* KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> [2010-07-27 16:51:55]: > > From a view of patch management, this set is a mixture of a few features for > memcg, and I should divide them to some groups. But, at first, I'd like to > show the total view. This set is consists from 5 sets. Main purpose is > create a room in page_cgroup for I/O tracking and add light-weight access method > for file-cache related accounting. > > 1. An virtual-indexed array. > 2,3. Use virtual-indexed array for id-to-memory_cgroup detection. > 4. modify page_cgroup to use ID instead of pointer, this gives us enough > spaces for further memory tracking. Yes, this is good, I've been meaning to merge the flags and the pointer. Thanks for looking into this. > 5,6 Use light-weight locking mechanism for file related accounting. > 7. use spin_lock instead of bit_spinlock. > > > As a function, patch 5,6 can be an independent patch and I'll accept > reordering series of patch if someone requests. > But we'll need all, I think. > (irq_save for patch 7 will be required later.) > -- Three Cheers, Balbir -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>