erofs: remove a superfluous check for encoded extents

It is possible when an inode is split into segments for multi-threaded
compression, and the tail extent of a segment could also be small.

Fixes: 1d191b4 ("erofs: implement encoded extent metadata")
Signed-off-by: Gao Xiang <[email protected]>
Link: https://lore.kernel.org/r/[email protected]

Author: hsiangkao

fs/erofs/zmap.c

@@ -639,12 +639,6 @@ static int z_erofs_map_blocks_ext(struct inode *inode,
 		}
 	}
 	map->m_llen = lend - map->m_la;
-	if (!last && map->m_llen < sb->s_blocksize) {
-		erofs_err(sb, "extent too small %llu @ offset %llu of nid %llu",
-			  map->m_llen, map->m_la, vi->nid);
-		DBG_BUGON(1);
-		return -EFSCORRUPTED;
-	}
 	return 0;
 }