Merge tag 'for-6.11-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:

 - followup fix for direct io and fsync under some conditions, reported
   by QEMU users

 - fix a potential leak when disabling quotas while some extent tracking
   work can still happen

 - in zoned mode handle unexpected change of zone write pointer in
   RAID1-like block groups, turn the zones to read-only

* tag 'for-6.11-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: fix race between direct IO write and fsync when using same fd
  btrfs: zoned: handle broken write pointer on zones
  btrfs: qgroup: don't use extent changeset when not needed

Author: torvalds

fs/btrfs/ctree.h

@@ -459,7 +459,6 @@ struct btrfs_file_private {
 	void *filldir_buf;
 	u64 last_index;
 	struct extent_state *llseek_cached_state;
-	bool fsync_skip_inode_lock;
 };
 
 static inline u32 BTRFS_LEAF_DATA_SIZE(const struct btrfs_fs_info *info)

fs/btrfs/direct-io.c

@@ -864,13 +864,6 @@ ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 	if (IS_ERR_OR_NULL(dio)) {
 		ret = PTR_ERR_OR_ZERO(dio);
 	} else {
-		struct btrfs_file_private stack_private = { 0 };
-		struct btrfs_file_private *private;
-		const bool have_private = (file->private_data != NULL);
-
-		if (!have_private)
-			file->private_data = &stack_private;
-
 		/*
 		 * If we have a synchronous write, we must make sure the fsync
 		 * triggered by the iomap_dio_complete() call below doesn't
@@ -879,13 +872,10 @@ ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 		 * partial writes due to the input buffer (or parts of it) not
 		 * being already faulted in.
 		 */
-		private = file->private_data;
-		private->fsync_skip_inode_lock = true;
+		ASSERT(current->journal_info == NULL);
+		current->journal_info = BTRFS_TRANS_DIO_WRITE_STUB;
 		ret = iomap_dio_complete(dio);
-		private->fsync_skip_inode_lock = false;
-
-		if (!have_private)
-			file->private_data = NULL;
+		current->journal_info = NULL;
 	}
 
 	/* No increment (+=) because iomap returns a cumulative value. */

fs/btrfs/file.c

@@ -1603,7 +1603,6 @@ static inline bool skip_inode_logging(const struct btrfs_log_ctx *ctx)
  */
 int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 {
-	struct btrfs_file_private *private = file->private_data;
 	struct dentry *dentry = file_dentry(file);
 	struct btrfs_inode *inode = BTRFS_I(d_inode(dentry));
 	struct btrfs_root *root = inode->root;
@@ -1613,7 +1612,13 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	int ret = 0, err;
 	u64 len;
 	bool full_sync;
-	const bool skip_ilock = (private ? private->fsync_skip_inode_lock : false);
+	bool skip_ilock = false;
+
+	if (current->journal_info == BTRFS_TRANS_DIO_WRITE_STUB) {
+		skip_ilock = true;
+		current->journal_info = NULL;
+		lockdep_assert_held(&inode->vfs_inode.i_rwsem);
+	}
 
 	trace_btrfs_sync_file(file, datasync);
 

fs/btrfs/qgroup.c

@@ -4346,10 +4346,9 @@ static int __btrfs_qgroup_release_data(struct btrfs_inode *inode,
 	int ret;
 
 	if (btrfs_qgroup_mode(inode->root->fs_info) == BTRFS_QGROUP_MODE_DISABLED) {
-		extent_changeset_init(&changeset);
 		return clear_record_extent_bits(&inode->io_tree, start,
 						start + len - 1,
-						EXTENT_QGROUP_RESERVED, &changeset);
+						EXTENT_QGROUP_RESERVED, NULL);
 	}
 
 	/* In release case, we shouldn't have @reserved */

fs/btrfs/transaction.h

@@ -27,6 +27,12 @@ struct btrfs_root_item;
 struct btrfs_root;
 struct btrfs_path;
 
+/*
+ * Signal that a direct IO write is in progress, to avoid deadlock for sync
+ * direct IO writes when fsync is called during the direct IO write path.
+ */
+#define BTRFS_TRANS_DIO_WRITE_STUB	((void *) 1)
+
 /* Radix-tree tag for roots that are part of the trasaction. */
 #define BTRFS_ROOT_TRANS_TAG			0
 

fs/btrfs/zoned.c

@@ -1406,6 +1406,8 @@ static int btrfs_load_block_group_dup(struct btrfs_block_group *bg,
 		return -EINVAL;
 	}
 
+	bg->zone_capacity = min_not_zero(zone_info[0].capacity, zone_info[1].capacity);
+
 	if (zone_info[0].alloc_offset == WP_MISSING_DEV) {
 		btrfs_err(bg->fs_info,
 			  "zoned: cannot recover write pointer for zone %llu",
@@ -1432,7 +1434,6 @@ static int btrfs_load_block_group_dup(struct btrfs_block_group *bg,
 	}
 
 	bg->alloc_offset = zone_info[0].alloc_offset;
-	bg->zone_capacity = min(zone_info[0].capacity, zone_info[1].capacity);
 	return 0;
 }
 
@@ -1450,6 +1451,9 @@ static int btrfs_load_block_group_raid1(struct btrfs_block_group *bg,
 		return -EINVAL;
 	}
 
+	/* In case a device is missing we have a cap of 0, so don't use it. */
+	bg->zone_capacity = min_not_zero(zone_info[0].capacity, zone_info[1].capacity);
+
 	for (i = 0; i < map->num_stripes; i++) {
 		if (zone_info[i].alloc_offset == WP_MISSING_DEV ||
 		    zone_info[i].alloc_offset == WP_CONVENTIONAL)
@@ -1471,9 +1475,6 @@ static int btrfs_load_block_group_raid1(struct btrfs_block_group *bg,
 			if (test_bit(0, active))
 				set_bit(BLOCK_GROUP_FLAG_ZONE_IS_ACTIVE, &bg->runtime_flags);
 		}
-		/* In case a device is missing we have a cap of 0, so don't use it. */
-		bg->zone_capacity = min_not_zero(zone_info[0].capacity,
-						 zone_info[1].capacity);
 	}
 
 	if (zone_info[0].alloc_offset != WP_MISSING_DEV)
@@ -1563,6 +1564,7 @@ int btrfs_load_block_group_zone_info(struct btrfs_block_group *cache, bool new)
 	unsigned long *active = NULL;
 	u64 last_alloc = 0;
 	u32 num_sequential = 0, num_conventional = 0;
+	u64 profile;
 
 	if (!btrfs_is_zoned(fs_info))
 		return 0;
@@ -1623,7 +1625,8 @@ int btrfs_load_block_group_zone_info(struct btrfs_block_group *cache, bool new)
 		}
 	}
 
-	switch (map->type & BTRFS_BLOCK_GROUP_PROFILE_MASK) {
+	profile = map->type & BTRFS_BLOCK_GROUP_PROFILE_MASK;
+	switch (profile) {
 	case 0: /* single */
 		ret = btrfs_load_block_group_single(cache, &zone_info[0], active);
 		break;
@@ -1650,6 +1653,23 @@ int btrfs_load_block_group_zone_info(struct btrfs_block_group *cache, bool new)
 		goto out;
 	}
 
+	if (ret == -EIO && profile != 0 && profile != BTRFS_BLOCK_GROUP_RAID0 &&
+	    profile != BTRFS_BLOCK_GROUP_RAID10) {
+		/*
+		 * Detected broken write pointer.  Make this block group
+		 * unallocatable by setting the allocation pointer at the end of
+		 * allocatable region. Relocating this block group will fix the
+		 * mismatch.
+		 *
+		 * Currently, we cannot handle RAID0 or RAID10 case like this
+		 * because we don't have a proper zone_capacity value. But,
+		 * reading from this block group won't work anyway by a missing
+		 * stripe.
+		 */
+		cache->alloc_offset = cache->zone_capacity;
+		ret = 0;
+	}
+
 out:
 	/* Reject non SINGLE data profiles without RST */
 	if ((map->type & BTRFS_BLOCK_GROUP_DATA) &&