btrfs: make csum_tree_block() handle node smaller than page

For subpage size support, metadata blocks of nodesize are smaller than
one page and this needs to be handled when calculating the checksum.

The checksummed start and length need to be adjusted but only for the
first page:

- start is simply offset in the page

- length is nodesize (subpage) or PAGE_SIZE for all other cases
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

fs/btrfs/disk-io.c

@@ -212,15 +212,16 @@ static void csum_tree_block(struct extent_buffer *buf, u8 *result)
 {
 	struct btrfs_fs_info *fs_info = buf->fs_info;
 	const int num_pages = fs_info->nodesize >> PAGE_SHIFT;
+	const int first_page_part = min_t(u32, PAGE_SIZE, fs_info->nodesize);
 	SHASH_DESC_ON_STACK(shash, fs_info->csum_shash);
 	char *kaddr;
 	int i;
 
 	shash->tfm = fs_info->csum_shash;
 	crypto_shash_init(shash);
-	kaddr = page_address(buf->pages[0]);
+	kaddr = page_address(buf->pages[0]) + offset_in_page(buf->start);
 	crypto_shash_update(shash, kaddr + BTRFS_CSUM_SIZE,
-			    PAGE_SIZE - BTRFS_CSUM_SIZE);
+			    first_page_part - BTRFS_CSUM_SIZE);
 
 	for (i = 1; i < num_pages; i++) {
 		kaddr = page_address(buf->pages[i]);