lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJD7tkZRH5-y-ux2hDt70JpmL5=YwRwrg_otiTbAei09x-GAJQ@mail.gmail.com>
Date:   Tue, 3 Oct 2023 16:22:10 -0700
From:   Yosry Ahmed <yosryahmed@...gle.com>
To:     Nhat Pham <nphamcs@...il.com>
Cc:     akpm@...ux-foundation.org, riel@...riel.com, hannes@...xchg.org,
        mhocko@...nel.org, roman.gushchin@...ux.dev, shakeelb@...gle.com,
        muchun.song@...ux.dev, tj@...nel.org, lizefan.x@...edance.com,
        shuah@...nel.org, mike.kravetz@...cle.com, fvdl@...gle.com,
        linux-mm@...ck.org, kernel-team@...a.com,
        linux-kernel@...r.kernel.org, cgroups@...r.kernel.org
Subject: Re: [PATCH] memcontrol: only transfer the memcg data for migration

On Tue, Oct 3, 2023 at 4:14 PM Nhat Pham <nphamcs@...il.com> wrote:
>
> For most migration use cases, only transfer the memcg data from the old
> folio to the new folio, and clear the old folio's memcg data. No
> charging and uncharging will be done. These use cases include the new
> hugetlb memcg accounting behavior (which was not previously handled).
>
> This shaves off some work on the migration path, and avoids the
> temporary double charging of a folio during its migration.
>
> The only exception is replace_page_cache_folio(), which will use the old
> mem_cgroup_migrate() (now renamed to mem_cgroup_replace_folio). In that
> context, the isolation of the old page isn't quite as thorough as with
> migration, so we cannot use our new implementation directly.
>
> This patch is the result of the following discussion on the new hugetlb
> memcg accounting behavior:
>
> https://lore.kernel.org/lkml/20231003171329.GB314430@monkey/
>
> Reported-by: Mike Kravetz <mike.kravetz@...cle.com>
> Closes: https://lore.kernel.org/lkml/20231003171329.GB314430@monkey/
> Suggested-by: Johannes Weiner <hannes@...xchg.org>
> Signed-off-by: Nhat Pham <nphamcs@...il.com>

Does this patch fit before or after your series? In both cases I think
there might be a problem for bisectability.

> ---
>  include/linux/memcontrol.h |  7 ++++++
>  mm/filemap.c               |  2 +-
>  mm/memcontrol.c            | 45 +++++++++++++++++++++++++++++++++++---
>  mm/migrate.c               |  3 +--
>  4 files changed, 51 insertions(+), 6 deletions(-)
>
> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
> index a827e2129790..e3eaa123256b 100644
> --- a/include/linux/memcontrol.h
> +++ b/include/linux/memcontrol.h
> @@ -711,6 +711,8 @@ static inline void mem_cgroup_uncharge_list(struct list_head *page_list)
>
>  void mem_cgroup_cancel_charge(struct mem_cgroup *memcg, unsigned int nr_pages);
>
> +void mem_cgroup_replace_folio(struct folio *old, struct folio *new);
> +
>  void mem_cgroup_migrate(struct folio *old, struct folio *new);
>
>  /**
> @@ -1294,6 +1296,11 @@ static inline void mem_cgroup_cancel_charge(struct mem_cgroup *memcg,
>  {
>  }
>
> +static inline void mem_cgroup_replace_folio(struct folio *old,
> +               struct folio *new)
> +{
> +}
> +
>  static inline void mem_cgroup_migrate(struct folio *old, struct folio *new)
>  {
>  }
> diff --git a/mm/filemap.c b/mm/filemap.c
> index 9481ffaf24e6..673745219c82 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -819,7 +819,7 @@ void replace_page_cache_folio(struct folio *old, struct folio *new)
>         new->mapping = mapping;
>         new->index = offset;
>
> -       mem_cgroup_migrate(old, new);
> +       mem_cgroup_replace_folio(old, new);
>
>         xas_lock_irq(&xas);
>         xas_store(&xas, new);
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 6660684f6f97..cbaa26605b3d 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -7316,16 +7316,17 @@ void __mem_cgroup_uncharge_list(struct list_head *page_list)
>  }
>
>  /**
> - * mem_cgroup_migrate - Charge a folio's replacement.
> + * mem_cgroup_replace_folio - Charge a folio's replacement.
>   * @old: Currently circulating folio.
>   * @new: Replacement folio.
>   *
>   * Charge @new as a replacement folio for @old. @old will
> - * be uncharged upon free.
> + * be uncharged upon free. This is only used by the page cache
> + * (in replace_page_cache_folio()).
>   *
>   * Both folios must be locked, @new->mapping must be set up.
>   */
> -void mem_cgroup_migrate(struct folio *old, struct folio *new)
> +void mem_cgroup_replace_folio(struct folio *old, struct folio *new)
>  {
>         struct mem_cgroup *memcg;
>         long nr_pages = folio_nr_pages(new);
> @@ -7364,6 +7365,44 @@ void mem_cgroup_migrate(struct folio *old, struct folio *new)
>         local_irq_restore(flags);
>  }
>
> +/**
> + * mem_cgroup_migrate - Transfer the memcg data from the old to the new folio.
> + * @old: Currently circulating folio.
> + * @new: Replacement folio.
> + *
> + * Transfer the memcg data from the old folio to the new folio for migration.
> + * The old folio's data info will be cleared. Note that the memory counters
> + * will remain unchanged throughout the process.
> + *
> + * Both folios must be locked, @new->mapping must be set up.
> + */
> +void mem_cgroup_migrate(struct folio *old, struct folio *new)
> +{
> +       struct mem_cgroup *memcg;
> +
> +       VM_BUG_ON_FOLIO(!folio_test_locked(old), old);
> +       VM_BUG_ON_FOLIO(!folio_test_locked(new), new);
> +       VM_BUG_ON_FOLIO(folio_test_anon(old) != folio_test_anon(new), new);
> +       VM_BUG_ON_FOLIO(folio_nr_pages(old) != folio_nr_pages(new), new);
> +
> +       if (mem_cgroup_disabled())
> +               return;
> +
> +       memcg = folio_memcg(old);
> +       /*
> +        * Note that it is normal to see !memcg for a hugetlb folio.
> +        * It could have been allocated when memory_hugetlb_accounting was not
> +        * selected, for e.g.
> +        */
> +       VM_WARN_ON_ONCE_FOLIO(!memcg, old);
> +       if (!memcg)
> +               return;
> +
> +       /* Transfer the charge and the css ref */
> +       commit_charge(new, memcg);
> +       old->memcg_data = 0;
> +}
> +
>  DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key);
>  EXPORT_SYMBOL(memcg_sockets_enabled_key);
>
> diff --git a/mm/migrate.c b/mm/migrate.c
> index 7d1804c4a5d9..6034c7ed1d65 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -633,8 +633,7 @@ void folio_migrate_flags(struct folio *newfolio, struct folio *folio)
>
>         folio_copy_owner(newfolio, folio);
>
> -       if (!folio_test_hugetlb(folio))
> -               mem_cgroup_migrate(folio, newfolio);
> +       mem_cgroup_migrate(folio, newfolio);
>  }
>  EXPORT_SYMBOL(folio_migrate_flags);
>
> --
> 2.34.1
>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ