On Mon, Feb 08, 2021 at 01:11:00PM -0800, Andrew Morton wrote:

On Mon,  8 Feb 2021 13:08:20 +0200 Mike Rapoport [off-list ref] wrote:

quoted

There could be struct pages that are not backed by actual physical memory.
This can happen when the actual memory bank is not a multiple of
SECTION_SIZE or when an architecture does not register memory holes
reserved by the firmware as memblock.memory.

Such pages are currently initialized using init_unavailable_mem() function
that iterates through PFNs in holes in memblock.memory and if there is a
struct page corresponding to a PFN, the fields of this page are set to
default values and it is marked as Reserved.

init_unavailable_mem() does not take into account zone and node the page
belongs to and sets both zone and node links in struct page to zero.

On a system that has firmware reserved holes in a zone above ZONE_DMA, for
instance in a configuration below:

	# grep -A1 E820 /proc/iomem
	7a17b000-7a216fff : Unknown E820 type
	7a217000-7bffffff : System RAM

unset zone link in struct page will trigger

	VM_BUG_ON_PAGE(!zone_spans_pfn(page_zone(page), pfn), page);

because there are pages in both ZONE_DMA32 and ZONE_DMA (unset zone link
in struct page) in the same pageblock.

...


Fixes: 73a6e474cb37 ("mm: memmap_init: iterate over memblock regions rather
that check each PFN")

What are your thoughts on the priority of this (rather large!) fix? 
Are such systems sufficiently common to warrant a 5.11 merge?  -stable?

I don't know how common are such systems, but the bug is exposed only for
builds with DEBUG_VM=y, so after problems with previous versions discovered
by various CI systems I'd say to hold it off till 5.11 is out.

If this time the fix works it'll make it to -stable anyway :)

-- 
Sincerely yours,
Mike.