The technical change looks good, but this comment not only has quite a few of grammar issues, but also seems rather pointless: > + /* > + * New open() will be failed since disk becomes not alive, and old > + * open() has either grabbed the module refcnt or been failed in > + * case of deleting from module_exit(), so disk->fops->owner won't > + * be unloaded if the disk is opened. > + */ > mutex_lock(&disk->open_mutex); > remove_inode_hash(disk->part0->bd_inode); > blk_drop_partitions(disk); > -- > 2.31.1 ---end quoted