Merge tag 'bcachefs-2024-10-05' of git://evilpiepirate.org/bcachefs

Pull bcachefs fixes from Kent Overstreet:
 "A lot of little fixes, bigger ones include:

   - bcachefs's __wait_on_freeing_inode() was broken in rc1 due to vfs
     changes, now fixed along with another lost wakeup

   - fragmentation LRU fixes; fsck now repairs successfully (this is the
     data structure copygc uses); along with some nice simplification.

   - Rework logged op error handling, so that if logged op replay errors
     (due to another filesystem error) we delete the logged op instead
     of going into an infinite loop)

   - Various small filesystem connectivitity repair fixes"

* tag 'bcachefs-2024-10-05' of git://evilpiepirate.org/bcachefs:
  bcachefs: Rework logged op error handling
  bcachefs: Add warn param to subvol_get_snapshot, peek_inode
  bcachefs: Kill snapshot arg to fsck_write_inode()
  bcachefs: Check for unlinked, non-empty dirs in check_inode()
  bcachefs: Check for unlinked inodes with dirents
  bcachefs: Check for directories with no backpointers
  bcachefs: Kill alloc_v4.fragmentation_lru
  bcachefs: minor lru fsck fixes
  bcachefs: Mark more errors AUTOFIX
  bcachefs: Make sure we print error that causes fsck to bail out
  bcachefs: bkey errors are only AUTOFIX during read
  bcachefs: Create lost+found in correct snapshot
  bcachefs: Fix reattach_inode()
  bcachefs: Add missing wakeup to bch2_inode_hash_remove()
  bcachefs: Fix trans_commit disk accounting revert
  bcachefs: Fix bch2_inode_is_open() check
  bcachefs: Fix return type of dirent_points_to_inode_nowarn()
  bcachefs: Fix bad shift in bch2_read_flag_list()

Author: torvalds

fs/bcachefs/alloc_background.c

@@ -332,7 +332,6 @@ void bch2_alloc_v4_swab(struct bkey_s k)
 	a->io_time[1]		= swab64(a->io_time[1]);
 	a->stripe		= swab32(a->stripe);
 	a->nr_external_backpointers = swab32(a->nr_external_backpointers);
-	a->fragmentation_lru	= swab64(a->fragmentation_lru);
 	a->stripe_sectors	= swab32(a->stripe_sectors);
 
 	bps = alloc_v4_backpointers(a);
@@ -347,6 +346,7 @@ void bch2_alloc_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c
 {
 	struct bch_alloc_v4 _a;
 	const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &_a);
+	struct bch_dev *ca = c ? bch2_dev_bucket_tryget_noerror(c, k.k->p) : NULL;
 
 	prt_newline(out);
 	printbuf_indent_add(out, 2);
@@ -364,9 +364,13 @@ void bch2_alloc_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c
 	prt_printf(out, "stripe_redundancy %u\n",	a->stripe_redundancy);
 	prt_printf(out, "io_time[READ]     %llu\n",	a->io_time[READ]);
 	prt_printf(out, "io_time[WRITE]    %llu\n",	a->io_time[WRITE]);
-	prt_printf(out, "fragmentation     %llu\n",	a->fragmentation_lru);
+
+	if (ca)
+		prt_printf(out, "fragmentation     %llu\n",	alloc_lru_idx_fragmentation(*a, ca));
 	prt_printf(out, "bp_start          %llu\n", BCH_ALLOC_V4_BACKPOINTERS_START(a));
 	printbuf_indent_sub(out, 2);
+
+	bch2_dev_put(ca);
 }
 
 void __bch2_alloc_to_v4(struct bkey_s_c k, struct bch_alloc_v4 *out)
@@ -882,12 +886,13 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 				goto err;
 		}
 
-		new_a->fragmentation_lru = alloc_lru_idx_fragmentation(*new_a, ca);
-		if (old_a->fragmentation_lru != new_a->fragmentation_lru) {
+		old_lru = alloc_lru_idx_fragmentation(*old_a, ca);
+		new_lru = alloc_lru_idx_fragmentation(*new_a, ca);
+		if (old_lru != new_lru) {
 			ret = bch2_lru_change(trans,
 					BCH_LRU_FRAGMENTATION_START,
 					bucket_to_u64(new.k->p),
-					old_a->fragmentation_lru, new_a->fragmentation_lru);
+					old_lru, new_lru);
 			if (ret)
 				goto err;
 		}
@@ -1629,18 +1634,22 @@ static int bch2_check_alloc_to_lru_ref(struct btree_trans *trans,
 	if (ret)
 		return ret;
 
+	struct bch_dev *ca = bch2_dev_tryget_noerror(c, alloc_k.k->p.inode);
+	if (!ca)
+		return 0;
+
 	a = bch2_alloc_to_v4(alloc_k, &a_convert);
 
-	if (a->fragmentation_lru) {
+	u64 lru_idx = alloc_lru_idx_fragmentation(*a, ca);
+	if (lru_idx) {
 		ret = bch2_lru_check_set(trans, BCH_LRU_FRAGMENTATION_START,
-					 a->fragmentation_lru,
-					 alloc_k, last_flushed);
+					 lru_idx, alloc_k, last_flushed);
 		if (ret)
-			return ret;
+			goto err;
 	}
 
 	if (a->data_type != BCH_DATA_cached)
-		return 0;
+		goto err;
 
 	if (fsck_err_on(!a->io_time[READ],
 			trans, alloc_key_cached_but_read_time_zero,
@@ -1669,6 +1678,7 @@ static int bch2_check_alloc_to_lru_ref(struct btree_trans *trans,
 		goto err;
 err:
 fsck_err:
+	bch2_dev_put(ca);
 	printbuf_exit(&buf);
 	return ret;
 }

fs/bcachefs/alloc_background_format.h

@@ -70,7 +70,7 @@ struct bch_alloc_v4 {
 	__u32			stripe;
 	__u32			nr_external_backpointers;
 	/* end of fields in original version of alloc_v4 */
-	__u64			fragmentation_lru;
+	__u64			_fragmentation_lru; /* obsolete */
 	__u32			stripe_sectors;
 	__u32			pad;
 } __packed __aligned(8);

fs/bcachefs/btree_gc.c

@@ -828,8 +828,6 @@ static int bch2_alloc_write_key(struct btree_trans *trans,
 			return ret;
 	}
 
-	gc.fragmentation_lru = alloc_lru_idx_fragmentation(gc, ca);
-
 	if (fsck_err_on(new.data_type != gc.data_type,
 			trans, alloc_key_data_type_wrong,
 			"bucket %llu:%llu gen %u has wrong data_type"
@@ -857,7 +855,6 @@ static int bch2_alloc_write_key(struct btree_trans *trans,
 	copy_bucket_field(alloc_key_cached_sectors_wrong,	cached_sectors);
 	copy_bucket_field(alloc_key_stripe_wrong,		stripe);
 	copy_bucket_field(alloc_key_stripe_redundancy_wrong,	stripe_redundancy);
-	copy_bucket_field(alloc_key_fragmentation_lru_wrong,	fragmentation_lru);
 #undef copy_bucket_field
 
 	if (!bch2_alloc_v4_cmp(*old, new))

fs/bcachefs/btree_trans_commit.c

@@ -832,7 +832,8 @@ bch2_trans_commit_write_locked(struct btree_trans *trans, unsigned flags,
 	for (struct jset_entry *entry2 = trans->journal_entries;
 	     entry2 != entry;
 	     entry2 = vstruct_next(entry2))
-		if (jset_entry_is_key(entry2) && entry2->start->k.type == KEY_TYPE_accounting) {
+		if (entry2->type == BCH_JSET_ENTRY_write_buffer_keys &&
+		    entry2->start->k.type == KEY_TYPE_accounting) {
 			struct bkey_s_accounting a = bkey_i_to_s_accounting(entry2->start);
 
 			bch2_accounting_neg(a);

fs/bcachefs/error.c

@@ -393,16 +393,22 @@ int __bch2_fsck_err(struct bch_fs *c,
 	     !(flags & FSCK_CAN_IGNORE)))
 		ret = -BCH_ERR_fsck_errors_not_fixed;
 
+	bool exiting =
+		test_bit(BCH_FS_fsck_running, &c->flags) &&
+		(ret != -BCH_ERR_fsck_fix &&
+		 ret != -BCH_ERR_fsck_ignore);
+
+	if (exiting)
+		print = true;
+
 	if (print) {
 		if (bch2_fs_stdio_redirect(c))
 			bch2_print(c, "%s\n", out->buf);
 		else
 			bch2_print_string_as_lines(KERN_ERR, out->buf);
 	}
 
-	if (test_bit(BCH_FS_fsck_running, &c->flags) &&
-	    (ret != -BCH_ERR_fsck_fix &&
-	     ret != -BCH_ERR_fsck_ignore))
+	if (exiting)
 		bch_err(c, "Unable to continue, halting");
 	else if (suppressing)
 		bch_err(c, "Ratelimiting new instances of previous error");
@@ -430,10 +436,17 @@ int __bch2_fsck_err(struct bch_fs *c,
 
 int __bch2_bkey_fsck_err(struct bch_fs *c,
 			 struct bkey_s_c k,
-			 enum bch_fsck_flags flags,
+			 enum bch_validate_flags validate_flags,
 			 enum bch_sb_error_id err,
 			 const char *fmt, ...)
 {
+	if (validate_flags & BCH_VALIDATE_silent)
+		return -BCH_ERR_fsck_delete_bkey;
+
+	unsigned fsck_flags = 0;
+	if (!(validate_flags & (BCH_VALIDATE_write|BCH_VALIDATE_commit)))
+		fsck_flags |= FSCK_AUTOFIX|FSCK_CAN_FIX;
+
 	struct printbuf buf = PRINTBUF;
 	va_list args;
 
@@ -445,7 +458,7 @@ int __bch2_bkey_fsck_err(struct bch_fs *c,
 	va_end(args);
 	prt_str(&buf, ": delete?");
 
-	int ret = __bch2_fsck_err(c, NULL, flags, err, "%s", buf.buf);
+	int ret = __bch2_fsck_err(c, NULL, fsck_flags, err, "%s", buf.buf);
 	printbuf_exit(&buf);
 	return ret;
 }

fs/bcachefs/error.h

@@ -167,10 +167,11 @@ void bch2_flush_fsck_errs(struct bch_fs *);
 #define fsck_err_on(cond, c, _err_type, ...)				\
 	__fsck_err_on(cond, c, FSCK_CAN_FIX|FSCK_CAN_IGNORE, _err_type, __VA_ARGS__)
 
+enum bch_validate_flags;
 __printf(5, 6)
 int __bch2_bkey_fsck_err(struct bch_fs *,
 			 struct bkey_s_c,
-			 enum bch_fsck_flags,
+			 enum bch_validate_flags,
 			 enum bch_sb_error_id,
 			 const char *, ...);
 
@@ -180,11 +181,7 @@ int __bch2_bkey_fsck_err(struct bch_fs *,
  */
 #define bkey_fsck_err(c, _err_type, _err_msg, ...)			\
 do {									\
-	if ((flags & BCH_VALIDATE_silent)) {				\
-		ret = -BCH_ERR_fsck_delete_bkey;			\
-		goto fsck_err;						\
-	}								\
-	int _ret = __bch2_bkey_fsck_err(c, k, FSCK_CAN_FIX|FSCK_AUTOFIX,\
+	int _ret = __bch2_bkey_fsck_err(c, k, flags,			\
 				BCH_FSCK_ERR_##_err_type,		\
 				_err_msg, ##__VA_ARGS__);		\
 	if (_ret != -BCH_ERR_fsck_fix &&				\

fs/bcachefs/fs.c

@@ -174,20 +174,24 @@ static const struct rhashtable_params bch2_vfs_inodes_params = {
 	.automatic_shrinking	= true,
 };
 
-static void __wait_on_freeing_inode(struct inode *inode)
+struct bch_inode_info *__bch2_inode_hash_find(struct bch_fs *c, subvol_inum inum)
 {
-	wait_queue_head_t *wq;
-	DEFINE_WAIT_BIT(wait, &inode->i_state, __I_NEW);
-	wq = bit_waitqueue(&inode->i_state, __I_NEW);
-	prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE);
-	spin_unlock(&inode->i_lock);
-	schedule();
-	finish_wait(wq, &wait.wq_entry);
+	return rhashtable_lookup_fast(&c->vfs_inodes_table, &inum, bch2_vfs_inodes_params);
 }
 
-struct bch_inode_info *__bch2_inode_hash_find(struct bch_fs *c, subvol_inum inum)
+static void __wait_on_freeing_inode(struct bch_fs *c,
+				    struct bch_inode_info *inode,
+				    subvol_inum inum)
 {
-	return rhashtable_lookup_fast(&c->vfs_inodes_table, &inum, bch2_vfs_inodes_params);
+	wait_queue_head_t *wq;
+	DEFINE_WAIT_BIT(wait, &inode->v.i_state, __I_NEW);
+	wq = inode_bit_waitqueue(&wait, &inode->v, __I_NEW);
+	prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE);
+	spin_unlock(&inode->v.i_lock);
+
+	if (__bch2_inode_hash_find(c, inum) == inode)
+		schedule_timeout(HZ * 10);
+	finish_wait(wq, &wait.wq_entry);
 }
 
 static struct bch_inode_info *bch2_inode_hash_find(struct bch_fs *c, struct btree_trans *trans,
@@ -204,10 +208,10 @@ static struct bch_inode_info *bch2_inode_hash_find(struct bch_fs *c, struct btre
 		}
 		if ((inode->v.i_state & (I_FREEING|I_WILL_FREE))) {
 			if (!trans) {
-				__wait_on_freeing_inode(&inode->v);
+				__wait_on_freeing_inode(c, inode, inum);
 			} else {
 				bch2_trans_unlock(trans);
-				__wait_on_freeing_inode(&inode->v);
+				__wait_on_freeing_inode(c, inode, inum);
 				int ret = bch2_trans_relock(trans);
 				if (ret)
 					return ERR_PTR(ret);
@@ -232,6 +236,11 @@ static void bch2_inode_hash_remove(struct bch_fs *c, struct bch_inode_info *inod
 					&inode->hash, bch2_vfs_inodes_params);
 		BUG_ON(ret);
 		inode->v.i_hash.pprev = NULL;
+		/*
+		 * This pairs with the bch2_inode_hash_find() ->
+		 * __wait_on_freeing_inode() path
+		 */
+		inode_wake_up_bit(&inode->v, __I_NEW);
 	}
 }