public inbox for [email protected]
 help / color / mirror / Atom feed
From: "Matthew Wilcox (Oracle)" <[email protected]>
To: Andrew Morton <[email protected]>
Cc: "Matthew Wilcox (Oracle)" <[email protected]>,
	Jens Axboe <[email protected]>,
	[email protected], [email protected]
Subject: [PATCH v2 04/13] mm: Convert free_transhuge_folio() to folio_undo_large_rmappable()
Date: Wed, 16 Aug 2023 16:11:52 +0100	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

Indirect calls are expensive, thanks to Spectre.  Test for
TRANSHUGE_PAGE_DTOR and destroy the folio appropriately.  Move the
free_compound_page() call into destroy_large_folio() to simplify later
patches.

Signed-off-by: Matthew Wilcox (Oracle) <[email protected]>
---
 include/linux/huge_mm.h |  2 --
 include/linux/mm.h      |  2 --
 mm/huge_memory.c        | 22 +++++++++++-----------
 mm/internal.h           |  2 ++
 mm/page_alloc.c         |  9 ++++++---
 5 files changed, 19 insertions(+), 18 deletions(-)

diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 20284387b841..f351c3f9d58b 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -144,8 +144,6 @@ unsigned long thp_get_unmapped_area(struct file *filp, unsigned long addr,
 		unsigned long len, unsigned long pgoff, unsigned long flags);
 
 void prep_transhuge_page(struct page *page);
-void free_transhuge_page(struct page *page);
-
 bool can_split_folio(struct folio *folio, int *pextra_pins);
 int split_huge_page_to_list(struct page *page, struct list_head *list);
 static inline int split_huge_page(struct page *page)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 19493d6a2bb8..6c338b65b86b 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1281,9 +1281,7 @@ enum compound_dtor_id {
 #ifdef CONFIG_HUGETLB_PAGE
 	HUGETLB_PAGE_DTOR,
 #endif
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
 	TRANSHUGE_PAGE_DTOR,
-#endif
 	NR_COMPOUND_DTORS,
 };
 
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 8480728fa220..9598bbe6c792 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2779,10 +2779,9 @@ int split_huge_page_to_list(struct page *page, struct list_head *list)
 	return ret;
 }
 
-void free_transhuge_page(struct page *page)
+void folio_undo_large_rmappable(struct folio *folio)
 {
-	struct folio *folio = (struct folio *)page;
-	struct deferred_split *ds_queue = get_deferred_split_queue(folio);
+	struct deferred_split *ds_queue;
 	unsigned long flags;
 
 	/*
@@ -2790,15 +2789,16 @@ void free_transhuge_page(struct page *page)
 	 * deferred_list. If folio is not in deferred_list, it's safe
 	 * to check without acquiring the split_queue_lock.
 	 */
-	if (data_race(!list_empty(&folio->_deferred_list))) {
-		spin_lock_irqsave(&ds_queue->split_queue_lock, flags);
-		if (!list_empty(&folio->_deferred_list)) {
-			ds_queue->split_queue_len--;
-			list_del(&folio->_deferred_list);
-		}
-		spin_unlock_irqrestore(&ds_queue->split_queue_lock, flags);
+	if (data_race(list_empty(&folio->_deferred_list)))
+		return;
+
+	ds_queue = get_deferred_split_queue(folio);
+	spin_lock_irqsave(&ds_queue->split_queue_lock, flags);
+	if (!list_empty(&folio->_deferred_list)) {
+		ds_queue->split_queue_len--;
+		list_del(&folio->_deferred_list);
 	}
-	free_compound_page(page);
+	spin_unlock_irqrestore(&ds_queue->split_queue_lock, flags);
 }
 
 void deferred_split_folio(struct folio *folio)
diff --git a/mm/internal.h b/mm/internal.h
index 5a03bc4782a2..1e98c867f0de 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -413,6 +413,8 @@ static inline void folio_set_order(struct folio *folio, unsigned int order)
 #endif
 }
 
+void folio_undo_large_rmappable(struct folio *folio);
+
 static inline void prep_compound_head(struct page *page, unsigned int order)
 {
 	struct folio *folio = (struct folio *)page;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index b569fd5562aa..0dbc2ecdefa5 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -287,9 +287,6 @@ const char * const migratetype_names[MIGRATE_TYPES] = {
 static compound_page_dtor * const compound_page_dtors[NR_COMPOUND_DTORS] = {
 	[NULL_COMPOUND_DTOR] = NULL,
 	[COMPOUND_PAGE_DTOR] = free_compound_page,
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-	[TRANSHUGE_PAGE_DTOR] = free_transhuge_page,
-#endif
 };
 
 int min_free_kbytes = 1024;
@@ -624,6 +621,12 @@ void destroy_large_folio(struct folio *folio)
 		return;
 	}
 
+	if (folio_test_transhuge(folio) && dtor == TRANSHUGE_PAGE_DTOR) {
+		folio_undo_large_rmappable(folio);
+		free_compound_page(&folio->page);
+		return;
+	}
+
 	VM_BUG_ON_FOLIO(dtor >= NR_COMPOUND_DTORS, folio);
 	compound_page_dtors[dtor](&folio->page);
 }
-- 
2.40.1


  parent reply	other threads:[~2023-08-16 15:13 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-16 15:11 [PATCH v2 00/13] Remove _folio_dtor and _folio_order Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 01/13] io_uring: Stop calling free_compound_page() Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 02/13] mm: Call free_huge_page() directly Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 03/13] mm: Convert free_huge_page() to free_huge_folio() Matthew Wilcox (Oracle)
2023-08-16 20:14   ` Sidhartha Kumar
2023-08-17  3:31   ` Yanteng Si
2023-08-16 15:11 ` Matthew Wilcox (Oracle) [this message]
2023-08-16 15:11 ` [PATCH v2 05/13] mm; Convert prep_transhuge_page() to folio_prep_large_rmappable() Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 06/13] mm: Remove free_compound_page() and the compound_page_dtors array Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 07/13] mm: Remove HUGETLB_PAGE_DTOR Matthew Wilcox (Oracle)
2023-08-16 21:45   ` Sidhartha Kumar
2023-08-22  3:13   ` Mike Kravetz
2023-08-22  3:32     ` Matthew Wilcox
2023-08-22 17:19       ` Mike Kravetz
2023-12-08 17:54   ` Vlastimil Babka
2023-12-08 18:31     ` Matthew Wilcox
2023-12-13 12:23       ` David Hildenbrand
2023-08-16 15:11 ` [PATCH v2 08/13] mm: Add large_rmappable page flag Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 09/13] mm: Rearrange page flags Matthew Wilcox (Oracle)
2023-08-16 15:11 ` [PATCH v2 10/13] mm: Free up a word in the first tail page Matthew Wilcox (Oracle)
2023-08-22 23:17   ` Mike Kravetz
2023-08-23  0:29     ` Matthew Wilcox
2023-08-16 15:11 ` [PATCH v2 11/13] mm: Remove folio_test_transhuge() Matthew Wilcox (Oracle)
2023-08-16 15:12 ` [PATCH v2 12/13] mm: Add tail private fields to struct folio Matthew Wilcox (Oracle)
2023-08-16 15:21   ` David Hildenbrand
2023-08-16 15:12 ` [PATCH v2 13/13] mm: Convert split_huge_pages_pid() to use a folio Matthew Wilcox (Oracle)

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 \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /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