Linux-mm Archive on lore.kernel.org
 help / color / Atom feed
From: Vlastimil Babka <vbabka@suse.cz>
To: Michal Hocko <mhocko@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	David Rientjes <rientjes@google.com>,
	Andrea Arcangeli <aarcange@redhat.com>,
	mm-commits@vger.kernel.org, Linux-MM <linux-mm@kvack.org>
Subject: Re: incoming
Date: Tue, 24 Sep 2019 21:55:09 +0200
Message-ID: <cad8a751-362c-a3b9-ced4-494c4ff20e7f@suse.cz> (raw)
In-Reply-To: <20190924074843.GC23050@dhcp22.suse.cz>

On 9/24/19 9:48 AM, Michal Hocko wrote:
> On Mon 23-09-19 21:31:53, Andrew Morton wrote:
>> On Mon, 23 Sep 2019 17:55:24 -0700 Linus Torvalds
>> <torvalds@linux-foundation.org> wrote:
>> 
>>> On Mon, Sep 23, 2019 at 3:31 PM Andrew Morton
>>> <akpm@linux-foundation.org> wrote:
>>>> 
>>>> - almost all of -mm, as below.
>>> 
>>> I was hoping that we could at least test the THP locality thing?
>>> Is it in your queue at all, or am I supposed to just do it
>>> myself?
>>> 
>> 
>> Confused.  I saw a privately emailed patch from David which nobody 
>> seems to have tested yet.  I parked that for consideration after
>> -rc1. Or are you referring to something else?
>> 
>> This thing keeps stalling.  It would be nice to push this along and
>> get something nailed down which we can at least get into 5.4-rc,
>> perhaps with a backport-this tag?
> 
> The patch proposed by David is really non trivial wrt. potential
> side effects. I have provided my review feedback [1] and it didn't
> get any reaction. I really believe that we need to debug this
> properly. A reproducer would be useful for others to work on that.
> 
> There is a more fundamental problem here and we need to address it 
> rather than to duck tape it and whack a mole afterwards.

I believe we found a problem when investigating over-reclaim in this
thread [1] where it seems madvised THP allocation attempt can result in
4MB reclaimed, if there is a small zone such as ZONE_DMA on the node. As
it happens, the patch "[patch 090/134] mm, reclaim: make
should_continue_reclaim perform dryrun detection" in Andrew's pile
should change this 4MB to 32 pages reclaimed (as a side-effect), but
that has to be tested. I'm also working on a patch to not reclaim even
those few pages. Of course there might be more fundamental issues with
reclaim/compaction interaction, but this one seems to become hopefully
clear now.

[1]
https://lore.kernel.org/linux-mm/4b4ba042-3741-7b16-2292-198c569da2aa@profihost.ag/

> [1] http://lkml.kernel.org/r/20190909193020.GD2063@dhcp22.suse.cz
> 



  parent reply index

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-23 22:31 incoming Andrew Morton
2019-09-24  0:55 ` incoming Linus Torvalds
2019-09-24  4:31   ` incoming Andrew Morton
2019-09-24  7:48     ` incoming Michal Hocko
2019-09-24 15:34       ` incoming Linus Torvalds
2019-09-25  6:36         ` incoming Michal Hocko
2019-09-24 19:55       ` Vlastimil Babka [this message]
  -- strict thread matches above, loose matches on Subject: below --
2020-02-21  4:00 incoming Andrew Morton
2020-02-21  4:03 ` incoming Andrew Morton
2020-02-21 18:21 ` incoming Linus Torvalds
2020-02-21 18:32   ` incoming Konstantin Ryabitsev
2020-02-21 19:33   ` incoming Linus Torvalds
2020-02-04  1:33 incoming Andrew Morton
2020-02-04  2:27 ` incoming Linus Torvalds
2020-02-04  2:46   ` incoming Andrew Morton
2020-02-04  3:11     ` incoming Linus Torvalds
2020-01-31  6:10 incoming Andrew Morton
2020-01-14  0:28 incoming Andrew Morton
2020-01-04 20:55 incoming Andrew Morton
2019-12-18  4:50 incoming Andrew Morton
2019-12-05  0:48 incoming Andrew Morton
2019-12-01  1:47 incoming Andrew Morton
2019-12-01  5:17 ` incoming James Bottomley
2019-12-01 21:07 ` incoming Linus Torvalds
2019-12-02  8:21   ` incoming Steven Price
2019-11-22  1:53 incoming Andrew Morton
2019-11-16  1:34 incoming Andrew Morton
2019-11-06  5:16 incoming Andrew Morton
2019-10-19  3:19 incoming Andrew Morton
2019-10-14 21:11 incoming Andrew Morton
2019-10-07  0:57 incoming Andrew Morton
2019-09-25 23:45 incoming Andrew Morton
2019-08-30 23:04 incoming Andrew Morton
2019-08-25  0:54 incoming Andrew Morton
     [not found] <20190716162536.bb52b8f34a8ecf5331a86a42@linux-foundation.org>
2019-07-17  8:47 ` incoming Vlastimil Babka
2019-07-17  8:57   ` incoming Bhaskar Chowdhury
2019-07-17 16:13   ` incoming Linus Torvalds
2019-07-17 17:09     ` incoming Christian Brauner
2019-07-17 18:13     ` incoming Vlastimil Babka
2007-05-02 22:02 incoming Andrew Morton
2007-05-02 22:31 ` incoming Benjamin Herrenschmidt
2007-05-03  7:55 ` incoming Russell King
2007-05-03  8:05   ` incoming Andrew Morton
2007-05-04 13:37 ` incoming Greg KH
2007-05-04 16:14   ` incoming Andrew Morton
2007-05-04 17:02     ` incoming Greg KH
2007-05-04 18:57     ` incoming Roland McGrath
2007-05-04 19:24       ` incoming Greg KH
2007-05-04 19:29         ` incoming Roland McGrath

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=cad8a751-362c-a3b9-ced4-494c4ff20e7f@suse.cz \
    --to=vbabka@suse.cz \
    --cc=aarcange@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@kernel.org \
    --cc=mm-commits@vger.kernel.org \
    --cc=rientjes@google.com \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-mm Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-mm/0 linux-mm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-mm linux-mm/ https://lore.kernel.org/linux-mm \
		linux-mm@kvack.org
	public-inbox-index linux-mm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kvack.linux-mm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git