From: "Matthew Wilcox (Oracle)" <willy@infradead.org> To: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>, "Darrick J . Wong" <darrick.wong@oracle.com>, Christoph Hellwig <hch@infradead.org>, linux-nvdimm@lists.01.org, linux-kernel@vger.kernel.org, Dave Kleikamp <shaggy@kernel.org>, jfs-discussion@lists.sourceforge.net, Dave Chinner <dchinner@redhat.com> Subject: [PATCH v2 5/9] iomap: Support arbitrarily many blocks per page Date: Fri, 11 Sep 2020 00:47:03 +0100 [thread overview] Message-ID: <20200910234707.5504-6-willy@infradead.org> (raw) In-Reply-To: <20200910234707.5504-1-willy@infradead.org> Size the uptodate array dynamically to support larger pages in the page cache. With a 64kB page, we're only saving 8 bytes per page today, but with a 2MB maximum page size, we'd have to allocate more than 4kB per page. Add a few debugging assertions. Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> Reviewed-by: Dave Chinner <dchinner@redhat.com> --- fs/iomap/buffered-io.c | 22 +++++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c index 7fc0e02d27b0..9670c096b83e 100644 --- a/fs/iomap/buffered-io.c +++ b/fs/iomap/buffered-io.c @@ -22,18 +22,25 @@ #include "../internal.h" /* - * Structure allocated for each page when block size < PAGE_SIZE to track - * sub-page uptodate status and I/O completions. + * Structure allocated for each page or THP when block size < page size + * to track sub-page uptodate status and I/O completions. */ struct iomap_page { atomic_t read_count; atomic_t write_count; spinlock_t uptodate_lock; - DECLARE_BITMAP(uptodate, PAGE_SIZE / 512); + unsigned long uptodate[]; }; static inline struct iomap_page *to_iomap_page(struct page *page) { + /* + * per-block data is stored in the head page. Callers should + * not be dealing with tail pages (and if they are, they can + * call thp_head() first. + */ + VM_BUG_ON_PGFLAGS(PageTail(page), page); + if (page_has_private(page)) return (struct iomap_page *)page_private(page); return NULL; @@ -45,11 +52,13 @@ static struct iomap_page * iomap_page_create(struct inode *inode, struct page *page) { struct iomap_page *iop = to_iomap_page(page); + unsigned int nr_blocks = i_blocks_per_page(inode, page); - if (iop || i_blocks_per_page(inode, page) <= 1) + if (iop || nr_blocks <= 1) return iop; - iop = kzalloc(sizeof(*iop), GFP_NOFS | __GFP_NOFAIL); + iop = kzalloc(struct_size(iop, uptodate, BITS_TO_LONGS(nr_blocks)), + GFP_NOFS | __GFP_NOFAIL); spin_lock_init(&iop->uptodate_lock); attach_page_private(page, iop); return iop; @@ -59,11 +68,14 @@ static void iomap_page_release(struct page *page) { struct iomap_page *iop = detach_page_private(page); + unsigned int nr_blocks = i_blocks_per_page(page->mapping->host, page); if (!iop) return; WARN_ON_ONCE(atomic_read(&iop->read_count)); WARN_ON_ONCE(atomic_read(&iop->write_count)); + WARN_ON_ONCE(bitmap_full(iop->uptodate, nr_blocks) != + PageUptodate(page)); kfree(iop); } -- 2.28.0 _______________________________________________ Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
WARNING: multiple messages have this Message-ID (diff)
From: "Matthew Wilcox (Oracle)" <willy@infradead.org> To: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>, "Darrick J . Wong" <darrick.wong@oracle.com>, Christoph Hellwig <hch@infradead.org>, linux-nvdimm@lists.01.org, linux-kernel@vger.kernel.org, Dave Kleikamp <shaggy@kernel.org>, jfs-discussion@lists.sourceforge.net, Dave Chinner <dchinner@redhat.com> Subject: [PATCH v2 5/9] iomap: Support arbitrarily many blocks per page Date: Fri, 11 Sep 2020 00:47:03 +0100 [thread overview] Message-ID: <20200910234707.5504-6-willy@infradead.org> (raw) In-Reply-To: <20200910234707.5504-1-willy@infradead.org> Size the uptodate array dynamically to support larger pages in the page cache. With a 64kB page, we're only saving 8 bytes per page today, but with a 2MB maximum page size, we'd have to allocate more than 4kB per page. Add a few debugging assertions. Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> Reviewed-by: Dave Chinner <dchinner@redhat.com> --- fs/iomap/buffered-io.c | 22 +++++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c index 7fc0e02d27b0..9670c096b83e 100644 --- a/fs/iomap/buffered-io.c +++ b/fs/iomap/buffered-io.c @@ -22,18 +22,25 @@ #include "../internal.h" /* - * Structure allocated for each page when block size < PAGE_SIZE to track - * sub-page uptodate status and I/O completions. + * Structure allocated for each page or THP when block size < page size + * to track sub-page uptodate status and I/O completions. */ struct iomap_page { atomic_t read_count; atomic_t write_count; spinlock_t uptodate_lock; - DECLARE_BITMAP(uptodate, PAGE_SIZE / 512); + unsigned long uptodate[]; }; static inline struct iomap_page *to_iomap_page(struct page *page) { + /* + * per-block data is stored in the head page. Callers should + * not be dealing with tail pages (and if they are, they can + * call thp_head() first. + */ + VM_BUG_ON_PGFLAGS(PageTail(page), page); + if (page_has_private(page)) return (struct iomap_page *)page_private(page); return NULL; @@ -45,11 +52,13 @@ static struct iomap_page * iomap_page_create(struct inode *inode, struct page *page) { struct iomap_page *iop = to_iomap_page(page); + unsigned int nr_blocks = i_blocks_per_page(inode, page); - if (iop || i_blocks_per_page(inode, page) <= 1) + if (iop || nr_blocks <= 1) return iop; - iop = kzalloc(sizeof(*iop), GFP_NOFS | __GFP_NOFAIL); + iop = kzalloc(struct_size(iop, uptodate, BITS_TO_LONGS(nr_blocks)), + GFP_NOFS | __GFP_NOFAIL); spin_lock_init(&iop->uptodate_lock); attach_page_private(page, iop); return iop; @@ -59,11 +68,14 @@ static void iomap_page_release(struct page *page) { struct iomap_page *iop = detach_page_private(page); + unsigned int nr_blocks = i_blocks_per_page(page->mapping->host, page); if (!iop) return; WARN_ON_ONCE(atomic_read(&iop->read_count)); WARN_ON_ONCE(atomic_read(&iop->write_count)); + WARN_ON_ONCE(bitmap_full(iop->uptodate, nr_blocks) != + PageUptodate(page)); kfree(iop); } -- 2.28.0
next prev parent reply other threads:[~2020-09-10 23:47 UTC|newest] Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top 2020-09-10 23:46 [PATCH v2 0/9] THP iomap patches for 5.10 Matthew Wilcox (Oracle) 2020-09-10 23:46 ` Matthew Wilcox (Oracle) 2020-09-10 23:46 ` [PATCH v2 1/9] iomap: Fix misplaced page flushing Matthew Wilcox (Oracle) 2020-09-10 23:46 ` Matthew Wilcox (Oracle) 2020-09-10 23:47 ` [PATCH v2 2/9] fs: Introduce i_blocks_per_page Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-15 14:58 ` Dave Kleikamp 2020-09-15 14:58 ` Dave Kleikamp 2020-09-15 15:40 ` David Laight 2020-09-15 15:40 ` David Laight 2020-09-15 15:49 ` Matthew Wilcox 2020-09-15 15:49 ` Matthew Wilcox 2020-09-10 23:47 ` [PATCH v2 3/9] iomap: Use kzalloc to allocate iomap_page Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-10 23:47 ` [PATCH v2 4/9] iomap: Use bitmap ops to set uptodate bits Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) [this message] 2020-09-10 23:47 ` [PATCH v2 5/9] iomap: Support arbitrarily many blocks per page Matthew Wilcox (Oracle) 2020-09-11 5:36 ` Christoph Hellwig 2020-09-11 5:36 ` Christoph Hellwig 2020-09-17 22:00 ` Darrick J. Wong 2020-09-17 22:00 ` Darrick J. Wong 2020-09-22 16:23 ` Qian Cai 2020-09-22 16:23 ` Qian Cai 2020-09-22 17:05 ` Matthew Wilcox 2020-09-22 17:05 ` Matthew Wilcox 2020-09-22 17:25 ` Qian Cai 2020-09-22 17:25 ` Qian Cai 2020-09-23 1:06 ` Qian Cai 2020-09-23 1:06 ` Qian Cai 2020-09-23 2:48 ` Matthew Wilcox 2020-09-23 2:48 ` Matthew Wilcox 2020-09-23 5:00 ` Darrick J. Wong 2020-09-23 5:00 ` Darrick J. Wong 2020-09-23 18:59 ` Matthew Wilcox 2020-09-23 18:59 ` Matthew Wilcox 2020-09-23 16:55 ` Qian Cai 2020-09-23 16:55 ` Qian Cai 2020-09-24 1:07 ` Matthew Wilcox 2020-09-24 1:07 ` Matthew Wilcox 2020-09-10 23:47 ` [PATCH v2 6/9] iomap: Convert read_count to read_bytes_pending Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-11 5:36 ` Christoph Hellwig 2020-09-11 5:36 ` Christoph Hellwig 2020-09-17 22:02 ` Darrick J. Wong 2020-09-17 22:02 ` Darrick J. Wong 2020-09-10 23:47 ` [PATCH v2 7/9] iomap: Convert write_count to write_bytes_pending Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-17 22:02 ` Darrick J. Wong 2020-09-17 22:02 ` Darrick J. Wong 2020-09-10 23:47 ` [PATCH v2 8/9] iomap: Convert iomap_write_end types Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-17 22:03 ` Darrick J. Wong 2020-09-17 22:03 ` Darrick J. Wong 2020-09-10 23:47 ` [PATCH v2 9/9] iomap: Change calling convention for zeroing Matthew Wilcox (Oracle) 2020-09-10 23:47 ` Matthew Wilcox (Oracle) 2020-09-11 6:42 ` Christoph Hellwig 2020-09-11 6:42 ` Christoph Hellwig 2020-09-17 22:05 ` Darrick J. Wong 2020-09-17 22:05 ` Darrick J. Wong 2020-09-17 22:11 ` Matthew Wilcox 2020-09-17 22:11 ` Matthew Wilcox 2020-09-17 22:18 ` Darrick J. Wong 2020-09-17 22:18 ` Darrick J. Wong
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=20200910234707.5504-6-willy@infradead.org \ --to=willy@infradead.org \ --cc=darrick.wong@oracle.com \ --cc=dchinner@redhat.com \ --cc=hch@infradead.org \ --cc=jfs-discussion@lists.sourceforge.net \ --cc=linux-fsdevel@vger.kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-nvdimm@lists.01.org \ --cc=linux-xfs@vger.kernel.org \ --cc=shaggy@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.