[patch 0/8] sched: Clean up might_sleep() and make it RT aware

[patch 0/8] sched: Clean up might_sleep() and make it RT aware

[Thomas Gleixner]

The might sleep checks have grown some warts over time and especially the
underscore zoo is hard to follow.

Also the debug output is slightly confusing especially when PREEMPT_RCU is
enabled as it does not provide information about the RCU nesting depth
which might be the actual culprit. In such a case it might even print
confusing information about the place which disabled preemption despite the
fact that preemption is actually enabled.

RT in turn had it's own nasty hack to ignore RCU nesting depth because
that's required for spin/rw lock held sections.

This series addresses this by:

  - Cleaning up the underscore zoo

  - Enhancing the debug output

  - Making RCU nest depth and preempt count checks distinct

  - Fixing up the relevant places to hand in the correct expected values
    for RT and !RT kernels.

The series is also available from

    git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git sched

Thanks,

	tglx
---
 include/linux/kernel.h       |   13 +++-----
 include/linux/preempt.h      |    5 +--
 include/linux/sched.h        |   39 ++++++++++++++++++-------
 kernel/locking/spinlock_rt.c |   17 ++++++++--
 kernel/sched/core.c          |   67 +++++++++++++++++++++++++++----------------
 mm/memory.c                  |    2 -
 6 files changed, 96 insertions(+), 47 deletions(-)

[patch 1/8] sched: Clean up the might_sleep() underscore zoo

[Thomas Gleixner]

__might_sleep() vs. ___might_sleep() is hard to distinguish. Aside of that
the three underscore variant is exposed to provide a checkpoint for
rescheduling points which are distinct from blocking points.

They are semantically a preemption point which means that scheduling is
state preserving. A real blocking operation, e.g. mutex_lock(), wait*(),
which cannot preserve a task state which is not equal to RUNNING.

While technically blocking on a "sleeping" spinlock in RT enabled kernels
falls into the voluntary scheduling category because it has to wait until
the contended spin/rw lock becomes available, the RT lock substitution code
can semantically be mapped to a voluntary preemption because the RT lock
substitution code and the scheduler are providing mechanisms to preserve
the task state and to take regular non-lock related wakeups into account.

Rename ___might_sleep() to __might_resched() to make the distinction of
these functions clear.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/kernel.h       |    6 +++---
 include/linux/sched.h        |    8 ++++----
 kernel/locking/spinlock_rt.c |    6 +++---
 kernel/sched/core.c          |    6 +++---
 4 files changed, 13 insertions(+), 13 deletions(-)

--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -111,7 +111,7 @@ static __always_inline void might_resche
 #endif /* CONFIG_PREEMPT_* */
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
-extern void ___might_sleep(const char *file, int line, int preempt_offset);
+extern void __might_resched(const char *file, int line, int preempt_offset);
 extern void __might_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_migrate(const char *file, int line);
@@ -168,8 +168,8 @@ extern void __cant_migrate(const char *f
  */
 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
 #else
-  static inline void ___might_sleep(const char *file, int line,
-				   int preempt_offset) { }
+  static inline void __might_resched(const char *file, int line,
+				     int preempt_offset) { }
   static inline void __might_sleep(const char *file, int line,
 				   int preempt_offset) { }
 # define might_sleep() do { might_resched(); } while (0)
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2041,7 +2041,7 @@ static inline int _cond_resched(void) {
 #endif /* !defined(CONFIG_PREEMPTION) || defined(CONFIG_PREEMPT_DYNAMIC) */
 
 #define cond_resched() ({			\
-	___might_sleep(__FILE__, __LINE__, 0);	\
+	__might_resched(__FILE__, __LINE__, 0);	\
 	_cond_resched();			\
 })
 
@@ -2049,9 +2049,9 @@ extern int __cond_resched_lock(spinlock_
 extern int __cond_resched_rwlock_read(rwlock_t *lock);
 extern int __cond_resched_rwlock_write(rwlock_t *lock);
 
-#define cond_resched_lock(lock) ({				\
-	___might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);\
-	__cond_resched_lock(lock);				\
+#define cond_resched_lock(lock) ({					\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
+	__cond_resched_lock(lock);					\
 })
 
 #define cond_resched_rwlock_read(lock) ({			\
--- a/kernel/locking/spinlock_rt.c
+++ b/kernel/locking/spinlock_rt.c
@@ -32,7 +32,7 @@ static __always_inline void rtlock_lock(
 
 static __always_inline void __rt_spin_lock(spinlock_t *lock)
 {
-	___might_sleep(__FILE__, __LINE__, 0);
+	__might_resched(__FILE__, __LINE__, 0);
 	rtlock_lock(&lock->lock);
 	rcu_read_lock();
 	migrate_disable();
@@ -210,7 +210,7 @@ EXPORT_SYMBOL(rt_write_trylock);
 
 void __sched rt_read_lock(rwlock_t *rwlock)
 {
-	___might_sleep(__FILE__, __LINE__, 0);
+	__might_resched(__FILE__, __LINE__, 0);
 	rwlock_acquire_read(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_read_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();
@@ -220,7 +220,7 @@ EXPORT_SYMBOL(rt_read_lock);
 
 void __sched rt_write_lock(rwlock_t *rwlock)
 {
-	___might_sleep(__FILE__, __LINE__, 0);
+	__might_resched(__FILE__, __LINE__, 0);
 	rwlock_acquire(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_write_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9446,11 +9446,11 @@ void __might_sleep(const char *file, int
 			(void *)current->task_state_change,
 			(void *)current->task_state_change);
 
-	___might_sleep(file, line, preempt_offset);
+	__might_resched(file, line, preempt_offset);
 }
 EXPORT_SYMBOL(__might_sleep);
 
-void ___might_sleep(const char *file, int line, int preempt_offset)
+void __might_resched(const char *file, int line, int preempt_offset)
 {
 	/* Ratelimiting timestamp: */
 	static unsigned long prev_jiffy;
@@ -9495,7 +9495,7 @@ void ___might_sleep(const char *file, in
 	dump_stack();
 	add_taint(TAINT_WARN, LOCKDEP_STILL_OK);
 }
-EXPORT_SYMBOL(___might_sleep);
+EXPORT_SYMBOL(__might_resched);
 
 void __cant_sleep(const char *file, int line, int preempt_offset)
 {

[patch 2/8] sched: Make cond_resched_*lock() variants consistent vs. might_sleep()

[Thomas Gleixner]

Commit 3427445afd26 ("sched: Exclude cond_resched() from nested sleep
test") removed the task state check of __might_sleep() for
cond_resched_lock() because cond_resched_lock() is not a voluntary
scheduling point which blocks. It's a preemption point which requires the
lock holder to release the spin lock.

The same rationale applies to cond_resched_rwlock_read/write(), but those
were not touched.

Make it consistent and use the non-state checking __might_resched() there
as well.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/sched.h |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2054,14 +2054,14 @@ extern int __cond_resched_rwlock_write(r
 	__cond_resched_lock(lock);					\
 })
 
-#define cond_resched_rwlock_read(lock) ({			\
-	__might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_read(lock);			\
+#define cond_resched_rwlock_read(lock) ({				\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
+	__cond_resched_rwlock_read(lock);				\
 })
 
-#define cond_resched_rwlock_write(lock) ({			\
-	__might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_write(lock);			\
+#define cond_resched_rwlock_write(lock) ({				\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
+	__cond_resched_rwlock_write(lock);				\
 })
 
 static inline void cond_resched_rcu(void)

[patch 3/8] sched: Remove preempt_offset argument from __might_sleep()

[Thomas Gleixner]

All callers hand in 0 and never will hand in anything else.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/kernel.h |    7 +++----
 kernel/sched/core.c    |    4 ++--
 mm/memory.c            |    2 +-
 3 files changed, 6 insertions(+), 7 deletions(-)

--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -112,7 +112,7 @@ static __always_inline void might_resche
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
 extern void __might_resched(const char *file, int line, int preempt_offset);
-extern void __might_sleep(const char *file, int line, int preempt_offset);
+extern void __might_sleep(const char *file, int line);
 extern void __cant_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_migrate(const char *file, int line);
 
@@ -129,7 +129,7 @@ extern void __cant_migrate(const char *f
  * supposed to.
  */
 # define might_sleep() \
-	do { __might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0)
+	do { __might_sleep(__FILE__, __LINE__); might_resched(); } while (0)
 /**
  * cant_sleep - annotation for functions that cannot sleep
  *
@@ -170,8 +170,7 @@ extern void __cant_migrate(const char *f
 #else
   static inline void __might_resched(const char *file, int line,
 				     int preempt_offset) { }
-  static inline void __might_sleep(const char *file, int line,
-				   int preempt_offset) { }
+static inline void __might_sleep(const char *file, int line) { }
 # define might_sleep() do { might_resched(); } while (0)
 # define cant_sleep() do { } while (0)
 # define cant_migrate()		do { } while (0)
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9432,7 +9432,7 @@ static inline int preempt_count_equals(i
 	return (nested == preempt_offset);
 }
 
-void __might_sleep(const char *file, int line, int preempt_offset)
+void __might_sleep(const char *file, int line)
 {
 	unsigned int state = get_current_state();
 	/*
@@ -9446,7 +9446,7 @@ void __might_sleep(const char *file, int
 			(void *)current->task_state_change,
 			(void *)current->task_state_change);
 
-	__might_resched(file, line, preempt_offset);
+	__might_resched(file, line, 0);
 }
 EXPORT_SYMBOL(__might_sleep);
 
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -5255,7 +5255,7 @@ void __might_fault(const char *file, int
 		return;
 	if (pagefault_disabled())
 		return;
-	__might_sleep(file, line, 0);
+	__might_sleep(file, line);
 #if defined(CONFIG_DEBUG_ATOMIC_SLEEP)
 	if (current->mm)
 		might_lock_read(&current->mm->mmap_lock);

[patch 4/8] sched: Cleanup might_sleep() printks

[Thomas Gleixner]

Convert them to pr_*(). No functional change.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 kernel/sched/core.c |   14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9473,16 +9473,14 @@ void __might_resched(const char *file, i
 	/* Save this before calling printk(), since that will clobber it: */
 	preempt_disable_ip = get_preempt_disable_ip(current);
 
-	printk(KERN_ERR
-		"BUG: sleeping function called from invalid context at %s:%d\n",
-			file, line);
-	printk(KERN_ERR
-		"in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n",
-			in_atomic(), irqs_disabled(), current->non_block_count,
-			current->pid, current->comm);
+	pr_err("BUG: sleeping function called from invalid context at %s:%d\n",
+	       file, line);
+	pr_err("in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n",
+	       in_atomic(), irqs_disabled(), current->non_block_count,
+	       current->pid, current->comm);
 
 	if (task_stack_end_corrupted(current))
-		printk(KERN_EMERG "Thread overran stack, or stack corrupted\n");
+		pr_emerg("Thread overran stack, or stack corrupted\n");
 
 	debug_show_held_locks(current);
 	if (irqs_disabled())

[patch 5/8] sched: Make might_sleep() output less confusing

[Thomas Gleixner]

might_sleep() output is pretty informative, but can be confusing at times
especially with PREEMPT_RCU when the check triggers due to a voluntary
sleep inside a RCU read side critical section:

 BUG: sleeping function called from invalid context at kernel/test.c:110
 in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 Preemption disabled at: migrate_disable+0x33/0xa0

in_atomic() is 0, but it still tells that preemption was disabled at
migrate_disable(), which is completely useless because preemption is not
disabled. But the interesting information to decode the above, i.e. the RCU
nesting depth, is not printed.

That becomes even more confusing when might_sleep() is invoked from
cond_resched_lock() within a RCU read side critical section. Here the
expected preemption count is 1 and not 0.

 BUG: sleeping function called from invalid context at kernel/test.c:131
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 Preemption disabled at: test_cond_lock+0xf3/0x1c0

So in_atomic() is set, which is expected as the caller holds a spinlock,
but it's unclear why this is broken and the preempt disable IP is just
pointing at the correct place, i.e. spin_lock(), which is obviously not
helpful either.

Make that more useful in general:

 - Print preempt_count() and the expected value

and for the CONFIG_PREEMPT_RCU case:

 - Print the RCU read side critical section nesting depth

 - Print the preempt disable IP only when preempt count
   does not have the expected value.

So the might_sleep() dump from a within a preemptible RCU read side
critical section becomes:

 BUG: sleeping function called from invalid context at kernel/test.c:110
 in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 preempt_count: 0, expected: 0
 RCU nest depth: 1, expected: 0

and the cond_resched_lock() case becomes:

 BUG: sleeping function called from invalid context at kernel/test.c:141
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 preempt_count: 1, expected: 1
 RCU nest depth: 1, expected: 0

which makes is pretty obvious what's going on. For all other cases the
preempt disable IP is still printed as before:

 BUG: sleeping function called from invalid context at kernel/test.c: 156
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1, name: swapper/0
 preempt_count: 1, expected: 0
 RCU nest depth: 0, expected: 0
 Preemption disabled at:
 [<ffffffff82b48326>] test_might_sleep+0xbe/0xf8

 BUG: sleeping function called from invalid context at kernel/test.c: 163
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1, name: swapper/0
 preempt_count: 1, expected: 0
 RCU nest depth: 1, expected: 0
 Preemption disabled at:
 [<ffffffff82b48326>] test_might_sleep+0x1e4/0x280

This also prepares to provide a better debugging output for RT enabled
kernels and their spinlock substitutions.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 kernel/sched/core.c |   27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9450,6 +9450,18 @@ void __might_sleep(const char *file, int
 }
 EXPORT_SYMBOL(__might_sleep);
 
+static void print_preempt_disable_ip(int preempt_offset, unsigned long ip)
+{
+	if (!IS_ENABLED(CONFIG_DEBUG_PREEMPT))
+		return;
+
+	if (preempt_count() == preempt_offset)
+		return;
+
+	pr_err("Preemption disabled at:");
+	print_ip_sym(KERN_ERR, ip);
+}
+
 void __might_resched(const char *file, int line, int preempt_offset)
 {
 	/* Ratelimiting timestamp: */
@@ -9478,6 +9490,13 @@ void __might_resched(const char *file, i
 	pr_err("in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n",
 	       in_atomic(), irqs_disabled(), current->non_block_count,
 	       current->pid, current->comm);
+	pr_err("preempt_count: %x, expected: %x\n", preempt_count(),
+	       preempt_offset);
+
+	if (IS_ENABLED(CONFIG_PREEMPT_RCU)) {
+		pr_err("RCU nest depth: %d, expected: 0\n",
+		       rcu_preempt_depth());
+	}
 
 	if (task_stack_end_corrupted(current))
 		pr_emerg("Thread overran stack, or stack corrupted\n");
@@ -9485,11 +9504,9 @@ void __might_resched(const char *file, i
 	debug_show_held_locks(current);
 	if (irqs_disabled())
 		print_irqtrace_events(current);
-	if (IS_ENABLED(CONFIG_DEBUG_PREEMPT)
-	    && !preempt_count_equals(preempt_offset)) {
-		pr_err("Preemption disabled at:");
-		print_ip_sym(KERN_ERR, preempt_disable_ip);
-	}
+
+	print_preempt_disable_ip(preempt_offset, preempt_disable_ip);
+
 	dump_stack();
 	add_taint(TAINT_WARN, LOCKDEP_STILL_OK);
 }

[patch 6/8] sched: Make RCU nest depth distinct in __might_resched()

[Thomas Gleixner]

For !RT kernels RCU nest depth in __might_resched() is always expected to
be 0, but on RT kernels it can be non zero while the preempt count is
expected to be always 0.

Instead of playing magic games in interpreting the 'preempt_offset'
argument, rename it to 'offsets' and use the lower 8 bits for the expected
preempt count, allow to hand in the expected RCU nest depth in the upper
bits and adopt the __might_resched() code and related checks and printks.

The affected call sites are updated in subsequent steps.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/kernel.h |    4 ++--
 include/linux/sched.h  |    3 +++
 kernel/sched/core.c    |   28 ++++++++++++++++------------
 3 files changed, 21 insertions(+), 14 deletions(-)

--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -111,7 +111,7 @@ static __always_inline void might_resche
 #endif /* CONFIG_PREEMPT_* */
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
-extern void __might_resched(const char *file, int line, int preempt_offset);
+extern void __might_resched(const char *file, int line, unsigned int offsets);
 extern void __might_sleep(const char *file, int line);
 extern void __cant_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_migrate(const char *file, int line);
@@ -169,7 +169,7 @@ extern void __cant_migrate(const char *f
 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
 #else
   static inline void __might_resched(const char *file, int line,
-				     int preempt_offset) { }
+				     unsigned int offsets) { }
 static inline void __might_sleep(const char *file, int line) { }
 # define might_sleep() do { might_resched(); } while (0)
 # define cant_sleep() do { } while (0)
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2049,6 +2049,9 @@ extern int __cond_resched_lock(spinlock_
 extern int __cond_resched_rwlock_read(rwlock_t *lock);
 extern int __cond_resched_rwlock_write(rwlock_t *lock);
 
+#define MIGHT_RESCHED_RCU_SHIFT		8
+#define MIGHT_RESCHED_PREEMPT_MASK	((1U << MIGHT_RESCHED_RCU_SHIFT) - 1)
+
 #define cond_resched_lock(lock) ({					\
 	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
 	__cond_resched_lock(lock);					\
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9425,12 +9425,6 @@ void __init sched_init(void)
 }
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
-static inline int preempt_count_equals(int preempt_offset)
-{
-	int nested = preempt_count() + rcu_preempt_depth();
-
-	return (nested == preempt_offset);
-}
 
 void __might_sleep(const char *file, int line)
 {
@@ -9462,7 +9456,16 @@ static void print_preempt_disable_ip(int
 	print_ip_sym(KERN_ERR, ip);
 }
 
-void __might_resched(const char *file, int line, int preempt_offset)
+static inline bool resched_offsets_ok(unsigned int offsets)
+{
+	unsigned int nested = preempt_count();
+
+	nested += rcu_preempt_depth() << MIGHT_RESCHED_RCU_SHIFT;
+
+	return nested == offsets;
+}
+
+void __might_resched(const char *file, int line, unsigned int offsets)
 {
 	/* Ratelimiting timestamp: */
 	static unsigned long prev_jiffy;
@@ -9472,7 +9475,7 @@ void __might_resched(const char *file, i
 	/* WARN_ON_ONCE() by default, no rate limit required: */
 	rcu_sleep_check();
 
-	if ((preempt_count_equals(preempt_offset) && !irqs_disabled() &&
+	if ((resched_offsets_ok(offsets) && !irqs_disabled() &&
 	     !is_idle_task(current) && !current->non_block_count) ||
 	    system_state == SYSTEM_BOOTING || system_state > SYSTEM_RUNNING ||
 	    oops_in_progress)
@@ -9491,11 +9494,11 @@ void __might_resched(const char *file, i
 	       in_atomic(), irqs_disabled(), current->non_block_count,
 	       current->pid, current->comm);
 	pr_err("preempt_count: %x, expected: %x\n", preempt_count(),
-	       preempt_offset);
+	       offsets & MIGHT_RESCHED_PREEMPT_MASK);
 
 	if (IS_ENABLED(CONFIG_PREEMPT_RCU)) {
-		pr_err("RCU nest depth: %d, expected: 0\n",
-		       rcu_preempt_depth());
+		pr_err("RCU nest depth: %d, expected: %u\n",
+		       rcu_preempt_depth(), offsets >> MIGHT_RESCHED_RCU_SHIFT);
 	}
 
 	if (task_stack_end_corrupted(current))
@@ -9505,7 +9508,8 @@ void __might_resched(const char *file, i
 	if (irqs_disabled())
 		print_irqtrace_events(current);
 
-	print_preempt_disable_ip(preempt_offset, preempt_disable_ip);
+	print_preempt_disable_ip(offsets & MIGHT_RESCHED_PREEMPT_MASK,
+				 preempt_disable_ip);
 
 	dump_stack();
 	add_taint(TAINT_WARN, LOCKDEP_STILL_OK);

[patch 7/8] sched: Make cond_resched_lock() variants RT aware

[Thomas Gleixner]

The __might_resched() checks in the cond_resched_lock() variants use
PREEMPT_LOCK_OFFSET for preempt count offset checking which takes the
preemption disable by the spin_lock() which is still held at that point
into account.

On PREEMPT_RT enabled kernels spin/rw_lock held sections stay preemptible
which means PREEMPT_LOCK_OFFSET is 0, but that still triggers the
__might_resched() check because that takes RCU read side nesting into
account.

On RT enabled kernels spin/read/write_lock() issue rcu_read_lock() to
resemble the !RT semantics, which means in cond_resched_lock() the might
resched check will see preempt_count() == 0 and rcu_preempt_depth() == 1.

Introduce PREEMPT_LOCK_SCHED_OFFSET for those might resched checks and map
them depending on CONFIG_PREEMPT_RT.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/preempt.h |    5 +++--
 include/linux/sched.h   |   34 +++++++++++++++++++++++++---------
 2 files changed, 28 insertions(+), 11 deletions(-)

--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -122,9 +122,10 @@
  * The preempt_count offset after spin_lock()
  */
 #if !defined(CONFIG_PREEMPT_RT)
-#define PREEMPT_LOCK_OFFSET	PREEMPT_DISABLE_OFFSET
+#define PREEMPT_LOCK_OFFSET		PREEMPT_DISABLE_OFFSET
 #else
-#define PREEMPT_LOCK_OFFSET	0
+/* Locks on RT do not disable preemption */
+#define PREEMPT_LOCK_OFFSET		0
 #endif
 
 /*
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2052,19 +2052,35 @@ extern int __cond_resched_rwlock_write(r
 #define MIGHT_RESCHED_RCU_SHIFT		8
 #define MIGHT_RESCHED_PREEMPT_MASK	((1U << MIGHT_RESCHED_RCU_SHIFT) - 1)
 
-#define cond_resched_lock(lock) ({					\
-	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_lock(lock);					\
+#ifndef CONFIG_PREEMPT_RT
+/*
+ * Non RT kernels have an elevated preempt count due to the held lock,
+ * but are not allowed to be inside a RCU read side critical section
+ */
+# define PREEMPT_LOCK_RESCHED_OFFSETS	PREEMPT_LOCK_OFFSET
+#else
+/*
+ * spin/rw_lock() on RT implies rcu_read_lock(). The might_sleep() check in
+ * cond_resched*lock() has to take that into account because it checks for
+ * preempt_count() and rcu_preempt_depth().
+ */
+# define PREEMPT_LOCK_RESCHED_OFFSETS	\
+	(PREEMPT_LOCK_OFFSET + (1U << MIGHT_RESCHED_RCU_SHIFT))
+#endif
+
+#define cond_resched_lock(lock) ({						\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_RESCHED_OFFSETS);	\
+	__cond_resched_lock(lock);						\
 })
 
-#define cond_resched_rwlock_read(lock) ({				\
-	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_read(lock);				\
+#define cond_resched_rwlock_read(lock) ({					\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_RESCHED_OFFSETS);	\
+	__cond_resched_rwlock_read(lock);					\
 })
 
-#define cond_resched_rwlock_write(lock) ({				\
-	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_write(lock);				\
+#define cond_resched_rwlock_write(lock) ({					\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_RESCHED_OFFSETS);	\
+	__cond_resched_rwlock_write(lock);					\
 })
 
 static inline void cond_resched_rcu(void)

[patch 8/8] locking/rt: Take RCU nesting into account for __might_resched()

[Thomas Gleixner]

The general rule that rcu_read_lock() held sections cannot voluntary sleep
does apply even on RT kernels. Though the substitution of spin/rw locks on
RT enabled kernels has to be exempt from that rule. On !RT a spin_lock()
can obviously nest inside a RCU read side critical section as the lock
acquisition is not going to block, but on RT this is not longer the case
due to the 'sleeping' spinlock substitution.

The RT patches contained a cheap hack to ignore the RCU nesting depth in
might_sleep() checks, which was a pragmatic but incorrect workaround.

Instead of generally ignoring the RCU nesting depth in __might_sleep() and
__might_resched() checks, pass the rcu_preempt_depth() via the offsets
argument to __might_resched() from spin/read/write_lock() which makes the
checks work correctly even in RCU read side critical sections.

The actual blocking on such a substituted lock within a RCU read side
critical section is already handled correctly in __schedule() by treating
it as a "preemption" of the RCU read side critical section.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 kernel/locking/spinlock_rt.c |   17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

--- a/kernel/locking/spinlock_rt.c
+++ b/kernel/locking/spinlock_rt.c
@@ -24,6 +24,17 @@
 #define RT_MUTEX_BUILD_SPINLOCKS
 #include "rtmutex.c"
 
+/*
+ * __might_resched() skips the state check as rtlocks are state
+ * preserving. Take RCU nesting into account as spin/read/write_lock() can
+ * legitimately nest into an RCU read side critical section.
+ */
+#define RTLOCK_RESCHED_OFFSETS						\
+	(rcu_preempt_depth() << MIGHT_RESCHED_RCU_SHIFT)
+
+#define rtlock_might_resched()						\
+	__might_resched(__FILE__, __LINE__, RTLOCK_RESCHED_OFFSETS)
+
 static __always_inline void rtlock_lock(struct rt_mutex_base *rtm)
 {
 	if (unlikely(!rt_mutex_cmpxchg_acquire(rtm, NULL, current)))
@@ -32,7 +43,7 @@ static __always_inline void rtlock_lock(
 
 static __always_inline void __rt_spin_lock(spinlock_t *lock)
 {
-	__might_resched(__FILE__, __LINE__, 0);
+	rtlock_might_resched();
 	rtlock_lock(&lock->lock);
 	rcu_read_lock();
 	migrate_disable();
@@ -210,7 +221,7 @@ EXPORT_SYMBOL(rt_write_trylock);
 
 void __sched rt_read_lock(rwlock_t *rwlock)
 {
-	__might_resched(__FILE__, __LINE__, 0);
+	rtlock_might_resched();
 	rwlock_acquire_read(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_read_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();
@@ -220,7 +231,7 @@ EXPORT_SYMBOL(rt_read_lock);
 
 void __sched rt_write_lock(rwlock_t *rwlock)
 {
-	__might_resched(__FILE__, __LINE__, 0);
+	rtlock_might_resched();
 	rwlock_acquire(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_write_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();

[tip: locking/core] locking/rt: Take RCU nesting into account for __might_resched()

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     ef1f4804b27a54da34de6984d16f1fe8f2cc7011
Gitweb:        https://git.kernel.org/tip/ef1f4804b27a54da34de6984d16f1fe8f2cc7011
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:46 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:51 +02:00

locking/rt: Take RCU nesting into account for __might_resched()

The general rule that rcu_read_lock() held sections cannot voluntary sleep
does apply even on RT kernels. Though the substitution of spin/rw locks on
RT enabled kernels has to be exempt from that rule. On !RT a spin_lock()
can obviously nest inside a RCU read side critical section as the lock
acquisition is not going to block, but on RT this is not longer the case
due to the 'sleeping' spinlock substitution.

The RT patches contained a cheap hack to ignore the RCU nesting depth in
might_sleep() checks, which was a pragmatic but incorrect workaround.

Instead of generally ignoring the RCU nesting depth in __might_sleep() and
__might_resched() checks, pass the rcu_preempt_depth() via the offsets
argument to __might_resched() from spin/read/write_lock() which makes the
checks work correctly even in RCU read side critical sections.

The actual blocking on such a substituted lock within a RCU read side
critical section is already handled correctly in __schedule() by treating
it as a "preemption" of the RCU read side critical section.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165358.368305497@linutronix.de
---
 kernel/locking/spinlock_rt.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/kernel/locking/spinlock_rt.c b/kernel/locking/spinlock_rt.c
index c528924..b2e553f 100644
--- a/kernel/locking/spinlock_rt.c
+++ b/kernel/locking/spinlock_rt.c
@@ -24,6 +24,17 @@
 #define RT_MUTEX_BUILD_SPINLOCKS
 #include "rtmutex.c"
 
+/*
+ * __might_resched() skips the state check as rtlocks are state
+ * preserving. Take RCU nesting into account as spin/read/write_lock() can
+ * legitimately nest into an RCU read side critical section.
+ */
+#define RTLOCK_RESCHED_OFFSETS						\
+	(rcu_preempt_depth() << MIGHT_RESCHED_RCU_SHIFT)
+
+#define rtlock_might_resched()						\
+	__might_resched(__FILE__, __LINE__, RTLOCK_RESCHED_OFFSETS)
+
 static __always_inline void rtlock_lock(struct rt_mutex_base *rtm)
 {
 	if (unlikely(!rt_mutex_cmpxchg_acquire(rtm, NULL, current)))
@@ -32,7 +43,7 @@ static __always_inline void rtlock_lock(struct rt_mutex_base *rtm)
 
 static __always_inline void __rt_spin_lock(spinlock_t *lock)
 {
-	__might_resched(__FILE__, __LINE__, 0);
+	rtlock_might_resched();
 	rtlock_lock(&lock->lock);
 	rcu_read_lock();
 	migrate_disable();
@@ -210,7 +221,7 @@ EXPORT_SYMBOL(rt_write_trylock);
 
 void __sched rt_read_lock(rwlock_t *rwlock)
 {
-	__might_resched(__FILE__, __LINE__, 0);
+	rtlock_might_resched();
 	rwlock_acquire_read(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_read_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();
@@ -220,7 +231,7 @@ EXPORT_SYMBOL(rt_read_lock);
 
 void __sched rt_write_lock(rwlock_t *rwlock)
 {
-	__might_resched(__FILE__, __LINE__, 0);
+	rtlock_might_resched();
 	rwlock_acquire(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_write_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();

[tip: locking/core] sched: Make cond_resched_lock() variants RT aware

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     3e9cc688e56cc2abb9b6067f57c8397f6c96d42c
Gitweb:        https://git.kernel.org/tip/3e9cc688e56cc2abb9b6067f57c8397f6c96d42c
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:44 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:51 +02:00

sched: Make cond_resched_lock() variants RT aware

The __might_resched() checks in the cond_resched_lock() variants use
PREEMPT_LOCK_OFFSET for preempt count offset checking which takes the
preemption disable by the spin_lock() which is still held at that point
into account.

On PREEMPT_RT enabled kernels spin/rw_lock held sections stay preemptible
which means PREEMPT_LOCK_OFFSET is 0, but that still triggers the
__might_resched() check because that takes RCU read side nesting into
account.

On RT enabled kernels spin/read/write_lock() issue rcu_read_lock() to
resemble the !RT semantics, which means in cond_resched_lock() the might
resched check will see preempt_count() == 0 and rcu_preempt_depth() == 1.

Introduce PREEMPT_LOCK_SCHED_OFFSET for those might resched checks and map
them depending on CONFIG_PREEMPT_RT.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165358.305969211@linutronix.de
---
 include/linux/preempt.h |  5 +++--
 include/linux/sched.h   | 34 +++++++++++++++++++++++++---------
 2 files changed, 28 insertions(+), 11 deletions(-)

diff --git a/include/linux/preempt.h b/include/linux/preempt.h
index 4d244e2..031898b 100644
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -122,9 +122,10 @@
  * The preempt_count offset after spin_lock()
  */
 #if !defined(CONFIG_PREEMPT_RT)
-#define PREEMPT_LOCK_OFFSET	PREEMPT_DISABLE_OFFSET
+#define PREEMPT_LOCK_OFFSET		PREEMPT_DISABLE_OFFSET
 #else
-#define PREEMPT_LOCK_OFFSET	0
+/* Locks on RT do not disable preemption */
+#define PREEMPT_LOCK_OFFSET		0
 #endif
 
 /*
diff --git a/include/linux/sched.h b/include/linux/sched.h
index b448c74..21b7cd0 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2049,19 +2049,35 @@ extern int __cond_resched_rwlock_write(rwlock_t *lock);
 #define MIGHT_RESCHED_RCU_SHIFT		8
 #define MIGHT_RESCHED_PREEMPT_MASK	((1U << MIGHT_RESCHED_RCU_SHIFT) - 1)
 
-#define cond_resched_lock(lock) ({					\
-	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_lock(lock);					\
+#ifndef CONFIG_PREEMPT_RT
+/*
+ * Non RT kernels have an elevated preempt count due to the held lock,
+ * but are not allowed to be inside a RCU read side critical section
+ */
+# define PREEMPT_LOCK_RESCHED_OFFSETS	PREEMPT_LOCK_OFFSET
+#else
+/*
+ * spin/rw_lock() on RT implies rcu_read_lock(). The might_sleep() check in
+ * cond_resched*lock() has to take that into account because it checks for
+ * preempt_count() and rcu_preempt_depth().
+ */
+# define PREEMPT_LOCK_RESCHED_OFFSETS	\
+	(PREEMPT_LOCK_OFFSET + (1U << MIGHT_RESCHED_RCU_SHIFT))
+#endif
+
+#define cond_resched_lock(lock) ({						\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_RESCHED_OFFSETS);	\
+	__cond_resched_lock(lock);						\
 })
 
-#define cond_resched_rwlock_read(lock) ({				\
-	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_read(lock);				\
+#define cond_resched_rwlock_read(lock) ({					\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_RESCHED_OFFSETS);	\
+	__cond_resched_rwlock_read(lock);					\
 })
 
-#define cond_resched_rwlock_write(lock) ({				\
-	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_write(lock);				\
+#define cond_resched_rwlock_write(lock) ({					\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_RESCHED_OFFSETS);	\
+	__cond_resched_rwlock_write(lock);					\
 })
 
 static inline void cond_resched_rcu(void)

[tip: locking/core] sched: Make RCU nest depth distinct in __might_resched()

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     50e081b96e35e43b65591f40f7376204decd1cb5
Gitweb:        https://git.kernel.org/tip/50e081b96e35e43b65591f40f7376204decd1cb5
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:43 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:51 +02:00

sched: Make RCU nest depth distinct in __might_resched()

For !RT kernels RCU nest depth in __might_resched() is always expected to
be 0, but on RT kernels it can be non zero while the preempt count is
expected to be always 0.

Instead of playing magic games in interpreting the 'preempt_offset'
argument, rename it to 'offsets' and use the lower 8 bits for the expected
preempt count, allow to hand in the expected RCU nest depth in the upper
bits and adopt the __might_resched() code and related checks and printks.

The affected call sites are updated in subsequent steps.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165358.243232823@linutronix.de
---
 include/linux/kernel.h |  4 ++--
 include/linux/sched.h  |  3 +++
 kernel/sched/core.c    | 28 ++++++++++++++++------------
 3 files changed, 21 insertions(+), 14 deletions(-)

diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index f95ee78..e8696e4 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -111,7 +111,7 @@ static __always_inline void might_resched(void)
 #endif /* CONFIG_PREEMPT_* */
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
-extern void __might_resched(const char *file, int line, int preempt_offset);
+extern void __might_resched(const char *file, int line, unsigned int offsets);
 extern void __might_sleep(const char *file, int line);
 extern void __cant_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_migrate(const char *file, int line);
@@ -169,7 +169,7 @@ extern void __cant_migrate(const char *file, int line);
 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
 #else
   static inline void __might_resched(const char *file, int line,
-				     int preempt_offset) { }
+				     unsigned int offsets) { }
 static inline void __might_sleep(const char *file, int line) { }
 # define might_sleep() do { might_resched(); } while (0)
 # define cant_sleep() do { } while (0)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 7a989f2..b448c74 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2046,6 +2046,9 @@ extern int __cond_resched_lock(spinlock_t *lock);
 extern int __cond_resched_rwlock_read(rwlock_t *lock);
 extern int __cond_resched_rwlock_write(rwlock_t *lock);
 
+#define MIGHT_RESCHED_RCU_SHIFT		8
+#define MIGHT_RESCHED_PREEMPT_MASK	((1U << MIGHT_RESCHED_RCU_SHIFT) - 1)
+
 #define cond_resched_lock(lock) ({					\
 	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
 	__cond_resched_lock(lock);					\
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 0a27cb8..8d3fa07 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9468,12 +9468,6 @@ void __init sched_init(void)
 }
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
-static inline int preempt_count_equals(int preempt_offset)
-{
-	int nested = preempt_count() + rcu_preempt_depth();
-
-	return (nested == preempt_offset);
-}
 
 void __might_sleep(const char *file, int line)
 {
@@ -9505,7 +9499,16 @@ static void print_preempt_disable_ip(int preempt_offset, unsigned long ip)
 	print_ip_sym(KERN_ERR, ip);
 }
 
-void __might_resched(const char *file, int line, int preempt_offset)
+static inline bool resched_offsets_ok(unsigned int offsets)
+{
+	unsigned int nested = preempt_count();
+
+	nested += rcu_preempt_depth() << MIGHT_RESCHED_RCU_SHIFT;
+
+	return nested == offsets;
+}
+
+void __might_resched(const char *file, int line, unsigned int offsets)
 {
 	/* Ratelimiting timestamp: */
 	static unsigned long prev_jiffy;
@@ -9515,7 +9518,7 @@ void __might_resched(const char *file, int line, int preempt_offset)
 	/* WARN_ON_ONCE() by default, no rate limit required: */
 	rcu_sleep_check();
 
-	if ((preempt_count_equals(preempt_offset) && !irqs_disabled() &&
+	if ((resched_offsets_ok(offsets) && !irqs_disabled() &&
 	     !is_idle_task(current) && !current->non_block_count) ||
 	    system_state == SYSTEM_BOOTING || system_state > SYSTEM_RUNNING ||
 	    oops_in_progress)
@@ -9534,11 +9537,11 @@ void __might_resched(const char *file, int line, int preempt_offset)
 	       in_atomic(), irqs_disabled(), current->non_block_count,
 	       current->pid, current->comm);
 	pr_err("preempt_count: %x, expected: %x\n", preempt_count(),
-	       preempt_offset);
+	       offsets & MIGHT_RESCHED_PREEMPT_MASK);
 
 	if (IS_ENABLED(CONFIG_PREEMPT_RCU)) {
-		pr_err("RCU nest depth: %d, expected: 0\n",
-		       rcu_preempt_depth());
+		pr_err("RCU nest depth: %d, expected: %u\n",
+		       rcu_preempt_depth(), offsets >> MIGHT_RESCHED_RCU_SHIFT);
 	}
 
 	if (task_stack_end_corrupted(current))
@@ -9548,7 +9551,8 @@ void __might_resched(const char *file, int line, int preempt_offset)
 	if (irqs_disabled())
 		print_irqtrace_events(current);
 
-	print_preempt_disable_ip(preempt_offset, preempt_disable_ip);
+	print_preempt_disable_ip(offsets & MIGHT_RESCHED_PREEMPT_MASK,
+				 preempt_disable_ip);
 
 	dump_stack();
 	add_taint(TAINT_WARN, LOCKDEP_STILL_OK);

[tip: locking/core] sched: Make might_sleep() output less confusing

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     8d713b699e84aade6b64e241a35f22e166fc8174
Gitweb:        https://git.kernel.org/tip/8d713b699e84aade6b64e241a35f22e166fc8174
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:41 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:50 +02:00

sched: Make might_sleep() output less confusing

might_sleep() output is pretty informative, but can be confusing at times
especially with PREEMPT_RCU when the check triggers due to a voluntary
sleep inside a RCU read side critical section:

 BUG: sleeping function called from invalid context at kernel/test.c:110
 in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 Preemption disabled at: migrate_disable+0x33/0xa0

in_atomic() is 0, but it still tells that preemption was disabled at
migrate_disable(), which is completely useless because preemption is not
disabled. But the interesting information to decode the above, i.e. the RCU
nesting depth, is not printed.

That becomes even more confusing when might_sleep() is invoked from
cond_resched_lock() within a RCU read side critical section. Here the
expected preemption count is 1 and not 0.

 BUG: sleeping function called from invalid context at kernel/test.c:131
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 Preemption disabled at: test_cond_lock+0xf3/0x1c0

So in_atomic() is set, which is expected as the caller holds a spinlock,
but it's unclear why this is broken and the preempt disable IP is just
pointing at the correct place, i.e. spin_lock(), which is obviously not
helpful either.

Make that more useful in general:

 - Print preempt_count() and the expected value

and for the CONFIG_PREEMPT_RCU case:

 - Print the RCU read side critical section nesting depth

 - Print the preempt disable IP only when preempt count
   does not have the expected value.

So the might_sleep() dump from a within a preemptible RCU read side
critical section becomes:

 BUG: sleeping function called from invalid context at kernel/test.c:110
 in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 preempt_count: 0, expected: 0
 RCU nest depth: 1, expected: 0

and the cond_resched_lock() case becomes:

 BUG: sleeping function called from invalid context at kernel/test.c:141
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 415, name: kworker/u112:52
 preempt_count: 1, expected: 1
 RCU nest depth: 1, expected: 0

which makes is pretty obvious what's going on. For all other cases the
preempt disable IP is still printed as before:

 BUG: sleeping function called from invalid context at kernel/test.c: 156
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1, name: swapper/0
 preempt_count: 1, expected: 0
 RCU nest depth: 0, expected: 0
 Preemption disabled at:
 [<ffffffff82b48326>] test_might_sleep+0xbe/0xf8

 BUG: sleeping function called from invalid context at kernel/test.c: 163
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1, name: swapper/0
 preempt_count: 1, expected: 0
 RCU nest depth: 1, expected: 0
 Preemption disabled at:
 [<ffffffff82b48326>] test_might_sleep+0x1e4/0x280

This also prepares to provide a better debugging output for RT enabled
kernels and their spinlock substitutions.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165358.181022656@linutronix.de
---
 kernel/sched/core.c | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index a7c6069..0a27cb8 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9493,6 +9493,18 @@ void __might_sleep(const char *file, int line)
 }
 EXPORT_SYMBOL(__might_sleep);
 
+static void print_preempt_disable_ip(int preempt_offset, unsigned long ip)
+{
+	if (!IS_ENABLED(CONFIG_DEBUG_PREEMPT))
+		return;
+
+	if (preempt_count() == preempt_offset)
+		return;
+
+	pr_err("Preemption disabled at:");
+	print_ip_sym(KERN_ERR, ip);
+}
+
 void __might_resched(const char *file, int line, int preempt_offset)
 {
 	/* Ratelimiting timestamp: */
@@ -9521,6 +9533,13 @@ void __might_resched(const char *file, int line, int preempt_offset)
 	pr_err("in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n",
 	       in_atomic(), irqs_disabled(), current->non_block_count,
 	       current->pid, current->comm);
+	pr_err("preempt_count: %x, expected: %x\n", preempt_count(),
+	       preempt_offset);
+
+	if (IS_ENABLED(CONFIG_PREEMPT_RCU)) {
+		pr_err("RCU nest depth: %d, expected: 0\n",
+		       rcu_preempt_depth());
+	}
 
 	if (task_stack_end_corrupted(current))
 		pr_emerg("Thread overran stack, or stack corrupted\n");
@@ -9528,11 +9547,9 @@ void __might_resched(const char *file, int line, int preempt_offset)
 	debug_show_held_locks(current);
 	if (irqs_disabled())
 		print_irqtrace_events(current);
-	if (IS_ENABLED(CONFIG_DEBUG_PREEMPT)
-	    && !preempt_count_equals(preempt_offset)) {
-		pr_err("Preemption disabled at:");
-		print_ip_sym(KERN_ERR, preempt_disable_ip);
-	}
+
+	print_preempt_disable_ip(preempt_offset, preempt_disable_ip);
+
 	dump_stack();
 	add_taint(TAINT_WARN, LOCKDEP_STILL_OK);
 }

[tip: locking/core] sched: Cleanup might_sleep() printks

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     a45ed302b6e6fe5b03166321c08b4f2ad4a92a35
Gitweb:        https://git.kernel.org/tip/a45ed302b6e6fe5b03166321c08b4f2ad4a92a35
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:40 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:50 +02:00

sched: Cleanup might_sleep() printks

Convert them to pr_*(). No functional change.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165358.117496067@linutronix.de
---
 kernel/sched/core.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 2d790df..a7c6069 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9516,16 +9516,14 @@ void __might_resched(const char *file, int line, int preempt_offset)
 	/* Save this before calling printk(), since that will clobber it: */
 	preempt_disable_ip = get_preempt_disable_ip(current);
 
-	printk(KERN_ERR
-		"BUG: sleeping function called from invalid context at %s:%d\n",
-			file, line);
-	printk(KERN_ERR
-		"in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n",
-			in_atomic(), irqs_disabled(), current->non_block_count,
-			current->pid, current->comm);
+	pr_err("BUG: sleeping function called from invalid context at %s:%d\n",
+	       file, line);
+	pr_err("in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n",
+	       in_atomic(), irqs_disabled(), current->non_block_count,
+	       current->pid, current->comm);
 
 	if (task_stack_end_corrupted(current))
-		printk(KERN_EMERG "Thread overran stack, or stack corrupted\n");
+		pr_emerg("Thread overran stack, or stack corrupted\n");
 
 	debug_show_held_locks(current);
 	if (irqs_disabled())

[tip: locking/core] sched: Remove preempt_offset argument from __might_sleep()

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     42a387566c567603bafa1ec0c5b71c35cba83e86
Gitweb:        https://git.kernel.org/tip/42a387566c567603bafa1ec0c5b71c35cba83e86
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:38 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:50 +02:00

sched: Remove preempt_offset argument from __might_sleep()

All callers hand in 0 and never will hand in anything else.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165358.054321586@linutronix.de
---
 include/linux/kernel.h | 7 +++----
 kernel/sched/core.c    | 4 ++--
 mm/memory.c            | 2 +-
 3 files changed, 6 insertions(+), 7 deletions(-)

diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 5e4ae54..f95ee78 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -112,7 +112,7 @@ static __always_inline void might_resched(void)
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
 extern void __might_resched(const char *file, int line, int preempt_offset);
-extern void __might_sleep(const char *file, int line, int preempt_offset);
+extern void __might_sleep(const char *file, int line);
 extern void __cant_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_migrate(const char *file, int line);
 
@@ -129,7 +129,7 @@ extern void __cant_migrate(const char *file, int line);
  * supposed to.
  */
 # define might_sleep() \
-	do { __might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0)
+	do { __might_sleep(__FILE__, __LINE__); might_resched(); } while (0)
 /**
  * cant_sleep - annotation for functions that cannot sleep
  *
@@ -170,8 +170,7 @@ extern void __cant_migrate(const char *file, int line);
 #else
   static inline void __might_resched(const char *file, int line,
 				     int preempt_offset) { }
-  static inline void __might_sleep(const char *file, int line,
-				   int preempt_offset) { }
+static inline void __might_sleep(const char *file, int line) { }
 # define might_sleep() do { might_resched(); } while (0)
 # define cant_sleep() do { } while (0)
 # define cant_migrate()		do { } while (0)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index c3943aa..2d790df 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9475,7 +9475,7 @@ static inline int preempt_count_equals(int preempt_offset)
 	return (nested == preempt_offset);
 }
 
-void __might_sleep(const char *file, int line, int preempt_offset)
+void __might_sleep(const char *file, int line)
 {
 	unsigned int state = get_current_state();
 	/*
@@ -9489,7 +9489,7 @@ void __might_sleep(const char *file, int line, int preempt_offset)
 			(void *)current->task_state_change,
 			(void *)current->task_state_change);
 
-	__might_resched(file, line, preempt_offset);
+	__might_resched(file, line, 0);
 }
 EXPORT_SYMBOL(__might_sleep);
 
diff --git a/mm/memory.c b/mm/memory.c
index 25fc46e..1cd1792 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -5255,7 +5255,7 @@ void __might_fault(const char *file, int line)
 		return;
 	if (pagefault_disabled())
 		return;
-	__might_sleep(file, line, 0);
+	__might_sleep(file, line);
 #if defined(CONFIG_DEBUG_ATOMIC_SLEEP)
 	if (current->mm)
 		might_lock_read(&current->mm->mmap_lock);

[tip: locking/core] sched: Make cond_resched_*lock() variants consistent vs. might_sleep()

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     7b5ff4bb9adc53cfbf7ac9ba7820ccf0cd7c070a
Gitweb:        https://git.kernel.org/tip/7b5ff4bb9adc53cfbf7ac9ba7820ccf0cd7c070a
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:37 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:50 +02:00

sched: Make cond_resched_*lock() variants consistent vs. might_sleep()

Commit 3427445afd26 ("sched: Exclude cond_resched() from nested sleep
test") removed the task state check of __might_sleep() for
cond_resched_lock() because cond_resched_lock() is not a voluntary
scheduling point which blocks. It's a preemption point which requires the
lock holder to release the spin lock.

The same rationale applies to cond_resched_rwlock_read/write(), but those
were not touched.

Make it consistent and use the non-state checking __might_resched() there
as well.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165357.991262778@linutronix.de
---
 include/linux/sched.h | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index b38f002..7a989f2 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2051,14 +2051,14 @@ extern int __cond_resched_rwlock_write(rwlock_t *lock);
 	__cond_resched_lock(lock);					\
 })
 
-#define cond_resched_rwlock_read(lock) ({			\
-	__might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_read(lock);			\
+#define cond_resched_rwlock_read(lock) ({				\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
+	__cond_resched_rwlock_read(lock);				\
 })
 
-#define cond_resched_rwlock_write(lock) ({			\
-	__might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
-	__cond_resched_rwlock_write(lock);			\
+#define cond_resched_rwlock_write(lock) ({				\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
+	__cond_resched_rwlock_write(lock);				\
 })
 
 static inline void cond_resched_rcu(void)

[tip: locking/core] sched: Clean up the might_sleep() underscore zoo

[tip-bot2 for Thomas Gleixner]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     874f670e6088d3bff3972ecd44c1cb00610f9183
Gitweb:        https://git.kernel.org/tip/874f670e6088d3bff3972ecd44c1cb00610f9183
Author:        Thomas Gleixner <tglx@linutronix.de>
AuthorDate:    Thu, 23 Sep 2021 18:54:35 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Fri, 01 Oct 2021 13:57:49 +02:00

sched: Clean up the might_sleep() underscore zoo

__might_sleep() vs. ___might_sleep() is hard to distinguish. Aside of that
the three underscore variant is exposed to provide a checkpoint for
rescheduling points which are distinct from blocking points.

They are semantically a preemption point which means that scheduling is
state preserving. A real blocking operation, e.g. mutex_lock(), wait*(),
which cannot preserve a task state which is not equal to RUNNING.

While technically blocking on a "sleeping" spinlock in RT enabled kernels
falls into the voluntary scheduling category because it has to wait until
the contended spin/rw lock becomes available, the RT lock substitution code
can semantically be mapped to a voluntary preemption because the RT lock
substitution code and the scheduler are providing mechanisms to preserve
the task state and to take regular non-lock related wakeups into account.

Rename ___might_sleep() to __might_resched() to make the distinction of
these functions clear.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210923165357.928693482@linutronix.de
---
 include/linux/kernel.h       | 6 +++---
 include/linux/sched.h        | 8 ++++----
 kernel/locking/spinlock_rt.c | 6 +++---
 kernel/sched/core.c          | 6 +++---
 4 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 2776423..5e4ae54 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -111,7 +111,7 @@ static __always_inline void might_resched(void)
 #endif /* CONFIG_PREEMPT_* */
 
 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
-extern void ___might_sleep(const char *file, int line, int preempt_offset);
+extern void __might_resched(const char *file, int line, int preempt_offset);
 extern void __might_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_sleep(const char *file, int line, int preempt_offset);
 extern void __cant_migrate(const char *file, int line);
@@ -168,8 +168,8 @@ extern void __cant_migrate(const char *file, int line);
  */
 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
 #else
-  static inline void ___might_sleep(const char *file, int line,
-				   int preempt_offset) { }
+  static inline void __might_resched(const char *file, int line,
+				     int preempt_offset) { }
   static inline void __might_sleep(const char *file, int line,
 				   int preempt_offset) { }
 # define might_sleep() do { might_resched(); } while (0)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index e12b524..b38f002 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2038,7 +2038,7 @@ static inline int _cond_resched(void) { return 0; }
 #endif /* !defined(CONFIG_PREEMPTION) || defined(CONFIG_PREEMPT_DYNAMIC) */
 
 #define cond_resched() ({			\
-	___might_sleep(__FILE__, __LINE__, 0);	\
+	__might_resched(__FILE__, __LINE__, 0);	\
 	_cond_resched();			\
 })
 
@@ -2046,9 +2046,9 @@ extern int __cond_resched_lock(spinlock_t *lock);
 extern int __cond_resched_rwlock_read(rwlock_t *lock);
 extern int __cond_resched_rwlock_write(rwlock_t *lock);
 
-#define cond_resched_lock(lock) ({				\
-	___might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);\
-	__cond_resched_lock(lock);				\
+#define cond_resched_lock(lock) ({					\
+	__might_resched(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);	\
+	__cond_resched_lock(lock);					\
 })
 
 #define cond_resched_rwlock_read(lock) ({			\
diff --git a/kernel/locking/spinlock_rt.c b/kernel/locking/spinlock_rt.c
index d2912e4..c528924 100644
--- a/kernel/locking/spinlock_rt.c
+++ b/kernel/locking/spinlock_rt.c
@@ -32,7 +32,7 @@ static __always_inline void rtlock_lock(struct rt_mutex_base *rtm)
 
 static __always_inline void __rt_spin_lock(spinlock_t *lock)
 {
-	___might_sleep(__FILE__, __LINE__, 0);
+	__might_resched(__FILE__, __LINE__, 0);
 	rtlock_lock(&lock->lock);
 	rcu_read_lock();
 	migrate_disable();
@@ -210,7 +210,7 @@ EXPORT_SYMBOL(rt_write_trylock);
 
 void __sched rt_read_lock(rwlock_t *rwlock)
 {
-	___might_sleep(__FILE__, __LINE__, 0);
+	__might_resched(__FILE__, __LINE__, 0);
 	rwlock_acquire_read(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_read_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();
@@ -220,7 +220,7 @@ EXPORT_SYMBOL(rt_read_lock);
 
 void __sched rt_write_lock(rwlock_t *rwlock)
 {
-	___might_sleep(__FILE__, __LINE__, 0);
+	__might_resched(__FILE__, __LINE__, 0);
 	rwlock_acquire(&rwlock->dep_map, 0, 0, _RET_IP_);
 	rwbase_write_lock(&rwlock->rwbase, TASK_RTLOCK_WAIT);
 	rcu_read_lock();
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 1bba412..c3943aa 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9489,11 +9489,11 @@ void __might_sleep(const char *file, int line, int preempt_offset)
 			(void *)current->task_state_change,
 			(void *)current->task_state_change);
 
-	___might_sleep(file, line, preempt_offset);
+	__might_resched(file, line, preempt_offset);
 }
 EXPORT_SYMBOL(__might_sleep);
 
-void ___might_sleep(const char *file, int line, int preempt_offset)
+void __might_resched(const char *file, int line, int preempt_offset)
 {
 	/* Ratelimiting timestamp: */
 	static unsigned long prev_jiffy;
@@ -9538,7 +9538,7 @@ void ___might_sleep(const char *file, int line, int preempt_offset)
 	dump_stack();
 	add_taint(TAINT_WARN, LOCKDEP_STILL_OK);
 }
-EXPORT_SYMBOL(___might_sleep);
+EXPORT_SYMBOL(__might_resched);
 
 void __cant_sleep(const char *file, int line, int preempt_offset)
 {