btrfs: defer splitting of ordered extents until I/O completion

The btrfs zoned completion code currently needs an ordered_extent and
extent_map per bio so that it can account for the non-predictable
write location from Zone Append.  To archive that it currently splits
the ordered_extent and extent_map at I/O submission time, and then
records the actual physical address in the ->physical field of the
ordered_extent.

This patch instead switches to record the "original" physical address
that the btrfs allocator assigned in spare space in the btrfs_bio,
and then rewrites the logical address in the btrfs_ordered_sum
structure at I/O completion time.  This allows the ordered extent
completion handler to simply walk the list of ordered csums and
split the ordered extent as needed.  This removes an extra ordered
extent and extent_map lookup and manipulation during the I/O
submission path, and instead batches it in the I/O completion path
where we need to touch these anyway.

Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: David Sterba <dsterba@suse.com>

1 files changed, 58 insertions, 12 deletions

diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index b55b0d4ee86f..b4dd018ba332 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -15,6 +15,7 @@
 #include "transaction.h"
 #include "dev-replace.h"
 #include "space-info.h"
+#include "super.h"
 #include "fs.h"
 #include "accessors.h"
 #include "bio.h"
@@ -1665,17 +1666,11 @@ void btrfs_record_physical_zoned(struct btrfs_bio *bbio)
 		sum->logical += physical - bbio->orig_physical;
 }
 
-void btrfs_rewrite_logical_zoned(struct btrfs_ordered_extent *ordered)
+static void btrfs_rewrite_logical_zoned(struct btrfs_ordered_extent *ordered,
+					u64 logical)
 {
-	struct btrfs_inode *inode = BTRFS_I(ordered->inode);
-	struct extent_map_tree *em_tree = &inode->extent_tree;
+	struct extent_map_tree *em_tree = &BTRFS_I(ordered->inode)->extent_tree;
 	struct extent_map *em;
-	struct btrfs_ordered_sum *sum =
-		list_first_entry(&ordered->list, typeof(*sum), list);
-	u64 logical = sum->logical;
-
-	if (ordered->disk_bytenr == logical)
-		goto out;
 
 	ordered->disk_bytenr = logical;
 
@@ -1685,6 +1680,54 @@ void btrfs_rewrite_logical_zoned(struct btrfs_ordered_extent *ordered)
 	em->block_start = logical;
 	free_extent_map(em);
 	write_unlock(&em_tree->lock);
+}
+
+static bool btrfs_zoned_split_ordered(struct btrfs_ordered_extent *ordered,
+				      u64 logical, u64 len)
+{
+	struct btrfs_ordered_extent *new;
+
+	if (!test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags) &&
+	    split_extent_map(BTRFS_I(ordered->inode), ordered->file_offset,
+			     ordered->num_bytes, len))
+		return false;
+
+	new = btrfs_split_ordered_extent(ordered, len);
+	if (IS_ERR(new))
+		return false;
+
+	if (new->disk_bytenr != logical)
+		btrfs_rewrite_logical_zoned(new, logical);
+	btrfs_finish_one_ordered(new);
+	return true;
+}
+
+void btrfs_finish_ordered_zoned(struct btrfs_ordered_extent *ordered)
+{
+	struct btrfs_inode *inode = BTRFS_I(ordered->inode);
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
+	struct btrfs_ordered_sum *sum =
+		list_first_entry(&ordered->list, typeof(*sum), list);
+	u64 logical = sum->logical;
+	u64 len = sum->len;
+
+	while (len < ordered->disk_num_bytes) {
+		sum = list_next_entry(sum, list);
+		if (sum->logical == logical + len) {
+			len += sum->len;
+			continue;
+		}
+		if (!btrfs_zoned_split_ordered(ordered, logical, len)) {
+			set_bit(BTRFS_ORDERED_IOERR, &ordered->flags);
+			btrfs_err(fs_info, "failed to split ordered extent");
+			goto out;
+		}
+		logical = sum->logical;
+		len = sum->len;
+	}
+
+	if (ordered->disk_bytenr != logical)
+		btrfs_rewrite_logical_zoned(ordered, logical);
 
 out:
 	/*
@@ -1694,9 +1737,12 @@ out:
 	 * here so that we don't attempt to log the csums later.
 	 */
 	if ((inode->flags & BTRFS_INODE_NODATASUM) ||
-	    test_bit(BTRFS_FS_STATE_NO_CSUMS, &inode->root->fs_info->fs_state)) {
-		list_del(&sum->list);
-		kfree(sum);
+	    test_bit(BTRFS_FS_STATE_NO_CSUMS, &fs_info->fs_state)) {
+		while ((sum = list_first_entry_or_null(&ordered->list,
+						       typeof(*sum), list))) {
+			list_del(&sum->list);
+			kfree(sum);
+		}
 	}
 }