On Wed, Jan 13, 2021 at 12:29 AM David Hildenbrand [off-list ref] wrote:

On 13.01.21 08:35, Dan Williams wrote:

quoted

pfn_section_valid() determines pfn validity on subsection granularity
where pfn_valid() may be limited to coarse section granularity.
Explicitly validate subsections after pfn_valid() succeeds.

Fixes: b13bc35193d9 ("mm/hotplug: invalid PFNs from pfn_to_online_page()")
Cc: Qian Cai <redacted>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Reported-by: David Hildenbrand <redacted>
Signed-off-by: Dan Williams <redacted>
---
 mm/memory_hotplug.c |   24 ++++++++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 55a69d4396e7..9f37f8a68da4 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c

@@ -308,11 +308,27 @@ static int check_hotplug_memory_addressable(unsigned long pfn,
 struct page *pfn_to_online_page(unsigned long pfn)
 {
      unsigned long nr = pfn_to_section_nr(pfn);
+     struct mem_section *ms;
+
+     if (nr >= NR_MEM_SECTIONS)
+             return NULL;
+
+     ms = __nr_to_section(nr);
+     if (!online_section(ms))
+             return NULL;
+
+     /*
+      * Save some code text when online_section() +
+      * pfn_section_valid() are sufficient.
+      */
+     if (IS_ENABLED(CONFIG_HAVE_ARCH_PFN_VALID))
+             if (!pfn_valid(pfn))
+                     return NULL;

Nit:

if (IS_ENABLED(CONFIG_HAVE_ARCH_PFN_VALID) &&
    !pfn_valid(pfn))

Ok... I'll do a final resend "To: akpm" after the kbuild robot
finishes chewing on this series.