1021 */ 1021 */
1022static inline int 1022static inline int
1023wait_on_bit_timeout(void *word, int bit, unsigned mode, unsigned long timeout) 1023wait_on_bit_timeout(unsigned long *word, int bit, unsigned mode,
1024 unsigned long timeout)
1024{ 1025{
1025 might_sleep(); 1026 might_sleep();
1026 if (!test_bit(bit, word)) 1027 if (!test_bit(bit, word))
@@ -1047,7 +1048,8 @@ wait_on_bit_timeout(void *word, int bit, unsigned mode, unsigned long timeout)
1047 * on that signal. 1048 * on that signal.
1048 */ 1049 */
1049static inline int 1050static inline int
1050wait_on_bit_action(void *word, int bit, wait_bit_action_f *action, unsigned mode) 1051wait_on_bit_action(unsigned long *word, int bit, wait_bit_action_f *action,
1052 unsigned mode)
1051{ 1053{
1052 might_sleep(); 1054 might_sleep();
1053 if (!test_bit(bit, word)) 1055 if (!test_bit(bit, word))
@@ -1075,7 +1077,7 @@ wait_on_bit_action(void *word, int bit, wait_bit_action_f *action, unsigned mode
1075 * the @mode allows that signal to wake the process. 1077 * the @mode allows that signal to wake the process.
1076 */ 1078 */
1077static inline int 1079static inline int
1078wait_on_bit_lock(void *word, int bit, unsigned mode) 1080wait_on_bit_lock(unsigned long *word, int bit, unsigned mode)
1079{ 1081{
1080 might_sleep(); 1082 might_sleep();
1081 if (!test_and_set_bit(bit, word)) 1083 if (!test_and_set_bit(bit, word))
@@ -1099,7 +1101,7 @@ wait_on_bit_lock(void *word, int bit, unsigned mode)
1099 * the @mode allows that signal to wake the process. 1101 * the @mode allows that signal to wake the process.
1100 */ 1102 */
1101static inline int 1103static inline int
1102wait_on_bit_lock_io(void *word, int bit, unsigned mode) 1104wait_on_bit_lock_io(unsigned long *word, int bit, unsigned mode)
1103{ 1105{
1104 might_sleep(); 1106 might_sleep();
1105 if (!test_and_set_bit(bit, word)) 1107 if (!test_and_set_bit(bit, word))
@@ -1125,7 +1127,8 @@ wait_on_bit_lock_io(void *word, int bit, unsigned mode)
1125 * the @mode allows that signal to wake the process. 1127 * the @mode allows that signal to wake the process.
1126 */ 1128 */
1127static inline int 1129static inline int
1128wait_on_bit_lock_action(void *word, int bit, wait_bit_action_f *action, unsigned mode) 1130wait_on_bit_lock_action(unsigned long *word, int bit, wait_bit_action_f *action,
1131 unsigned mode)
1129{ 1132{
1130 might_sleep(); 1133 might_sleep();
1131 if (!test_and_set_bit(bit, word)) 1134 if (!test_and_set_bit(bit, word))
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index 497bc14cdb85..0320bbb7d7b5 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -98,7 +98,8 @@ struct inet_connection_sock {
98 const struct tcp_congestion_ops *icsk_ca_ops; 98 const struct tcp_congestion_ops *icsk_ca_ops;
99 const struct inet_connection_sock_af_ops *icsk_af_ops; 99 const struct inet_connection_sock_af_ops *icsk_af_ops;
100 unsigned int (*icsk_sync_mss)(struct sock *sk, u32 pmtu); 100 unsigned int (*icsk_sync_mss)(struct sock *sk, u32 pmtu);
101 __u8 icsk_ca_state:7, 101 __u8 icsk_ca_state:6,
102 icsk_ca_setsockopt:1,
102 icsk_ca_dst_locked:1; 103 icsk_ca_dst_locked:1;
103 __u8 icsk_retransmits; 104 __u8 icsk_retransmits;
104 __u8 icsk_pending; 105 __u8 icsk_pending;
diff --git a/include/net/mac80211.h b/include/net/mac80211.h
index 8e3668b44c29..fc57f6b82fc5 100644
--- a/include/net/mac80211.h
+++ b/include/net/mac80211.h
@@ -354,7 +354,7 @@ enum ieee80211_rssi_event_data {
354}; 354};
355 355
356/** 356/**
357 * enum ieee80211_rssi_event - data attached to an %RSSI_EVENT 357 * struct ieee80211_rssi_event - data attached to an %RSSI_EVENT
358 * @data: See &enum ieee80211_rssi_event_data 358 * @data: See &enum ieee80211_rssi_event_data
359 */ 359 */
360struct ieee80211_rssi_event { 360struct ieee80211_rssi_event {
@@ -388,7 +388,7 @@ enum ieee80211_mlme_event_status {
388}; 388};
389 389
390/** 390/**
391 * enum ieee80211_mlme_event - data attached to an %MLME_EVENT 391 * struct ieee80211_mlme_event - data attached to an %MLME_EVENT
392 * @data: See &enum ieee80211_mlme_event_data 392 * @data: See &enum ieee80211_mlme_event_data
393 * @status: See &enum ieee80211_mlme_event_status 393 * @status: See &enum ieee80211_mlme_event_status
394 * @reason: the reason code if applicable 394 * @reason: the reason code if applicable
@@ -401,9 +401,10 @@ struct ieee80211_mlme_event {
401 401
402/** 402/**
403 * struct ieee80211_event - event to be sent to the driver 403 * struct ieee80211_event - event to be sent to the driver
404 * @type The event itself. See &enum ieee80211_event_type. 404 * @type: The event itself. See &enum ieee80211_event_type.
405 * @rssi: relevant if &type is %RSSI_EVENT 405 * @rssi: relevant if &type is %RSSI_EVENT
406 * @mlme: relevant if &type is %AUTH_EVENT 406 * @mlme: relevant if &type is %AUTH_EVENT
407 * @u: union holding the above two fields
407 */ 408 */
408struct ieee80211_event { 409struct ieee80211_event {
409 enum ieee80211_event_type type; 410 enum ieee80211_event_type type;
diff --git a/include/net/sctp/sctp.h b/include/net/sctp/sctp.h
index c56a438c3a1e..ce13cf20f625 100644
--- a/include/net/sctp/sctp.h
+++ b/include/net/sctp/sctp.h
@@ -574,11 +574,14 @@ static inline void sctp_v6_map_v4(union sctp_addr *addr)
574/* Map v4 address to v4-mapped v6 address */ 574/* Map v4 address to v4-mapped v6 address */
575static inline void sctp_v4_map_v6(union sctp_addr *addr) 575static inline void sctp_v4_map_v6(union sctp_addr *addr)
576{ 576{
577 __be16 port;
578
579 port = addr->v4.sin_port;
580 addr->v6.sin6_addr.s6_addr32[3] = addr->v4.sin_addr.s_addr;
581 addr->v6.sin6_port = port;
577 addr->v6.sin6_family = AF_INET6; 582 addr->v6.sin6_family = AF_INET6;
578 addr->v6.sin6_flowinfo = 0; 583 addr->v6.sin6_flowinfo = 0;
579 addr->v6.sin6_scope_id = 0; 584 addr->v6.sin6_scope_id = 0;
580 addr->v6.sin6_port = addr->v4.sin_port;
581 addr->v6.sin6_addr.s6_addr32[3] = addr->v4.sin_addr.s_addr;
582 addr->v6.sin6_addr.s6_addr32[0] = 0; 585 addr->v6.sin6_addr.s6_addr32[0] = 0;
583 addr->v6.sin6_addr.s6_addr32[1] = 0; 586 addr->v6.sin6_addr.s6_addr32[1] = 0;
584 addr->v6.sin6_addr.s6_addr32[2] = htonl(0x0000ffff); 587 addr->v6.sin6_addr.s6_addr32[2] = htonl(0x0000ffff);
diff --git a/include/sound/hda_regmap.h b/include/sound/hda_regmap.h
index 53a18b3635e2..df705908480a 100644
--- a/include/sound/hda_regmap.h
+++ b/include/sound/hda_regmap.h
@@ -9,6 +9,8 @@
9#include <sound/core.h> 9#include <sound/core.h>
10#include <sound/hdaudio.h> 10#include <sound/hdaudio.h>
11 11
12#define AC_AMP_FAKE_MUTE 0x10 /* fake mute bit set to amp verbs */
13
12int snd_hdac_regmap_init(struct hdac_device *codec); 14int snd_hdac_regmap_init(struct hdac_device *codec);
13void snd_hdac_regmap_exit(struct hdac_device *codec); 15void snd_hdac_regmap_exit(struct hdac_device *codec);
14int snd_hdac_regmap_add_vendor_verb(struct hdac_device *codec, 16int snd_hdac_regmap_add_vendor_verb(struct hdac_device *codec,
diff --git a/include/target/target_core_backend.h b/include/target/target_core_backend.h
index d61be7297b2c..5f1225706993 100644
--- a/include/target/target_core_backend.h
+++ b/include/target/target_core_backend.h
@@ -1,9 +1,7 @@
1#ifndef TARGET_CORE_BACKEND_H 1#ifndef TARGET_CORE_BACKEND_H
2#define TARGET_CORE_BACKEND_H 2#define TARGET_CORE_BACKEND_H
3 3
4#define TRANSPORT_PLUGIN_PHBA_PDEV 1 4#define TRANSPORT_FLAG_PASSTHROUGH 1
5#define TRANSPORT_PLUGIN_VHBA_PDEV 2
6#define TRANSPORT_PLUGIN_VHBA_VDEV 3
7 5
8struct target_backend_cits { 6struct target_backend_cits {
9 struct config_item_type tb_dev_cit; 7 struct config_item_type tb_dev_cit;
@@ -22,7 +20,7 @@ struct se_subsystem_api {
22 char inquiry_rev[4]; 20 char inquiry_rev[4];
23 struct module *owner; 21 struct module *owner;
24 22
25 u8 transport_type; 23 u8 transport_flags;
26 24
27 int (*attach_hba)(struct se_hba *, u32); 25 int (*attach_hba)(struct se_hba *, u32);
28 void (*detach_hba)(struct se_hba *); 26 void (*detach_hba)(struct se_hba *);
@@ -138,5 +136,7 @@ int se_dev_set_queue_depth(struct se_device *, u32);
138int se_dev_set_max_sectors(struct se_device *, u32); 136int se_dev_set_max_sectors(struct se_device *, u32);
139int se_dev_set_optimal_sectors(struct se_device *, u32); 137int se_dev_set_optimal_sectors(struct se_device *, u32);
140int se_dev_set_block_size(struct se_device *, u32); 138int se_dev_set_block_size(struct se_device *, u32);
139sense_reason_t passthrough_parse_cdb(struct se_cmd *cmd,
140 sense_reason_t (*exec_cmd)(struct se_cmd *cmd));
141 141
142#endif /* TARGET_CORE_BACKEND_H */ 142#endif /* TARGET_CORE_BACKEND_H */
diff --git a/include/target/target_core_configfs.h b/include/target/target_core_configfs.h
index 25bb04c4209e..b99c01170392 100644
--- a/include/target/target_core_configfs.h
+++ b/include/target/target_core_configfs.h
@@ -40,8 +40,6 @@ struct target_fabric_configfs {
40 struct config_item *tf_fabric; 40 struct config_item *tf_fabric;
41 /* Passed from fabric modules */ 41 /* Passed from fabric modules */
42 struct config_item_type *tf_fabric_cit; 42 struct config_item_type *tf_fabric_cit;
43 /* Pointer to target core subsystem */
44 struct configfs_subsystem *tf_subsys;
45 /* Pointer to fabric's struct module */ 43 /* Pointer to fabric's struct module */
46 struct module *tf_module; 44 struct module *tf_module;
47 struct target_core_fabric_ops tf_ops; 45 struct target_core_fabric_ops tf_ops;
diff --git a/include/target/target_core_fabric.h b/include/target/target_core_fabric.h
index 17c7f5ac7ea0..0f4dc3768587 100644
--- a/include/target/target_core_fabric.h
+++ b/include/target/target_core_fabric.h
@@ -4,7 +4,6 @@
4struct target_core_fabric_ops { 4struct target_core_fabric_ops {
5 struct module *module; 5 struct module *module;
6 const char *name; 6 const char *name;
7 struct configfs_subsystem *tf_subsys;
8 char *(*get_fabric_name)(void); 7 char *(*get_fabric_name)(void);
9 u8 (*get_fabric_proto_ident)(struct se_portal_group *); 8 u8 (*get_fabric_proto_ident)(struct se_portal_group *);
10 char *(*tpg_get_wwn)(struct se_portal_group *); 9 char *(*tpg_get_wwn)(struct se_portal_group *);
@@ -109,6 +108,9 @@ struct target_core_fabric_ops {
109int target_register_template(const struct target_core_fabric_ops *fo); 108int target_register_template(const struct target_core_fabric_ops *fo);
110void target_unregister_template(const struct target_core_fabric_ops *fo); 109void target_unregister_template(const struct target_core_fabric_ops *fo);
111 110
111int target_depend_item(struct config_item *item);
112void target_undepend_item(struct config_item *item);
113
112struct se_session *transport_init_session(enum target_prot_op); 114struct se_session *transport_init_session(enum target_prot_op);
113int transport_alloc_session_tags(struct se_session *, unsigned int, 115int transport_alloc_session_tags(struct se_session *, unsigned int,
114 unsigned int); 116 unsigned int);
diff --git a/include/trace/events/kmem.h b/include/trace/events/kmem.h
index 81ea59812117..f7554fd7fc62 100644
--- a/include/trace/events/kmem.h
+++ b/include/trace/events/kmem.h
@@ -140,19 +140,42 @@ DEFINE_EVENT(kmem_free, kfree,
140 TP_ARGS(call_site, ptr) 140 TP_ARGS(call_site, ptr)
141); 141);
142 142
143DEFINE_EVENT(kmem_free, kmem_cache_free, 143DEFINE_EVENT_CONDITION(kmem_free, kmem_cache_free,
144 144
145 TP_PROTO(unsigned long call_site, const void *ptr), 145 TP_PROTO(unsigned long call_site, const void *ptr),
146 146
147 TP_ARGS(call_site, ptr) 147 TP_ARGS(call_site, ptr),
148
149 /*
150 * This trace can be potentially called from an offlined cpu.
151 * Since trace points use RCU and RCU should not be used from
152 * offline cpus, filter such calls out.
153 * While this trace can be called from a preemptable section,
154 * it has no impact on the condition since tasks can migrate
155 * only from online cpus to other online cpus. Thus its safe
156 * to use raw_smp_processor_id.
157 */
158 TP_CONDITION(cpu_online(raw_smp_processor_id()))
148); 159);
149 160
150TRACE_EVENT(mm_page_free, 161TRACE_EVENT_CONDITION(mm_page_free,
151 162
152 TP_PROTO(struct page *page, unsigned int order), 163 TP_PROTO(struct page *page, unsigned int order),
153 164
154 TP_ARGS(page, order), 165 TP_ARGS(page, order),
155 166
167
168 /*
169 * This trace can be potentially called from an offlined cpu.
170 * Since trace points use RCU and RCU should not be used from
171 * offline cpus, filter such calls out.
172 * While this trace can be called from a preemptable section,
173 * it has no impact on the condition since tasks can migrate
174 * only from online cpus to other online cpus. Thus its safe
175 * to use raw_smp_processor_id.
176 */
177 TP_CONDITION(cpu_online(raw_smp_processor_id())),
178
156 TP_STRUCT__entry( 179 TP_STRUCT__entry(
157 __field( unsigned long, pfn ) 180 __field( unsigned long, pfn )
158 __field( unsigned int, order ) 181 __field( unsigned int, order )
@@ -253,12 +276,35 @@ DEFINE_EVENT(mm_page, mm_page_alloc_zone_locked,
253 TP_ARGS(page, order, migratetype) 276 TP_ARGS(page, order, migratetype)
254); 277);
255 278
256DEFINE_EVENT_PRINT(mm_page, mm_page_pcpu_drain, 279TRACE_EVENT_CONDITION(mm_page_pcpu_drain,
257 280
258 TP_PROTO(struct page *page, unsigned int order, int migratetype), 281 TP_PROTO(struct page *page, unsigned int order, int migratetype),
259 282
260 TP_ARGS(page, order, migratetype), 283 TP_ARGS(page, order, migratetype),
261 284
285 /*
286 * This trace can be potentially called from an offlined cpu.
287 * Since trace points use RCU and RCU should not be used from
288 * offline cpus, filter such calls out.
289 * While this trace can be called from a preemptable section,
290 * it has no impact on the condition since tasks can migrate
291 * only from online cpus to other online cpus. Thus its safe
292 * to use raw_smp_processor_id.
293 */
294 TP_CONDITION(cpu_online(raw_smp_processor_id())),
295
296 TP_STRUCT__entry(
297 __field( unsigned long, pfn )
298 __field( unsigned int, order )
299 __field( int, migratetype )
300 ),
301
302 TP_fast_assign(
303 __entry->pfn = page ? page_to_pfn(page) : -1UL;
304 __entry->order = order;
305 __entry->migratetype = migratetype;
306 ),
307
262 TP_printk("page=%p pfn=%lu order=%d migratetype=%d", 308 TP_printk("page=%p pfn=%lu order=%d migratetype=%d",
263 pfn_to_page(__entry->pfn), __entry->pfn, 309 pfn_to_page(__entry->pfn), __entry->pfn,
264 __entry->order, __entry->migratetype) 310 __entry->order, __entry->migratetype)
diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index 30fedaf3e56a..d57a575fe31f 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -147,7 +147,8 @@ TRACE_EVENT(sched_switch,
147 __print_flags(__entry->prev_state & (TASK_STATE_MAX-1), "|", 147 __print_flags(__entry->prev_state & (TASK_STATE_MAX-1), "|",
148 { 1, "S"} , { 2, "D" }, { 4, "T" }, { 8, "t" }, 148 { 1, "S"} , { 2, "D" }, { 4, "T" }, { 8, "t" },
149 { 16, "Z" }, { 32, "X" }, { 64, "x" }, 149 { 16, "Z" }, { 32, "X" }, { 64, "x" },
150 { 128, "K" }, { 256, "W" }, { 512, "P" }) : "R", 150 { 128, "K" }, { 256, "W" }, { 512, "P" },
151 { 1024, "N" }) : "R",
151 __entry->prev_state & TASK_STATE_MAX ? "+" : "", 152 __entry->prev_state & TASK_STATE_MAX ? "+" : "",
152 __entry->next_comm, __entry->next_pid, __entry->next_prio) 153 __entry->next_comm, __entry->next_pid, __entry->next_prio)
153); 154);
diff --git a/include/trace/events/timer.h b/include/trace/events/timer.h
index 68c2c2000f02..073b9ac245ba 100644
--- a/include/trace/events/timer.h
+++ b/include/trace/events/timer.h
@@ -43,15 +43,18 @@ DEFINE_EVENT(timer_class, timer_init,
43 */ 43 */
44TRACE_EVENT(timer_start, 44TRACE_EVENT(timer_start,
45 45
46 TP_PROTO(struct timer_list *timer, unsigned long expires), 46 TP_PROTO(struct timer_list *timer,
47 unsigned long expires,
48 unsigned int flags),
47 49
48 TP_ARGS(timer, expires), 50 TP_ARGS(timer, expires, flags),
49 51
50 TP_STRUCT__entry( 52 TP_STRUCT__entry(
51 __field( void *, timer ) 53 __field( void *, timer )
52 __field( void *, function ) 54 __field( void *, function )
53 __field( unsigned long, expires ) 55 __field( unsigned long, expires )
54 __field( unsigned long, now ) 56 __field( unsigned long, now )
57 __field( unsigned int, flags )
55 ), 58 ),
56 59
57 TP_fast_assign( 60 TP_fast_assign(
@@ -59,11 +62,12 @@ TRACE_EVENT(timer_start,
59 __entry->function = timer->function; 62 __entry->function = timer->function;
60 __entry->expires = expires; 63 __entry->expires = expires;
61 __entry->now = jiffies; 64 __entry->now = jiffies;
65 __entry->flags = flags;
62 ), 66 ),
63 67
64 TP_printk("timer=%p function=%pf expires=%lu [timeout=%ld]", 68 TP_printk("timer=%p function=%pf expires=%lu [timeout=%ld] flags=0x%08x",
65 __entry->timer, __entry->function, __entry->expires, 69 __entry->timer, __entry->function, __entry->expires,
66 (long)__entry->expires - __entry->now) 70 (long)__entry->expires - __entry->now, __entry->flags)
67); 71);
68 72
69/** 73/**
diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h
index 880dd7437172..c178d13d6f4c 100644
--- a/include/trace/events/writeback.h
+++ b/include/trace/events/writeback.h
@@ -250,7 +250,6 @@ DEFINE_EVENT(writeback_class, name, \
250DEFINE_WRITEBACK_EVENT(writeback_nowork); 250DEFINE_WRITEBACK_EVENT(writeback_nowork);
251DEFINE_WRITEBACK_EVENT(writeback_wake_background); 251DEFINE_WRITEBACK_EVENT(writeback_wake_background);
252DEFINE_WRITEBACK_EVENT(writeback_bdi_register); 252DEFINE_WRITEBACK_EVENT(writeback_bdi_register);
253DEFINE_WRITEBACK_EVENT(writeback_bdi_unregister);
254 253
255DECLARE_EVENT_CLASS(wbc_class, 254DECLARE_EVENT_CLASS(wbc_class,
256 TP_PROTO(struct writeback_control *wbc, struct backing_dev_info *bdi), 255 TP_PROTO(struct writeback_control *wbc, struct backing_dev_info *bdi),
diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h
index 871e73f99a4d..94d44ab2fda1 100644
--- a/include/uapi/drm/radeon_drm.h
+++ b/include/uapi/drm/radeon_drm.h
@@ -1038,6 +1038,7 @@ struct drm_radeon_cs {
1038#define RADEON_INFO_CURRENT_GPU_SCLK 0x22 1038#define RADEON_INFO_CURRENT_GPU_SCLK 0x22
1039#define RADEON_INFO_CURRENT_GPU_MCLK 0x23 1039#define RADEON_INFO_CURRENT_GPU_MCLK 0x23
1040#define RADEON_INFO_READ_REG 0x24 1040#define RADEON_INFO_READ_REG 0x24
1041#define RADEON_INFO_VA_UNMAP_WORKING 0x25
1041 1042
1042struct drm_radeon_info { 1043struct drm_radeon_info {
1043 uint32_t request; 1044 uint32_t request;
diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
index 309211b3eb67..d97f84c080da 100644
--- a/include/uapi/linux/perf_event.h
+++ b/include/uapi/linux/perf_event.h
@@ -167,6 +167,7 @@ enum perf_branch_sample_type_shift {
167 PERF_SAMPLE_BRANCH_COND_SHIFT = 10, /* conditional branches */ 167 PERF_SAMPLE_BRANCH_COND_SHIFT = 10, /* conditional branches */
168 168
169 PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT = 11, /* call/ret stack */ 169 PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT = 11, /* call/ret stack */
170 PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT = 12, /* indirect jumps */
170 171
171 PERF_SAMPLE_BRANCH_MAX_SHIFT /* non-ABI */ 172 PERF_SAMPLE_BRANCH_MAX_SHIFT /* non-ABI */
172}; 173};
@@ -186,6 +187,7 @@ enum perf_branch_sample_type {
186 PERF_SAMPLE_BRANCH_COND = 1U << PERF_SAMPLE_BRANCH_COND_SHIFT, 187 PERF_SAMPLE_BRANCH_COND = 1U << PERF_SAMPLE_BRANCH_COND_SHIFT,
187 188
188 PERF_SAMPLE_BRANCH_CALL_STACK = 1U << PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT, 189 PERF_SAMPLE_BRANCH_CALL_STACK = 1U << PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT,
190 PERF_SAMPLE_BRANCH_IND_JUMP = 1U << PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT,
189 191
190 PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT, 192 PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT,
191}; 193};
@@ -564,6 +566,10 @@ struct perf_event_mmap_page {
564#define PERF_RECORD_MISC_GUEST_USER (5 << 0) 566#define PERF_RECORD_MISC_GUEST_USER (5 << 0)
565 567
566/* 568/*
569 * Indicates that /proc/PID/maps parsing are truncated by time out.
570 */
571#define PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT (1 << 12)
572/*
567 * PERF_RECORD_MISC_MMAP_DATA and PERF_RECORD_MISC_COMM_EXEC are used on 573 * PERF_RECORD_MISC_MMAP_DATA and PERF_RECORD_MISC_COMM_EXEC are used on
568 * different events so can reuse the same bit position. 574 * different events so can reuse the same bit position.
569 */ 575 */
@@ -800,6 +806,18 @@ enum perf_event_type {
800 */ 806 */
801 PERF_RECORD_ITRACE_START = 12, 807 PERF_RECORD_ITRACE_START = 12,
802 808
809 /*
810 * Records the dropped/lost sample number.
811 *
812 * struct {
813 * struct perf_event_header header;
814 *
815 * u64 lost;
816 * struct sample_id sample_id;
817 * };
818 */
819 PERF_RECORD_LOST_SAMPLES = 13,
820
803 PERF_RECORD_MAX, /* non-ABI */ 821 PERF_RECORD_MAX, /* non-ABI */
804}; 822};
805 823
diff --git a/include/uapi/linux/virtio_balloon.h b/include/uapi/linux/virtio_balloon.h
index 984169a819ee..d7f1cbc3766c 100644
--- a/include/uapi/linux/virtio_balloon.h
+++ b/include/uapi/linux/virtio_balloon.h
@@ -26,6 +26,7 @@
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE. */ 27 * SUCH DAMAGE. */
28#include <linux/types.h> 28#include <linux/types.h>
29#include <linux/virtio_types.h>
29#include <linux/virtio_ids.h> 30#include <linux/virtio_ids.h>
30#include <linux/virtio_config.h> 31#include <linux/virtio_config.h>
31 32
generated by cgit v1.2.2 (git 2.25.0) at 2025-11-02 12:02:57 -0500