From: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
To: Michal Hocko <mhocko@kernel.org>
Cc: Gautham R Shenoy <ego@linux.vnet.ibm.com>,
David Hildenbrand <david@redhat.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
Satheesh Rajendran <sathnaga@linux.vnet.ibm.com>,
Mel Gorman <mgorman@suse.de>,
"Kirill A. Shutemov" <kirill@shutemov.name>,
Andrew Morton <akpm@linux-foundation.org>,
linuxppc-dev@lists.ozlabs.org, Christopher Lameter <cl@linux.com>,
Vlastimil Babka <vbabka@suse.cz>
Subject: Re: [PATCH v5 3/3] mm/page_alloc: Keep memoryless cpuless node 0 offline
Date: Thu, 2 Jul 2020 20:02:27 +0530 [thread overview]
Message-ID: <20200702143227.GE17918@linux.vnet.ibm.com> (raw)
In-Reply-To: <20200702084123.GC18446@dhcp22.suse.cz>
* Michal Hocko <mhocko@kernel.org> [2020-07-02 10:41:23]:
> On Thu 02-07-20 12:14:08, Srikar Dronamraju wrote:
> > * Michal Hocko <mhocko@kernel.org> [2020-07-01 14:21:10]:
> >
> > > > >>>>> The autonuma problem sounds interesting but again this patch doesn't
> > > > >>>>> really solve the underlying problem because I strongly suspect that the
> > > > >>>>> problem is still there when a numa node gets all its memory offline as
> > > > >>>>> mentioned above.
> > >
> > > I would really appreciate a feedback to these two as well.
> >
> > 1. Its not just numactl that's to be fixed but all tools/utilities that
> > depend on /sys/devices/system/node/online. Are we saying to not rely/believe
> > in the output given by the kernel but do further verification?
>
> No, what we are saying is that even an online node might have zero
> number of online pages/cpus. So the online status is not really
> something that matters. If people are confused by that output then user
> space tools can make their confusion go away. I really do not understand
> why the kernel should do any logic there.
The user facing teams are saying they are getting queries from the users who
are unable to understand from the tools/sysfs files why a node is online and
but has no attached resources. Its the amount of time that is being spent on
these issues that triggered the patch. Initially even I was skeptical that
this was a non-issue.
>
> > Also how would the user space differentiate between the case where the
> > Kernel missed marking a node as offline to the case where the memory was
> > offlined on a cpuless node but node wasn't offline?.
>
> What I am arguing is that those two shouldn't be any different. Really!
>
> > 2. Regarding the autonuma, the case of offline memory is user/admin driven,
> > so if there is a performance hit, its something that's driven by his
> > user/admin actions. Also how often do we see users offline complete memory
> > of cpuless node on a 2 node system?
>
> How often do we see crippled HW configurations like that? Really if
> autonuma should be made more clever for one case it should recognize the
> other as well.
>
Lets take a 16 socket PowerVM system and assume that 32 lpars are created
on that socket, i.e 2 lpars for each socket. (PowerVM has the final say on
how the lpars are created.) In such a case, we can expect 30 out of the 32
lpars to face this problem, with the only 2 lpars that actually run on
socket 0 having the correct configuration.
> > >
> > > This begs a question whether ppc can do the same thing?
> >
> > Certainly ppc can be made to adapt to this situation but that would be a
> > workaround. Do we have a reason why we think node 0 is unique and special?
>
> It is not. As replied in other email in this thread. I would hope for
> having less hacks in the numa initialization. Cleaning up the mess is
> would be a lot of work and testing on all NUMA capable architectures.
> This is a heritage from the past I am afraid. All that I am arguing here
> is that your touch to the generic code with a very simple looking patch
> might have side effects which are pretty much impossible to review.
> Moreover it seems that nothing but ppc really needs this treatment.
> So fixing it in ppc specific code sounds much more safe.
>
> Normally I would really push for a generic solution but after getting
> burned several times in this area I do not dare anymore. The problem is
> not in the code complexity but in how spread it is in places where you
> do not expect side effects.
>
I do understand and respect your viewpoint.
> --
> Michal Hocko
> SUSE Labs
--
Thanks and Regards
Srikar Dronamraju
next prev parent reply other threads:[~2020-07-02 14:45 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-06-24 9:28 [PATCH v5 0/3] Offline memoryless cpuless node 0 Srikar Dronamraju
2020-06-24 9:28 ` [PATCH v5 1/3] powerpc/numa: Set numa_node for all possible cpus Srikar Dronamraju
2020-06-24 9:48 ` Gautham R Shenoy
2020-06-24 9:28 ` [PATCH v5 2/3] powerpc/numa: Prefer node id queried from vphn Srikar Dronamraju
2020-06-24 10:29 ` Gautham R Shenoy
2020-06-24 9:28 ` [PATCH v5 3/3] mm/page_alloc: Keep memoryless cpuless node 0 offline Srikar Dronamraju
2020-06-29 14:58 ` Christopher Lameter
2020-06-30 4:01 ` Srikar Dronamraju
2020-07-01 12:23 ` Michal Hocko
2020-07-01 8:42 ` Michal Hocko
2020-07-01 10:04 ` Srikar Dronamraju
2020-07-01 10:15 ` David Hildenbrand
2020-07-01 11:01 ` Srikar Dronamraju
2020-07-01 11:06 ` David Hildenbrand
2020-07-01 11:30 ` David Hildenbrand
2020-07-01 12:21 ` Michal Hocko
2020-07-02 6:44 ` Srikar Dronamraju
2020-07-02 8:41 ` Michal Hocko
2020-07-02 14:32 ` Srikar Dronamraju [this message]
2020-07-03 9:10 ` Michal Suchánek
2020-07-03 9:24 ` Michal Hocko
2020-07-03 10:59 ` Michal Hocko
2020-07-03 11:32 ` David Hildenbrand
2020-07-03 11:46 ` Michal Hocko
2020-07-03 12:58 ` Srikar Dronamraju
2020-08-07 4:32 ` Andrew Morton
2020-08-07 6:58 ` David Hildenbrand
2020-08-07 10:04 ` Michal Suchánek
2020-08-12 6:01 ` Srikar Dronamraju
2020-08-18 7:32 ` David Hildenbrand
2020-08-18 7:37 ` Michal Hocko
2020-08-18 7:49 ` Srikar Dronamraju
2020-07-06 16:08 ` Andi Kleen
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=20200702143227.GE17918@linux.vnet.ibm.com \
--to=srikar@linux.vnet.ibm.com \
--cc=akpm@linux-foundation.org \
--cc=cl@linux.com \
--cc=david@redhat.com \
--cc=ego@linux.vnet.ibm.com \
--cc=kirill@shutemov.name \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=mgorman@suse.de \
--cc=mhocko@kernel.org \
--cc=sathnaga@linux.vnet.ibm.com \
--cc=torvalds@linux-foundation.org \
--cc=vbabka@suse.cz \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).