[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <dwkcsytrcauf24634bsx6dm2wxofaxxaa4jwsu5xszmtje3gin@7dzzzn6opjor>
Date: Tue, 5 Aug 2025 13:40:47 +0200
From: "Pankaj Raghav (Samsung)" <kernel@...kajraghav.com>
To: David Hildenbrand <david@...hat.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@...cle.com>,
Suren Baghdasaryan <surenb@...gle.com>, Ryan Roberts <ryan.roberts@....com>,
Baolin Wang <baolin.wang@...ux.alibaba.com>, Borislav Petkov <bp@...en8.de>, Ingo Molnar <mingo@...hat.com>,
"H . Peter Anvin" <hpa@...or.com>, Vlastimil Babka <vbabka@...e.cz>, Zi Yan <ziy@...dia.com>,
Mike Rapoport <rppt@...nel.org>, Dave Hansen <dave.hansen@...ux.intel.com>,
Michal Hocko <mhocko@...e.com>, Andrew Morton <akpm@...ux-foundation.org>,
Thomas Gleixner <tglx@...utronix.de>, Nico Pache <npache@...hat.com>, Dev Jain <dev.jain@....com>,
"Liam R . Howlett" <Liam.Howlett@...cle.com>, Jens Axboe <axboe@...nel.dk>, linux-kernel@...r.kernel.org,
linux-mm@...ck.org, willy@...radead.org, x86@...nel.org, linux-block@...r.kernel.org,
Ritesh Harjani <ritesh.list@...il.com>, linux-fsdevel@...r.kernel.org,
"Darrick J . Wong" <djwong@...nel.org>, mcgrof@...nel.org, gost.dev@...sung.com, hch@....de,
Pankaj Raghav <p.raghav@...sung.com>
Subject: Re: [PATCH 3/5] mm: add static huge zero folio
Thanks a lot Lorenzo and David for the feedback and quick iteration on
the patchset. I really like the number of lines of code has been
steadily reducing since the first version :)
I will fold the changes in the next series.
<snip>
> > > @@ -866,9 +866,14 @@ static int __init thp_shrinker_init(void)
> > > huge_zero_folio_shrinker->scan_objects = shrink_huge_zero_folio_scan;
> > > shrinker_register(huge_zero_folio_shrinker);
> > > - deferred_split_shrinker->count_objects = deferred_split_count;
> > > - deferred_split_shrinker->scan_objects = deferred_split_scan;
> > > - shrinker_register(deferred_split_shrinker);
> > > + if (IS_ENABLED(CONFIG_STATIC_HUGE_ZERO_FOLIO)) {
> > > + if (!get_huge_zero_folio())
> > > + pr_warn("Allocating static huge zero folio failed\n");
> > > + } else {
> > > + deferred_split_shrinker->count_objects = deferred_split_count;
> > > + deferred_split_shrinker->scan_objects = deferred_split_scan;
> > > + shrinker_register(deferred_split_shrinker);
> > > + }
> > > return 0;
> > > }
> > > --
> > > 2.50.1
> > >
> > >
> > > Now, one thing I do not like is that we have "ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO" but
> > > then have a user-selectable option.
> > >
> > > Should we just get rid of ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO?
> >
One of the early feedbacks from Lorenzo was that there might be some
architectures that has PMD size > 2M might enable this by mistake. So
the ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO was introduced as an extra
precaution apart from user selectable CONFIG_STATIC_HUGE_ZERO_FOLIO.
Isn't it better to have an extra knob per-arch to be on the safer side
or you think it is too excessive?
> > Yeah, though I guess we probably need to make it need CONFIG_MMU if so?
> > Probably don't want to provide it if it might somehow break things?
>
> It would still depend on THP, and THP is !MMU. So that should just work.
>
> We could go one step further and special case in mm_get_huge_zero_folio() +
> mm_put_huge_zero_folio() on CONFIG_STATIC_HUGE_ZERO_FOLIO.
>
> Something like
>
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 9c38a95e9f091..9b87884e5f299 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -248,6 +248,9 @@ static void put_huge_zero_page(void)
>
> struct folio *mm_get_huge_zero_folio(struct mm_struct *mm)
> {
> + if (IS_ENABLED(CONFIG_STATIC_HUGE_ZERO_FOLIO))
> + return huge_zero_folio;
> +
> if (test_bit(MMF_HUGE_ZERO_PAGE, &mm->flags))
> return READ_ONCE(huge_zero_folio);
>
> @@ -262,6 +265,9 @@ struct folio *mm_get_huge_zero_folio(struct mm_struct
> *mm)
>
> void mm_put_huge_zero_folio(struct mm_struct *mm)
> {
> + if (IS_ENABLED(CONFIG_STATIC_HUGE_ZERO_FOLIO))
> + return huge_zero_folio;
> +
> if (test_bit(MMF_HUGE_ZERO_PAGE, &mm->flags))
> put_huge_zero_page();
> }
>
>
--
Pankaj
Powered by blists - more mailing lists