From: Zhiguo Jiang <justinjiang@vivo.com>
To: Andrew Morton <akpm@linux-foundation.org>,
David Hildenbrand <david@redhat.com>,
Matthew Wilcox <willy@infradead.org>,
Chris Li <chrisl@kernel.org>, Michal Hocko <mhocko@suse.com>,
Kefeng Wang <wangkefeng.wang@huawei.com>,
Dan Schatzberg <schatzberg.dan@gmail.com>,
Yosry Ahmed <yosryahmed@google.com>,
Yue Zhao <findns94@gmail.com>, Hugh Dickins <hughd@google.com>,
Johannes Weiner <hannes@cmpxchg.org>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org
Cc: opensource.kernel@vivo.com, Zhiguo Jiang <justinjiang@vivo.com>
Subject: [PATCH v1 2/2] mm:vmscan: fix shrink sc parameters issue
Date: Thu, 11 Jan 2024 20:24:51 +0800 [thread overview]
Message-ID: <20240111122451.682-3-justinjiang@vivo.com> (raw)
In-Reply-To: <20240111122451.682-1-justinjiang@vivo.com>
The values of struct scan_control's members obtained by
prepare_scan_control() are always from root_mem_cgroup, rather than
the memcg where the shrinked folio is actually located, such as
sc->anon_cost, sc->file_cost, sc->may_deactivate and so on. This is
an inaccurate sc values for the actual situation of the current
shrinking memcg. so we need fix the values of struct scan_control's
members are corresponding to the current shrinking memcg.
Signed-off-by: Zhiguo Jiang <justinjiang@vivo.com>
---
mm/vmscan.c | 35 ++++++++++++++++++-----------------
1 file changed, 18 insertions(+), 17 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 8a1fbdaca042..31001123e8f1 100755
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2229,29 +2229,30 @@ enum scan_balance {
SCAN_FILE,
};
-static void prepare_scan_control(pg_data_t *pgdat, struct scan_control *sc)
+static void prepare_scan_control(pg_data_t *pgdat, struct scan_control *sc,
+ struct mem_cgroup *memcg)
{
unsigned long file;
- struct lruvec *target_lruvec;
+ struct lruvec *lruvec;
if (lru_gen_enabled())
return;
- target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat);
+ lruvec = mem_cgroup_lruvec(memcg, pgdat);
/*
* Flush the memory cgroup stats, so that we read accurate per-memcg
* lruvec stats for heuristics.
*/
- mem_cgroup_flush_stats(sc->target_mem_cgroup);
+ mem_cgroup_flush_stats(memcg);
/*
* Determine the scan balance between anon and file LRUs.
*/
- spin_lock_irq(&target_lruvec->lru_lock);
- sc->anon_cost = target_lruvec->anon_cost;
- sc->file_cost = target_lruvec->file_cost;
- spin_unlock_irq(&target_lruvec->lru_lock);
+ spin_lock_irq(&lruvec->lru_lock);
+ sc->anon_cost = lruvec->anon_cost;
+ sc->file_cost = lruvec->file_cost;
+ spin_unlock_irq(&lruvec->lru_lock);
/*
* Target desirable inactive:active list ratios for the anon
@@ -2265,18 +2266,18 @@ static void prepare_scan_control(pg_data_t *pgdat, struct scan_control *sc)
* workingset is being established. Deactivate to get
* rid of any stale active pages quickly.
*/
- refaults = lruvec_page_state(target_lruvec,
+ refaults = lruvec_page_state(lruvec,
WORKINGSET_ACTIVATE_ANON);
- if (refaults != target_lruvec->refaults[WORKINGSET_ANON] ||
- inactive_is_low(target_lruvec, LRU_INACTIVE_ANON))
+ if (refaults != lruvec->refaults[WORKINGSET_ANON] ||
+ inactive_is_low(lruvec, LRU_INACTIVE_ANON))
sc->may_deactivate |= DEACTIVATE_ANON;
else
sc->may_deactivate &= ~DEACTIVATE_ANON;
- refaults = lruvec_page_state(target_lruvec,
+ refaults = lruvec_page_state(lruvec,
WORKINGSET_ACTIVATE_FILE);
- if (refaults != target_lruvec->refaults[WORKINGSET_FILE] ||
- inactive_is_low(target_lruvec, LRU_INACTIVE_FILE))
+ if (refaults != lruvec->refaults[WORKINGSET_FILE] ||
+ inactive_is_low(lruvec, LRU_INACTIVE_FILE))
sc->may_deactivate |= DEACTIVATE_FILE;
else
sc->may_deactivate &= ~DEACTIVATE_FILE;
@@ -2288,7 +2289,7 @@ static void prepare_scan_control(pg_data_t *pgdat, struct scan_control *sc)
* thrashing, try to reclaim those first before touching
* anonymous pages.
*/
- file = lruvec_page_state(target_lruvec, NR_INACTIVE_FILE);
+ file = lruvec_page_state(lruvec, NR_INACTIVE_FILE);
if (file >> sc->priority && !(sc->may_deactivate & DEACTIVATE_FILE))
sc->cache_trim_mode = 1;
else
@@ -5885,6 +5886,8 @@ static void shrink_node_memcgs(pg_data_t *pgdat, struct scan_control *sc)
reclaimed = sc->nr_reclaimed;
scanned = sc->nr_scanned;
+ prepare_scan_control(pgdat, sc, memcg);
+
shrink_lruvec(lruvec, sc);
shrink_slab(sc->gfp_mask, pgdat->node_id, memcg,
@@ -5918,8 +5921,6 @@ static void shrink_node(pg_data_t *pgdat, struct scan_control *sc)
nr_reclaimed = sc->nr_reclaimed;
nr_scanned = sc->nr_scanned;
- prepare_scan_control(pgdat, sc);
-
shrink_node_memcgs(pgdat, sc);
flush_reclaim_state(sc);
--
2.39.0
next prev parent reply other threads:[~2024-01-11 12:25 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-11 12:24 [PATCH v1 0/2] *** SUBJECT HERE *** Zhiguo Jiang
2024-01-11 12:24 ` [PATCH v1 1/2] mm:vmscan: fix workingset eviction memcg issue Zhiguo Jiang
2024-01-11 13:41 ` Johannes Weiner
2024-01-11 14:15 ` zhiguojiang
2024-01-11 12:24 ` Zhiguo Jiang [this message]
2024-01-11 13:43 ` [PATCH v1 2/2] mm:vmscan: fix shrink sc parameters issue Johannes Weiner
2024-01-12 1:27 ` zhiguojiang
2024-01-11 12:51 ` [PATCH v1 0/2] *** SUBJECT HERE *** Christophe Leroy
2024-01-11 13:10 ` zhiguojiang
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=20240111122451.682-3-justinjiang@vivo.com \
--to=justinjiang@vivo.com \
--cc=akpm@linux-foundation.org \
--cc=chrisl@kernel.org \
--cc=david@redhat.com \
--cc=findns94@gmail.com \
--cc=hannes@cmpxchg.org \
--cc=hughd@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=opensource.kernel@vivo.com \
--cc=schatzberg.dan@gmail.com \
--cc=wangkefeng.wang@huawei.com \
--cc=willy@infradead.org \
--cc=yosryahmed@google.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