This adds a method to tell lockdep not to check lock ordering within a
lock class - but to still check lock ordering w.r.t. other lock types.

This is for bcachefs, where for btree node locks we have our own
deadlock avoidance strategy w.r.t. other btree node locks (cycle
detection), but we still want lockdep to check lock ordering w.r.t.
other lock types.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Will Deacon <will@kernel.org>
Cc: Waiman Long <longman@redhat.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
---
 include/linux/lockdep.h       |  6 ++++++
 include/linux/lockdep_types.h |  2 +-
 kernel/locking/lockdep.c      | 26 ++++++++++++++++++++++++++
 3 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
index fc86557d2a21..1bc7eb7cb548 100644
--- a/include/linux/lockdep.h
+++ b/include/linux/lockdep.h
@@ -700,4 +700,10 @@ lockdep_rcu_suspicious(const char *file, const int line, const char *s)
 }
 #endif
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+void lockdep_set_no_check_recursion(struct lockdep_map *map);
+#else
+static inline void lockdep_set_no_check_recursion(struct lockdep_map *map) {}
+#endif
+
 #endif /* __LINUX_LOCKDEP_H */
diff --git a/include/linux/lockdep_types.h b/include/linux/lockdep_types.h
index 2ebc323d345a..aa6bddac2a64 100644
--- a/include/linux/lockdep_types.h
+++ b/include/linux/lockdep_types.h
@@ -137,7 +137,7 @@ struct lock_class {
 	u8				wait_type_inner;
 	u8				wait_type_outer;
 	u8				lock_type;
-	/* u8				hole; */
+	u8				no_check_recursion;
 
 #ifdef CONFIG_LOCK_STAT
 	unsigned long			contention_point[LOCKSTAT_POINTS];
diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c
index 38924d90da85..50a0a1ea960a 100644
--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3048,6 +3048,9 @@ check_deadlock(struct task_struct *curr, struct held_lock *next)
 
 		class = hlock_class(prev);
 
+		if (class->no_check_recursion)
+			continue;
+
 		if (class->cmp_fn &&
 		    class->cmp_fn(prev->instance, next->instance) < 0)
 			continue;
@@ -3113,6 +3116,10 @@ check_prev_add(struct task_struct *curr, struct held_lock *prev,
 		return 2;
 	}
 
+	if (hlock_class(prev) == hlock_class(next) &&
+	    hlock_class(prev)->no_check_recursion)
+		return 2;
+
 	if (prev->class_idx == next->class_idx) {
 		struct lock_class *class = hlock_class(prev);
 
@@ -6732,3 +6739,22 @@ void lockdep_rcu_suspicious(const char *file, const int line, const char *s)
 	warn_rcu_exit(rcu);
 }
 EXPORT_SYMBOL_GPL(lockdep_rcu_suspicious);
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+void lockdep_set_no_check_recursion(struct lockdep_map *lock)
+{
+	struct lock_class *class = lock->class_cache[0];
+	unsigned long flags;
+
+	raw_local_irq_save(flags);
+	lockdep_recursion_inc();
+
+	if (!class)
+		class = register_lock_class(lock, 0, 0);
+	if (class)
+		class->no_check_recursion = true;
+	lockdep_recursion_finish();
+	raw_local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(lockdep_set_no_check_recursion);
+#endif
-- 
2.42.0

On Wed, Nov 22, 2023 at 06:51:09PM -0500, Kent Overstreet wrote:
> This adds a method to tell lockdep not to check lock ordering within a
> lock class - but to still check lock ordering w.r.t. other lock types.
> 
> This is for bcachefs, where for btree node locks we have our own
> deadlock avoidance strategy w.r.t. other btree node locks (cycle
> detection), but we still want lockdep to check lock ordering w.r.t.
> other lock types.

So earlier you added custom sort order.

Additionally there is the wound-wait mutexes that also have semantics
similar to what you describe.

Explain why you can't use either your own added feature or the existing
infrastructure to solve this?

On Fri, Nov 24, 2023 at 10:58:04AM +0100, Peter Zijlstra wrote:
> On Wed, Nov 22, 2023 at 06:51:09PM -0500, Kent Overstreet wrote:
> > This adds a method to tell lockdep not to check lock ordering within a
> > lock class - but to still check lock ordering w.r.t. other lock types.
> > 
> > This is for bcachefs, where for btree node locks we have our own
> > deadlock avoidance strategy w.r.t. other btree node locks (cycle
> > detection), but we still want lockdep to check lock ordering w.r.t.
> > other lock types.
> 
> So earlier you added custom sort order.

That was never merged? I've been meaning to revisit that though and
convert some other things to it, the original patchset just converted
bcache, not bcachefs.

But it's not applicable here, bcachefs doesn't have lock ordering, it
has cycle detection.

> Additionally there is the wound-wait mutexes that also have semantics
> similar to what you describe.

Yeah, we talked about that as well. Wait/wound is similar in principle
but aborts much more frequently, since it's just comparing transaction
start time and not doing full cycle detection.