From: Jaegeuk Kim <jaegeuk@kernel.org> To: linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net, kernel-team@android.com Subject: Re: [PATCH v2] f2fs: avoid readahead race condition Date: Mon, 29 Jun 2020 08:03:23 -0700 [thread overview] Message-ID: <20200629150323.GA3293033@google.com> (raw) In-Reply-To: <20200624012148.180050-1-jaegeuk@kernel.org> If two readahead threads having same offset enter in readpages, every read IOs are split and issued to the disk which giving lower bandwidth. This patch tries to avoid redundant readahead calls. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> --- v2: - add missing code to bypass read fs/f2fs/data.c | 18 +++++++++++++++++- fs/f2fs/f2fs.h | 1 + fs/f2fs/super.c | 2 ++ 3 files changed, 20 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index d6094b9f3916..9b69a159cc6c 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2403,6 +2403,7 @@ int f2fs_mpage_readpages(struct address_space *mapping, #endif unsigned max_nr_pages = nr_pages; int ret = 0; + bool drop_ra = false; map.m_pblk = 0; map.m_lblk = 0; @@ -2413,13 +2414,25 @@ int f2fs_mpage_readpages(struct address_space *mapping, map.m_seg_type = NO_CHECK_TYPE; map.m_may_create = false; + /* + * Two readahead threads for same address range can cause race condition + * which fragments sequential read IOs. So let's avoid each other. + */ + if (pages && is_readahead) { + page = list_last_entry(pages, struct page, lru); + if (F2FS_I(inode)->ra_offset == page_index(page)) + drop_ra = true; + else + F2FS_I(inode)->ra_offset = page_index(page); + } + for (; nr_pages; nr_pages--) { if (pages) { page = list_last_entry(pages, struct page, lru); prefetchw(&page->flags); list_del(&page->lru); - if (add_to_page_cache_lru(page, mapping, + if (drop_ra || add_to_page_cache_lru(page, mapping, page_index(page), readahead_gfp_mask(mapping))) goto next_page; @@ -2484,6 +2497,9 @@ int f2fs_mpage_readpages(struct address_space *mapping, BUG_ON(pages && !list_empty(pages)); if (bio) __f2fs_submit_read_bio(F2FS_I_SB(inode), bio, DATA); + + if (pages && is_readahead && !drop_ra) + F2FS_I(inode)->ra_offset = -1; return pages ? 0 : ret; } diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 35afa13124b8..a95f84d72a55 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -806,6 +806,7 @@ struct f2fs_inode_info { struct list_head inmem_pages; /* inmemory pages managed by f2fs */ struct task_struct *inmem_task; /* store inmemory task */ struct mutex inmem_lock; /* lock for inmemory pages */ + pgoff_t ra_offset; /* ongoing readahead offset */ struct extent_tree *extent_tree; /* cached extent_tree entry */ /* avoid racing between foreground op and gc */ diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 0e860186a9c5..6fd2ad43d9e4 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1011,6 +1011,8 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb) /* Will be used by directory only */ fi->i_dir_level = F2FS_SB(sb)->dir_level; + fi->ra_offset = -1; + return &fi->vfs_inode; } -- 2.27.0.111.gc72c7da667-goog
WARNING: multiple messages have this Message-ID (diff)
From: Jaegeuk Kim <jaegeuk@kernel.org> To: linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net, kernel-team@android.com Subject: Re: [f2fs-dev] [PATCH v2] f2fs: avoid readahead race condition Date: Mon, 29 Jun 2020 08:03:23 -0700 [thread overview] Message-ID: <20200629150323.GA3293033@google.com> (raw) In-Reply-To: <20200624012148.180050-1-jaegeuk@kernel.org> If two readahead threads having same offset enter in readpages, every read IOs are split and issued to the disk which giving lower bandwidth. This patch tries to avoid redundant readahead calls. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> --- v2: - add missing code to bypass read fs/f2fs/data.c | 18 +++++++++++++++++- fs/f2fs/f2fs.h | 1 + fs/f2fs/super.c | 2 ++ 3 files changed, 20 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index d6094b9f3916..9b69a159cc6c 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2403,6 +2403,7 @@ int f2fs_mpage_readpages(struct address_space *mapping, #endif unsigned max_nr_pages = nr_pages; int ret = 0; + bool drop_ra = false; map.m_pblk = 0; map.m_lblk = 0; @@ -2413,13 +2414,25 @@ int f2fs_mpage_readpages(struct address_space *mapping, map.m_seg_type = NO_CHECK_TYPE; map.m_may_create = false; + /* + * Two readahead threads for same address range can cause race condition + * which fragments sequential read IOs. So let's avoid each other. + */ + if (pages && is_readahead) { + page = list_last_entry(pages, struct page, lru); + if (F2FS_I(inode)->ra_offset == page_index(page)) + drop_ra = true; + else + F2FS_I(inode)->ra_offset = page_index(page); + } + for (; nr_pages; nr_pages--) { if (pages) { page = list_last_entry(pages, struct page, lru); prefetchw(&page->flags); list_del(&page->lru); - if (add_to_page_cache_lru(page, mapping, + if (drop_ra || add_to_page_cache_lru(page, mapping, page_index(page), readahead_gfp_mask(mapping))) goto next_page; @@ -2484,6 +2497,9 @@ int f2fs_mpage_readpages(struct address_space *mapping, BUG_ON(pages && !list_empty(pages)); if (bio) __f2fs_submit_read_bio(F2FS_I_SB(inode), bio, DATA); + + if (pages && is_readahead && !drop_ra) + F2FS_I(inode)->ra_offset = -1; return pages ? 0 : ret; } diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 35afa13124b8..a95f84d72a55 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -806,6 +806,7 @@ struct f2fs_inode_info { struct list_head inmem_pages; /* inmemory pages managed by f2fs */ struct task_struct *inmem_task; /* store inmemory task */ struct mutex inmem_lock; /* lock for inmemory pages */ + pgoff_t ra_offset; /* ongoing readahead offset */ struct extent_tree *extent_tree; /* cached extent_tree entry */ /* avoid racing between foreground op and gc */ diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 0e860186a9c5..6fd2ad43d9e4 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1011,6 +1011,8 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb) /* Will be used by directory only */ fi->i_dir_level = F2FS_SB(sb)->dir_level; + fi->ra_offset = -1; + return &fi->vfs_inode; } -- 2.27.0.111.gc72c7da667-goog _______________________________________________ Linux-f2fs-devel mailing list Linux-f2fs-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
next prev parent reply other threads:[~2020-06-29 21:34 UTC|newest] Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top 2020-06-24 1:21 [PATCH] f2fs: avoid readahead race condition Jaegeuk Kim 2020-06-24 1:21 ` [f2fs-dev] " Jaegeuk Kim 2020-06-28 2:35 ` Chao Yu 2020-06-28 2:35 ` Chao Yu 2020-06-29 15:03 ` Jaegeuk Kim [this message] 2020-06-29 15:03 ` [f2fs-dev] [PATCH v2] " Jaegeuk Kim 2020-06-29 16:09 ` Eric Biggers 2020-06-29 16:09 ` Eric Biggers 2020-06-29 18:24 ` Jaegeuk Kim 2020-06-29 18:24 ` Jaegeuk Kim 2020-06-29 18:25 ` Eric Biggers 2020-06-29 18:25 ` [f2fs-dev] " Eric Biggers 2020-06-29 20:27 ` [f2fs-dev] [PATCH v3] " Jaegeuk Kim 2020-06-29 20:27 ` Jaegeuk Kim 2020-06-30 20:43 ` Nathan Chancellor 2020-06-30 20:43 ` Nathan Chancellor 2020-06-30 20:56 ` Jaegeuk Kim 2020-06-30 20:56 ` Jaegeuk Kim 2020-07-01 1:59 ` Chao Yu 2020-07-01 1:59 ` Chao Yu 2020-07-01 7:47 ` Chao Yu 2020-07-01 7:47 ` Chao Yu 2020-07-01 16:14 ` Jaegeuk Kim 2020-07-01 16:14 ` Jaegeuk Kim 2020-07-03 1:34 ` Chao Yu 2020-07-03 1:34 ` Chao Yu 2020-07-01 16:14 ` [f2fs-dev] [PATCH v4] " Jaegeuk Kim 2020-07-01 16:14 ` Jaegeuk Kim
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=20200629150323.GA3293033@google.com \ --to=jaegeuk@kernel.org \ --cc=kernel-team@android.com \ --cc=linux-f2fs-devel@lists.sourceforge.net \ --cc=linux-kernel@vger.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.