linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Ravi Jonnalagadda <ravis.opensrc@gmail.com>
To: damon@lists.linux.dev, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org
Cc: sj@kernel.org, akpm@linux-foundation.org, corbet@lwn.net,
	bijan311@gmail.com, ajayjoshi@micron.com,
	Ravi Jonnalagadda <ravis.opensrc@gmail.com>
Subject: [RFC PATCH 4/5] mm/damon/paddr: capacity clamp and directional early-exit for node_sys_bp
Date: Thu, 22 Jan 2026 20:57:27 -0800	[thread overview]
Message-ID: <20260123045733.6954-5-ravis.opensrc@gmail.com> (raw)
In-Reply-To: <20260123045733.6954-1-ravis.opensrc@gmail.com>

Clamp effective target to node capacity (bp) and skip in-migration if
the node already meets/exceeds it. This avoids oscillation and
unnecessary work in two-context DRAM/CXL setups when quota goals
(e.g., node_sys_bp) are met.

Signed-off-by: Ravi Jonnalagadda <ravis.opensrc@gmail.com>
---
 mm/damon/paddr.c | 44 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 44 insertions(+)

diff --git a/mm/damon/paddr.c b/mm/damon/paddr.c
index 30e4e5663dcb..64dbdd2196a5 100644
--- a/mm/damon/paddr.c
+++ b/mm/damon/paddr.c
@@ -300,10 +300,54 @@ static unsigned long damon_pa_deactivate_pages(struct damon_region *r,
 			sz_filter_passed);
 }
 
+static unsigned long damon_pa_node_capacity_bp(int nid)
+{
+	struct pglist_data *pgdat;
+	unsigned long sys_total = damon_pa_totalram_bytes();
+	unsigned long node_pages, node_total;
+
+	if (nid < 0 || !sys_total)
+		return 0;
+	pgdat = NODE_DATA(nid);
+	if (!pgdat)
+		return 0;
+	node_pages = pgdat->node_spanned_pages;
+	node_total = node_pages << PAGE_SHIFT;
+	return div64_u64((u64)node_total * 10000ULL, sys_total);
+}
+
 static unsigned long damon_pa_migrate(struct damon_region *r,
 		unsigned long addr_unit, struct damos *s,
 		unsigned long *sz_filter_passed)
 {
+	/*
+	 * Capacity clamp + directional early-exit for node_sys_bp goals:
+	 * If we are migrating INTO g->nid and the current bp for that node is
+	 * already >= min(target_bp, capacity_bp), skip work this interval.
+	 */
+	{
+		struct damos_quota_goal *g;
+
+		list_for_each_entry(g, &s->quota.goals, list) {
+			unsigned long cap_bp, effective_target_bp;
+
+			if (g->metric != DAMOS_QUOTA_NODE_SYS_BP)
+				continue;
+			if (g->nid < 0)
+				continue;
+
+			cap_bp = damon_pa_node_capacity_bp(g->nid);
+			if (!cap_bp)
+				break;
+
+			effective_target_bp = min(g->target_value, cap_bp);
+			if (s->target_nid == g->nid &&
+			    g->current_value >= effective_target_bp)
+				return 0;
+			break;
+		}
+	}
+
 	phys_addr_t addr, applied;
 	LIST_HEAD(folio_list);
 	struct folio *folio;
-- 
2.43.0



  parent reply	other threads:[~2026-01-23  4:57 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-23  4:57 [RFC PATCH 0/5] mm/damon: Add node_sys_bp quota goal metric for PA-based migration control Ravi Jonnalagadda
2026-01-23  4:57 ` [RFC PATCH 1/5] mm/damon/core: add DAMOS_QUOTA_NODE_SYS_BP metric Ravi Jonnalagadda
2026-01-23  4:57 ` [RFC PATCH 2/5] mm/damon: add get_goal_metric() op and PA provider Ravi Jonnalagadda
2026-01-23  4:57 ` [RFC PATCH 3/5] mm/damon/core: add new ops-specific goal metric Ravi Jonnalagadda
2026-01-23  4:57 ` Ravi Jonnalagadda [this message]
2026-01-23  4:57 ` [RFC PATCH 5/5] mm/damon/sysfs-schemes: accept "node_sys_bp" in goal's target_metric Ravi Jonnalagadda
2026-01-24  1:50 ` [RFC PATCH 0/5] mm/damon: Add node_sys_bp quota goal metric for PA-based migration control SeongJae Park
2026-01-27 18:52   ` Ravi Jonnalagadda
2026-01-28  1:25     ` SeongJae Park
2026-02-12  6:27       ` SeongJae Park
2026-02-04  2:25   ` [RFC PATCH 0/5] mm/damon: Add node_sys_bp quota goal metric for Yunjeong Mun
2026-02-04  6:06     ` SeongJae Park
2026-02-12  6:30       ` SeongJae Park

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=20260123045733.6954-5-ravis.opensrc@gmail.com \
    --to=ravis.opensrc@gmail.com \
    --cc=ajayjoshi@micron.com \
    --cc=akpm@linux-foundation.org \
    --cc=bijan311@gmail.com \
    --cc=corbet@lwn.net \
    --cc=damon@lists.linux.dev \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=sj@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