From: Vlastimil Babka <vbabka@suse.cz>
To: Michal Hocko <mhocko@kernel.org>, linux-mm@kvack.org
Cc: Mel Gorman <mgorman@suse.de>,
Andrew Morton <akpm@linux-foundation.org>,
Joonsoo Kim <js1304@gmail.com>, Jia He <hejianet@gmail.com>,
Taku Izumi <izumi.taku@jp.fujitsu.com>,
Johannes Weiner <hannes@cmpxchg.org>,
Michal Hocko <mhocko@suse.com>
Subject: Re: [PATCH 1/2] mm: fix remote numa hits statistics
Date: Mon, 2 Jan 2017 15:16:23 +0100 [thread overview]
Message-ID: <1d9e466b-dc87-eb41-113f-e7737a4bb7ef@suse.cz> (raw)
In-Reply-To: <20161221080653.29437-1-mhocko@kernel.org>
On 12/21/2016 09:06 AM, Michal Hocko wrote:
> From: Michal Hocko <mhocko@suse.com>
>
> Jia He has noticed that b9f00e147f27 ("mm, page_alloc: reduce branches
> in zone_statistics") has an unintentional side effect that remote node
> allocation requests are accounted as NUMA_MISS rathat than NUMA_HIT and
> NUMA_OTHER if such a request doesn't use __GFP_OTHER_NODE. There are
> many of these potentially because the flag is used very rarely while
> we have many users of __alloc_pages_node.
>
> Fix this by simply ignoring __GFP_OTHER_NODE (it can be removed in a
> follow up patch) and treat all allocations that were satisfied from the
> preferred zone's node as NUMA_HITS because this is the same node we
> requested the allocation from in most cases. If this is not the local
> node then we just account it as NUMA_OTHER rather than NUMA_LOCAL.
>
> One downsize would be that an allocation request for a node which is
> outside of the mempolicy nodemask would be reported as a hit which is a
> bit weird but that was the case before b9f00e147f27 already.
>
> Reported-by: Jia He <hejianet@gmail.com>
> Fixes: b9f00e147f27 ("mm, page_alloc: reduce branches in zone_statistics")
> Signed-off-by: Michal Hocko <mhocko@suse.com>
cbmc tells me that this patch is not equal to pre-commit b9f00e147f27
(in situation where __GFP_OTHER_NODE is not passed, as that's the only
relevant scenario after your patch), which seems unintended.
counter example:
numa_node_id() == 1
preferred_zone on node 0
allocated from zone on node 1
pre-b9f00e147f27:
allocated zone (node 1) increased NUMA_MISS and NUMA_LOCAL
preferred zone (node 0) increased NUMA_FOREIGN
(that looks correct to me)
your patch:
allocated zone (node 1) increased NUMA_MISS
preferred zone (node 0) increased NUMA_FOREIGN
i.e. it's missing NUMA_LOCAL on node 1, which is IMHO wrong as this was
an allocation local to the CPU (albeit a MISS wrt the preferred node).
Vlastimil
> ---
> mm/page_alloc.c | 16 ++++------------
> 1 file changed, 4 insertions(+), 12 deletions(-)
>
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index f6d5b73e1d7c..506946a902c5 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -2583,25 +2583,17 @@ int __isolate_free_page(struct page *page, unsigned int order)
> * Update NUMA hit/miss statistics
> *
> * Must be called with interrupts disabled.
> - *
> - * When __GFP_OTHER_NODE is set assume the node of the preferred
> - * zone is the local node. This is useful for daemons who allocate
> - * memory on behalf of other processes.
> */
> static inline void zone_statistics(struct zone *preferred_zone, struct zone *z,
> gfp_t flags)
> {
> #ifdef CONFIG_NUMA
> - int local_nid = numa_node_id();
> - enum zone_stat_item local_stat = NUMA_LOCAL;
> -
> - if (unlikely(flags & __GFP_OTHER_NODE)) {
> - local_stat = NUMA_OTHER;
> - local_nid = preferred_zone->node;
> - }
> + if (z->node == preferred_zone->node) {
> + enum zone_stat_item local_stat = NUMA_LOCAL;
>
> - if (z->node == local_nid) {
> __inc_zone_state(z, NUMA_HIT);
> + if (z->node != numa_node_id())
> + local_stat = NUMA_OTHER;
> __inc_zone_state(z, local_stat);
> } else {
> __inc_zone_state(z, NUMA_MISS);
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2017-01-02 14:16 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-12 5:59 [PATCH RFC 0/1] mm, page_alloc: fix incorrect zone_statistics data Jia He
2016-12-12 5:59 ` [PATCH RFC 1/1] " Jia He
2016-12-20 9:18 ` Michal Hocko
2016-12-20 13:10 ` Mel Gorman
2016-12-20 13:26 ` Michal Hocko
2016-12-20 14:28 ` Mel Gorman
2016-12-20 14:35 ` Michal Hocko
2016-12-20 14:49 ` Vlastimil Babka
2016-12-20 14:54 ` Mel Gorman
2016-12-21 7:57 ` Michal Hocko
2016-12-21 8:06 ` [PATCH 1/2] mm: fix remote numa hits statistics Michal Hocko
2016-12-21 8:06 ` [PATCH 2/2] mm: get rid of __GFP_OTHER_NODE Michal Hocko
2017-01-02 14:18 ` Vlastimil Babka
2016-12-29 11:46 ` [PATCH 1/2] mm: fix remote numa hits statistics Mel Gorman
2016-12-29 12:28 ` Michal Hocko
2017-01-02 14:16 ` Vlastimil Babka [this message]
2017-01-02 14:46 ` Michal Hocko
2017-01-02 15:07 ` Vlastimil Babka
2016-12-20 14:42 ` [PATCH RFC 1/1] mm, page_alloc: fix incorrect zone_statistics data Mel Gorman
2016-12-20 15:13 ` Vlastimil Babka
2016-12-21 3:01 ` hejianet
2016-12-20 12:31 ` Mel Gorman
2016-12-21 3:07 ` hejianet
2017-01-02 15:30 [PATCH 0/2] numa node stats alternative fix Michal Hocko
2017-01-02 15:30 ` [PATCH 1/2] mm: fix remote numa hits statistics 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=1d9e466b-dc87-eb41-113f-e7737a4bb7ef@suse.cz \
--to=vbabka@suse.cz \
--cc=akpm@linux-foundation.org \
--cc=hannes@cmpxchg.org \
--cc=hejianet@gmail.com \
--cc=izumi.taku@jp.fujitsu.com \
--cc=js1304@gmail.com \
--cc=linux-mm@kvack.org \
--cc=mgorman@suse.de \
--cc=mhocko@kernel.org \
--cc=mhocko@suse.com \
/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).