linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Claudio Imbrenda <imbrenda@linux.ibm.com>
To: David Hildenbrand <david@redhat.com>
Cc: kvm@vger.kernel.org, cohuck@redhat.com, borntraeger@de.ibm.com,
	frankja@linux.ibm.com, thuth@redhat.com, pasic@linux.ibm.com,
	linux-s390@vger.kernel.org, linux-kernel@vger.kernel.org,
	Ulrich.Weigand@de.ibm.com,
	"linux-mm@kvack.org" <linux-mm@kvack.org>,
	Michal Hocko <mhocko@kernel.org>
Subject: Re: [PATCH v3 00/14] KVM: s390: pv: implement lazy destroy
Date: Fri, 6 Aug 2021 15:44:00 +0200	[thread overview]
Message-ID: <20210806154400.2ca55563@p-imbrenda> (raw)
In-Reply-To: <ada27c6d-4dc9-04c3-d5b9-566e65359701@redhat.com>

On Fri, 6 Aug 2021 13:30:21 +0200
David Hildenbrand <david@redhat.com> wrote:

[...]

> >>> When the system runs out of memory, if a guest has terminated and
> >>> its memory is being cleaned asynchronously, the OOM killer will
> >>> wait a little and then see if memory has been freed. This has the
> >>> practical effect of slowing down memory allocations when the
> >>> system is out of memory to give the cleanup thread time to
> >>> cleanup and free memory, and avoid an actual OOM situation.  
> >>
> >> ... and this sound like the kind of arch MM hacks that will bite us
> >> in the long run. Of course, I might be wrong, but already doing
> >> excessive GFP_ATOMIC allocations or messing with the OOM killer
> >> that  
> > 
> > they are GFP_ATOMIC but they should not put too much weight on the
> > memory and can also fail without consequences, I used:
> > 
> > GFP_ATOMIC | __GFP_NOMEMALLOC | __GFP_NOWARN
> > 
> > also notice that after every page allocation a page gets freed, so
> > this is only temporary.  
> 
> Correct me if I'm wrong: you're allocating unmovable pages for
> tracking (e.g., ZONE_DMA, ZONE_NORMAL) from atomic reserves and will
> free a movable process page, correct? Or which page will you be
> freeing?

we are transforming ALL moveable pages belonging to userspace into
unmoveable pages. every ~500 pages one page gets actually
allocated (unmoveable), and another (moveable) one gets freed.

> > 
> > I would not call it "messing with the OOM killer", I'm using the
> > same interface used by virtio-baloon  
> 
> Right, and for virtio-balloon it's actually a workaround to restore
> the original behavior of a rarely used feature: deflate-on-oom.
> Commit da10329cb057 ("virtio-balloon: switch back to OOM handler for 
> VIRTIO_BALLOON_F_DEFLATE_ON_OOM") tried to document why we switched
> back from a shrinker to VIRTIO_BALLOON_F_DEFLATE_ON_OOM:
> 
> "The name "deflate on OOM" makes it pretty clear when deflation should
>   happen - after other approaches to reclaim memory failed, not while
>   reclaiming. This allows to minimize the footprint of a guest -
> memory will only be taken out of the balloon when really needed."
> 
> Note some subtle differences:
> 
> a) IIRC, before running into the OOM killer, will try reclaiming
>     anything  else. This is what we want for deflate-on-oom, it might
> not be what you want for your feature (e.g., flushing other
> processes/VMs to disk/swap instead of waiting for a single process to
> stop).

we are already reclaiming the memory of the dead secure guest.

> b) Migration of movable balloon inflated pages continues working
> because we are dealing with non-lru page migration.
> 
> Will page reclaim, page migration, compaction, ... of these movable
> LRU pages still continue working while they are sitting around
> waiting to be cleaned up? I can see that we're grabbing an extra
> reference when we put them onto the list, that might be a problem:
> for example, we can most certainly not swap out these pages or write
> them back to disk on memory pressure.

this is true. on the other hand, swapping a moveable page would be even
slower, because those pages would need to be exported and not destroyed.

> >   
> >> way for a pure (shutdown) optimization is an alarm signal. Of
> >> course, I might be wrong.
> >>
> >> You should at least CC linux-mm. I'll do that right now and also CC
> >> Michal. He might have time to have a quick glimpse at patch #11 and
> >> #13.
> >>
> >> https://lkml.kernel.org/r/20210804154046.88552-12-imbrenda@linux.ibm.com
> >> https://lkml.kernel.org/r/20210804154046.88552-14-imbrenda@linux.ibm.com
> >>
> >> IMHO, we should proceed with patch 1-10, as they solve a really
> >> important problem ("slow reboots") in a nice way, whereby patch 11
> >> handles a case that can be worked around comparatively easily by
> >> management tools -- my 2 cents.  
> > 
> > how would management tools work around the issue that a shutdown can
> > take very long?  
> 
> The traditional approach is to wait starting a new VM on another 
> hypervisor instead until memory has been freed up, or start it on 
> another hypervisor. That raises the question about the target use
> case.
> 
> What I don't get is that we have to pay the price for freeing up that 
> memory. Why isn't it sufficient to keep the process running and let 
> ordinary MM do it's thing?

what price?

you mean let mm do the slowest possible thing when tearing down a dead
guest?

without this, the dying guest would still take up all the memory. and
swapping it would not be any faster (it would be slower, in fact). the
system would OOM anyway.

> Maybe you should clearly spell out what the target use case for the
> fast shutdown (fast quitting of the process?) is?. I assume it is,
> starting a new VM / process / whatsoever on the same host
> immediately, and then
> 
> a) Eventually slowing down other processes due heavy reclaim.

for each dying guest, only one CPU is used by the reclaim; depending on
the total load of the system, this might not even be noticeable

> b) Slowing down the new process because you have to pay the price of 
> cleaning up memory.

do you prefer to OOM because the dying guest will need ages to clean up
its memory?

> I think I am missing why we need the lazy destroy at all when killing
> a process. Couldn't you instead teach the OOM killer "hey, we're
> currently quitting a heavy process that is just *very* slow to free
> up memory, please wait for that before starting shooting around" ?

isn't this ^ exactly what the OOM notifier does?


another note here:

when the process quits, the mm starts the tear down. at this point, the
mm has no idea that this is a dying KVM guest, so the best it can do is
exporting (which is significantly slower than destroy page)

kvm comes into play long after the mm is gone, and at this point it
can't do anything anymore. the memory is already gone (very slowly).

if I kill -9 qemu (or if qemu segfaults), KVM will never notice until
the mm is gone.


  reply	other threads:[~2021-08-06 13:44 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-04 15:40 Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 01/14] KVM: s390: pv: add macros for UVC CC values Claudio Imbrenda
2021-08-06  7:26   ` David Hildenbrand
2021-08-06  9:34     ` Claudio Imbrenda
2021-08-06 15:15     ` Janosch Frank
2021-08-04 15:40 ` [PATCH v3 02/14] KVM: s390: pv: avoid stall notifications for some UVCs Claudio Imbrenda
2021-08-06  7:30   ` David Hildenbrand
2021-08-06  9:33     ` Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 03/14] KVM: s390: pv: leak the ASCE page when destroy fails Claudio Imbrenda
2021-08-06  7:31   ` David Hildenbrand
2021-08-06  9:32     ` Claudio Imbrenda
2021-08-06 11:39       ` David Hildenbrand
2021-08-04 15:40 ` [PATCH v3 04/14] KVM: s390: pv: properly handle page flags for protected guests Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 05/14] KVM: s390: pv: handle secure storage violations " Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 06/14] KVM: s390: pv: handle secure storage exceptions for normal guests Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 07/14] KVM: s390: pv: refactor s390_reset_acc Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 08/14] KVM: s390: pv: usage counter instead of flag Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 09/14] KVM: s390: pv: add export before import Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 10/14] KVM: s390: pv: lazy destroy for reboot Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 11/14] KVM: s390: pv: extend lazy destroy to handle shutdown Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 12/14] KVM: s390: pv: module parameter to fence lazy destroy Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 13/14] KVM: s390: pv: add OOM notifier for " Claudio Imbrenda
2021-08-04 15:40 ` [PATCH v3 14/14] KVM: s390: pv: avoid export before import if possible Claudio Imbrenda
2021-08-06  7:10 ` [PATCH v3 00/14] KVM: s390: pv: implement lazy destroy David Hildenbrand
2021-08-06  9:30   ` Claudio Imbrenda
2021-08-06 11:30     ` David Hildenbrand
2021-08-06 13:44       ` Claudio Imbrenda [this message]
2021-08-09  8:50         ` David Hildenbrand

Reply instructions:

You may reply publicly 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=20210806154400.2ca55563@p-imbrenda \
    --to=imbrenda@linux.ibm.com \
    --cc=Ulrich.Weigand@de.ibm.com \
    --cc=borntraeger@de.ibm.com \
    --cc=cohuck@redhat.com \
    --cc=david@redhat.com \
    --cc=frankja@linux.ibm.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=mhocko@kernel.org \
    --cc=pasic@linux.ibm.com \
    --cc=thuth@redhat.com \
    --subject='Re: [PATCH v3 00/14] KVM: s390: pv: implement lazy destroy' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).