On Tue, Sep 03, 2019 at 05:00:35PM +0200, Johannes Thumshirn wrote: > The callers of csum_tree_block_size() blindly assume we're only having > crc32c as a possible checksum and thus pass in > btrfs_csum_sizes[BTRFS_CSUM_TYPE_CRC32] for the size argument of > csum_tree_block_size(). > > Signed-off-by: Johannes Thumshirn <jthumsh...@suse.de> > Reviewed-by: Nikolay Borisov <nbori...@suse.com> > --- > mkfs/common.c | 14 +++++++------- > mkfs/common.h | 2 ++ > 2 files changed, 9 insertions(+), 7 deletions(-) > > diff --git a/mkfs/common.c b/mkfs/common.c > index caca5e707233..b6e549b19272 100644 > --- a/mkfs/common.c > +++ b/mkfs/common.c > @@ -101,7 +101,7 @@ static int btrfs_create_tree_root(int fd, struct > btrfs_mkfs_config *cfg, > } > > /* generate checksum */ > - csum_tree_block_size(buf, btrfs_csum_sizes[BTRFS_CSUM_TYPE_CRC32], 0); > + csum_tree_block_size(buf, btrfs_csum_sizes[cfg->csum_type], 0);
I don't see where cfg->csum_type is initialized. The tests pass so there's probably some implicit initialization to 0 that makes it work.
