On Tue, 2011-09-27 at 06:52 +0800, Andrew Morton wrote: > On Mon, 26 Sep 2011 15:23:20 +0200 > Michal Hocko <mhocko@xxxxxxx> wrote: > > > On Mon 26-09-11 16:11:52, Shaohua Li wrote: > > > I saw DMA zone is always unreclaimable in my system. > > > zone->all_unreclaimable isn't initialized till a page from the zone is > > > freed. This isn't a big problem normally, but a little confused, so > > > fix here. > > > > The value is initialized when a node is allocated. setup_node_data uses > > alloc_remap which memsets the whole structure or memblock allocation > > which is initialized to 0 as well AFAIK and memory hotplug uses > > arch_alloc_nodedata which is kzalloc. > > setup_node_data() does memset(NODE_DATA(nid), 0, sizeof(pg_data_t)) just > to be sure. > > However, Shaohua reports that "DMA zone is always unreclaimable in my system", > and presumably this patch fixed it. So we don't know what's going on? > > > > Presumably all the other "zone->foo = 0" assignments in free_area_init_core() > are unneeded. Looks I didn't run my test correctly, sorry. I just check it, and this is a vmscan bug, I'll work out a new patch. -- 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/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>