linux-kernel - Re: [PATCH 3/3] fs, mm: account buffer

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]
Date:   Fri, 22 Jun 2018 16:33:00 -0700
From:   Shakeel Butt <shakeelb@...gle.com>
To:     Roman Gushchin <guro@...com>
Cc:     Johannes Weiner <hannes@...xchg.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...nel.org>,
        Vladimir Davydov <vdavydov.dev@...il.com>,
        Jan Kara <jack@...e.com>, Greg Thelen <gthelen@...gle.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Cgroups <cgroups@...r.kernel.org>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>,
        Linux MM <linux-mm@...ck.org>, Jan Kara <jack@...e.cz>,
        Alexander Viro <viro@...iv.linux.org.uk>
Subject: Re: [PATCH 3/3] fs, mm: account buffer_head to kmemcg

On Tue, Jun 19, 2018 at 12:55 PM Roman Gushchin <guro@...com> wrote:
>
> On Tue, Jun 19, 2018 at 12:51:15PM -0700, Shakeel Butt wrote:
> > On Tue, Jun 19, 2018 at 10:41 AM Roman Gushchin <guro@...com> wrote:
> > >
> > > On Tue, Jun 19, 2018 at 12:27:41PM -0400, Johannes Weiner wrote:
> > > > On Mon, Jun 18, 2018 at 10:13:27PM -0700, Shakeel Butt wrote:
> > > > > The buffer_head can consume a significant amount of system memory and
> > > > > is directly related to the amount of page cache. In our production
> > > > > environment we have observed that a lot of machines are spending a
> > > > > significant amount of memory as buffer_head and can not be left as
> > > > > system memory overhead.
> > > > >
> > > > > Charging buffer_head is not as simple as adding __GFP_ACCOUNT to the
> > > > > allocation. The buffer_heads can be allocated in a memcg different from
> > > > > the memcg of the page for which buffer_heads are being allocated. One
> > > > > concrete example is memory reclaim. The reclaim can trigger I/O of pages
> > > > > of any memcg on the system. So, the right way to charge buffer_head is
> > > > > to extract the memcg from the page for which buffer_heads are being
> > > > > allocated and then use targeted memcg charging API.
> > > > >
> > > > > Signed-off-by: Shakeel Butt <shakeelb@...gle.com>
> > > > > Cc: Jan Kara <jack@...e.cz>
> > > > > Cc: Greg Thelen <gthelen@...gle.com>
> > > > > Cc: Michal Hocko <mhocko@...nel.org>
> > > > > Cc: Johannes Weiner <hannes@...xchg.org>
> > > > > Cc: Vladimir Davydov <vdavydov.dev@...il.com>
> > > > > Cc: Alexander Viro <viro@...iv.linux.org.uk>
> > > > > Cc: Andrew Morton <akpm@...ux-foundation.org>
> > > > > ---
> > > > >  fs/buffer.c                | 14 +++++++++++++-
> > > > >  include/linux/memcontrol.h |  7 +++++++
> > > > >  mm/memcontrol.c            | 21 +++++++++++++++++++++
> > > > >  3 files changed, 41 insertions(+), 1 deletion(-)
> > > > >
> > > > > diff --git a/fs/buffer.c b/fs/buffer.c
> > > > > index 8194e3049fc5..26389b7a3cab 100644
> > > > > --- a/fs/buffer.c
> > > > > +++ b/fs/buffer.c
> > > > > @@ -815,10 +815,17 @@ struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size,
> > > > >     struct buffer_head *bh, *head;
> > > > >     gfp_t gfp = GFP_NOFS;
> > > > >     long offset;
> > > > > +   struct mem_cgroup *old_memcg;
> > > > > +   struct mem_cgroup *memcg = get_mem_cgroup_from_page(page);
> > > > >
> > > > >     if (retry)
> > > > >             gfp |= __GFP_NOFAIL;
> > > > >
> > > > > +   if (memcg) {
> > > > > +           gfp |= __GFP_ACCOUNT;
> > > > > +           old_memcg = memalloc_memcg_save(memcg);
> > > > > +   }
> > > >
> > > > Please move the get_mem_cgroup_from_page() call out of the
> > > > declarations and down to right before the if (memcg) branch.
> > > >
> > > > >     head = NULL;
> > > > >     offset = PAGE_SIZE;
> > > > >     while ((offset -= size) >= 0) {
> > > > > @@ -835,6 +842,11 @@ struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size,
> > > > >             /* Link the buffer to its page */
> > > > >             set_bh_page(bh, page, offset);
> > > > >     }
> > > > > +out:
> > > > > +   if (memcg) {
> > > > > +           memalloc_memcg_restore(old_memcg);
> > > > > +#ifdef CONFIG_MEMCG
> > > > > +           css_put(&memcg->css);
> > > > > +#endif
> > > >
> > > > Please add a put_mem_cgroup() ;)
> > >
> > > I've added such helper by commit 8a34a8b7fd62 ("mm, oom: cgroup-aware OOM killer").
> > > It's in the mm tree.
> > >
> >
> > I was using mem_cgroup_put() defined by Roman's patch but there were a
> > lot of build failure reports where someone was taking this series
> > without Roman's series or applying the series out of order. Andrew
> > asked me to keep it like this and then he will convert these callsites
> > into mem_cgroup_put() after making making sure Roman's series is
> > applied in mm tree. I will recheck with him, how he wants to handle it
> > now.
>
> I can also split the introduction of mem_cgroup_put() into a separate commit,
> as it seems to be usable not only by the cgroup oom stuff.
> Please, let me know, if it's a preferred way to go.
>

Oh I forgot to reply. Yes, let's do that, a separate patch to
introduce mem_cgroup_put() which can used by remote charging and memcg
aware oom-killer patches.

Shakeel