linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Johannes Weiner <hannes@cmpxchg.org>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Vlastimil Babka <vbabka@suse.cz>,
	Mel Gorman <mgorman@techsingularity.net>, Zi Yan <ziy@nvidia.com>,
	"Huang, Ying" <ying.huang@intel.com>,
	David Hildenbrand <david@redhat.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: [PATCH 02/10] mm: page_alloc: optimize free_unref_folios()
Date: Wed, 20 Mar 2024 14:02:07 -0400	[thread overview]
Message-ID: <20240320180429.678181-3-hannes@cmpxchg.org> (raw)
In-Reply-To: <20240320180429.678181-1-hannes@cmpxchg.org>

Move direct freeing of isolated pages to the lock-breaking block in
the second loop. This saves an unnecessary migratetype reassessment.

Minor comment and local variable scoping cleanups.

Suggested-by: Vlastimil Babka <vbabka@suse.cz>
Tested-by: "Huang, Ying" <ying.huang@intel.com>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
---
 mm/page_alloc.c | 32 +++++++++++++++++++++++---------
 1 file changed, 23 insertions(+), 9 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 60a632b7c9f6..994e4f790e92 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2524,7 +2524,7 @@ void free_unref_folios(struct folio_batch *folios)
 	unsigned long __maybe_unused UP_flags;
 	struct per_cpu_pages *pcp = NULL;
 	struct zone *locked_zone = NULL;
-	int i, j, migratetype;
+	int i, j;
 
 	/* Prepare folios for freeing */
 	for (i = 0, j = 0; i < folios->nr; i++) {
@@ -2536,14 +2536,15 @@ void free_unref_folios(struct folio_batch *folios)
 			folio_undo_large_rmappable(folio);
 		if (!free_pages_prepare(&folio->page, order))
 			continue;
-
 		/*
-		 * Free isolated folios and orders not handled on the PCP
-		 * directly to the allocator, see comment in free_unref_page.
+		 * Free orders not handled on the PCP directly to the
+		 * allocator.
 		 */
-		migratetype = get_pfnblock_migratetype(&folio->page, pfn);
-		if (!pcp_allowed_order(order) ||
-		    is_migrate_isolate(migratetype)) {
+		if (!pcp_allowed_order(order)) {
+			int migratetype;
+
+			migratetype = get_pfnblock_migratetype(&folio->page,
+							       pfn);
 			free_one_page(folio_zone(folio), &folio->page, pfn,
 					order, migratetype, FPI_NONE);
 			continue;
@@ -2560,15 +2561,29 @@ void free_unref_folios(struct folio_batch *folios)
 		struct zone *zone = folio_zone(folio);
 		unsigned long pfn = folio_pfn(folio);
 		unsigned int order = (unsigned long)folio->private;
+		int migratetype;
 
 		folio->private = NULL;
 		migratetype = get_pfnblock_migratetype(&folio->page, pfn);
 
 		/* Different zone requires a different pcp lock */
-		if (zone != locked_zone) {
+		if (zone != locked_zone ||
+		    is_migrate_isolate(migratetype)) {
 			if (pcp) {
 				pcp_spin_unlock(pcp);
 				pcp_trylock_finish(UP_flags);
+				locked_zone = NULL;
+				pcp = NULL;
+			}
+
+			/*
+			 * Free isolated pages directly to the
+			 * allocator, see comment in free_unref_page.
+			 */
+			if (is_migrate_isolate(migratetype)) {
+				free_one_page(zone, &folio->page, pfn,
+					      order, migratetype, FPI_NONE);
+				continue;
 			}
 
 			/*
@@ -2581,7 +2596,6 @@ void free_unref_folios(struct folio_batch *folios)
 				pcp_trylock_finish(UP_flags);
 				free_one_page(zone, &folio->page, pfn,
 					      order, migratetype, FPI_NONE);
-				locked_zone = NULL;
 				continue;
 			}
 			locked_zone = zone;
-- 
2.44.0



  parent reply	other threads:[~2024-03-20 18:04 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-20 18:02 [PATCH V4 00/10] mm: page_alloc: freelist migratetype hygiene Johannes Weiner
2024-03-20 18:02 ` [PATCH 01/10] mm: page_alloc: remove pcppage migratetype caching Johannes Weiner
2024-03-20 18:02 ` Johannes Weiner [this message]
2024-03-25 15:56   ` [PATCH 02/10] mm: page_alloc: optimize free_unref_folios() Vlastimil Babka
2024-03-20 18:02 ` [PATCH 03/10] mm: page_alloc: fix up block types when merging compatible blocks Johannes Weiner
2024-03-20 18:02 ` [PATCH 04/10] mm: page_alloc: move free pages when converting block during isolation Johannes Weiner
2024-03-20 18:02 ` [PATCH 05/10] mm: page_alloc: fix move_freepages_block() range error Johannes Weiner
2024-03-25 16:22   ` Vlastimil Babka
2024-03-20 18:02 ` [PATCH 06/10] mm: page_alloc: fix freelist movement during block conversion Johannes Weiner
2024-03-26 11:28   ` Vlastimil Babka
2024-03-26 12:34     ` Johannes Weiner
2024-04-05 12:11   ` Baolin Wang
2024-04-05 16:56     ` Johannes Weiner
2024-04-07  6:58       ` Baolin Wang
2024-04-08  7:24       ` Vlastimil Babka
2024-04-09  6:21       ` Vlastimil Babka
2024-03-20 18:02 ` [PATCH 07/10] mm: page_alloc: close migratetype race between freeing and stealing Johannes Weiner
2024-03-26 15:25   ` Vlastimil Babka
2024-03-20 18:02 ` [PATCH 08/10] mm: page_alloc: set migratetype inside move_freepages() Johannes Weiner
2024-03-26 15:40   ` Vlastimil Babka
2024-03-20 18:02 ` [PATCH 09/10] mm: page_isolation: prepare for hygienic freelists Johannes Weiner
2024-03-21 13:13   ` kernel test robot
2024-03-21 14:24     ` Johannes Weiner
2024-03-21 15:03       ` Zi Yan
2024-03-27  8:06   ` Vlastimil Babka
2024-03-20 18:02 ` [PATCH 10/10] mm: page_alloc: consolidate free page accounting Johannes Weiner
2024-03-27  8:54   ` Vlastimil Babka
2024-03-27 14:32     ` Johannes Weiner
2024-03-27 18:57     ` [PATCH 1/3] mm: page_alloc: consolidate free page accounting fix Johannes Weiner
2024-03-27 18:58     ` [PATCH 2/3] mm: page_alloc: consolidate free page accounting fix 2 Johannes Weiner
2024-03-27 19:01     ` [PATCH 3/3] mm: page_alloc: batch vmstat updates in expand() Johannes Weiner
2024-03-27 20:35       ` Vlastimil Babka
2024-04-07 10:19   ` [PATCH 10/10] mm: page_alloc: consolidate free page accounting Baolin Wang
2024-04-08  7:38     ` Vlastimil Babka
2024-04-08  9:13       ` Baolin Wang
2024-04-08 14:23       ` Johannes Weiner
2024-04-09  6:23         ` Vlastimil Babka
2024-04-09  7:48           ` [PATCH] mm: page_alloc: consolidate free page accounting fix 3 Baolin Wang
2024-04-09 21:15             ` kernel test robot
2024-04-09 22:36               ` Johannes Weiner
2024-04-09 21:25             ` kernel test robot
2024-04-09  7:56           ` [PATCH 10/10] mm: page_alloc: consolidate free page accounting Baolin Wang
2024-04-09  8:41             ` Vlastimil Babka
2024-04-09  9:31         ` Baolin Wang
2024-04-09 14:46           ` Zi Yan
2024-04-10  8:49             ` Baolin Wang
2024-03-27  9:30 ` [PATCH V4 00/10] mm: page_alloc: freelist migratetype hygiene Vlastimil Babka
2024-03-27 13:10   ` Zi Yan
2024-03-27 14:29   ` Johannes Weiner
2024-04-08  9:30 ` Baolin Wang
2024-04-08 14:24   ` Johannes Weiner
2024-05-11  5:14 ` Yu Zhao
2024-05-13 16:03   ` Johannes Weiner
2024-05-13 18:10     ` Yu Zhao
2024-05-13 19:04       ` Johannes Weiner
2024-06-05  4:53         ` Yu Zhao
2024-06-10 15:28           ` Johannes Weiner
2024-06-12 18:52             ` Yu Zhao
2024-06-13 15:39               ` Johannes Weiner

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=20240320180429.678181-3-hannes@cmpxchg.org \
    --to=hannes@cmpxchg.org \
    --cc=akpm@linux-foundation.org \
    --cc=david@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mgorman@techsingularity.net \
    --cc=vbabka@suse.cz \
    --cc=ying.huang@intel.com \
    --cc=ziy@nvidia.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