Re: [PATCH 1/2] mm/memremap.c: Add pfn_to_devmap_page() to get page in ZONE_DEVICE

[Xiongwei Song <sxwjean@gmail.com>]

Hi David,

On Mon, Jan 10, 2022 at 4:16 PM David Hildenbrand <david@redhat.com> wrote:
>
> On 09.01.22 14:05, sxwjean@me.com wrote:
> > From: Xiongwei Song <sxwjean@gmail.com>
> >
> > when requesting page information by /proc/kpage*, the pages in ZONE_DEVICE
> > were missed. We need a function to help on this.
> >
> > The pfn_to_devmap_page() function like pfn_to_online_page(), but only
> > concerns the pages in ZONE_DEVICE.
> >
> > Signed-off-by: Xiongwei Song <sxwjean@gmail.com>
> > ---
> >  include/linux/memremap.h |  8 ++++++++
> >  mm/memremap.c            | 42 ++++++++++++++++++++++++++++++++++++++++
> >  2 files changed, 50 insertions(+)
> >
> > diff --git a/include/linux/memremap.h b/include/linux/memremap.h
> > index c0e9d35889e8..621723e9c4a5 100644
> > --- a/include/linux/memremap.h
> > +++ b/include/linux/memremap.h
> > @@ -137,6 +137,8 @@ void *devm_memremap_pages(struct device *dev, struct dev_pagemap *pgmap);
> >  void devm_memunmap_pages(struct device *dev, struct dev_pagemap *pgmap);
> >  struct dev_pagemap *get_dev_pagemap(unsigned long pfn,
> >               struct dev_pagemap *pgmap);
> > +struct page *pfn_to_devmap_page(unsigned long pfn,
> > +             struct dev_pagemap **pgmap);
> >  bool pgmap_pfn_valid(struct dev_pagemap *pgmap, unsigned long pfn);
> >
> >  unsigned long vmem_altmap_offset(struct vmem_altmap *altmap);
> > @@ -166,6 +168,12 @@ static inline struct dev_pagemap *get_dev_pagemap(unsigned long pfn,
> >       return NULL;
> >  }
> >
> > +static inline struct page *pfn_to_devmap_page(unsigned long pfn,
> > +             struct dev_pagemap **pgmap)
> > +{
> > +     return NULL;
> > +}
> > +
> >  static inline bool pgmap_pfn_valid(struct dev_pagemap *pgmap, unsigned long pfn)
> >  {
> >       return false;
> > diff --git a/mm/memremap.c b/mm/memremap.c
> > index 5a66a71ab591..072dbe6ab81c 100644
> > --- a/mm/memremap.c
> > +++ b/mm/memremap.c
> > @@ -494,6 +494,48 @@ struct dev_pagemap *get_dev_pagemap(unsigned long pfn,
> >  }
> >  EXPORT_SYMBOL_GPL(get_dev_pagemap);
> >
> > +/**
> > + * pfn_to_devmap_page - get page pointer which belongs to dev_pagemap by @pfn
> > + * @pfn: page frame number to lookup page_map
> > + * @pgmap: to save pgmap address which is for putting reference
> > + *
> > + * If @pgmap is non-NULL, then pfn is on ZONE_DEVICE and return page pointer.
> > + */
> > +struct page *pfn_to_devmap_page(unsigned long pfn, struct dev_pagemap **pgmap)
> > +{
> > +     unsigned long nr = pfn_to_section_nr(pfn);
> > +     struct mem_section *ms;
> > +     struct page *page = NULL;
> > +
> > +     if (nr >= NR_MEM_SECTIONS)
> > +             return NULL;
> > +
> > +     if (IS_ENABLED(CONFIG_HAVE_ARCH_PFN_VALID) && !pfn_valid(pfn))
> > +             return NULL;
> > +
> > +     ms = __nr_to_section(nr);
> > +     if (!valid_section(ms))
> > +             return NULL;
> > +     if (!pfn_section_valid(ms, pfn))
> > +             return NULL;
> > +
> > +     /*
> > +      * Two types of sections may include valid pfns:
> > +      * - The pfns of section belong to ZONE_DEVICE and ZONE_{NORMAL,MOVABLE}
> > +      *   at the same time.
> > +      * - All pfns in one section are offline but valid.
> > +      */
> > +     if (!online_device_section(ms) && online_section(ms))
> > +             return NULL;
> > +
> > +     *pgmap = get_dev_pagemap(pfn, NULL);
> > +     if (*pgmap)
> > +             page = pfn_to_page(pfn);
> > +
> > +     return page;
> > +}
> > +EXPORT_SYMBOL_GPL(pfn_to_devmap_page);
>
> Is this complexity really required?
>
> Take a look at mm/memory-failure.c
>
> p = pfn_to_online_page(pfn);
> if (!p) {
>         if (pfn_valid(pfn)) {
>                 pgmap = get_dev_pagemap(pfn, NULL);
>                 if (pgmap)
>                         // success
>                 // error
>         }
>         // error
> }

Yes, this method is much simpler than mine. Will do this in v2.

>
>
> Also, why do we need the export?

Ah, no strong reason. Will remove that in v2.

Would you mind adding "Suggested-by" for this patch?

Thank you for your time.

Regards,
Xiongwei