On 19.07.19 13:36, Michal Hocko wrote: > On Fri 19-07-19 11:20:43, David Hildenbrand wrote: >> On 19.07.19 11:13, Michal Hocko wrote: >>> On Fri 19-07-19 11:05:51, David Hildenbrand wrote: >>>> On 19.07.19 10:42, Michal Hocko wrote: >>>>> On Thu 18-07-19 16:22:39, David Hildenbrand wrote: >>>>>> We don't allow to offline memory block devices that belong to multiple >>>>>> numa nodes. Therefore, such devices can never get removed. It is >>>>>> sufficient to process a single node when removing the memory block. >>>>>> >>>>>> Remember for each memory block if it belongs to no, a single, or mixed >>>>>> nodes, so we can use that information to skip unregistering or print a >>>>>> warning (essentially a safety net to catch BUGs). >>>>> >>>>> I do not really like NUMA_NO_NODE - 1 thing. This is yet another invalid >>>>> node that is magic. Why should we even care? In other words why is this >>>>> patch an improvement? >>>> >>>> Oh, and to answer that part of the question: >>>> >>>> We no longer have to iterate over each pfn of a memory block to be removed. >>> >>> Is it possible that we are overzealous when unregistering syfs files and >>> we should simply skip the pfn walk even without this change? >>> >> >> I assume you mean something like v1 without the warning/"NUMA_NO_NODE -1"? >> >> See what I have right now below. > > Yes. I didn'g get to look closely but you caught the idea. Thanks! > Will do a quick test and resent later this day, thanks for having a look! -- Thanks, David / dhildenb
