Re: [PATCH -rc7] mm/huge_memory: mark huge_zero_page reserved

[Yang Shi <shy828301@gmail.com>]

On Tue, May 14, 2024 at 3:14 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Sat, 11 May 2024 11:54:35 +0800 Miaohe Lin <linmiaohe@huawei.com> wrote:
>
> > When I did memory failure tests recently, below panic occurs:
> >
> >  kernel BUG at include/linux/mm.h:1135!
> >  invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
> >  CPU: 9 PID: 137 Comm: kswapd1 Not tainted 6.9.0-rc4-00491-gd5ce28f156fe-dirty #14
> >
> > ...
> >
> > --- a/mm/huge_memory.c
> > +++ b/mm/huge_memory.c
> > @@ -208,6 +208,7 @@ static bool get_huge_zero_page(void)
> >               __free_pages(zero_page, compound_order(zero_page));
> >               goto retry;
> >       }
> > +     __SetPageReserved(zero_page);
> >       WRITE_ONCE(huge_zero_pfn, page_to_pfn(zero_page));
> >
> >       /* We take additional reference here. It will be put back by shrinker */
> > @@ -260,6 +261,7 @@ static unsigned long shrink_huge_zero_page_scan(struct shrinker *shrink,
> >               struct page *zero_page = xchg(&huge_zero_page, NULL);
> >               BUG_ON(zero_page == NULL);
> >               WRITE_ONCE(huge_zero_pfn, ~0UL);
> > +             __ClearPageReserved(zero_page);
> >               __free_pages(zero_page, compound_order(zero_page));
> >               return HPAGE_PMD_NR;
> >       }
>
> This causes a bit of a mess when staged ahead of mm-stable.  So to
> avoid disruption I staged it behind mm-stable.  This means that when
> the -stable maintainers try to merge it, they will ask for a fixed up
> version for older kernels so you can please just send them this
> version.

Can you please drop this from mm-unstable since both I and David
nack'ed a similar patch in another thread.
https://lore.kernel.org/linux-mm/20240511032801.1295023-1-linmiaohe@huawei.com/

Both patches actually do the same thing, just this one uses page, the
other one uses folio.

>
> To facilitate this I added the below adjustment:
>
> (btw, shouldn't get_huge_zero_page() and shrink_huge_zero_page_scan()
> be renamed to *_folio_*?)
>
>
> From: Andrew Morton <akpm@linux-foundation.org>
> Subject: mm-huge_memory-mark-huge_zero_page-reserved-fix
> Date: Tue May 14 01:53:37 PM PDT 2024
>
> Update it for 5691753d73a2 ("mm: convert huge_zero_page to huge_zero_folio")
>
> Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
> Cc: David Hildenbrand <david@redhat.com>
> Cc: Miaohe Lin <linmiaohe@huawei.com>
> Cc: Naoya Horiguchi <nao.horiguchi@gmail.com>
> Cc: Xu Yu <xuyu@linux.alibaba.com>
> Cc: Yang Shi <shy828301@gmail.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
>
>  mm/huge_memory.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> --- a/mm/huge_memory.c~mm-huge_memory-mark-huge_zero_page-reserved-fix
> +++ a/mm/huge_memory.c
> @@ -212,7 +212,7 @@ retry:
>                 folio_put(zero_folio);
>                 goto retry;
>         }
> -       __SetPageReserved(zero_page);
> +       __folio_set_reserved(zero_folio);
>         WRITE_ONCE(huge_zero_pfn, folio_pfn(zero_folio));
>
>         /* We take additional reference here. It will be put back by shrinker */
> @@ -265,7 +265,7 @@ static unsigned long shrink_huge_zero_pa
>                 struct folio *zero_folio = xchg(&huge_zero_folio, NULL);
>                 BUG_ON(zero_folio == NULL);
>                 WRITE_ONCE(huge_zero_pfn, ~0UL);
> -               __ClearPageReserved(zero_page);
> +               __folio_clear_reserved(zero_folio);
>                 folio_put(zero_folio);
>                 return HPAGE_PMD_NR;
>         }
> _
>