Received-SPF: pass (google.com: domain of linux-kernel+bounces-200591-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Tue, 4 Jun 2024 08:29:27 -0400
From: Johannes Weiner <hannes@cmpxchg.org>
To: Byungchul Park <byungchul@sk.com>
Cc: "Huang, Ying" <ying.huang@intel.com>, akpm@linux-foundation.org,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	kernel_team@skhynix.com, iamjoonsoo.kim@lge.com,
	rientjes@google.com
Subject: Re: [PATCH v2] mm: let kswapd work again for node that used to be
 hopeless but may not now
Message-ID: <20240604122927.GA1992@cmpxchg.org>
References: <20240604072323.10886-1-byungchul@sk.com>
 <87bk4hcf7h.fsf@yhuang6-desk2.ccr.corp.intel.com>
 <20240604084533.GA68919@system.software.com>
 <8734ptccgi.fsf@yhuang6-desk2.ccr.corp.intel.com>
 <20240604091221.GA28034@system.software.com>
 <20240604102516.GB28034@system.software.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240604102516.GB28034@system.software.com>

On Tue, Jun 04, 2024 at 07:25:16PM +0900, Byungchul Park wrote:
> On Tue, Jun 04, 2024 at 06:12:22PM +0900, Byungchul Park wrote:
> > On Tue, Jun 04, 2024 at 04:57:17PM +0800, Huang, Ying wrote:
> > > Byungchul Park <byungchul@sk.com> writes:
> > > 
> > > > On Tue, Jun 04, 2024 at 03:57:54PM +0800, Huang, Ying wrote:
> > > >> Byungchul Park <byungchul@sk.com> writes:
> > > >> 
> > > >> > Changes from v1:
> > > >> > 	1. Don't allow to resume kswapd if the system is under memory
> > > >> > 	   pressure that might affect direct reclaim by any chance, like
> > > >> > 	   if NR_FREE_PAGES is less than (low wmark + min wmark)/2.
> > > >> >
> > > >> > --->8---
> > > >> > From 6c73fc16b75907f5da9e6b33aff86bf7d7c9dd64 Mon Sep 17 00:00:00 2001
> > > >> > From: Byungchul Park <byungchul@sk.com>
> > > >> > Date: Tue, 4 Jun 2024 15:27:56 +0900
> > > >> > Subject: [PATCH v2] mm: let kswapd work again for node that used to be hopeless but may not now
> > > >> >
> > > >> > A system should run with kswapd running in background when under memory
> > > >> > pressure, such as when the available memory level is below the low water
> > > >> > mark and there are reclaimable folios.
> > > >> >
> > > >> > However, the current code let the system run with kswapd stopped if
> > > >> > kswapd has been stopped due to more than MAX_RECLAIM_RETRIES failures
> > > >> > until direct reclaim will do for that, even if there are reclaimable
> > > >> > folios that can be reclaimed by kswapd.  This case was observed in the
> > > >> > following scenario:
> > > >> >
> > > >> >    CONFIG_NUMA_BALANCING enabled
> > > >> >    sysctl_numa_balancing_mode set to NUMA_BALANCING_MEMORY_TIERING
> > > >> >    numa node0 (500GB local DRAM, 128 CPUs)
> > > >> >    numa node1 (100GB CXL memory, no CPUs)
> > > >> >    swap off
> > > >> >
> > > >> >    1) Run a workload with big anon pages e.g. mmap(200GB).
> > > >> >    2) Continue adding the same workload to the system.
> > > >> >    3) The anon pages are placed in node0 by promotion/demotion.
> > > >> >    4) kswapd0 stops because of the unreclaimable anon pages in node0.
> > > >> >    5) Kill the memory hoggers to restore the system.
> > > >> >
> > > >> > After restoring the system at 5), the system starts to run without
> > > >> > kswapd.  Even worse, tiering mechanism is no longer able to work since
> > > >> > the mechanism relies on kswapd for demotion.
> > > >> 
> > > >> We have run into the situation that kswapd is kept in failure state for
> > > >> long in a multiple tiers system.  I think that your solution is too
> > > >
> > > > My solution just gives a chance for kswapd to work again even if
> > > > kswapd_failures >= MAX_RECLAIM_RETRIES, if there are potential
> > > > reclaimable folios.  That's it.
> > > >
> > > >> limited, because OOM killing may not happen, while the access pattern of
> > > >
> > > > I don't get this.  OOM will happen as is, through direct reclaim.
> > > 
> > > A system that fails to reclaim via kswapd may succeed to reclaim via
> > > direct reclaim, because more CPUs are used to scanning the page tables.
> > > 
> > > In a system with NUMA balancing based page promotion and page demotion
> > > enabled, page promotion will wake up kswapd, but kswapd may fail in some
> > > situations.  But page promotion will no trigger direct reclaim or OOM.
> > > 
> > > >> the workloads may change.  We have a preliminary and simple solution for
> > > >> this as follows,
> > > >> 
> > > >> https://git.kernel.org/pub/scm/linux/kernel/git/vishal/tiering.git/commit/?h=tiering-0.8&id=17a24a354e12d4d4675d78481b358f668d5a6866
> > > >
> > > > Whether tiering is involved or not, the same problem can arise if
> > > > kswapd gets stopped due to kswapd_failures >= MAX_RECLAIM_RETRIES.
> > > 
> > > Your description is about tiering too.  Can you describe a situation
> > 
> > I mentioned "tiering" while I described how to reproduce because I ran
> > into the situation while testing with tiering system but I don't think
> > it's the necessary condition.
> > 
> > Let me ask you back, why the logic to stop kswapd was considered in the
> > first place?  That's because the problem was already observed anyway
> 
> To be clear..
> 
> The problem, kswapd_failures >= MAX_RECLAIM_RETRIES, can happen whether
> tiering is involved not not.  Once kswapd stops, the system should run
> without kswapd even after recovered e.g. by killing the hoggers.  *Even
> worse*, tiering mechanism doesn't work in this situation.

But like Ying said, in other situations it's direct reclaim that kicks
in and clears the flag.

The failure-sleep and direct reclaim triggered recovery have been in
place since 2017. Both parties who observed an issue with it recently
did so in tiering scenarios. IMO a tiering-specific solution makes the
most sense.