for-5.14-rc2-tag

-----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCgAdFiEE8rQSAMVO+zA4DBdWxWXV+ddtWDsFAmD7GZUACgkQxWXV+ddt
 WDs+BA/+OHDY2ROYEnysAqF1qaDENVVnUavnDYYa+Uk61KVvx0pm/mHY9SllsuH4
 WCIwCwH7LZs11cRYp3vD80t4OdVGBKaDvEfX+znMCQYuoBm6G5eT3n5jhsVFr1jJ
 EqUVzUY+S44IWAEhzkVnSAD4xnMsan8b+YnngIFSMEqJlH+on6w8FhyP0QXwInxk
 1kfjl8tDMiryKFaekfGX5WaeflEeWGoHNf2xYokzPD/Oq6TCaoLycar1YXH+80FM
 05Jl0+jfEWbaHouMNd8bW9nHnSxh30i7gorY17Q6KLOFDCThNiKZuypZsQcCi/df
 TbjQDNTZjSsReFvrFeFlEdGv3dFHBGxz1Ns7RFPfVeNgmN0WnOLmzS+4rmfGyi8L
 +3TQ6MGqgG0DppPwfB9caDvxYsbN23uA1v5J1B+Dsbo47lFWWIoBQBtDvErAiHEy
 KF7B4jIOWrx3ZYwv3pkE3D+D19sKkB9+wLnlwVSF77npKO1up8W0h4mPdMLZaznW
 TGBXxwqI4105MSX5UatBpX+HYATpEWG5tmeZz5ERGFNC/piILmY4iVz/c5Vguh9/
 iUQwjSudIDWgGxcL7VClqrdF7sucsml6Svb+ZrxckmK7pa97TG2bIlzJDg0eFcle
 NBcw8RBcBMUay/Y04cKHLJAj6OOjBiXnxKjjHrhvtaBmOV2SHpc=
 =kj2e
 -----END PGP SIGNATURE-----

Merge tag 'for-5.14-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "A few fixes and one patch to help some block layer API cleanups:

   - skip missing device when running fstrim

   - fix unpersisted i_size on fsync after expanding truncate

   - fix lock inversion problem when doing qgroup extent tracing

   - replace bdgrab/bdput usage, replace gendisk by block_device"

* tag 'for-5.14-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: store a block_device in struct btrfs_ordered_extent
  btrfs: fix lock inversion problem when doing qgroup extent tracing
  btrfs: check for missing device in btrfs_trim_fs
  btrfs: fix unpersisted i_size on fsync after expanding truncate

fs/btrfs/backref.c

@@ -1488,15 +1488,15 @@ static int btrfs_find_all_roots_safe(struct btrfs_trans_handle *trans,
 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
 			 struct btrfs_fs_info *fs_info, u64 bytenr,
 			 u64 time_seq, struct ulist **roots,
-			 bool ignore_offset)
+			 bool ignore_offset, bool skip_commit_root_sem)
 {
 	int ret;
 
-	if (!trans)
+	if (!trans && !skip_commit_root_sem)
 		down_read(&fs_info->commit_root_sem);
 	ret = btrfs_find_all_roots_safe(trans, fs_info, bytenr,
 					time_seq, roots, ignore_offset);
-	if (!trans)
+	if (!trans && !skip_commit_root_sem)
 		up_read(&fs_info->commit_root_sem);
 	return ret;
 }

fs/btrfs/backref.h

@@ -47,7 +47,8 @@ int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
 			 const u64 *extent_item_pos, bool ignore_offset);
 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
 			 struct btrfs_fs_info *fs_info, u64 bytenr,
-			 u64 time_seq, struct ulist **roots, bool ignore_offset);
+			 u64 time_seq, struct ulist **roots, bool ignore_offset,
+			 bool skip_commit_root_sem);
 char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
 			u32 name_len, unsigned long name_off,
 			struct extent_buffer *eb_in, u64 parent,

fs/btrfs/delayed-ref.c

@@ -974,7 +974,7 @@ int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
 		kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
 
 	if (qrecord_inserted)
-		btrfs_qgroup_trace_extent_post(fs_info, record);
+		btrfs_qgroup_trace_extent_post(trans, record);
 
 	return 0;
 }
@@ -1069,7 +1069,7 @@ int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans,
 
 
 	if (qrecord_inserted)
-		return btrfs_qgroup_trace_extent_post(fs_info, record);
+		return btrfs_qgroup_trace_extent_post(trans, record);
 	return 0;
 }
 

fs/btrfs/extent-tree.c

@@ -6019,6 +6019,9 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 	mutex_lock(&fs_info->fs_devices->device_list_mutex);
 	devices = &fs_info->fs_devices->devices;
 	list_for_each_entry(device, devices, dev_list) {
+		if (test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state))
+			continue;
+
 		ret = btrfs_trim_free_extents(device, &group_trimmed);
 		if (ret) {
 			dev_failed++;

fs/btrfs/inode.c

@@ -2992,7 +2992,7 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent)
 		goto out;
 	}
 
-	if (ordered_extent->disk)
+	if (ordered_extent->bdev)
 		btrfs_rewrite_logical_zoned(ordered_extent);
 
 	btrfs_free_io_failure_record(inode, start, end);

fs/btrfs/ordered-data.c

@@ -190,8 +190,6 @@ static int __btrfs_add_ordered_extent(struct btrfs_inode *inode, u64 file_offset
 	entry->truncated_len = (u64)-1;
 	entry->qgroup_rsv = ret;
 	entry->physical = (u64)-1;
-	entry->disk = NULL;
-	entry->partno = (u8)-1;
 
 	ASSERT(type == BTRFS_ORDERED_REGULAR ||
 	       type == BTRFS_ORDERED_NOCOW ||

fs/btrfs/ordered-data.h

@@ -145,8 +145,7 @@ struct btrfs_ordered_extent {
 	 * command in a workqueue context
 	 */
 	u64 physical;
-	struct gendisk *disk;
-	u8 partno;
+	struct block_device *bdev;
 };
 
 /*

fs/btrfs/qgroup.c

@@ -1704,17 +1704,39 @@ int btrfs_qgroup_trace_extent_nolock(struct btrfs_fs_info *fs_info,
 	return 0;
 }
 
-int btrfs_qgroup_trace_extent_post(struct btrfs_fs_info *fs_info,
+int btrfs_qgroup_trace_extent_post(struct btrfs_trans_handle *trans,
 				   struct btrfs_qgroup_extent_record *qrecord)
 {
 	struct ulist *old_root;
 	u64 bytenr = qrecord->bytenr;
 	int ret;
 
-	ret = btrfs_find_all_roots(NULL, fs_info, bytenr, 0, &old_root, false);
+	/*
+	 * We are always called in a context where we are already holding a
+	 * transaction handle. Often we are called when adding a data delayed
+	 * reference from btrfs_truncate_inode_items() (truncating or unlinking),
+	 * in which case we will be holding a write lock on extent buffer from a
+	 * subvolume tree. In this case we can't allow btrfs_find_all_roots() to
+	 * acquire fs_info->commit_root_sem, because that is a higher level lock
+	 * that must be acquired before locking any extent buffers.
+	 *
+	 * So we want btrfs_find_all_roots() to not acquire the commit_root_sem
+	 * but we can't pass it a non-NULL transaction handle, because otherwise
+	 * it would not use commit roots and would lock extent buffers, causing
+	 * a deadlock if it ends up trying to read lock the same extent buffer
+	 * that was previously write locked at btrfs_truncate_inode_items().
+	 *
+	 * So pass a NULL transaction handle to btrfs_find_all_roots() and
+	 * explicitly tell it to not acquire the commit_root_sem - if we are
+	 * holding a transaction handle we don't need its protection.
+	 */
+	ASSERT(trans != NULL);
+
+	ret = btrfs_find_all_roots(NULL, trans->fs_info, bytenr, 0, &old_root,
+				   false, true);
 	if (ret < 0) {
-		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
-		btrfs_warn(fs_info,
+		trans->fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
+		btrfs_warn(trans->fs_info,
 "error accounting new delayed refs extent (err code: %d), quota inconsistent",
 			ret);
 		return 0;
@@ -1758,7 +1780,7 @@ int btrfs_qgroup_trace_extent(struct btrfs_trans_handle *trans, u64 bytenr,
 		kfree(record);
 		return 0;
 	}
-	return btrfs_qgroup_trace_extent_post(fs_info, record);
+	return btrfs_qgroup_trace_extent_post(trans, record);
 }
 
 int btrfs_qgroup_trace_leaf_items(struct btrfs_trans_handle *trans,
@@ -2629,7 +2651,7 @@ int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans)
 				/* Search commit root to find old_roots */
 				ret = btrfs_find_all_roots(NULL, fs_info,
 						record->bytenr, 0,
-						&record->old_roots, false);
+						&record->old_roots, false, false);
 				if (ret < 0)
 					goto cleanup;
 			}
@@ -2645,7 +2667,7 @@ int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans)
 			 * current root. It's safe inside commit_transaction().
 			 */
 			ret = btrfs_find_all_roots(trans, fs_info,
-				record->bytenr, BTRFS_SEQ_LAST, &new_roots, false);
+			   record->bytenr, BTRFS_SEQ_LAST, &new_roots, false, false);
 			if (ret < 0)
 				goto cleanup;
 			if (qgroup_to_skip) {
@@ -3179,7 +3201,7 @@ static int qgroup_rescan_leaf(struct btrfs_trans_handle *trans,
 			num_bytes = found.offset;
 
 		ret = btrfs_find_all_roots(NULL, fs_info, found.objectid, 0,
-					   &roots, false);
+					   &roots, false, false);
 		if (ret < 0)
 			goto out;
 		/* For rescan, just pass old_roots as NULL */

fs/btrfs/qgroup.h

@@ -298,7 +298,7 @@ int btrfs_qgroup_trace_extent_nolock(
  * using current root, then we can move all expensive backref walk out of
  * transaction committing, but not now as qgroup accounting will be wrong again.
  */
-int btrfs_qgroup_trace_extent_post(struct btrfs_fs_info *fs_info,
+int btrfs_qgroup_trace_extent_post(struct btrfs_trans_handle *trans,
 				   struct btrfs_qgroup_extent_record *qrecord);
 
 /*

fs/btrfs/tests/qgroup-tests.c

@@ -224,7 +224,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root,
 	 * quota.
 	 */
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		test_err("couldn't find old roots: %d", ret);
@@ -237,7 +237,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root,
 		return ret;
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		ulist_free(new_roots);
@@ -261,7 +261,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root,
 	new_roots = NULL;
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		test_err("couldn't find old roots: %d", ret);
@@ -273,7 +273,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root,
 		return -EINVAL;
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		ulist_free(new_roots);
@@ -325,7 +325,7 @@ static int test_multiple_refs(struct btrfs_root *root,
 	}
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		test_err("couldn't find old roots: %d", ret);
@@ -338,7 +338,7 @@ static int test_multiple_refs(struct btrfs_root *root,
 		return ret;
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		ulist_free(new_roots);
@@ -360,7 +360,7 @@ static int test_multiple_refs(struct btrfs_root *root,
 	}
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		test_err("couldn't find old roots: %d", ret);
@@ -373,7 +373,7 @@ static int test_multiple_refs(struct btrfs_root *root,
 		return ret;
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		ulist_free(new_roots);
@@ -401,7 +401,7 @@ static int test_multiple_refs(struct btrfs_root *root,
 	}
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		test_err("couldn't find old roots: %d", ret);
@@ -414,7 +414,7 @@ static int test_multiple_refs(struct btrfs_root *root,
 		return ret;
 
 	ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots,
-			false);
+			false, false);
 	if (ret) {
 		ulist_free(old_roots);
 		ulist_free(new_roots);

fs/btrfs/tree-log.c

@@ -5526,16 +5526,29 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 		spin_lock(&inode->lock);
 		inode->logged_trans = trans->transid;
 		/*
-		 * Don't update last_log_commit if we logged that an inode exists
-		 * after it was loaded to memory (full_sync bit set).
-		 * This is to prevent data loss when we do a write to the inode,
-		 * then the inode gets evicted after all delalloc was flushed,
-		 * then we log it exists (due to a rename for example) and then
-		 * fsync it. This last fsync would do nothing (not logging the
-		 * extents previously written).
+		 * Don't update last_log_commit if we logged that an inode exists.
+		 * We do this for two reasons:
+		 *
+		 * 1) We might have had buffered writes to this inode that were
+		 *    flushed and had their ordered extents completed in this
+		 *    transaction, but we did not previously log the inode with
+		 *    LOG_INODE_ALL. Later the inode was evicted and after that
+		 *    it was loaded again and this LOG_INODE_EXISTS log operation
+		 *    happened. We must make sure that if an explicit fsync against
+		 *    the inode is performed later, it logs the new extents, an
+		 *    updated inode item, etc, and syncs the log. The same logic
+		 *    applies to direct IO writes instead of buffered writes.
+		 *
+		 * 2) When we log the inode with LOG_INODE_EXISTS, its inode item
+		 *    is logged with an i_size of 0 or whatever value was logged
+		 *    before. If later the i_size of the inode is increased by a
+		 *    truncate operation, the log is synced through an fsync of
+		 *    some other inode and then finally an explicit fsync against
+		 *    this inode is made, we must make sure this fsync logs the
+		 *    inode with the new i_size, the hole between old i_size and
+		 *    the new i_size, and syncs the log.
 		 */
-		if (inode_only != LOG_INODE_EXISTS ||
-		    !test_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &inode->runtime_flags))
+		if (inode_only != LOG_INODE_EXISTS)
 			inode->last_log_commit = inode->last_sub_trans;
 		spin_unlock(&inode->lock);
 	}

fs/btrfs/zoned.c

@@ -1349,8 +1349,7 @@ void btrfs_record_physical_zoned(struct inode *inode, u64 file_offset,
 		return;
 
 	ordered->physical = physical;
-	ordered->disk = bio->bi_bdev->bd_disk;
-	ordered->partno = bio->bi_bdev->bd_partno;
+	ordered->bdev = bio->bi_bdev;
 
 	btrfs_put_ordered_extent(ordered);
 }
@@ -1362,18 +1361,16 @@ void btrfs_rewrite_logical_zoned(struct btrfs_ordered_extent *ordered)
 	struct extent_map_tree *em_tree;
 	struct extent_map *em;
 	struct btrfs_ordered_sum *sum;
-	struct block_device *bdev;
 	u64 orig_logical = ordered->disk_bytenr;
 	u64 *logical = NULL;
 	int nr, stripe_len;
 
 	/* Zoned devices should not have partitions. So, we can assume it is 0 */
-	ASSERT(ordered->partno == 0);
-	bdev = bdgrab(ordered->disk->part0);
-	if (WARN_ON(!bdev))
+	ASSERT(!bdev_is_partition(ordered->bdev));
+	if (WARN_ON(!ordered->bdev))
 		return;
 
-	if (WARN_ON(btrfs_rmap_block(fs_info, orig_logical, bdev,
+	if (WARN_ON(btrfs_rmap_block(fs_info, orig_logical, ordered->bdev,
 				     ordered->physical, &logical, &nr,
 				     &stripe_len)))
 		goto out;
@@ -1402,7 +1399,6 @@ void btrfs_rewrite_logical_zoned(struct btrfs_ordered_extent *ordered)
 
 out:
 	kfree(logical);
-	bdput(bdev);
 }
 
 bool btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,