bcachefs: Fix race between trans_put() and btree_transactions_read()

debug.c was using closure_get() on a different thread's closure where
the we don't know if the object being refcounted is alive.

We keep btree_trans objects on a list so they can be printed by debug
code, and because it is cost prohibitive to touch the btree_trans list
every time we allocate and free btree_trans objects, cached objects are
also on this list.

However, we do not want the debug code to see cached but not in use
btree_trans objects - critically because the btree_paths array will have
been freed (if it was reallocated).

closure_get() is also incorrect to use when that get may race with it
hitting zero, i.e. we must already have a ref on the object or know the
ref can't currently hit 0 for other reasons (as used in the cycle
detector).

to fix this, use the previously introduced closure_get_not_zero(),
closure_return_sync(), and closure_init_stack_release(); the debug code
now can only take a ref on a trans object if it's alive and in use.

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

1 files changed, 4 insertions, 6 deletions

diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c
index 3a1419d17888..15c1c7cfefe6 100644
--- a/fs/bcachefs/btree_iter.c
+++ b/fs/bcachefs/btree_iter.c
@@ -3130,7 +3130,6 @@ struct btree_trans *__bch2_trans_get(struct bch_fs *c, unsigned fn_idx)
 
 	trans = mempool_alloc(&c->btree_trans_pool, GFP_NOFS);
 	memset(trans, 0, sizeof(*trans));
-	closure_init_stack(&trans->ref);
 
 	seqmutex_lock(&c->btree_trans_lock);
 	if (IS_ENABLED(CONFIG_BCACHEFS_DEBUG)) {
@@ -3161,7 +3160,6 @@ struct btree_trans *__bch2_trans_get(struct bch_fs *c, unsigned fn_idx)
 list_add_done:
 	seqmutex_unlock(&c->btree_trans_lock);
 got_trans:
-	trans->ref.closure_get_happened = false;
 	trans->c		= c;
 	trans->last_begin_time	= local_clock();
 	trans->fn_idx		= fn_idx;
@@ -3200,6 +3198,8 @@ got_trans:
 	trans->srcu_idx		= srcu_read_lock(&c->btree_trans_barrier);
 	trans->srcu_lock_time	= jiffies;
 	trans->srcu_held	= true;
+
+	closure_init_stack_release(&trans->ref);
 	return trans;
 }
 
@@ -3257,10 +3257,10 @@ void bch2_trans_put(struct btree_trans *trans)
 		bch2_journal_keys_put(c);
 
 	/*
-	 * trans->ref protects trans->locking_wait.task, btree_paths arary; used
+	 * trans->ref protects trans->locking_wait.task, btree_paths array; used
 	 * by cycle detector
 	 */
-	closure_sync(&trans->ref);
+	closure_return_sync(&trans->ref);
 	trans->locking_wait.task = NULL;
 
 	unsigned long *paths_allocated = trans->paths_allocated;
@@ -3385,8 +3385,6 @@ void bch2_fs_btree_iter_exit(struct bch_fs *c)
 				per_cpu_ptr(c->btree_trans_bufs, cpu)->trans;
 
 			if (trans) {
-				closure_sync(&trans->ref);
-
 				seqmutex_lock(&c->btree_trans_lock);
 				list_del(&trans->list);
 				seqmutex_unlock(&c->btree_trans_lock);