summaryrefslogtreecommitdiffstats
path: root/fs/btrfs/inode.c
diff options
context:
space:
mode:
authorQu Wenruo <wqu@suse.com>2022-09-16 15:28:38 +0800
committerDavid Sterba <dsterba@suse.com>2022-12-05 18:00:48 +0100
commit280f15cb96a61e4122bb28cdc316343ff4918b7d (patch)
treec810207bda1679687401644eb1776527916917a2 /fs/btrfs/inode.c
parenta196a8944f77b7b762795a0862d8aa4a005625a4 (diff)
downloadlinux-280f15cb96a61e4122bb28cdc316343ff4918b7d.tar.bz2
btrfs: remove new_inline argument from btrfs_extent_item_to_extent_map()
The argument @new_inline changes the following members of extent_map: - em->compress_type - EXTENT_FLAG_COMPRESSED of em->flags However neither members makes a difference for inline extents: - Inline extent read never use above em members As inside btrfs_get_extent() we directly use the file extent item to do the read. - Inline extents are never to be split Thus code really needs em->compress_type or that flag will never be executed on inlined extents. (btrfs_drop_extent_cache() would be one example) - Fiemap no longer relies on extent maps Recent fiemap optimization makes fiemap to search subvolume tree directly, without using any extent map at all. Thus those members make no difference for inline extents any more. Furthermore such exception without much explanation is really a source of confusion. Thus this patch will completely remove the argument, and always set the involved members, unifying the behavior. Signed-off-by: Qu Wenruo <wqu@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
Diffstat (limited to 'fs/btrfs/inode.c')
-rw-r--r--fs/btrfs/inode.c2
1 files changed, 1 insertions, 1 deletions
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 5b45eb141215..0ab899132dbf 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -7058,7 +7058,7 @@ next:
goto insert;
}
- btrfs_extent_item_to_extent_map(inode, path, item, !page, em);
+ btrfs_extent_item_to_extent_map(inode, path, item, em);
if (extent_type == BTRFS_FILE_EXTENT_REG ||
extent_type == BTRFS_FILE_EXTENT_PREALLOC) {