Date:   Thu, 26 Jan 2023 15:14:25 -0300
From:   Marcelo Tosatti <mtosatti@redhat.com>
To:     Michal Hocko <mhocko@suse.com>
Cc:     Leonardo =?iso-8859-1?Q?Br=E1s?= <leobras@redhat.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Roman Gushchin <roman.gushchin@linux.dev>,
        Shakeel Butt <shakeelb@google.com>,
        Muchun Song <muchun.song@linux.dev>,
        Andrew Morton <akpm@linux-foundation.org>,
        cgroups@vger.kernel.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 0/5] Introduce memcg_stock_pcp remote draining
Message-ID: <Y9LDAZmApLeffrT8@tpad>
References: <20230125073502.743446-1-leobras@redhat.com>
 <Y9DpbVF+JR/G+5Or@dhcp22.suse.cz>
 <9e61ab53e1419a144f774b95230b789244895424.camel@redhat.com>
 <Y9FzSBw10MGXm2TK@tpad>
 <Y9IvoDJbLbFcitTc@dhcp22.suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y9IvoDJbLbFcitTc@dhcp22.suse.cz>
Precedence: bulk

On Thu, Jan 26, 2023 at 08:45:36AM +0100, Michal Hocko wrote:
> On Wed 25-01-23 15:22:00, Marcelo Tosatti wrote:
> [...]
> > Remote draining reduces interruptions whether CPU 
> > is marked as isolated or not:
> > 
> > - Allows isolated CPUs from benefiting of pcp caching.
> > - Removes the interruption to non isolated CPUs. See for example 
> > 
> > https://lkml.org/lkml/2022/6/13/2769
> 
> This is talking about page allocato per cpu caches, right? In this patch
> we are talking about memcg pcp caches. Are you sure the same applies
> here?

Both can stall the users of the drain operation.

"Minchan Kim tested this independently and reported;

	My workload is not NOHZ CPUs but run apps under heavy memory
	pressure so they goes to direct reclaim and be stuck on
	drain_all_pages until work on workqueue run."

Therefore using a workqueue to drain memcg pcps also depends on the 
remote CPU executing that work item in time (which can stall
the following). No?

===

   7   3141  mm/memory.c <<wp_page_copy>>
             if (mem_cgroup_charge(page_folio(new_page), mm, GFP_KERNEL))
   8   4118  mm/memory.c <<do_anonymous_page>>
             if (mem_cgroup_charge(page_folio(page), vma->vm_mm, GFP_KERNEL))
   9   4577  mm/memory.c <<do_cow_fault>>
             if (mem_cgroup_charge(page_folio(vmf->cow_page), vma->vm_mm,
  10    621  mm/migrate_device.c <<migrate_vma_insert_page>>
             if (mem_cgroup_charge(page_folio(page), vma->vm_mm, GFP_KERNEL))
  11    710  mm/shmem.c <<shmem_add_to_page_cache>>
             error = mem_cgroup_charge(folio, charge_mm, gfp);