bcachefs: Add lockdep support for btree node locks

This adds lockdep tracking for held btree locks with a single dep_map in
btree_trans, i.e. tracking all held btree locks as one object.

This is more practical and more useful than having lockdep track held
btree locks individually, because
 - we can take more locks than lockdep can track (unbounded, now that we
   have dynamically resizable btree paths)
 - there's no lock ordering between btree locks for lockdep to track (we
   do cycle detection)
 - and this makes it easy to teach lockdep that btree locks are not safe
   to hold while invoking memory reclaim.

The last rule is one that lockdep would never learn, because we only do
trylock() from within shrinkers - but we very much do not want to be
invoking memory reclaim while holding btree node locks.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c
index 9485208b6758bf3c9d21f491a3f20417d9042f60..803cc58ff577e7a5655e8c8f4ad737aad1519046 100644 (file)
--- a/fs/bcachefs/btree_iter.c
+++ b/fs/bcachefs/btree_iter.c
@@ -3171,6 +3171,9 @@ got_trans:
 
        trans->paths_allocated[0] = 1;
 
+       static struct lock_class_key lockdep_key;
+       lockdep_init_map(&trans->dep_map, "bcachefs_btree", &lockdep_key, 0);
+
        if (fn_idx < BCH_TRANSACTIONS_NR) {
                trans->fn = bch2_btree_transaction_fns[fn_idx];
 
@@ -3440,7 +3443,22 @@ int bch2_fs_btree_iter_init(struct bch_fs *c)
                mempool_init_kmalloc_pool(&c->btree_trans_mem_pool, 1,
                                          BTREE_TRANS_MEM_MAX) ?:
                init_srcu_struct(&c->btree_trans_barrier);
-       if (!ret)
-               c->btree_trans_barrier_initialized = true;
-       return ret;
+       if (ret)
+               return ret;
+
+       /*
+        * static annotation (hackily done) for lock ordering of reclaim vs.
+        * btree node locks:
+        */
+#ifdef CONFIG_LOCKDEP
+       fs_reclaim_acquire(GFP_KERNEL);
+       struct btree_trans *trans = bch2_trans_get(c);
+       trans_set_locked(trans);
+       bch2_trans_put(trans);
+       fs_reclaim_release(GFP_KERNEL);
+#endif
+
+       c->btree_trans_barrier_initialized = true;
+       return 0;
+
 }

diff --git a/fs/bcachefs/btree_locking.h b/fs/bcachefs/btree_locking.h
index 8f5f1973c7d8e51f2b56c8bd7ff83d39d39fbf8f..8dbceec8ec25753749f3e6a57b7dc3a41c65d008 100644 (file)
--- a/fs/bcachefs/btree_locking.h
+++ b/fs/bcachefs/btree_locking.h
@@ -197,6 +197,7 @@ int bch2_six_check_for_deadlock(struct six_lock *lock, void *p);
 static inline void trans_set_locked(struct btree_trans *trans)
 {
        if (!trans->locked) {
+               lock_acquire_exclusive(&trans->dep_map, 0, 0, NULL, _THIS_IP_);
                trans->locked = true;
                trans->last_unlock_ip = 0;
 
@@ -208,6 +209,7 @@ static inline void trans_set_locked(struct btree_trans *trans)
 static inline void trans_set_unlocked(struct btree_trans *trans)
 {
        if (trans->locked) {
+               lock_release(&trans->dep_map, _THIS_IP_);
                trans->locked = false;
                trans->last_unlock_ip = _RET_IP_;
 

diff --git a/fs/bcachefs/btree_types.h b/fs/bcachefs/btree_types.h
index 4fe77d7f7242bac4935746f53aeec2ca18613ae2..79898f68777264dcc4bad6ad684cac5e48dc525c 100644 (file)
--- a/fs/bcachefs/btree_types.h
+++ b/fs/bcachefs/btree_types.h
@@ -523,6 +523,9 @@ struct btree_trans {
        unsigned                journal_u64s;
        unsigned                extra_disk_res; /* XXX kill */
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+       struct lockdep_map      dep_map;
+#endif
        /* Entries before this are zeroed out on every bch2_trans_get() call */
 
        struct list_head        list;