On Wed, May 23, 2018 at 04:16:08PM +0200, Michal Hocko wrote: > On Wed 23-05-18 15:54:03, Oscar Salvador wrote: > > On Thu, May 10, 2018 at 02:02:00PM +0200, Michal Hocko wrote: > > > On Fri 04-05-18 17:50:51, Jonathan Cameron wrote: > > > [...] > > > > Exact path to the problem is as follows: > > > > > > > > mm/memory_hotplug.c : add_memory_resource > > > > The node is not online so we enter the > > > > if (new_node) twice, on the second such block there is a call to > > > > link_mem_sections which calls into > > > > drivers/node.c: link_mem_sections which calls > > > > drivers/node.c: register_mem_sect_under_node which calls > > > > get_nid_for_pfn and keeps trying until the output of that matches > > > > the expected node (passed all the way down from add_memory_resource) > > > > > > I am sorry but I am still confused. Why don't we create sysfs files from > > > __add_pages > > > __add_section > > > hotplug_memory_register > > > register_mem_sect_under_node > > > > IIUC the problem is that at the point we are calling register_mem_sect_under_node(), > > pages are not initialized yet. > > Ahh, of course. I keep forgetting the latest hotplug optimizations that > we do not initialize even nid for struct pages. Which is the whole point > of this patch... Sigh. > > I think the whole sysfs initialization needs to be refactored to be more > sane. The way how we depend on things silently is just not maintainable. I will try to work that out. I also want to see if we can get rid of link_mem_sections() since it shares almost all the code with walk_memory_range(). Maybe we can pass register_mem_sect_under_node() as a callback of walk_memory_range(). Oscar Salvador
