[GIT PULL] locking fixes

[Ingo Molnar]

Linus,

Please pull the latest locking-urgent-for-linus git tree from:

   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git locking-urgent-for-linus

   # HEAD: 318b1dedcd39012624f466d281627553e9fa2570 locking/refcounts: Add missing kernel.h header to have UINT_MAX defined

The main change is the uninlining of large refcount_t APIs, plus a header 
dependency fix.

Note that the uninlining allowed us to enable the underflow/overflow warnings 
unconditionally and remove the debug Kconfig switch: this might trigger new 
warnings in buggy code and turn crashes/use-after-free bugs into less harmful 
memory leaks.

 Thanks,

	Ingo

------------------>
Elena Reshetova (1):
      locking/refcounts: Add missing kernel.h header to have UINT_MAX defined

Peter Zijlstra (1):
      locking/refcounts: Out-of-line everything


 include/linux/refcount.h | 278 +++--------------------------------------------
 lib/Kconfig.debug        |  13 ---
 lib/Makefile             |   2 +-
 lib/refcount.c           | 267 +++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 281 insertions(+), 279 deletions(-)
 create mode 100644 lib/refcount.c

diff --git a/include/linux/refcount.h b/include/linux/refcount.h
index 600aadf9cca4..0023fee4bbbc 100644
--- a/include/linux/refcount.h
+++ b/include/linux/refcount.h
@@ -1,54 +1,10 @@
 #ifndef _LINUX_REFCOUNT_H
 #define _LINUX_REFCOUNT_H
 
-/*
- * Variant of atomic_t specialized for reference counts.
- *
- * The interface matches the atomic_t interface (to aid in porting) but only
- * provides the few functions one should use for reference counting.
- *
- * It differs in that the counter saturates at UINT_MAX and will not move once
- * there. This avoids wrapping the counter and causing 'spurious'
- * use-after-free issues.
- *
- * Memory ordering rules are slightly relaxed wrt regular atomic_t functions
- * and provide only what is strictly required for refcounts.
- *
- * The increments are fully relaxed; these will not provide ordering. The
- * rationale is that whatever is used to obtain the object we're increasing the
- * reference count on will provide the ordering. For locked data structures,
- * its the lock acquire, for RCU/lockless data structures its the dependent
- * load.
- *
- * Do note that inc_not_zero() provides a control dependency which will order
- * future stores against the inc, this ensures we'll never modify the object
- * if we did not in fact acquire a reference.
- *
- * The decrements will provide release order, such that all the prior loads and
- * stores will be issued before, it also provides a control dependency, which
- * will order us against the subsequent free().
- *
- * The control dependency is against the load of the cmpxchg (ll/sc) that
- * succeeded. This means the stores aren't fully ordered, but this is fine
- * because the 1->0 transition indicates no concurrency.
- *
- * Note that the allocator is responsible for ordering things between free()
- * and alloc().
- *
- */
-
 #include <linux/atomic.h>
-#include <linux/bug.h>
 #include <linux/mutex.h>
 #include <linux/spinlock.h>
-
-#ifdef CONFIG_DEBUG_REFCOUNT
-#define REFCOUNT_WARN(cond, str) WARN_ON(cond)
-#define __refcount_check	__must_check
-#else
-#define REFCOUNT_WARN(cond, str) (void)(cond)
-#define __refcount_check
-#endif
+#include <linux/kernel.h>
 
 typedef struct refcount_struct {
 	atomic_t refs;
@@ -66,229 +22,21 @@ static inline unsigned int refcount_read(const refcount_t *r)
 	return atomic_read(&r->refs);
 }
 
-static inline __refcount_check
-bool refcount_add_not_zero(unsigned int i, refcount_t *r)
-{
-	unsigned int old, new, val = atomic_read(&r->refs);
-
-	for (;;) {
-		if (!val)
-			return false;
-
-		if (unlikely(val == UINT_MAX))
-			return true;
-
-		new = val + i;
-		if (new < val)
-			new = UINT_MAX;
-		old = atomic_cmpxchg_relaxed(&r->refs, val, new);
-		if (old == val)
-			break;
-
-		val = old;
-	}
-
-	REFCOUNT_WARN(new == UINT_MAX, "refcount_t: saturated; leaking memory.\n");
-
-	return true;
-}
-
-static inline void refcount_add(unsigned int i, refcount_t *r)
-{
-	REFCOUNT_WARN(!refcount_add_not_zero(i, r), "refcount_t: addition on 0; use-after-free.\n");
-}
-
-/*
- * Similar to atomic_inc_not_zero(), will saturate at UINT_MAX and WARN.
- *
- * Provides no memory ordering, it is assumed the caller has guaranteed the
- * object memory to be stable (RCU, etc.). It does provide a control dependency
- * and thereby orders future stores. See the comment on top.
- */
-static inline __refcount_check
-bool refcount_inc_not_zero(refcount_t *r)
-{
-	unsigned int old, new, val = atomic_read(&r->refs);
-
-	for (;;) {
-		new = val + 1;
-
-		if (!val)
-			return false;
-
-		if (unlikely(!new))
-			return true;
-
-		old = atomic_cmpxchg_relaxed(&r->refs, val, new);
-		if (old == val)
-			break;
-
-		val = old;
-	}
-
-	REFCOUNT_WARN(new == UINT_MAX, "refcount_t: saturated; leaking memory.\n");
-
-	return true;
-}
-
-/*
- * Similar to atomic_inc(), will saturate at UINT_MAX and WARN.
- *
- * Provides no memory ordering, it is assumed the caller already has a
- * reference on the object, will WARN when this is not so.
- */
-static inline void refcount_inc(refcount_t *r)
-{
-	REFCOUNT_WARN(!refcount_inc_not_zero(r), "refcount_t: increment on 0; use-after-free.\n");
-}
-
-/*
- * Similar to atomic_dec_and_test(), it will WARN on underflow and fail to
- * decrement when saturated at UINT_MAX.
- *
- * Provides release memory ordering, such that prior loads and stores are done
- * before, and provides a control dependency such that free() must come after.
- * See the comment on top.
- */
-static inline __refcount_check
-bool refcount_sub_and_test(unsigned int i, refcount_t *r)
-{
-	unsigned int old, new, val = atomic_read(&r->refs);
-
-	for (;;) {
-		if (unlikely(val == UINT_MAX))
-			return false;
-
-		new = val - i;
-		if (new > val) {
-			REFCOUNT_WARN(new > val, "refcount_t: underflow; use-after-free.\n");
-			return false;
-		}
-
-		old = atomic_cmpxchg_release(&r->refs, val, new);
-		if (old == val)
-			break;
-
-		val = old;
-	}
-
-	return !new;
-}
-
-static inline __refcount_check
-bool refcount_dec_and_test(refcount_t *r)
-{
-	return refcount_sub_and_test(1, r);
-}
+extern __must_check bool refcount_add_not_zero(unsigned int i, refcount_t *r);
+extern void refcount_add(unsigned int i, refcount_t *r);
 
-/*
- * Similar to atomic_dec(), it will WARN on underflow and fail to decrement
- * when saturated at UINT_MAX.
- *
- * Provides release memory ordering, such that prior loads and stores are done
- * before.
- */
-static inline
-void refcount_dec(refcount_t *r)
-{
-	REFCOUNT_WARN(refcount_dec_and_test(r), "refcount_t: decrement hit 0; leaking memory.\n");
-}
-
-/*
- * No atomic_t counterpart, it attempts a 1 -> 0 transition and returns the
- * success thereof.
- *
- * Like all decrement operations, it provides release memory order and provides
- * a control dependency.
- *
- * It can be used like a try-delete operator; this explicit case is provided
- * and not cmpxchg in generic, because that would allow implementing unsafe
- * operations.
- */
-static inline __refcount_check
-bool refcount_dec_if_one(refcount_t *r)
-{
-	return atomic_cmpxchg_release(&r->refs, 1, 0) == 1;
-}
-
-/*
- * No atomic_t counterpart, it decrements unless the value is 1, in which case
- * it will return false.
- *
- * Was often done like: atomic_add_unless(&var, -1, 1)
- */
-static inline __refcount_check
-bool refcount_dec_not_one(refcount_t *r)
-{
-	unsigned int old, new, val = atomic_read(&r->refs);
+extern __must_check bool refcount_inc_not_zero(refcount_t *r);
+extern void refcount_inc(refcount_t *r);
 
-	for (;;) {
-		if (unlikely(val == UINT_MAX))
-			return true;
+extern __must_check bool refcount_sub_and_test(unsigned int i, refcount_t *r);
+extern void refcount_sub(unsigned int i, refcount_t *r);
 
-		if (val == 1)
-			return false;
+extern __must_check bool refcount_dec_and_test(refcount_t *r);
+extern void refcount_dec(refcount_t *r);
 
-		new = val - 1;
-		if (new > val) {
-			REFCOUNT_WARN(new > val, "refcount_t: underflow; use-after-free.\n");
-			return true;
-		}
-
-		old = atomic_cmpxchg_release(&r->refs, val, new);
-		if (old == val)
-			break;
-
-		val = old;
-	}
-
-	return true;
-}
-
-/*
- * Similar to atomic_dec_and_mutex_lock(), it will WARN on underflow and fail
- * to decrement when saturated at UINT_MAX.
- *
- * Provides release memory ordering, such that prior loads and stores are done
- * before, and provides a control dependency such that free() must come after.
- * See the comment on top.
- */
-static inline __refcount_check
-bool refcount_dec_and_mutex_lock(refcount_t *r, struct mutex *lock)
-{
-	if (refcount_dec_not_one(r))
-		return false;
-
-	mutex_lock(lock);
-	if (!refcount_dec_and_test(r)) {
-		mutex_unlock(lock);
-		return false;
-	}
-
-	return true;
-}
-
-/*
- * Similar to atomic_dec_and_lock(), it will WARN on underflow and fail to
- * decrement when saturated at UINT_MAX.
- *
- * Provides release memory ordering, such that prior loads and stores are done
- * before, and provides a control dependency such that free() must come after.
- * See the comment on top.
- */
-static inline __refcount_check
-bool refcount_dec_and_lock(refcount_t *r, spinlock_t *lock)
-{
-	if (refcount_dec_not_one(r))
-		return false;
-
-	spin_lock(lock);
-	if (!refcount_dec_and_test(r)) {
-		spin_unlock(lock);
-		return false;
-	}
-
-	return true;
-}
+extern __must_check bool refcount_dec_if_one(refcount_t *r);
+extern __must_check bool refcount_dec_not_one(refcount_t *r);
+extern __must_check bool refcount_dec_and_mutex_lock(refcount_t *r, struct mutex *lock);
+extern __must_check bool refcount_dec_and_lock(refcount_t *r, spinlock_t *lock);
 
 #endif /* _LINUX_REFCOUNT_H */
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index acedbe626d47..0dbce99d8433 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -716,19 +716,6 @@ source "lib/Kconfig.kmemcheck"
 
 source "lib/Kconfig.kasan"
 
-config DEBUG_REFCOUNT
-	bool "Verbose refcount checks"
-	help
-	  Say Y here if you want reference counters (refcount_t and kref) to
-	  generate WARNs on dubious usage. Without this refcount_t will still
-	  be a saturating counter and avoid Use-After-Free by turning it into
-	  a resource leak Denial-Of-Service.
-
-	  Use of this option will increase kernel text size but will alert the
-	  admin of potential abuse.
-
-	  If in doubt, say "N".
-
 endmenu # "Memory Debugging"
 
 config ARCH_HAS_KCOV
diff --git a/lib/Makefile b/lib/Makefile
index 19ea76149a37..192e4d03caf9 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -36,7 +36,7 @@ obj-y += bcd.o div64.o sort.o parser.o debug_locks.o random32.o \
 	 gcd.o lcm.o list_sort.o uuid.o flex_array.o iov_iter.o clz_ctz.o \
 	 bsearch.o find_bit.o llist.o memweight.o kfifo.o \
 	 percpu-refcount.o percpu_ida.o rhashtable.o reciprocal_div.o \
-	 once.o
+	 once.o refcount.o
 obj-y += string_helpers.o
 obj-$(CONFIG_TEST_STRING_HELPERS) += test-string_helpers.o
 obj-y += hexdump.o
diff --git a/lib/refcount.c b/lib/refcount.c
new file mode 100644
index 000000000000..1d33366189d1
--- /dev/null
+++ b/lib/refcount.c
@@ -0,0 +1,267 @@
+/*
+ * Variant of atomic_t specialized for reference counts.
+ *
+ * The interface matches the atomic_t interface (to aid in porting) but only
+ * provides the few functions one should use for reference counting.
+ *
+ * It differs in that the counter saturates at UINT_MAX and will not move once
+ * there. This avoids wrapping the counter and causing 'spurious'
+ * use-after-free issues.
+ *
+ * Memory ordering rules are slightly relaxed wrt regular atomic_t functions
+ * and provide only what is strictly required for refcounts.
+ *
+ * The increments are fully relaxed; these will not provide ordering. The
+ * rationale is that whatever is used to obtain the object we're increasing the
+ * reference count on will provide the ordering. For locked data structures,
+ * its the lock acquire, for RCU/lockless data structures its the dependent
+ * load.
+ *
+ * Do note that inc_not_zero() provides a control dependency which will order
+ * future stores against the inc, this ensures we'll never modify the object
+ * if we did not in fact acquire a reference.
+ *
+ * The decrements will provide release order, such that all the prior loads and
+ * stores will be issued before, it also provides a control dependency, which
+ * will order us against the subsequent free().
+ *
+ * The control dependency is against the load of the cmpxchg (ll/sc) that
+ * succeeded. This means the stores aren't fully ordered, but this is fine
+ * because the 1->0 transition indicates no concurrency.
+ *
+ * Note that the allocator is responsible for ordering things between free()
+ * and alloc().
+ *
+ */
+
+#include <linux/refcount.h>
+#include <linux/bug.h>
+
+bool refcount_add_not_zero(unsigned int i, refcount_t *r)
+{
+	unsigned int old, new, val = atomic_read(&r->refs);
+
+	for (;;) {
+		if (!val)
+			return false;
+
+		if (unlikely(val == UINT_MAX))
+			return true;
+
+		new = val + i;
+		if (new < val)
+			new = UINT_MAX;
+		old = atomic_cmpxchg_relaxed(&r->refs, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+
+	WARN(new == UINT_MAX, "refcount_t: saturated; leaking memory.\n");
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(refcount_add_not_zero);
+
+void refcount_add(unsigned int i, refcount_t *r)
+{
+	WARN(!refcount_add_not_zero(i, r), "refcount_t: addition on 0; use-after-free.\n");
+}
+EXPORT_SYMBOL_GPL(refcount_add);
+
+/*
+ * Similar to atomic_inc_not_zero(), will saturate at UINT_MAX and WARN.
+ *
+ * Provides no memory ordering, it is assumed the caller has guaranteed the
+ * object memory to be stable (RCU, etc.). It does provide a control dependency
+ * and thereby orders future stores. See the comment on top.
+ */
+bool refcount_inc_not_zero(refcount_t *r)
+{
+	unsigned int old, new, val = atomic_read(&r->refs);
+
+	for (;;) {
+		new = val + 1;
+
+		if (!val)
+			return false;
+
+		if (unlikely(!new))
+			return true;
+
+		old = atomic_cmpxchg_relaxed(&r->refs, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+
+	WARN(new == UINT_MAX, "refcount_t: saturated; leaking memory.\n");
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(refcount_inc_not_zero);
+
+/*
+ * Similar to atomic_inc(), will saturate at UINT_MAX and WARN.
+ *
+ * Provides no memory ordering, it is assumed the caller already has a
+ * reference on the object, will WARN when this is not so.
+ */
+void refcount_inc(refcount_t *r)
+{
+	WARN(!refcount_inc_not_zero(r), "refcount_t: increment on 0; use-after-free.\n");
+}
+EXPORT_SYMBOL_GPL(refcount_inc);
+
+bool refcount_sub_and_test(unsigned int i, refcount_t *r)
+{
+	unsigned int old, new, val = atomic_read(&r->refs);
+
+	for (;;) {
+		if (unlikely(val == UINT_MAX))
+			return false;
+
+		new = val - i;
+		if (new > val) {
+			WARN(new > val, "refcount_t: underflow; use-after-free.\n");
+			return false;
+		}
+
+		old = atomic_cmpxchg_release(&r->refs, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+
+	return !new;
+}
+EXPORT_SYMBOL_GPL(refcount_sub_and_test);
+
+/*
+ * Similar to atomic_dec_and_test(), it will WARN on underflow and fail to
+ * decrement when saturated at UINT_MAX.
+ *
+ * Provides release memory ordering, such that prior loads and stores are done
+ * before, and provides a control dependency such that free() must come after.
+ * See the comment on top.
+ */
+bool refcount_dec_and_test(refcount_t *r)
+{
+	return refcount_sub_and_test(1, r);
+}
+EXPORT_SYMBOL_GPL(refcount_dec_and_test);
+
+/*
+ * Similar to atomic_dec(), it will WARN on underflow and fail to decrement
+ * when saturated at UINT_MAX.
+ *
+ * Provides release memory ordering, such that prior loads and stores are done
+ * before.
+ */
+
+void refcount_dec(refcount_t *r)
+{
+	WARN(refcount_dec_and_test(r), "refcount_t: decrement hit 0; leaking memory.\n");
+}
+EXPORT_SYMBOL_GPL(refcount_dec);
+
+/*
+ * No atomic_t counterpart, it attempts a 1 -> 0 transition and returns the
+ * success thereof.
+ *
+ * Like all decrement operations, it provides release memory order and provides
+ * a control dependency.
+ *
+ * It can be used like a try-delete operator; this explicit case is provided
+ * and not cmpxchg in generic, because that would allow implementing unsafe
+ * operations.
+ */
+bool refcount_dec_if_one(refcount_t *r)
+{
+	return atomic_cmpxchg_release(&r->refs, 1, 0) == 1;
+}
+EXPORT_SYMBOL_GPL(refcount_dec_if_one);
+
+/*
+ * No atomic_t counterpart, it decrements unless the value is 1, in which case
+ * it will return false.
+ *
+ * Was often done like: atomic_add_unless(&var, -1, 1)
+ */
+bool refcount_dec_not_one(refcount_t *r)
+{
+	unsigned int old, new, val = atomic_read(&r->refs);
+
+	for (;;) {
+		if (unlikely(val == UINT_MAX))
+			return true;
+
+		if (val == 1)
+			return false;
+
+		new = val - 1;
+		if (new > val) {
+			WARN(new > val, "refcount_t: underflow; use-after-free.\n");
+			return true;
+		}
+
+		old = atomic_cmpxchg_release(&r->refs, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(refcount_dec_not_one);
+
+/*
+ * Similar to atomic_dec_and_mutex_lock(), it will WARN on underflow and fail
+ * to decrement when saturated at UINT_MAX.
+ *
+ * Provides release memory ordering, such that prior loads and stores are done
+ * before, and provides a control dependency such that free() must come after.
+ * See the comment on top.
+ */
+bool refcount_dec_and_mutex_lock(refcount_t *r, struct mutex *lock)
+{
+	if (refcount_dec_not_one(r))
+		return false;
+
+	mutex_lock(lock);
+	if (!refcount_dec_and_test(r)) {
+		mutex_unlock(lock);
+		return false;
+	}
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(refcount_dec_and_mutex_lock);
+
+/*
+ * Similar to atomic_dec_and_lock(), it will WARN on underflow and fail to
+ * decrement when saturated at UINT_MAX.
+ *
+ * Provides release memory ordering, such that prior loads and stores are done
+ * before, and provides a control dependency such that free() must come after.
+ * See the comment on top.
+ */
+bool refcount_dec_and_lock(refcount_t *r, spinlock_t *lock)
+{
+	if (refcount_dec_not_one(r))
+		return false;
+
+	spin_lock(lock);
+	if (!refcount_dec_and_test(r)) {
+		spin_unlock(lock);
+		return false;
+	}
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(refcount_dec_and_lock);
+