linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Matthew Wilcox <willy@infradead.org>
To: linux-mm@kvack.org, linux-fsdevel@vger.kernel.org
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>,
	Jan Kara <jack@suse.cz>,
	William Kucharski <william.kucharski@oracle.com>
Subject: [PATCH v3 10/11] mm: Use clear_bit_unlock_is_negative_byte for PageWriteback
Date: Thu, 16 Apr 2020 15:01:29 -0700	[thread overview]
Message-ID: <20200416220130.13343-11-willy@infradead.org> (raw)
In-Reply-To: <20200416220130.13343-1-willy@infradead.org>

From: "Matthew Wilcox (Oracle)" <willy@infradead.org>

We can use clear_bit_unlock_is_negative_byte() for writeback as well as
the lock bit.  wake_up_page() then has no more callers and can be removed.

Given the other code being executed between the clear and the test,
this is not going to be as dramatic a win as it was for PageLocked,
but symmetry between the two is nice and lets us remove some code.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
---
 include/linux/page-flags.h |  2 +-
 mm/filemap.c               | 12 ++----------
 mm/page-writeback.c        | 37 ++++++++++++++++++++-----------------
 3 files changed, 23 insertions(+), 28 deletions(-)

diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index af7c0ff5f517..96c7d220c8cf 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -545,7 +545,7 @@ static __always_inline void SetPageUptodate(struct page *page)
 
 CLEARPAGEFLAG(Uptodate, uptodate, PF_NO_TAIL)
 
-int test_clear_page_writeback(struct page *page);
+bool __clear_page_writeback(struct page *page);
 int __test_set_page_writeback(struct page *page, bool keep_write);
 
 #define test_set_page_writeback(page)			\
diff --git a/mm/filemap.c b/mm/filemap.c
index 401b24d980ba..c704d333d3bf 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1084,13 +1084,6 @@ static void wake_up_page_bit(struct page *page, int bit_nr)
 	spin_unlock_irqrestore(&q->lock, flags);
 }
 
-static void wake_up_page(struct page *page, int bit)
-{
-	if (!PageWaiters(page))
-		return;
-	wake_up_page_bit(page, bit);
-}
-
 /*
  * A choice of three behaviors for wait_on_page_bit_common():
  */
@@ -1293,9 +1286,8 @@ void end_page_writeback(struct page *page)
 		rotate_reclaimable_page(page);
 	}
 
-	test_clear_page_writeback(page);
-	smp_mb__after_atomic();
-	wake_up_page(page, PG_writeback);
+	if (__clear_page_writeback(page))
+		wake_up_page_bit(page, PG_writeback);
 }
 EXPORT_SYMBOL(end_page_writeback);
 
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index ebaf0d8263a6..d019d86fc21f 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -595,7 +595,7 @@ static void wb_domain_writeout_inc(struct wb_domain *dom,
 
 /*
  * Increment @wb's writeout completion count and the global writeout
- * completion count. Called from test_clear_page_writeback().
+ * completion count.
  */
 static inline void __wb_writeout_inc(struct bdi_writeback *wb)
 {
@@ -2711,12 +2711,19 @@ int clear_page_dirty_for_io(struct page *page)
 }
 EXPORT_SYMBOL(clear_page_dirty_for_io);
 
-int test_clear_page_writeback(struct page *page)
+#define clear_writeback_bit(page)	\
+	clear_bit_unlock_is_negative_byte(PG_writeback, &page->flags)
+
+/*
+ * The return value is whether there are waiters pending, not whether
+ * the flag was set.
+ */
+bool __clear_page_writeback(struct page *page)
 {
 	struct address_space *mapping = page_mapping(page);
 	struct mem_cgroup *memcg;
 	struct lruvec *lruvec;
-	int ret;
+	bool ret;
 
 	if (WARN_ON(!PageWriteback(page))) {
 		dump_page(page, "!writeback");
@@ -2731,16 +2738,14 @@ int test_clear_page_writeback(struct page *page)
 		unsigned long flags;
 
 		xa_lock_irqsave(&mapping->i_pages, flags);
-		ret = TestClearPageWriteback(page);
-		if (ret) {
-			__xa_clear_mark(&mapping->i_pages, page_index(page),
+		ret = clear_writeback_bit(page);
+		__xa_clear_mark(&mapping->i_pages, page_index(page),
 						PAGECACHE_TAG_WRITEBACK);
-			if (bdi_cap_account_writeback(bdi)) {
-				struct bdi_writeback *wb = inode_to_wb(inode);
+		if (bdi_cap_account_writeback(bdi)) {
+			struct bdi_writeback *wb = inode_to_wb(inode);
 
-				dec_wb_stat(wb, WB_WRITEBACK);
-				__wb_writeout_inc(wb);
-			}
+			dec_wb_stat(wb, WB_WRITEBACK);
+			__wb_writeout_inc(wb);
 		}
 
 		if (mapping->host && !mapping_tagged(mapping,
@@ -2749,7 +2754,7 @@ int test_clear_page_writeback(struct page *page)
 
 		xa_unlock_irqrestore(&mapping->i_pages, flags);
 	} else {
-		ret = TestClearPageWriteback(page);
+		ret = clear_writeback_bit(page);
 	}
 	/*
 	 * NOTE: Page might be free now! Writeback doesn't hold a page
@@ -2757,11 +2762,9 @@ int test_clear_page_writeback(struct page *page)
 	 * the clearing of PG_writeback. The below can only access
 	 * page state that is static across allocation cycles.
 	 */
-	if (ret) {
-		dec_lruvec_state(lruvec, NR_WRITEBACK);
-		dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
-		inc_node_page_state(page, NR_WRITTEN);
-	}
+	dec_lruvec_state(lruvec, NR_WRITEBACK);
+	dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
+	inc_node_page_state(page, NR_WRITTEN);
 	__unlock_page_memcg(memcg);
 	return ret;
 }
-- 
2.25.1



  parent reply	other threads:[~2020-04-16 22:01 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-16 22:01 [PATCH v3 00/11] Make PageWriteback use the PageLocked optimisation Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 01/11] alpha: Add clear_bit_unlock_is_negative_byte implementation Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 02/11] ia64: " Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 03/11] m68k: " Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 04/11] mips: " Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 05/11] riscv: " Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 06/11] s390: " Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 07/11] mm: Remove definition of clear_bit_unlock_is_negative_byte Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 08/11] mm: Move PG_writeback into the bottom byte Matthew Wilcox
2020-04-16 22:01 ` [PATCH v3 09/11] mm: Convert writeback BUG to WARN_ON Matthew Wilcox
2020-04-23 15:58   ` Jan Kara
2020-04-16 22:01 ` Matthew Wilcox [this message]
2020-04-16 22:01 ` [PATCH v3 11/11] mm: Remove TestClearPageWriteback Matthew Wilcox
2020-04-17  7:28 ` [PATCH v3 00/11] Make PageWriteback use the PageLocked optimisation Geert Uytterhoeven
2020-04-17 11:12   ` Matthew Wilcox

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=20200416220130.13343-11-willy@infradead.org \
    --to=willy@infradead.org \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=william.kucharski@oracle.com \
    --subject='Re: [PATCH v3 10/11] mm: Use clear_bit_unlock_is_negative_byte for PageWriteback' \
    /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

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).