Update bcachefs sources to 9a555a741e80 bcachefs: omit alignment attribute on big endian struct bkey

.bcachefs_revision

@@ -1 +1 @@
-39a84c99af2dbbd797eb2f0d1bed04111f04f245
+9a555a741e807275c320807babd3f42efb8fee90

include/linux/eytzinger.h

@@ -277,7 +277,7 @@ static inline ssize_t eytzinger0_find_le(void *base, size_t nr, size_t size,
 	int _res;							\
 									\
 	while (_i < _nr &&						\
-	       (_res = _cmp(_search, _base + _i * _size, _size)))	\
+	       (_res = _cmp(_search, _base + _i * _size)))		\
 		_i = eytzinger0_child(_i, _res > 0);			\
 	_i;								\
 })

libbcachefs/alloc_background.c

@@ -860,23 +860,28 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 			*bucket_gen(ca, new.k->p.offset) = new_a->gen;
 
 		bch2_dev_usage_update(c, ca, old_a, new_a, journal_seq, false);
+		percpu_up_read(&c->mark_lock);
 
-		if (new_a->data_type == BCH_DATA_free &&
+#define eval_state(_a, expr)		({ const struct bch_alloc_v4 *a = _a; expr; })
-		    (!new_a->journal_seq || new_a->journal_seq < c->journal.flushed_seq_ondisk))
+#define statechange(expr)		!eval_state(old_a, expr) && eval_state(new_a, expr)
+#define bucket_flushed(a)		(!a->journal_seq || a->journal_seq <= c->journal.flushed_seq_ondisk)
+
+		if (statechange(a->data_type == BCH_DATA_free &&
+				bucket_flushed(a)))
 			closure_wake_up(&c->freelist_wait);
 
-		if (new_a->data_type == BCH_DATA_need_discard &&
+		if (statechange(a->data_type == BCH_DATA_need_discard &&
-		    (!bucket_journal_seq || bucket_journal_seq < c->journal.flushed_seq_ondisk))
+				bucket_flushed(a)) &&
+		    !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset))
 			bch2_do_discards(c);
 
-		if (old_a->data_type != BCH_DATA_cached &&
+		if (statechange(a->data_type == BCH_DATA_cached) &&
-		    new_a->data_type == BCH_DATA_cached &&
+		    !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset) &&
 		    should_invalidate_buckets(ca, bch2_dev_usage_read(ca)))
 			bch2_do_invalidates(c);
 
-		if (new_a->data_type == BCH_DATA_need_gc_gens)
+		if (statechange(a->data_type == BCH_DATA_need_gc_gens))
 			bch2_do_gc_gens(c);
-		percpu_up_read(&c->mark_lock);
 	}
 
 	if ((flags & BTREE_TRIGGER_GC) &&

libbcachefs/bcachefs.h

@@ -267,6 +267,9 @@ do {									\
 
 #define bch2_fmt(_c, fmt)		bch2_log_msg(_c, fmt "\n")
 
+__printf(2, 3)
+void bch2_print_opts(struct bch_opts *, const char *, ...);
+
 __printf(2, 3)
 void __bch2_print(struct bch_fs *c, const char *fmt, ...);
 

libbcachefs/bcachefs_format.h

@@ -222,7 +222,36 @@ struct bkey {
 
 	__u8		pad[1];
 #endif
-} __packed __aligned(8);
+} __packed
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+/*
+ * The big-endian version of bkey can't be compiled by rustc with the "aligned"
+ * attr since it doesn't allow types to have both "packed" and "aligned" attrs.
+ * So for Rust compatibility, don't include this. It can be included in the LE
+ * version because the "packed" attr is redundant in that case.
+ *
+ * History: (quoting Kent)
+ *
+ * Specifically, when i was designing bkey, I wanted the header to be no
+ * bigger than necessary so that bkey_packed could use the rest. That means that
+ * decently offten extent keys will fit into only 8 bytes, instead of spilling over
+ * to 16.
+ *
+ * But packed_bkey treats the part after the header - the packed section -
+ * as a single multi word, variable length integer. And bkey, the unpacked
+ * version, is just a special case version of a bkey_packed; all the packed
+ * bkey code will work on keys in any packed format, the in-memory
+ * representation of an unpacked key also is just one type of packed key...
+ *
+ * So that constrains the key part of a bkig endian bkey to start right
+ * after the header.
+ *
+ * If we ever do a bkey_v2 and need to expand the hedaer by another byte for
+ * some reason - that will clean up this wart.
+ */
+__aligned(8)
+#endif
+;
 
 struct bkey_packed {
 	__u64		_data[0];
@@ -1426,17 +1455,14 @@ LE32_BITMASK(JSET_NO_FLUSH,	struct jset, flags, 5, 6);
 /* Btree: */
 
 enum btree_id_flags {
-	/* key size field is nonzero, btree iterators handle as ranges  */
 	BTREE_ID_EXTENTS	= BIT(0),
 	BTREE_ID_SNAPSHOTS	= BIT(1),
 	BTREE_ID_SNAPSHOT_FIELD	= BIT(2),
-	BTREE_ID_SNAPSHOTS_UNREFFED	= BIT(3),
 	BTREE_ID_DATA		= BIT(3),
 };
 
 #define BCH_BTREE_IDS()								\
-	x(extents,		0,	BTREE_ID_EXTENTS|BTREE_ID_SNAPSHOTS|	\
+	x(extents,		0,	BTREE_ID_EXTENTS|BTREE_ID_SNAPSHOTS|BTREE_ID_DATA,\
-					BTREE_ID_SNAPSHOTS_UNREFFED|BTREE_ID_DATA,\
 	  BIT_ULL(KEY_TYPE_whiteout)|						\
 	  BIT_ULL(KEY_TYPE_error)|						\
 	  BIT_ULL(KEY_TYPE_cookie)|						\
@@ -1454,7 +1480,7 @@ enum btree_id_flags {
 	  BIT_ULL(KEY_TYPE_whiteout)|						\
 	  BIT_ULL(KEY_TYPE_hash_whiteout)|					\
 	  BIT_ULL(KEY_TYPE_dirent))						\
-	x(xattrs,		3,	BTREE_ID_SNAPSHOTS|BTREE_ID_SNAPSHOTS_UNREFFED,\
+	x(xattrs,		3,	BTREE_ID_SNAPSHOTS,			\
 	  BIT_ULL(KEY_TYPE_whiteout)|						\
 	  BIT_ULL(KEY_TYPE_cookie)|						\
 	  BIT_ULL(KEY_TYPE_hash_whiteout)|					\

libbcachefs/bkey_methods.h

@@ -78,7 +78,6 @@ bool bch2_bkey_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
 
 enum btree_update_flags {
 	__BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE = __BTREE_ITER_FLAGS_END,
-	__BTREE_UPDATE_SNAPSHOT_WHITEOUT_CHECKS_DONE,
 	__BTREE_UPDATE_NOJOURNAL,
 	__BTREE_UPDATE_KEY_CACHE_RECLAIM,
 
@@ -92,8 +91,6 @@ enum btree_update_flags {
 };
 
 #define BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE (1U << __BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE)
-#define BTREE_UPDATE_SNAPSHOT_WHITEOUT_CHECKS_DONE	\
-					(1U << __BTREE_UPDATE_SNAPSHOT_WHITEOUT_CHECKS_DONE)
 #define BTREE_UPDATE_NOJOURNAL		(1U << __BTREE_UPDATE_NOJOURNAL)
 #define BTREE_UPDATE_KEY_CACHE_RECLAIM	(1U << __BTREE_UPDATE_KEY_CACHE_RECLAIM)
 

libbcachefs/btree_iter.c

@@ -1520,7 +1520,7 @@ static noinline void btree_paths_realloc(struct btree_trans *trans)
 {
 	unsigned nr = trans->nr_paths * 2;
 
-	void *p = kzalloc(BITS_TO_LONGS(nr) * sizeof(unsigned long) +
+	void *p = kvzalloc(BITS_TO_LONGS(nr) * sizeof(unsigned long) +
 			  sizeof(struct btree_trans_paths) +
 			  nr * sizeof(struct btree_path) +
 			  nr * sizeof(btree_path_idx_t) + 8 +
@@ -3087,7 +3087,7 @@ void bch2_trans_put(struct btree_trans *trans)
 	trans->paths		= NULL;
 
 	if (paths_allocated != trans->_paths_allocated)
-		kfree_rcu_mightsleep(paths_allocated);
+		kvfree_rcu_mightsleep(paths_allocated);
 
 	if (trans->mem_bytes == BTREE_TRANS_MEM_MAX)
 		mempool_free(trans->mem, &c->btree_trans_mem_pool);

libbcachefs/btree_types.h

@@ -705,17 +705,6 @@ static inline bool btree_type_has_snapshots(enum btree_id id)
 	return (1U << id) & mask;
 }
 
-static inline bool btree_type_snapshots_unreffed(enum btree_id id)
-{
-	const unsigned mask = 0
-#define x(name, nr, flags, ...)	|((!!((flags) & BTREE_ID_SNAPSHOTS_UNREFFED)) << nr)
-	BCH_BTREE_IDS()
-#undef x
-	;
-
-	return (1U << id) & mask;
-}
-
 static inline bool btree_type_has_snapshot_field(enum btree_id id)
 {
 	const unsigned mask = 0

libbcachefs/btree_update.c

@@ -82,168 +82,38 @@ static noinline int extent_back_merge(struct btree_trans *trans,
 	return 0;
 }
 
-static struct bkey_s_c peek_slot_including_whiteouts(struct btree_trans *trans, struct btree_iter *iter,
-						     enum btree_id btree, struct bpos pos)
-{
-	struct bkey_s_c k;
-	int ret;
-
-	for_each_btree_key_norestart(trans, *iter, btree, pos,
-			   BTREE_ITER_ALL_SNAPSHOTS|
-			   BTREE_ITER_NOPRESERVE, k, ret) {
-		if (!bkey_eq(k.k->p, pos))
-			break;
-		if (bch2_snapshot_is_ancestor(trans->c, pos.snapshot, k.k->p.snapshot))
-			return k;
-	}
-	bch2_trans_iter_exit(trans, iter);
-
-	return ret ? bkey_s_c_err(ret) : bkey_s_c_null;
-}
-
 /*
  * When deleting, check if we need to emit a whiteout (because we're overwriting
  * something in an ancestor snapshot)
  */
-static int need_whiteout_for_snapshot(struct btree_trans *trans, enum btree_id btree, struct bpos pos)
+static int need_whiteout_for_snapshot(struct btree_trans *trans,
+				      enum btree_id btree_id, struct bpos pos)
 {
-	pos.snapshot = bch2_snapshot_parent(trans->c, pos.snapshot);
-	if (!pos.snapshot)
-		return 0;
-
-	struct btree_iter iter;
-	struct bkey_s_c k = peek_slot_including_whiteouts(trans, &iter, btree, pos);
-	int ret = bkey_err(k) ?: k.k && !bkey_whiteout(k.k);
-	bch2_trans_iter_exit(trans, &iter);
-
-	return ret;
-}
-
-/*
- * We're overwriting a key at @pos in snapshot @snapshot, so we need to insert a
- * whiteout: that might be in @snapshot, or if there are overwites in sibling
- * snapshots, find the common ancestor where @pos is overwritten in every
- * descendent and insert the whiteout there - which might be at @pos.
- */
-static int delete_interior_snapshot_key(struct btree_trans *trans,
-					enum btree_id btree,
-					struct bpos whiteout, bool deleting,
-					struct bpos overwrite, bool old_is_whiteout)
-{
-	struct bch_fs *c = trans->c;
-	struct bpos orig_whiteout = whiteout, sib = whiteout;
 	struct btree_iter iter;
 	struct bkey_s_c k;
+	u32 snapshot = pos.snapshot;
 	int ret;
 
-	sib.snapshot = bch2_snapshot_sibling(c, sib.snapshot);
+	if (!bch2_snapshot_parent(trans->c, pos.snapshot))
-
-	for_each_btree_key_norestart(trans, iter, btree, sib,
-				     BTREE_ITER_ALL_SNAPSHOTS|BTREE_ITER_INTENT, k, ret) {
-		BUG_ON(bpos_gt(k.k->p, overwrite));
-
-		if (bpos_lt(k.k->p, sib)) /* unrelated branch - skip */
-			continue;
-		if (bpos_gt(k.k->p, sib)) /* did not find @sib */
-			break;
-
-		/* @overwrite is also written in @sib, now check parent */
-		whiteout.snapshot = bch2_snapshot_parent(c, whiteout.snapshot);
-		if (bpos_eq(whiteout, overwrite))
-			break;
-
-		sib = whiteout;
-		sib.snapshot = bch2_snapshot_sibling(c, sib.snapshot);
-	}
-
-	if (ret)
-		goto err;
-
-	if (!deleting && bpos_eq(whiteout, orig_whiteout))
-		goto out;
-
-	if (!bpos_eq(iter.pos, whiteout)) {
-		bch2_trans_iter_exit(trans, &iter);
-		bch2_trans_iter_init(trans, &iter, btree, whiteout, BTREE_ITER_INTENT);
-		k = bch2_btree_iter_peek_slot(&iter);
-		ret = bkey_err(k);
-		if (ret)
-			goto err;
-	}
-
-	iter.flags &= ~BTREE_ITER_ALL_SNAPSHOTS;
-	iter.flags |= BTREE_ITER_FILTER_SNAPSHOTS;
-
-	struct bkey_i *delete = bch2_trans_kmalloc(trans, sizeof(*delete));
-	ret = PTR_ERR_OR_ZERO(delete);
-	if (ret)
-		goto err;
-
-	bkey_init(&delete->k);
-	delete->k.p = whiteout;
-
-	ret = !bpos_eq(whiteout, overwrite)
-		? !old_is_whiteout
-		: need_whiteout_for_snapshot(trans, btree, whiteout);
-	if (ret < 0)
-		goto err;
-	if (ret)
-		delete->k.type = KEY_TYPE_whiteout;
-
-	ret = bch2_trans_update(trans, &iter, delete,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
-				BTREE_UPDATE_SNAPSHOT_WHITEOUT_CHECKS_DONE);
-out:
-err:
-	bch2_trans_iter_exit(trans, &iter);
-	return ret;
-}
-
-/*
- * We're overwriting a key in a snapshot that has ancestors: if we're
- * overwriting a key in a different snapshot, we need to check if it is now
- * fully overritten and can be deleted, and if we're deleting a key in the
- * current snapshot we need to check if we need to leave a whiteout.
- */
-static noinline int
-overwrite_interior_snapshot_key(struct btree_trans *trans,
-				struct btree_iter *iter,
-				struct bkey_i *k)
-{
-	struct bkey_s_c old = bch2_btree_iter_peek_slot(iter);
-
-	int ret = bkey_err(old);
-	if (ret)
-		return ret;
-
-	if (!bkey_deleted(old.k)) {
-		if (btree_type_snapshots_unreffed(iter->btree_id) &&
-		    old.k->p.snapshot != k->k.p.snapshot) {
-			/*
-			 * We're overwriting a key in a different snapshot:
-			 * check if it's also been overwritten in siblings
-			 */
-			ret = delete_interior_snapshot_key(trans, iter->btree_id,
-							   k->k.p,   bkey_deleted(&k->k),
-							   old.k->p, bkey_whiteout(old.k));
-			if (ret)
-				return ret;
-			if (bkey_deleted(&k->k))
-				return 1;
-		} else if (bkey_deleted(&k->k)) {
-			/*
-			 * We're deleting a key in the current snapshot:
-			 * check if we need to leave a whiteout
-			 */
-			ret = need_whiteout_for_snapshot(trans, iter->btree_id, k->k.p);
-			if (unlikely(ret < 0))
-				return ret;
-			if (ret)
-				k->k.type = KEY_TYPE_whiteout;
-		}
-	}
-
 		return 0;
+
+	pos.snapshot++;
+
+	for_each_btree_key_norestart(trans, iter, btree_id, pos,
+			   BTREE_ITER_ALL_SNAPSHOTS|
+			   BTREE_ITER_NOPRESERVE, k, ret) {
+		if (!bkey_eq(k.k->p, pos))
+			break;
+
+		if (bch2_snapshot_is_ancestor(trans->c, snapshot,
+					      k.k->p.snapshot)) {
+			ret = !bkey_whiteout(k.k);
+			break;
+		}
+	}
+	bch2_trans_iter_exit(trans, &iter);
+
+	return ret;
 }
 
 int __bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
@@ -633,29 +503,32 @@ static noinline int bch2_trans_update_get_key_cache(struct btree_trans *trans,
 int __must_check bch2_trans_update(struct btree_trans *trans, struct btree_iter *iter,
 				   struct bkey_i *k, enum btree_update_flags flags)
 {
+	btree_path_idx_t path_idx = iter->update_path ?: iter->path;
+	int ret;
+
 	if (iter->flags & BTREE_ITER_IS_EXTENTS)
 		return bch2_trans_update_extent(trans, iter, k, flags);
 
-	if (!(flags & (BTREE_UPDATE_SNAPSHOT_WHITEOUT_CHECKS_DONE|
+	if (bkey_deleted(&k->k) &&
-		       BTREE_UPDATE_KEY_CACHE_RECLAIM)) &&
+	    !(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
-	    (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS) &&
+	    (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)) {
-	    bch2_snapshot_parent(trans->c, k->k.p.snapshot)) {
+		ret = need_whiteout_for_snapshot(trans, iter->btree_id, k->k.p);
-		int ret = overwrite_interior_snapshot_key(trans, iter, k);
+		if (unlikely(ret < 0))
+			return ret;
+
 		if (ret)
-			return ret < 0 ? ret : 0;
+			k->k.type = KEY_TYPE_whiteout;
 	}
 
 	/*
 	 * Ensure that updates to cached btrees go to the key cache:
 	 */
-	btree_path_idx_t path_idx = iter->update_path ?: iter->path;
 	struct btree_path *path = trans->paths + path_idx;
-
 	if (!(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
 	    !path->cached &&
 	    !path->level &&
 	    btree_id_cached(trans->c, path->btree_id)) {
-		int ret = bch2_trans_update_get_key_cache(trans, iter, path);
+		ret = bch2_trans_update_get_key_cache(trans, iter, path);
 		if (ret)
 			return ret;
 

libbcachefs/chardev.c

@@ -165,6 +165,11 @@ static void bch2_fsck_offline_thread_fn(struct thread_with_stdio *stdio)
 		bch2_fs_stop(c);
 }
 
+static const struct thread_with_stdio_ops bch2_offline_fsck_ops = {
+	.exit		= bch2_fsck_thread_exit,
+	.fn		= bch2_fsck_offline_thread_fn,
+};
+
 static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_arg)
 {
 	struct bch_ioctl_fsck_offline arg;
@@ -217,9 +222,7 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a
 
 	opt_set(thr->opts, stdio, (u64)(unsigned long)&thr->thr.stdio);
 
-	ret = run_thread_with_stdio(&thr->thr,
+	ret = run_thread_with_stdio(&thr->thr, &bch2_offline_fsck_ops);
-			bch2_fsck_thread_exit,
-			bch2_fsck_offline_thread_fn);
 err:
 	if (ret < 0) {
 		if (thr)
@@ -794,6 +797,11 @@ static void bch2_fsck_online_thread_fn(struct thread_with_stdio *stdio)
 	bch2_ro_ref_put(c);
 }
 
+static const struct thread_with_stdio_ops bch2_online_fsck_ops = {
+	.exit		= bch2_fsck_thread_exit,
+	.fn		= bch2_fsck_online_thread_fn,
+};
+
 static long bch2_ioctl_fsck_online(struct bch_fs *c,
 				   struct bch_ioctl_fsck_online arg)
 {
@@ -834,9 +842,7 @@ static long bch2_ioctl_fsck_online(struct bch_fs *c,
 			goto err;
 	}
 
-	ret = run_thread_with_stdio(&thr->thr,
+	ret = run_thread_with_stdio(&thr->thr, &bch2_online_fsck_ops);
-			bch2_fsck_thread_exit,
-			bch2_fsck_online_thread_fn);
 err:
 	if (ret < 0) {
 		bch_err_fn(c, ret);

libbcachefs/dirent.c

@@ -512,7 +512,7 @@ u64 bch2_dirent_lookup(struct bch_fs *c, subvol_inum dir,
 	return ret;
 }
 
-int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot)
+int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 subvol, u32 snapshot)
 {
 	struct btree_iter iter;
 	struct bkey_s_c k;
@@ -522,6 +522,9 @@ int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot)
 			   SPOS(dir, 0, snapshot),
 			   POS(dir, U64_MAX), 0, k, ret)
 		if (k.k->type == KEY_TYPE_dirent) {
+			struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
+			if (d.v->d_type == DT_SUBVOL && le32_to_cpu(d.v->d_parent_subvol) != subvol)
+				continue;
 			ret = -BCH_ERR_ENOTEMPTY_dir_not_empty;
 			break;
 		}
@@ -535,7 +538,7 @@ 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);
+		bch2_empty_dir_snapshot(trans, dir.inum, dir.subvol, snapshot);
 }
 
 int bch2_readdir(struct bch_fs *c, subvol_inum inum, struct dir_context *ctx)

libbcachefs/dirent.h

@@ -69,7 +69,7 @@ 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_snapshot(struct btree_trans *, u64, u32, u32);
 int bch2_empty_dir_trans(struct btree_trans *, subvol_inum);
 int bch2_readdir(struct bch_fs *, subvol_inum, struct dir_context *);
 

libbcachefs/fs-common.c

@@ -486,9 +486,9 @@ int bch2_rename_trans(struct btree_trans *trans,
 			goto err;
 		}
 
-		if (S_ISDIR(dst_inode_u->bi_mode) &&
+		if (S_ISDIR(dst_inode_u->bi_mode)) {
-		    bch2_empty_dir_trans(trans, dst_inum)) {
+			ret = bch2_empty_dir_trans(trans, dst_inum);
-			ret = -ENOTEMPTY;
+			if (ret)
 				goto err;
 		}
 	}

libbcachefs/fs-io-direct.c

@@ -88,6 +88,8 @@ static int bch2_direct_IO_read(struct kiocb *req, struct iov_iter *iter)
 		return ret;
 
 	shorten = iov_iter_count(iter) - round_up(ret, block_bytes(c));
+	if (shorten >= iter->count)
+		shorten = 0;
 	iter->count -= shorten;
 
 	bio = bio_alloc_bioset(NULL,

libbcachefs/inode.c

@@ -1117,8 +1117,9 @@ static int may_delete_deleted_inode(struct btree_trans *trans,
 		goto out;
 
 	if (S_ISDIR(inode.bi_mode)) {
-		ret = bch2_empty_dir_snapshot(trans, pos.offset, pos.snapshot);
+		ret = bch2_empty_dir_snapshot(trans, pos.offset, 0, pos.snapshot);
-		if (fsck_err_on(ret == -ENOTEMPTY, c, deleted_inode_is_dir,
+		if (fsck_err_on(bch2_err_matches(ret, ENOTEMPTY),
+				c, deleted_inode_is_dir,
 				"non empty directory %llu:%u in deleted_inodes btree",
 				pos.offset, pos.snapshot))
 			goto delete;

libbcachefs/journal_io.c

@@ -1506,7 +1506,7 @@ static int journal_write_alloc(struct journal *j, struct journal_buf *w)
 		c->opts.foreground_target;
 	unsigned i, replicas = 0, replicas_want =
 		READ_ONCE(c->opts.metadata_replicas);
-	unsigned replicas_need = min(replicas_want,
+	unsigned replicas_need = min_t(unsigned, replicas_want,
 				       READ_ONCE(c->opts.metadata_replicas_required));
 
 	rcu_read_lock();

libbcachefs/printbuf.c

@@ -56,6 +56,7 @@ void bch2_prt_vprintf(struct printbuf *out, const char *fmt, va_list args)
 
 		va_copy(args2, args);
 		len = vsnprintf(out->buf + out->pos, printbuf_remaining(out), fmt, args2);
+		va_end(args2);
 	} while (len + 1 >= printbuf_remaining(out) &&
 		 !bch2_printbuf_make_room(out, len + 1));
 

libbcachefs/recovery.c

@@ -577,8 +577,9 @@ u64 bch2_recovery_passes_from_stable(u64 v)
 
 static bool check_version_upgrade(struct bch_fs *c)
 {
-	unsigned latest_compatible = bch2_latest_compatible_version(c->sb.version);
 	unsigned latest_version	= bcachefs_metadata_version_current;
+	unsigned latest_compatible = min(latest_version,
+					 bch2_latest_compatible_version(c->sb.version));
 	unsigned old_version = c->sb.version_upgrade_complete ?: c->sb.version;
 	unsigned new_version = 0;
 
@@ -597,7 +598,7 @@ static bool check_version_upgrade(struct bch_fs *c)
 			new_version = latest_version;
 			break;
 		case BCH_VERSION_UPGRADE_none:
-			new_version = old_version;
+			new_version = min(old_version, latest_version);
 			break;
 		}
 	}
@@ -774,7 +775,7 @@ int bch2_fs_recovery(struct bch_fs *c)
 		goto err;
 	}
 
-	if (!(c->opts.nochanges && c->opts.norecovery)) {
+	if (!c->opts.nochanges) {
 		mutex_lock(&c->sb_lock);
 		bool write_sb = false;
 
@@ -804,7 +805,7 @@ int bch2_fs_recovery(struct bch_fs *c)
 		if (bch2_check_version_downgrade(c)) {
 			struct printbuf buf = PRINTBUF;
 
-			prt_str(&buf, "Version downgrade required:\n");
+			prt_str(&buf, "Version downgrade required:");
 
 			__le64 passes = ext->recovery_passes_required[0];
 			bch2_sb_set_downgrade(c,
@@ -812,7 +813,7 @@ int bch2_fs_recovery(struct bch_fs *c)
 					BCH_VERSION_MINOR(c->sb.version));
 			passes = ext->recovery_passes_required[0] & ~passes;
 			if (passes) {
-				prt_str(&buf, "  running recovery passes: ");
+				prt_str(&buf, "\n  running recovery passes: ");
 				prt_bitflags(&buf, bch2_recovery_passes,
 					     bch2_recovery_passes_from_stable(le64_to_cpu(passes)));
 			}

libbcachefs/replicas.c

@@ -232,7 +232,7 @@ static inline int __replicas_entry_idx(struct bch_replicas_cpu *r,
 
 	verify_replicas_entry(search);
 
-#define entry_cmp(_l, _r, size)	memcmp(_l, _r, entry_size)
+#define entry_cmp(_l, _r)	memcmp(_l, _r, entry_size)
 	idx = eytzinger0_find(r->entries, r->nr, r->entry_size,
 			      entry_cmp, search);
 #undef entry_cmp

libbcachefs/super-io.c

@@ -717,6 +717,7 @@ retry:
 
 	if (IS_ERR(sb->bdev_handle)) {
 		ret = PTR_ERR(sb->bdev_handle);
+		prt_printf(&err, "error opening %s: %s", path, bch2_err_str(ret));
 		goto err;
 	}
 	sb->bdev = sb->bdev_handle->bdev;
@@ -743,9 +744,9 @@ retry:
 	prt_printf(&err2, "bcachefs (%s): error reading default superblock: %s\n",
 	       path, err.buf);
 	if (ret == -BCH_ERR_invalid_sb_magic && ignore_notbchfs_msg)
-		printk(KERN_INFO "%s", err2.buf);
+		bch2_print_opts(opts, KERN_INFO "%s", err2.buf);
 	else
-		printk(KERN_ERR "%s", err2.buf);
+		bch2_print_opts(opts, KERN_ERR "%s", err2.buf);
 
 	printbuf_exit(&err2);
 	printbuf_reset(&err);
@@ -808,7 +809,7 @@ got_super:
 
 	ret = bch2_sb_validate(sb, &err, READ);
 	if (ret) {
-		printk(KERN_ERR "bcachefs (%s): error validating superblock: %s\n",
+		bch2_print_opts(opts, KERN_ERR "bcachefs (%s): error validating superblock: %s\n",
 				path, err.buf);
 		goto err_no_print;
 	}
@@ -816,7 +817,7 @@ out:
 	printbuf_exit(&err);
 	return ret;
 err:
-	printk(KERN_ERR "bcachefs (%s): error reading superblock: %s\n",
+	bch2_print_opts(opts, KERN_ERR "bcachefs (%s): error reading superblock: %s\n",
 			path, err.buf);
 err_no_print:
 	bch2_free_super(sb);

libbcachefs/super.c

@@ -87,6 +87,23 @@ const char * const bch2_fs_flag_strs[] = {
 	NULL
 };
 
+void bch2_print_opts(struct bch_opts *opts, const char *fmt, ...)
+{
+	struct stdio_redirect *stdio = (void *)(unsigned long)opts->stdio;
+
+	va_list args;
+	va_start(args, fmt);
+	if (likely(!stdio)) {
+		vprintk(fmt, args);
+	} else {
+		if (fmt[0] == KERN_SOH[0])
+			fmt += 2;
+
+		stdio_redirect_vprintf(stdio, true, fmt, args);
+	}
+	va_end(args);
+}
+
 void __bch2_print(struct bch_fs *c, const char *fmt, ...)
 {
 	struct stdio_redirect *stdio = bch2_fs_stdio_redirect(c);
@@ -1592,27 +1609,27 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
 	__bch2_dev_read_only(c, ca);
 
 	ret = bch2_dev_data_drop(c, ca->dev_idx, flags);
-	bch_err_msg(ca, ret, "dropping data");
+	bch_err_msg(ca, ret, "bch2_dev_data_drop()");
 	if (ret)
 		goto err;
 
 	ret = bch2_dev_remove_alloc(c, ca);
-	bch_err_msg(ca, ret, "deleting alloc info");
+	bch_err_msg(ca, ret, "bch2_dev_remove_alloc()");
 	if (ret)
 		goto err;
 
 	ret = bch2_journal_flush_device_pins(&c->journal, ca->dev_idx);
-	bch_err_msg(ca, ret, "flushing journal");
+	bch_err_msg(ca, ret, "bch2_journal_flush_device_pins()");
 	if (ret)
 		goto err;
 
 	ret = bch2_journal_flush(&c->journal);
-	bch_err(ca, "journal error");
+	bch_err_msg(ca, ret, "bch2_journal_flush()");
 	if (ret)
 		goto err;
 
 	ret = bch2_replicas_gc2(c);
-	bch_err_msg(ca, ret, "in replicas_gc2()");
+	bch_err_msg(ca, ret, "bch2_replicas_gc2()");
 	if (ret)
 		goto err;