[PATCH] btrfs: Minor improvements to should_alloc_chunk

[PATCH] btrfs: Minor improvements to should_alloc_chunk

[Nikolay Borisov]

Since it's a predicate function make it explicitly return boolean. Also
the  'thresh' variable is only used when force is CHUNK_ALLOC_LIMITED so
reduce the scope of the variable as necessary. Finally, remove the + 2m
used in the final check. Given the granularity of btrfs' allocation I
doubt that the + 2m made a difference when making a decision whether to
allocate a chunk or not.

Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---
 fs/btrfs/block-group.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 3eecbc2b3dae..613527733fb2 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -3312,30 +3312,29 @@ static void force_metadata_allocation(struct btrfs_fs_info *info)
 	}
 }
 
-static int should_alloc_chunk(struct btrfs_fs_info *fs_info,
+static bool should_alloc_chunk(struct btrfs_fs_info *fs_info,
 			      struct btrfs_space_info *sinfo, int force)
 {
 	u64 bytes_used = btrfs_space_info_used(sinfo, false);
-	u64 thresh;
 
 	if (force == CHUNK_ALLOC_FORCE)
-		return 1;
+		return true;
 
 	/*
 	 * in limited mode, we want to have some free space up to
 	 * about 1% of the FS size.
 	 */
 	if (force == CHUNK_ALLOC_LIMITED) {
-		thresh = btrfs_super_total_bytes(fs_info->super_copy);
+		u64 thresh = btrfs_super_total_bytes(fs_info->super_copy);
 		thresh = max_t(u64, SZ_64M, div_factor_fine(thresh, 1));
 
 		if (sinfo->total_bytes - bytes_used < thresh)
-			return 1;
+			return true;
 	}
 
-	if (bytes_used + SZ_2M < div_factor(sinfo->total_bytes, 8))
-		return 0;
-	return 1;
+	if (bytes_used < div_factor(sinfo->total_bytes, 8))
+		return false;
+	return true;
 }
 
 int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, u64 type)
-- 
2.17.1

Re: [PATCH] btrfs: Minor improvements to should_alloc_chunk

[David Sterba]

On Fri, Jul 02, 2021 at 04:02:06PM +0300, Nikolay Borisov wrote:
> Since it's a predicate function make it explicitly return boolean. Also
> the  'thresh' variable is only used when force is CHUNK_ALLOC_LIMITED so
> reduce the scope of the variable as necessary. Finally, remove the + 2m
> used in the final check. Given the granularity of btrfs' allocation I
> doubt that the + 2m made a difference when making a decision whether to
> allocate a chunk or not.

This is mixing 2 cleanups and potentially a functional change in one
patch, "I doubt it made a difference" does not sound like a good
reasoning, I can say that I doubt you're right so what now? :)

Doing the cleanups (return value and moving the scope) in one is
probably fine, but the +2M change needs some explanation and should go
to a separate patch.