On Mon, Dec 16, 2013 at 01:04:13AM -0800, Hugh Dickins wrote: > It's silly to force the 64-bit CONFIG_GENERIC_LOCKBREAK architectures > to kmalloc eight bytes for an indirect page table lock: the lock needs > to fit in the space that a pointer to it would occupy, not into an int. Ah, no. A spinlock is very much assumed to be 32bit, any spinlock that's bigger than that is bloated. For the page-frame case we do indeed not care about the strict 32bit but more about not being larger than a pointer, however there are already other users. See for instance include/linux/lockref.h and lib/lockref.c, they very much require the spinlock to be 32bit and the below would break that. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>