Re: uninitialized pmem struct pages

[Michal Hocko]

On Mon 04-01-21 21:33:06, Dan Williams wrote:
> On Mon, Jan 4, 2021 at 7:59 AM Michal Hocko <mhocko@xxxxxxxx> wrote:
[...]
> > Not sure what exactly you are asking for but crash says
> > crash> kmem -p 6060000
> >       PAGE          PHYSICAL      MAPPING       INDEX CNT FLAGS
> > fffff8c600181800     6060000                0        0  0 fffffc0000000
> >
> > > I do wonder what hosts pfn_to_page(PHYS_PFN(0x6060000)) - is it actually
> > > part of the actual altmap (i.e. > 0x6060000) or maybe even self-hosted?
> >
> > I am not really familiar with the pmem so I would need more assistance
> > here. I've tried this (shot into the dark):
> > crash> struct page.pgmap fffff8c600181800
> >       pgmap = 0xfffff8c600181808
> 
> Does /proc/iomem show an active namespace in the range?

Any tips how I dig that out from the crash dump?

> You should be
> able to skip ahead to the first pfn in that namespace to find the
> first dev_pagemap. I would have expected pfn_to_online_page() to have
> saved you here. This address range is section aligned.

Well, the affected code in this case was 
	end_pfn = min(start_pfn + nr_pages,
                        zone_end_pfn(page_zone(pfn_to_page(start_pfn))));
where start_pfn was the first pfn of a memory section. This code was
completely unaware of zone device or dev_pagemap like most others pfn
walkers. It just wanted to get bounds for the zone but it stumbled over
uninitialized zone/node.
-- 
Michal Hocko
SUSE Labs