From: David Hildenbrand <david@xxxxxxxxxx> Subject: mm/memory_hotplug: enforce section granularity when onlining/offlining Already two people (including me) tried to offline subsections, because the function looks like it can deal with it. But we really can only online/offline full sections that are properly aligned (e.g., we can only mark full sections online/offline via SECTION_IS_ONLINE). Add a simple safety net to document the restriction now. Current users (core and powernv/memtrace) respect these restrictions. Link: https://lkml.kernel.org/r/20200819175957.28465-3-david@xxxxxxxxxx Signed-off-by: David Hildenbrand <david@xxxxxxxxxx> Acked-by: Michal Hocko <mhocko@xxxxxxxx> Reviewed-by: Oscar Salvador <osalvador@xxxxxxx> Cc: Wei Yang <richard.weiyang@xxxxxxxxxxxxxxxxx> Cc: Baoquan He <bhe@xxxxxxxxxx> Cc: Pankaj Gupta <pankaj.gupta.linux@xxxxxxxxx> Cc: Charan Teja Reddy <charante@xxxxxxxxxxxxxx> Cc: Dan Williams <dan.j.williams@xxxxxxxxx> Cc: Fenghua Yu <fenghua.yu@xxxxxxxxx> Cc: Logan Gunthorpe <logang@xxxxxxxxxxxx> Cc: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx> Cc: Mel Gorman <mgorman@xxxxxxx> Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx> Cc: Michel Lespinasse <walken@xxxxxxxxxx> Cc: Mike Rapoport <rppt@xxxxxxxxxx> Cc: Tony Luck <tony.luck@xxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- mm/memory_hotplug.c | 10 ++++++++++ 1 file changed, 10 insertions(+) --- a/mm/memory_hotplug.c~mm-memory_hotplug-enforce-section-granularity-when-onlining-offlining +++ a/mm/memory_hotplug.c @@ -809,6 +809,11 @@ int __ref online_pages(unsigned long pfn int ret; struct memory_notify arg; + /* We can only online full sections (e.g., SECTION_IS_ONLINE) */ + if (WARN_ON_ONCE(!nr_pages || + !IS_ALIGNED(pfn | nr_pages, PAGES_PER_SECTION))) + return -EINVAL; + mem_hotplug_begin(); /* associate pfn range with the zone */ @@ -1494,6 +1499,11 @@ int __ref offline_pages(unsigned long st struct memory_notify arg; char *reason; + /* We can only offline full sections (e.g., SECTION_IS_ONLINE) */ + if (WARN_ON_ONCE(!nr_pages || + !IS_ALIGNED(start_pfn | nr_pages, PAGES_PER_SECTION))) + return -EINVAL; + mem_hotplug_begin(); /* _
