btrfs: don't clear async_chunk->inode in async_cow_start
authorChristoph Hellwig <hch@lst.de>
Wed, 28 Jun 2023 15:31:33 +0000 (17:31 +0200)
committerDavid Sterba <dsterba@suse.com>
Mon, 21 Aug 2023 12:52:15 +0000 (14:52 +0200)
Now that the ->inode check isn't needed in submit_compressed_extents
any more, there is no reason to clear the field early.  Always keep
the inode around until the work item is finished and remove the special
casing, and the counting of compressed extents in compress_file_range.

Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/inode.c

index 27fdfd08a3ef4e46c951006c538c0816aee1fbfd..76496b084f5e6d0d108eb9e2fa5a0e64d1c1c573 100644 (file)
@@ -830,7 +830,7 @@ static inline void inode_should_defrag(struct btrfs_inode *inode,
  * are written in the same order that the flusher thread sent them
  * down.
  */
-static noinline int compress_file_range(struct async_chunk *async_chunk)
+static noinline void compress_file_range(struct async_chunk *async_chunk)
 {
        struct btrfs_inode *inode = async_chunk->inode;
        struct btrfs_fs_info *fs_info = inode->root->fs_info;
@@ -848,7 +848,6 @@ static noinline int compress_file_range(struct async_chunk *async_chunk)
        int i;
        int will_compress;
        int compress_type = fs_info->compress_type;
-       int compressed_extents = 0;
        int redirty = 0;
 
        inode_should_defrag(inode, start, end, end - start + 1, SZ_16K);
@@ -1025,7 +1024,7 @@ cont:
                                }
                                kfree(pages);
                        }
-                       return 0;
+                       return;
                }
        }
 
@@ -1044,8 +1043,6 @@ cont:
                 */
                total_in = round_up(total_in, fs_info->sectorsize);
                if (total_compressed + blocksize <= total_in) {
-                       compressed_extents++;
-
                        /*
                         * The async work queues will take care of doing actual
                         * allocation on disk for these compressed pages, and
@@ -1061,7 +1058,7 @@ cont:
                                cond_resched();
                                goto again;
                        }
-                       return compressed_extents;
+                       return;
                }
        }
        if (pages) {
@@ -1102,9 +1099,6 @@ cleanup_and_bail_uncompressed:
                extent_range_redirty_for_io(&inode->vfs_inode, start, end);
        add_async_extent(async_chunk, start, end - start + 1, 0, NULL, 0,
                         BTRFS_COMPRESS_NONE);
-       compressed_extents++;
-
-       return compressed_extents;
 }
 
 static void free_async_extent_pages(struct async_extent *async_extent)
@@ -1655,15 +1649,9 @@ out_unlock:
 static noinline void async_cow_start(struct btrfs_work *work)
 {
        struct async_chunk *async_chunk;
-       int compressed_extents;
 
        async_chunk = container_of(work, struct async_chunk, work);
-
-       compressed_extents = compress_file_range(async_chunk);
-       if (compressed_extents == 0) {
-               btrfs_add_delayed_iput(async_chunk->inode);
-               async_chunk->inode = NULL;
-       }
+       compress_file_range(async_chunk);
 }
 
 /*
@@ -1700,8 +1688,7 @@ static noinline void async_cow_free(struct btrfs_work *work)
        struct async_cow *async_cow;
 
        async_chunk = container_of(work, struct async_chunk, work);
-       if (async_chunk->inode)
-               btrfs_add_delayed_iput(async_chunk->inode);
+       btrfs_add_delayed_iput(async_chunk->inode);
        if (async_chunk->blkcg_css)
                css_put(async_chunk->blkcg_css);