From: Michal Hocko <mhocko@suse.com>
To: David Hildenbrand <david@redhat.com>
Cc: Alexey Makhalov <amakhalov@vmware.com>,
Dennis Zhou <dennis@kernel.org>,
Eric Dumazet <eric.dumazet@gmail.com>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
Andrew Morton <akpm@linux-foundation.org>,
Oscar Salvador <osalvador@suse.de>, Tejun Heo <tj@kernel.org>,
Christoph Lameter <cl@linux.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"stable@vger.kernel.org" <stable@vger.kernel.org>
Subject: Re: [PATCH v3] mm: fix panic in __alloc_pages
Date: Wed, 8 Dec 2021 09:34:37 +0100 [thread overview]
Message-ID: <YbBuHSkvd6fDdQ9d@dhcp22.suse.cz> (raw)
In-Reply-To: <5a44c44a-141c-363d-c23e-558edc23b9b4@redhat.com>
On Wed 08-12-21 09:24:39, David Hildenbrand wrote:
> On 08.12.21 09:12, Michal Hocko wrote:
> > On Tue 07-12-21 19:03:28, David Hildenbrand wrote:
> >> On 07.12.21 18:17, Alexey Makhalov wrote:
> >>>
> >>>
> >>>> On Dec 7, 2021, at 9:13 AM, David Hildenbrand <david@redhat.com> wrote:
> >>>>
> >>>> On 07.12.21 18:02, Alexey Makhalov wrote:
> >>>>>
> >>>>>
> >>>>>> On Dec 7, 2021, at 8:36 AM, Michal Hocko <mhocko@suse.com> wrote:
> >>>>>>
> >>>>>> On Tue 07-12-21 17:27:29, Michal Hocko wrote:
> >>>>>> [...]
> >>>>>>> So your proposal is to drop set_node_online from the patch and add it as
> >>>>>>> a separate one which handles
> >>>>>>> - sysfs part (i.e. do not register a node which doesn't span a
> >>>>>>> physical address space)
> >>>>>>> - hotplug side of (drop the pgd allocation, register node lazily
> >>>>>>> when a first memblocks are registered)
> >>>>>>
> >>>>>> In other words, the first stage
> >>>>>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> >>>>>> index c5952749ad40..f9024ba09c53 100644
> >>>>>> --- a/mm/page_alloc.c
> >>>>>> +++ b/mm/page_alloc.c
> >>>>>> @@ -6382,7 +6382,11 @@ static void __build_all_zonelists(void *data)
> >>>>>> if (self && !node_online(self->node_id)) {
> >>>>>> build_zonelists(self);
> >>>>>> } else {
> >>>>>> - for_each_online_node(nid) {
> >>>>>> + /*
> >>>>>> + * All possible nodes have pgdat preallocated
> >>>>>> + * free_area_init
> >>>>>> + */
> >>>>>> + for_each_node(nid) {
> >>>>>> pg_data_t *pgdat = NODE_DATA(nid);
> >>>>>>
> >>>>>> build_zonelists(pgdat);
> >>>>>
> >>>>> Will it blow up memory usage for the nodes which might never be onlined?
> >>>>> I prefer the idea of init on demand.
> >>>>>
> >>>>> Even now there is an existing problem.
> >>>>> In my experiments, I observed _huge_ memory consumption increase by increasing number
> >>>>> of possible numa nodes. I’m going to report it in separate mail thread.
> >>>>
> >>>> I already raised that PPC might be problematic in that regard. Which
> >>>> architecture / setup do you have in mind that can have a lot of possible
> >>>> nodes?
> >>>>
> >>> It is x86_64 VMware VM, not the regular one, but specially configured (1 vCPU per node,
> >>> with hot-plug support, 128 possible nodes)
> >>
> >> I thought the pgdat would be smaller but I just gave it a test:
> >
> > Yes, pgdat is quite large! Just embeded zones can eat a lot.
> >
> >> On my system, pgdata_t is 173824 bytes. So 128 nodes would correspond to
> >> 21 MiB, which is indeed a lot. I assume it's due to "struct zonelist",
> >> which has MAX_ZONES_PER_ZONELIST == (MAX_NUMNODES * MAX_NR_ZONES) zone
> >> references ...
> >
> > This is what pahole tells me
> > struct pglist_data {
> > struct zone node_zones[4] __attribute__((__aligned__(64))); /* 0 5632 */
> > /* --- cacheline 88 boundary (5632 bytes) --- */
> > struct zonelist node_zonelists[1]; /* 5632 80 */
> > [...]
> > /* size: 6400, cachelines: 100, members: 27 */
> > /* sum members: 6369, holes: 5, sum holes: 31 */
> >
> > with my particular config (which is !NUMA). I haven't really checked
> > whether there are other places which might scale with MAX_NUM_NODES or
> > something like that.
> >
> > Anyway, is 21MB of wasted space for 128 Node machine something really
> > note worthy?
> >
>
> I think we'll soon might see setups (again, CXL is an example, but als
> owhen providing a dynamic amount of performance differentiated memory
> via virtio-mem) where this will most probably matter. With performance
> differentiated memory we'll see a lot more nodes getting used in
> general, and a lot more nodes eventually getting hotplugged.
There are certainly machines with many nodes. E.g. SLES kernels are
build with CONFIG_NODES_SHIFT=10 which is a lot of potential nodes.
And I have seen really large machines with many nodes but those usually
come with a lot of memory and they do not tend to have non populated
nodes AFAIR.
> If 128 nodes is realistic, I cannot tell.
>
> We could optimize by allocating some members dynamically. For example
> we'll never need MAX_NUMNODES entries, but only the number of possible
> nodes.
Yes agreed. Scaling with MAX_NUMNODES is almost always wasteful.
--
Michal Hocko
SUSE Labs
next prev parent reply other threads:[~2021-12-08 8:34 UTC|newest]
Thread overview: 81+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-01 20:13 [PATCH] mm: fix panic in __alloc_pages Alexey Makhalov
2021-11-01 20:38 ` Matthew Wilcox
2021-11-02 7:47 ` Michal Hocko
2021-11-02 8:12 ` David Hildenbrand
2021-11-02 8:48 ` Alexey Makhalov
2021-11-02 9:04 ` Michal Hocko
2021-11-02 9:24 ` David Hildenbrand
2021-11-02 10:34 ` Alexey Makhalov
2021-11-02 11:00 ` David Hildenbrand
2021-11-02 11:44 ` Michal Hocko
2021-11-02 12:06 ` David Hildenbrand
2021-11-02 12:27 ` Michal Hocko
2021-11-02 12:39 ` David Hildenbrand
2021-11-02 13:25 ` Michal Hocko
2021-11-02 13:41 ` David Hildenbrand
2021-11-02 14:12 ` Michal Hocko
2021-11-02 14:44 ` David Hildenbrand
2021-11-02 13:52 ` Oscar Salvador
2021-11-02 14:35 ` Michal Hocko
2021-11-08 6:12 ` Alexey Makhalov
2021-11-08 6:36 ` [PATCH v2] " Alexey Makhalov
2021-11-08 8:32 ` David Hildenbrand
2021-11-08 20:23 ` [PATCH v3] " Alexey Makhalov
2021-11-09 2:08 ` Eric Dumazet
2021-11-09 7:03 ` David Hildenbrand
2021-11-09 16:55 ` Eric Dumazet
2021-11-09 17:15 ` Michal Hocko
2021-11-09 19:06 ` Dennis Zhou
2021-11-09 19:54 ` Michal Hocko
2021-11-16 1:31 ` Alexey Makhalov
2021-11-16 9:17 ` Michal Hocko
2021-11-16 20:22 ` Alexey Makhalov
2021-11-18 8:35 ` Michal Hocko
2021-12-07 10:54 ` Michal Hocko
2021-12-07 11:08 ` David Hildenbrand
2021-12-07 12:13 ` Michal Hocko
2021-12-07 12:28 ` David Hildenbrand
2021-12-07 13:23 ` Michal Hocko
2021-12-07 15:09 ` David Hildenbrand
2021-12-07 15:29 ` Michal Hocko
2021-12-07 15:34 ` David Hildenbrand
2021-12-07 15:56 ` Michal Hocko
2021-12-07 16:09 ` David Hildenbrand
2021-12-07 16:27 ` Michal Hocko
2021-12-07 16:36 ` Michal Hocko
2021-12-07 16:40 ` David Hildenbrand
2021-12-08 8:28 ` Michal Hocko
2021-12-07 17:02 ` Alexey Makhalov
2021-12-07 17:13 ` David Hildenbrand
2021-12-07 17:17 ` Alexey Makhalov
2021-12-07 18:03 ` David Hildenbrand
2021-12-08 8:12 ` Michal Hocko
2021-12-08 8:24 ` David Hildenbrand
2021-12-08 8:34 ` Michal Hocko [this message]
2021-12-08 8:38 ` David Hildenbrand
2021-12-08 8:04 ` Michal Hocko
2021-12-08 8:19 ` Alexey Makhalov
2021-12-08 8:30 ` Michal Hocko
2021-12-08 8:54 ` Michal Hocko
2021-12-08 8:57 ` Alexey Makhalov
2021-12-08 9:55 ` Michal Hocko
2021-12-09 2:16 ` Alexey Makhalov
2021-12-09 8:46 ` Michal Hocko
2021-12-09 9:28 ` Alexey Makhalov
2021-12-09 9:56 ` Michal Hocko
2021-12-09 10:23 ` Alexey Makhalov
2021-12-09 13:29 ` Michal Hocko
2021-12-09 19:01 ` Alexey Makhalov
2021-12-10 9:11 ` Michal Hocko
2021-12-17 12:53 ` Michal Hocko
2021-12-21 5:46 ` Alexey Makhalov
2021-12-21 9:46 ` Michal Hocko
2021-12-21 20:23 ` Alexey Makhalov
2021-12-22 11:41 ` Michal Hocko
2021-12-09 10:48 ` Michal Hocko
2021-12-13 15:06 ` Michal Hocko
2021-12-13 15:07 ` David Hildenbrand
2021-12-14 8:38 ` Michal Hocko
2021-11-08 10:37 ` [PATCH v2] " Michal Hocko
2021-11-02 9:40 ` [PATCH] " Alexey Makhalov
2021-11-02 9:40 ` Michal Hocko
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=YbBuHSkvd6fDdQ9d@dhcp22.suse.cz \
--to=mhocko@suse.com \
--cc=akpm@linux-foundation.org \
--cc=amakhalov@vmware.com \
--cc=cl@linux.com \
--cc=david@redhat.com \
--cc=dennis@kernel.org \
--cc=eric.dumazet@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=osalvador@suse.de \
--cc=stable@vger.kernel.org \
--cc=tj@kernel.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
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).