26 files changed, 154 insertions, 66 deletions

diff --git a/fs/bcachefs/alloc_foreground.c b/fs/bcachefs/alloc_foreground.c
index 1ba0eeb7552a..0e6157982607 100644
--- a/fs/bcachefs/alloc_foreground.c
+++ b/fs/bcachefs/alloc_foreground.c
@@ -1374,8 +1374,17 @@ retry:
 			goto alloc_done;
 
 		/* Don't retry from all devices if we're out of open buckets: */
-		if (bch2_err_matches(ret, BCH_ERR_open_buckets_empty))
-			goto allocate_blocking;
+		if (bch2_err_matches(ret, BCH_ERR_open_buckets_empty)) {
+			int ret = open_bucket_add_buckets(trans, &ptrs, wp, devs_have,
+					      target, erasure_code,
+					      nr_replicas, &nr_effective,
+					      &have_cache, watermark,
+					      flags, cl);
+			if (!ret ||
+			    bch2_err_matches(ret, BCH_ERR_transaction_restart) ||
+			    bch2_err_matches(ret, BCH_ERR_open_buckets_empty))
+				goto alloc_done;
+		}
 
 		/*
 		 * Only try to allocate cache (durability = 0 devices) from the
@@ -1389,7 +1398,6 @@ retry:
 					      &have_cache, watermark,
 					      flags, cl);
 	} else {
-allocate_blocking:
 		ret = open_bucket_add_buckets(trans, &ptrs, wp, devs_have,
 					      target, erasure_code,
 					      nr_replicas, &nr_effective,
diff --git a/fs/bcachefs/btree_cache.c b/fs/bcachefs/btree_cache.c
index 47e7770d0583..79495cd7a794 100644
--- a/fs/bcachefs/btree_cache.c
+++ b/fs/bcachefs/btree_cache.c
@@ -9,6 +9,7 @@
 #include "debug.h"
 #include "errcode.h"
 #include "error.h"
+#include "journal.h"
 #include "trace.h"
 
 #include <linux/prefetch.h>
@@ -424,14 +425,11 @@ void bch2_fs_btree_cache_exit(struct bch_fs *c)
 		BUG_ON(btree_node_read_in_flight(b) ||
 		       btree_node_write_in_flight(b));
 
-		if (btree_node_dirty(b))
-			bch2_btree_complete_write(c, b, btree_current_write(b));
-		clear_btree_node_dirty_acct(c, b);
-
 		btree_node_data_free(c, b);
 	}
 
-	BUG_ON(atomic_read(&c->btree_cache.dirty));
+	BUG_ON(!bch2_journal_error(&c->journal) &&
+	       atomic_read(&c->btree_cache.dirty));
 
 	list_splice(&bc->freed_pcpu, &bc->freed_nonpcpu);
 
diff --git a/fs/bcachefs/btree_io.c b/fs/bcachefs/btree_io.c
index 57c20390e10e..5a720f0cd5a6 100644
--- a/fs/bcachefs/btree_io.c
+++ b/fs/bcachefs/btree_io.c
@@ -1704,8 +1704,8 @@ int bch2_btree_root_read(struct bch_fs *c, enum btree_id id,
 	return bch2_trans_run(c, __bch2_btree_root_read(trans, id, k, level));
 }
 
-void bch2_btree_complete_write(struct bch_fs *c, struct btree *b,
-			      struct btree_write *w)
+static void bch2_btree_complete_write(struct bch_fs *c, struct btree *b,
+				      struct btree_write *w)
 {
 	unsigned long old, new, v = READ_ONCE(b->will_make_reachable);
 
diff --git a/fs/bcachefs/btree_io.h b/fs/bcachefs/btree_io.h
index 7e03dd76fb38..e0d7fa5b1dfb 100644
--- a/fs/bcachefs/btree_io.h
+++ b/fs/bcachefs/btree_io.h
@@ -134,9 +134,6 @@ void bch2_btree_node_read(struct bch_fs *, struct btree *, bool);
 int bch2_btree_root_read(struct bch_fs *, enum btree_id,
 			 const struct bkey_i *, unsigned);
 
-void bch2_btree_complete_write(struct bch_fs *, struct btree *,
-			      struct btree_write *);
-
 bool bch2_btree_post_write_cleanup(struct bch_fs *, struct btree *);
 
 enum btree_write_flags {
diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c
index 8e0fe65f6101..6be79129738d 100644
--- a/fs/bcachefs/btree_iter.c
+++ b/fs/bcachefs/btree_iter.c
@@ -3214,10 +3214,9 @@ void bch2_fs_btree_iter_exit(struct bch_fs *c)
 	mempool_exit(&c->btree_trans_pool);
 }
 
-int bch2_fs_btree_iter_init(struct bch_fs *c)
+void bch2_fs_btree_iter_init_early(struct bch_fs *c)
 {
 	struct btree_transaction_stats *s;
-	int ret;
 
 	for (s = c->btree_transaction_stats;
 	     s < c->btree_transaction_stats + ARRAY_SIZE(c->btree_transaction_stats);
@@ -3228,6 +3227,11 @@ int bch2_fs_btree_iter_init(struct bch_fs *c)
 
 	INIT_LIST_HEAD(&c->btree_trans_list);
 	seqmutex_init(&c->btree_trans_lock);
+}
+
+int bch2_fs_btree_iter_init(struct bch_fs *c)
+{
+	int ret;
 
 	c->btree_trans_bufs = alloc_percpu(struct btree_trans_buf);
 	if (!c->btree_trans_bufs)
diff --git a/fs/bcachefs/btree_iter.h b/fs/bcachefs/btree_iter.h
index 85e7cb52f6b6..eaffced4c132 100644
--- a/fs/bcachefs/btree_iter.h
+++ b/fs/bcachefs/btree_iter.h
@@ -938,6 +938,7 @@ unsigned bch2_trans_get_fn_idx(const char *);
 void bch2_btree_trans_to_text(struct printbuf *, struct btree_trans *);
 
 void bch2_fs_btree_iter_exit(struct bch_fs *);
+void bch2_fs_btree_iter_init_early(struct bch_fs *);
 int bch2_fs_btree_iter_init(struct bch_fs *);
 
 #endif /* _BCACHEFS_BTREE_ITER_H */
diff --git a/fs/bcachefs/btree_key_cache.c b/fs/bcachefs/btree_key_cache.c
index 37fbf22de8fc..1b7a5668df7c 100644
--- a/fs/bcachefs/btree_key_cache.c
+++ b/fs/bcachefs/btree_key_cache.c
@@ -992,8 +992,6 @@ void bch2_fs_btree_key_cache_exit(struct btree_key_cache *bc)
 	list_for_each_entry_safe(ck, n, &items, list) {
 		cond_resched();
 
-		bch2_journal_pin_drop(&c->journal, &ck->journal);
-
 		list_del(&ck->list);
 		kfree(ck->k);
 		six_lock_exit(&ck->c.lock);
diff --git a/fs/bcachefs/btree_update.c b/fs/bcachefs/btree_update.c
index 324767c0ddcc..25fdca00bf7b 100644
--- a/fs/bcachefs/btree_update.c
+++ b/fs/bcachefs/btree_update.c
@@ -554,6 +554,19 @@ int __must_check bch2_trans_update_seq(struct btree_trans *trans, u64 seq,
 						 BTREE_UPDATE_PREJOURNAL);
 }
 
+static noinline int bch2_btree_insert_clone_trans(struct btree_trans *trans,
+						  enum btree_id btree,
+						  struct bkey_i *k)
+{
+	struct bkey_i *n = bch2_trans_kmalloc(trans, bkey_bytes(&k->k));
+	int ret = PTR_ERR_OR_ZERO(n);
+	if (ret)
+		return ret;
+
+	bkey_copy(n, k);
+	return bch2_btree_insert_trans(trans, btree, n, 0);
+}
+
 int __must_check bch2_trans_update_buffered(struct btree_trans *trans,
 					    enum btree_id btree,
 					    struct bkey_i *k)
@@ -564,6 +577,9 @@ int __must_check bch2_trans_update_buffered(struct btree_trans *trans,
 	EBUG_ON(trans->nr_wb_updates > trans->wb_updates_size);
 	EBUG_ON(k->k.u64s > BTREE_WRITE_BUFERED_U64s_MAX);
 
+	if (unlikely(trans->journal_replay_not_finished))
+		return bch2_btree_insert_clone_trans(trans, btree, k);
+
 	trans_for_each_wb_update(trans, i) {
 		if (i->btree == btree && bpos_eq(i->k.k.p, k->k.p)) {
 			bkey_copy(&i->k, k);
diff --git a/fs/bcachefs/btree_update_interior.c b/fs/bcachefs/btree_update_interior.c
index 6697417273aa..239fcc3c7c99 100644
--- a/fs/bcachefs/btree_update_interior.c
+++ b/fs/bcachefs/btree_update_interior.c
@@ -99,7 +99,7 @@ static void btree_node_interior_verify(struct bch_fs *c, struct btree *b)
 
 /* Calculate ideal packed bkey format for new btree nodes: */
 
-void __bch2_btree_calc_format(struct bkey_format_state *s, struct btree *b)
+static void __bch2_btree_calc_format(struct bkey_format_state *s, struct btree *b)
 {
 	struct bkey_packed *k;
 	struct bset_tree *t;
@@ -125,21 +125,20 @@ static struct bkey_format bch2_btree_calc_format(struct btree *b)
 	return bch2_bkey_format_done(&s);
 }
 
-static size_t btree_node_u64s_with_format(struct btree *b,
+static size_t btree_node_u64s_with_format(struct btree_nr_keys nr,
+					  struct bkey_format *old_f,
 					  struct bkey_format *new_f)
 {
-	struct bkey_format *old_f = &b->format;
-
 	/* stupid integer promotion rules */
 	ssize_t delta =
 	    (((int) new_f->key_u64s - old_f->key_u64s) *
-	     (int) b->nr.packed_keys) +
+	     (int) nr.packed_keys) +
 	    (((int) new_f->key_u64s - BKEY_U64s) *
-	     (int) b->nr.unpacked_keys);
+	     (int) nr.unpacked_keys);
 
-	BUG_ON(delta + b->nr.live_u64s < 0);
+	BUG_ON(delta + nr.live_u64s < 0);
 
-	return b->nr.live_u64s + delta;
+	return nr.live_u64s + delta;
 }
 
 /**
@@ -147,16 +146,18 @@ static size_t btree_node_u64s_with_format(struct btree *b,
  *
  * @c:		filesystem handle
  * @b:		btree node to rewrite
+ * @nr:		number of keys for new node (i.e. b->nr)
  * @new_f:	bkey format to translate keys to
  *
  * Returns: true if all re-packed keys will be able to fit in a new node.
  *
  * Assumes all keys will successfully pack with the new format.
  */
-bool bch2_btree_node_format_fits(struct bch_fs *c, struct btree *b,
+static bool bch2_btree_node_format_fits(struct bch_fs *c, struct btree *b,
+				 struct btree_nr_keys nr,
 				 struct bkey_format *new_f)
 {
-	size_t u64s = btree_node_u64s_with_format(b, new_f);
+	size_t u64s = btree_node_u64s_with_format(nr, &b->format, new_f);
 
 	return __vstruct_bytes(struct btree_node, u64s) < btree_bytes(c);
 }
@@ -391,7 +392,7 @@ static struct btree *bch2_btree_node_alloc_replacement(struct btree_update *as,
 	 * The keys might expand with the new format - if they wouldn't fit in
 	 * the btree node anymore, use the old format for now:
 	 */
-	if (!bch2_btree_node_format_fits(as->c, b, &format))
+	if (!bch2_btree_node_format_fits(as->c, b, b->nr, &format))
 		format = b->format;
 
 	SET_BTREE_NODE_SEQ(n->data, BTREE_NODE_SEQ(b->data) + 1);
@@ -1056,6 +1057,17 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	flags &= ~BCH_WATERMARK_MASK;
 	flags |= watermark;
 
+	if (!(flags & BTREE_INSERT_JOURNAL_RECLAIM) &&
+	    watermark < c->journal.watermark) {
+		struct journal_res res = { 0 };
+
+		ret = drop_locks_do(trans,
+			bch2_journal_res_get(&c->journal, &res, 1,
+					     watermark|JOURNAL_RES_GET_CHECK));
+		if (ret)
+			return ERR_PTR(ret);
+	}
+
 	while (1) {
 		nr_nodes[!!update_level] += 1 + split;
 		update_level++;
@@ -1334,8 +1346,11 @@ static void __btree_split_node(struct btree_update *as,
 	struct bkey_packed *out[2];
 	struct bkey uk;
 	unsigned u64s, n1_u64s = (b->nr.live_u64s * 3) / 5;
+	struct { unsigned nr_keys, val_u64s; } nr_keys[2];
 	int i;
 
+	memset(&nr_keys, 0, sizeof(nr_keys));
+
 	for (i = 0; i < 2; i++) {
 		BUG_ON(n[i]->nsets != 1);
 
@@ -1357,6 +1372,9 @@ static void __btree_split_node(struct btree_update *as,
 		if (!i)
 			n1_pos = uk.p;
 		bch2_bkey_format_add_key(&format[i], &uk);
+
+		nr_keys[i].nr_keys++;
+		nr_keys[i].val_u64s += bkeyp_val_u64s(&b->format, k);
 	}
 
 	btree_set_min(n[0], b->data->min_key);
@@ -1369,6 +1387,12 @@ static void __btree_split_node(struct btree_update *as,
 		bch2_bkey_format_add_pos(&format[i], n[i]->data->max_key);
 
 		n[i]->data->format = bch2_bkey_format_done(&format[i]);
+
+		unsigned u64s = nr_keys[i].nr_keys * n[i]->data->format.key_u64s +
+			nr_keys[i].val_u64s;
+		if (__vstruct_bytes(struct btree_node, u64s) > btree_bytes(as->c))
+			n[i]->data->format = b->format;
+
 		btree_node_set_format(n[i], n[i]->data->format);
 	}
 
@@ -1811,8 +1835,8 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 	bch2_bkey_format_add_pos(&new_s, next->data->max_key);
 	new_f = bch2_bkey_format_done(&new_s);
 
-	sib_u64s = btree_node_u64s_with_format(b, &new_f) +
-		btree_node_u64s_with_format(m, &new_f);
+	sib_u64s = btree_node_u64s_with_format(b->nr, &b->format, &new_f) +
+		btree_node_u64s_with_format(m->nr, &m->format, &new_f);
 
 	if (sib_u64s > BTREE_FOREGROUND_MERGE_HYSTERESIS(c)) {
 		sib_u64s -= BTREE_FOREGROUND_MERGE_HYSTERESIS(c);
diff --git a/fs/bcachefs/btree_update_interior.h b/fs/bcachefs/btree_update_interior.h
index 031076e75fa1..a6668992a272 100644
--- a/fs/bcachefs/btree_update_interior.h
+++ b/fs/bcachefs/btree_update_interior.h
@@ -6,10 +6,6 @@
 #include "btree_locking.h"
 #include "btree_update.h"
 
-void __bch2_btree_calc_format(struct bkey_format_state *, struct btree *);
-bool bch2_btree_node_format_fits(struct bch_fs *c, struct btree *,
-				struct bkey_format *);
-
 #define BTREE_UPDATE_NODES_MAX		((BTREE_MAX_DEPTH - 2) * 2 + GC_MERGE_NODES)
 
 #define BTREE_UPDATE_JOURNAL_RES	(BTREE_UPDATE_NODES_MAX * (BKEY_BTREE_PTR_U64s_MAX + 1))
diff --git a/fs/bcachefs/data_update.c b/fs/bcachefs/data_update.c
index 71aa5e59787b..b05457d284a6 100644
--- a/fs/bcachefs/data_update.c
+++ b/fs/bcachefs/data_update.c
@@ -471,7 +471,7 @@ int bch2_extent_drop_ptrs(struct btree_trans *trans,
 	 * we aren't using the extent overwrite path to delete, we're
 	 * just using the normal key deletion path:
 	 */
-	if (bkey_deleted(&n->k))
+	if (bkey_deleted(&n->k) && !(iter->flags & BTREE_ITER_IS_EXTENTS))
 		n->k.size = 0;
 
 	return bch2_trans_relock(trans) ?:
@@ -560,7 +560,8 @@ int bch2_data_update_init(struct btree_trans *trans,
 				move_ctxt_wait_event(ctxt,
 						(locked = bch2_bucket_nocow_trylock(&c->nocow_locks,
 									  PTR_BUCKET_POS(c, &p.ptr), 0)) ||
-						!atomic_read(&ctxt->read_sectors));
+						(!atomic_read(&ctxt->read_sectors) &&
+						 !atomic_read(&ctxt->write_sectors)));
 
 				if (!locked)
 					bch2_bucket_nocow_lock(&c->nocow_locks,
@@ -591,7 +592,7 @@ int bch2_data_update_init(struct btree_trans *trans,
 		m->data_opts.rewrite_ptrs = 0;
 		/* if iter == NULL, it's just a promote */
 		if (iter)
-			ret = bch2_extent_drop_ptrs(trans, iter, k, data_opts);
+			ret = bch2_extent_drop_ptrs(trans, iter, k, m->data_opts);
 		goto done;
 	}
 
diff --git a/fs/bcachefs/dirent.c b/fs/bcachefs/dirent.c
index 1a0f2d571569..2bfff0da7000 100644
--- a/fs/bcachefs/dirent.c
+++ b/fs/bcachefs/dirent.c
@@ -485,20 +485,15 @@ retry:
 	return ret;
 }
 
-int bch2_empty_dir_trans(struct btree_trans *trans, subvol_inum dir)
+int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot)
 {
 	struct btree_iter iter;
 	struct bkey_s_c k;
-	u32 snapshot;
 	int ret;
 
-	ret = bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot);
-	if (ret)
-		return ret;
-
 	for_each_btree_key_upto_norestart(trans, iter, BTREE_ID_dirents,
-			   SPOS(dir.inum, 0, snapshot),
-			   POS(dir.inum, U64_MAX), 0, k, ret)
+			   SPOS(dir, 0, snapshot),
+			   POS(dir, U64_MAX), 0, k, ret)
 		if (k.k->type == KEY_TYPE_dirent) {
 			ret = -ENOTEMPTY;
 			break;
@@ -508,6 +503,14 @@ int bch2_empty_dir_trans(struct btree_trans *trans, subvol_inum dir)
 	return ret;
 }
 
+int bch2_empty_dir_trans(struct btree_trans *trans, subvol_inum dir)
+{
+	u32 snapshot;
+
+	return bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot) ?:
+		bch2_empty_dir_snapshot(trans, dir.inum, snapshot);
+}
+
 int bch2_readdir(struct bch_fs *c, subvol_inum inum, struct dir_context *ctx)
 {
 	struct btree_trans *trans = bch2_trans_get(c);
diff --git a/fs/bcachefs/dirent.h b/fs/bcachefs/dirent.h
index cd262bf4d9c5..1e3431990abd 100644
--- a/fs/bcachefs/dirent.h
+++ b/fs/bcachefs/dirent.h
@@ -64,6 +64,7 @@ u64 bch2_dirent_lookup(struct bch_fs *, subvol_inum,
 		       const struct bch_hash_info *,
 		       const struct qstr *, subvol_inum *);
 
+int bch2_empty_dir_snapshot(struct btree_trans *, u64, u32);
 int bch2_empty_dir_trans(struct btree_trans *, subvol_inum);
 int bch2_readdir(struct bch_fs *, subvol_inum, struct dir_context *);
 
diff --git a/fs/bcachefs/extents.c b/fs/bcachefs/extents.c
index f6c92df55270..9d8afcb5979a 100644
--- a/fs/bcachefs/extents.c
+++ b/fs/bcachefs/extents.c
@@ -1294,7 +1294,8 @@ unsigned bch2_bkey_ptrs_need_rebalance(struct bch_fs *c, struct bkey_s_c k,
 		unsigned i = 0;
 
 		bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
-			if (p.crc.compression_type == BCH_COMPRESSION_TYPE_incompressible) {
+			if (p.crc.compression_type == BCH_COMPRESSION_TYPE_incompressible ||
+			    p.ptr.unwritten) {
 				rewrite_ptrs = 0;
 				goto incompressible;
 			}
diff --git a/fs/bcachefs/fs-ioctl.c b/fs/bcachefs/fs-ioctl.c
index 5a39bcb597a3..a70b7a03057d 100644
--- a/fs/bcachefs/fs-ioctl.c
+++ b/fs/bcachefs/fs-ioctl.c
@@ -413,7 +413,7 @@ retry:
 
 	if ((arg.flags & BCH_SUBVOL_SNAPSHOT_CREATE) &&
 	    !arg.src_ptr)
-		snapshot_src.subvol = to_bch_ei(dir)->ei_inode.bi_subvol;
+		snapshot_src.subvol = inode_inum(to_bch_ei(dir)).subvol;
 
 	inode = __bch2_create(file_mnt_idmap(filp), to_bch_ei(dir),
 			      dst_dentry, arg.mode|S_IFDIR,
diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c
index 4d51be813509..ba93e32d7708 100644
--- a/fs/bcachefs/fs.c
+++ b/fs/bcachefs/fs.c
@@ -1143,24 +1143,33 @@ static int bch2_encode_fh(struct inode *vinode, u32 *fh, int *len,
 {
 	struct bch_inode_info *inode	= to_bch_ei(vinode);
 	struct bch_inode_info *dir	= to_bch_ei(vdir);
-
-	if (*len < sizeof(struct bcachefs_fid_with_parent) / sizeof(u32))
-		return FILEID_INVALID;
+	int min_len;
 
 	if (!S_ISDIR(inode->v.i_mode) && dir) {
 		struct bcachefs_fid_with_parent *fid = (void *) fh;
 
+		min_len = sizeof(*fid) / sizeof(u32);
+		if (*len < min_len) {
+			*len = min_len;
+			return FILEID_INVALID;
+		}
+
 		fid->fid = bch2_inode_to_fid(inode);
 		fid->dir = bch2_inode_to_fid(dir);
 
-		*len = sizeof(*fid) / sizeof(u32);
+		*len = min_len;
 		return FILEID_BCACHEFS_WITH_PARENT;
 	} else {
 		struct bcachefs_fid *fid = (void *) fh;
 
+		min_len = sizeof(*fid) / sizeof(u32);
+		if (*len < min_len) {
+			*len = min_len;
+			return FILEID_INVALID;
+		}
 		*fid = bch2_inode_to_fid(inode);
 
-		*len = sizeof(*fid) / sizeof(u32);
+		*len = min_len;
 		return FILEID_BCACHEFS_WITHOUT_PARENT;
 	}
 }
@@ -1733,6 +1742,9 @@ static int bch2_unfreeze(struct super_block *sb)
 	struct bch_fs *c = sb->s_fs_info;
 	int ret;
 
+	if (test_bit(BCH_FS_EMERGENCY_RO, &c->flags))
+		return 0;
+
 	down_write(&c->state_lock);
 	ret = bch2_fs_read_write(c);
 	up_write(&c->state_lock);
diff --git a/fs/bcachefs/inode.c b/fs/bcachefs/inode.c
index c7849b0753e7..9309cfeecd8d 100644
--- a/fs/bcachefs/inode.c
+++ b/fs/bcachefs/inode.c
@@ -7,6 +7,7 @@
 #include "btree_update.h"
 #include "buckets.h"
 #include "compress.h"
+#include "dirent.h"
 #include "error.h"
 #include "extents.h"
 #include "extent_update.h"
@@ -1093,11 +1094,15 @@ static int may_delete_deleted_inode(struct btree_trans *trans,
 	if (ret)
 		goto out;
 
-	if (fsck_err_on(S_ISDIR(inode.bi_mode), c,
-			deleted_inode_is_dir,
-			"directory %llu:%u in deleted_inodes btree",
-			pos.offset, pos.snapshot))
-		goto delete;
+	if (S_ISDIR(inode.bi_mode)) {
+		ret = bch2_empty_dir_snapshot(trans, pos.offset, pos.snapshot);
+		if (fsck_err_on(ret == -ENOTEMPTY, c, deleted_inode_is_dir,
+				"non empty directory %llu:%u in deleted_inodes btree",
+				pos.offset, pos.snapshot))
+			goto delete;
+		if (ret)
+			goto out;
+	}
 
 	if (fsck_err_on(!(inode.bi_flags & BCH_INODE_unlinked), c,
 			deleted_inode_not_unlinked,
diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c
index 489b34046e78..8cf238be6213 100644
--- a/fs/bcachefs/journal.c
+++ b/fs/bcachefs/journal.c
@@ -249,7 +249,7 @@ static bool journal_entry_want_write(struct journal *j)
 	return ret;
 }
 
-static bool journal_entry_close(struct journal *j)
+bool bch2_journal_entry_close(struct journal *j)
 {
 	bool ret;
 
@@ -383,7 +383,7 @@ static bool journal_quiesced(struct journal *j)
 	bool ret = atomic64_read(&j->seq) == j->seq_ondisk;
 
 	if (!ret)
-		journal_entry_close(j);
+		bch2_journal_entry_close(j);
 	return ret;
 }
 
@@ -436,7 +436,7 @@ retry:
 
 	/*
 	 * Recheck after taking the lock, so we don't race with another thread
-	 * that just did journal_entry_open() and call journal_entry_close()
+	 * that just did journal_entry_open() and call bch2_journal_entry_close()
 	 * unnecessarily
 	 */
 	if (journal_res_get_fast(j, res, flags)) {
@@ -1041,7 +1041,7 @@ void bch2_fs_journal_stop(struct journal *j)
 	bch2_journal_reclaim_stop(j);
 	bch2_journal_flush_all_pins(j);
 
-	wait_event(j->wait, journal_entry_close(j));
+	wait_event(j->wait, bch2_journal_entry_close(j));
 
 	/*
 	 * Always write a new journal entry, to make sure the clock hands are up
diff --git a/fs/bcachefs/journal.h b/fs/bcachefs/journal.h
index 4c513fca5ef2..2f768e11aec9 100644
--- a/fs/bcachefs/journal.h
+++ b/fs/bcachefs/journal.h
@@ -266,6 +266,7 @@ static inline union journal_res_state journal_state_buf_put(struct journal *j, u
 	return s;
 }
 
+bool bch2_journal_entry_close(struct journal *);
 void bch2_journal_buf_put_final(struct journal *, u64, bool);
 
 static inline void __bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq)
diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c
index 0f17fc5f8d68..5de1b68fb8af 100644
--- a/fs/bcachefs/journal_io.c
+++ b/fs/bcachefs/journal_io.c
@@ -1599,6 +1599,7 @@ static CLOSURE_CALLBACK(journal_write_done)
 	} while ((v = atomic64_cmpxchg(&j->reservations.counter,
 				       old.v, new.v)) != old.v);
 
+	bch2_journal_reclaim_fast(j);
 	bch2_journal_space_available(j);
 
 	closure_wake_up(&w->wait);
diff --git a/fs/bcachefs/journal_reclaim.c b/fs/bcachefs/journal_reclaim.c
index e63c6eda86af..ec712104addb 100644
--- a/fs/bcachefs/journal_reclaim.c
+++ b/fs/bcachefs/journal_reclaim.c
@@ -776,6 +776,9 @@ static int journal_flush_done(struct journal *j, u64 seq_to_flush,
 			       (1U << JOURNAL_PIN_btree), 0, 0, 0))
 		*did_work = true;
 
+	if (seq_to_flush > journal_cur_seq(j))
+		bch2_journal_entry_close(j);
+
 	spin_lock(&j->lock);
 	/*
 	 * If journal replay hasn't completed, the unreplayed journal entries
diff --git a/fs/bcachefs/recovery.c b/fs/bcachefs/recovery.c
index 770ced1c6285..c7d9074c82d9 100644
--- a/fs/bcachefs/recovery.c
+++ b/fs/bcachefs/recovery.c
@@ -144,7 +144,7 @@ static int bch2_journal_replay(struct bch_fs *c)
 	u64 start_seq	= c->journal_replay_seq_start;
 	u64 end_seq	= c->journal_replay_seq_start;
 	size_t i;
-	int ret;
+	int ret = 0;
 
 	move_gap(keys->d, keys->nr, keys->size, keys->gap, keys->nr);
 	keys->gap = keys->nr;
diff --git a/fs/bcachefs/recovery.h b/fs/bcachefs/recovery.h
index 852d30567da9..d266aae90200 100644
--- a/fs/bcachefs/recovery.h
+++ b/fs/bcachefs/recovery.h
@@ -10,6 +10,9 @@ extern const char * const bch2_recovery_passes[];
 static inline int bch2_run_explicit_recovery_pass(struct bch_fs *c,
 						  enum bch_recovery_pass pass)
 {
+	if (c->recovery_passes_explicit & BIT_ULL(pass))
+		return 0;
+
 	bch_info(c, "running explicit recovery pass %s (%u), currently at %s (%u)",
 		 bch2_recovery_passes[pass], pass,
 		 bch2_recovery_passes[c->curr_recovery_pass], c->curr_recovery_pass);
diff --git a/fs/bcachefs/reflink.c b/fs/bcachefs/reflink.c
index 6e1bfe9feb59..37d16e04e671 100644
--- a/fs/bcachefs/reflink.c
+++ b/fs/bcachefs/reflink.c
@@ -121,6 +121,14 @@ int bch2_trans_mark_reflink_v(struct btree_trans *trans,
 {
 	check_indirect_extent_deleting(new, &flags);
 
+	if (old.k->type == KEY_TYPE_reflink_v &&
+	    new->k.type == KEY_TYPE_reflink_v &&
+	    old.k->u64s == new->k.u64s &&
+	    !memcmp(bkey_s_c_to_reflink_v(old).v->start,
+		    bkey_i_to_reflink_v(new)->v.start,
+		    bkey_val_bytes(&new->k) - 8))
+		return 0;
+
 	return bch2_trans_mark_extent(trans, btree_id, level, old, new, flags);
 }
 
diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c
index f63474c5c5a2..818ec467a06b 100644
--- a/fs/bcachefs/super.c
+++ b/fs/bcachefs/super.c
@@ -72,6 +72,12 @@
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Kent Overstreet <kent.overstreet@gmail.com>");
 MODULE_DESCRIPTION("bcachefs filesystem");
+MODULE_SOFTDEP("pre: crc32c");
+MODULE_SOFTDEP("pre: crc64");
+MODULE_SOFTDEP("pre: sha256");
+MODULE_SOFTDEP("pre: chacha20");
+MODULE_SOFTDEP("pre: poly1305");
+MODULE_SOFTDEP("pre: xxhash");
 
 #define KTYPE(type)							\
 static const struct attribute_group type ## _group = {			\
@@ -714,6 +720,7 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 
 	bch2_fs_copygc_init(c);
 	bch2_fs_btree_key_cache_init_early(&c->btree_key_cache);
+	bch2_fs_btree_iter_init_early(c);
 	bch2_fs_btree_interior_update_init_early(c);
 	bch2_fs_allocator_background_init(c);
 	bch2_fs_allocator_foreground_init(c);
diff --git a/fs/bcachefs/sysfs.c b/fs/bcachefs/sysfs.c
index ab743115f169..f3cb7115b530 100644
--- a/fs/bcachefs/sysfs.c
+++ b/fs/bcachefs/sysfs.c
@@ -276,8 +276,8 @@ static int bch2_compression_stats_to_text(struct printbuf *out, struct bch_fs *c
 		if (!btree_type_has_ptrs(id))
 			continue;
 
-		for_each_btree_key(trans, iter, id, POS_MIN,
-				   BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
+		ret = for_each_btree_key2(trans, iter, id, POS_MIN,
+					  BTREE_ITER_ALL_SNAPSHOTS, k, ({
 			struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
 			const union bch_extent_entry *entry;
 			struct extent_ptr_decoded p;
@@ -309,8 +309,8 @@ static int bch2_compression_stats_to_text(struct printbuf *out, struct bch_fs *c
 				nr_uncompressed_extents++;
 			else if (compressed)
 				nr_compressed_extents++;
-		}
-		bch2_trans_iter_exit(trans, &iter);
+			0;
+		}));
 	}
 
 	bch2_trans_put(trans);