On Fri 25-01-19 09:19:24, Michal Hocko wrote:
> On Fri 25-01-19 08:37:04, Michal Hocko wrote:
> > On Fri 25-01-19 17:48:32, Linus Torvalds wrote:
> > > [ Just adding a lot of other people to the cc ]
> > >
> > > Robert, could you add a dmesg of a successful boot to that bugzilla,
> > > or just as an attachement in email to this group of people..
> > >
> > > This looks to be with the Fedora kernel config. Two people reporting
> > > it, it looks like similar machines.
> > >
> > > I assume it's some odd memory sizing detail that happens to trigger a
> > > particular case.
> >
> > Quite possible.
>
> Forgot to ask. Can we get a dmesg with 2830bf6f05fb ("mm,
> memory_hotplug: initialize struct pages for the full memory section")
> reverted and memblock=debug kernel command line parameter?
And one more thing which I have overlook until now and it is not really
clear to me. One of th comments says
: The relevant part was:
: kernel bug at mm/page_alloc.c=790
I suppose this is 4.19 stable kernel because that would be
VM_BUG_ON_PAGE(pfn & ((1 << order) - 1), page);
in __free_one_page. I do not really see how 2830bf6f05fb could make any
difference here. It simply zeroes out the rest of the mem section and
that is guaranteed to be allocated because we do not do subsections. The
above VM_BUG_ON says that we start allocating an unaligned pfn for its
order.
Or are there two issues reported in that bug?
--
Michal Hocko
SUSE Labs
