Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757128AbbBEJ6Z (ORCPT ); Thu, 5 Feb 2015 04:58:25 -0500 Received: from mailout4.samsung.com ([203.254.224.34]:31913 "EHLO mailout4.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752443AbbBEJ6X (ORCPT ); Thu, 5 Feb 2015 04:58:23 -0500 X-AuditID: cbfee61b-f79d76d0000024d6-70-54d33ebd24df From: Chao Yu To: Jaegeuk Kim , Changman Lee Cc: linux-f2fs-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org Subject: [PATCH v3 08/10] f2fs: enable rb-tree extent cache Date: Thu, 05 Feb 2015 17:57:31 +0800 Message-id: <000501d0412a$43487bc0$c9d97340$@samsung.com> MIME-version: 1.0 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit X-Mailer: Microsoft Outlook 14.0 Thread-index: AdA/kH4qj+5SqMYnQLaybY5tXwE6Ww== Content-language: zh-cn X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFrrELMWRmVeSWpSXmKPExsVy+t9jAd29dpdDDP4/MbS4tq+RyeLJ+lnM FpcWuVtc3jWHzYHFY9OqTjaP3Qs+M3n0bVnF6PF5k1wASxSXTUpqTmZZapG+XQJXxpw9XUwF Z/Uqzv3cx9bAeESti5GTQ0LAROLF5e8sELaYxIV769m6GLk4hAQWMUocPraNBcL5wSjR0X6G EaSKTUBFYnnHfyYQW0TAS2LS/hNg3cwCHhKNHd9ZQWxhAWuJW+tug9WzCKhKPH5zmB3E5hWw lPjycjIThC0o8WPyPaheLYn1O48zQdjyEpvXvGWGuEhBYsfZ10BzOIB26Uk8bBWFKBGX2Hjk FssERoFZSCbNQjJpFpJJs5C0LGBkWcUomlqQXFCclJ5rpFecmFtcmpeul5yfu4kRHNLPpHcw rmqwOMQowMGoxMNrse9SiBBrYllxZe4hRgkOZiURXk2OyyFCvCmJlVWpRfnxRaU5qcWHGKU5 WJTEeZXs20KEBNITS1KzU1MLUotgskwcnFINjPXtl41n8Ih1ydssXbC/OOecVnvunemsem3f OXfVrr9csMarT27NjvmXhAs5qoNsDH/WahZMLNm3c7mmvOUnlU3+L19N0F+kaNhc8+xcZtO3 uMcpPpmWTkH+E5bcmCXn9bPsd8W3Xv74VR4lzMdUmUwenyt72Tz73PXGRF6Veq+H1xy2rJub rsRSnJFoqMVcVJwIAGTBwURlAgAA Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 6026 Lines: 189 This patch enables rb-tree based extent cache in f2fs. When we mount with "-o extent_cache", f2fs will try to add recently accessed page-block mappings into rb-tree based extent cache as much as possible, instead of original one extent info cache. By this way, f2fs can support more effective cache between dnode page cache and disk. It will supply high hit ratio in the cache with fewer memory when dnode page cache are reclaimed in environment of low memory. Storage: Sandisk sd card 64g 1.append write file (offset: 0, size: 128M); 2.override write file (offset: 2M, size: 1M); 3.override write file (offset: 4M, size: 1M); ... 4.override write file (offset: 48M, size: 1M); ... 5.override write file (offset: 112M, size: 1M); 6.sync 7.echo 3 > /proc/sys/vm/drop_caches 8.read file (size:128M, unit: 4k, count: 32768) (time dd if=/mnt/f2fs/128m bs=4k count=32768) Extent Hit Ratio: before patched Hit Ratio 121 / 1071 1071 / 1071 Performance: before patched real 0m37.051s 0m35.556s user 0m0.040s 0m0.026s sys 0m2.990s 0m2.251s Memory Cost: before patched Tree Count: 0 1 (size: 24 bytes) Node Count: 0 45 (size: 1440 bytes) v3: o retest and given more details of test result. Signed-off-by: Chao Yu --- fs/f2fs/data.c | 13 +++++++++++++ fs/f2fs/f2fs.h | 5 +++++ fs/f2fs/inode.c | 1 + fs/f2fs/segment.c | 3 +++ fs/f2fs/super.c | 9 ++++++++- 5 files changed, 30 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 4b240fb..6c8119a 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -666,6 +666,9 @@ void f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink) void **slot; unsigned int found; + if (!test_opt(sbi, EXTENT_CACHE)) + return; + if (available_free_memory(sbi, EXTENT_CACHE)) return; @@ -714,6 +717,9 @@ void f2fs_destroy_extent_tree(struct inode *inode) struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct extent_tree *et; + if (!test_opt(sbi, EXTENT_CACHE)) + return; + down_read(&sbi->extent_tree_lock); et = radix_tree_lookup(&sbi->extent_tree_root, inode->i_ino); if (!et) { @@ -749,6 +755,9 @@ out: static bool f2fs_lookup_extent_cache(struct inode *inode, pgoff_t pgofs, struct extent_info *ei) { + if (test_opt(F2FS_I_SB(inode), EXTENT_CACHE)) + return f2fs_lookup_extent_tree(inode, pgofs, ei); + return lookup_extent_info(inode, pgofs, ei); } @@ -765,6 +774,10 @@ void f2fs_update_extent_cache(struct dnode_of_data *dn) fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) + dn->ofs_in_node; + if (test_opt(F2FS_I_SB(dn->inode), EXTENT_CACHE)) + return f2fs_update_extent_tree(dn->inode, fofs, + dn->data_blkaddr); + if (update_extent_info(dn->inode, fofs, dn->data_blkaddr)) sync_inode_page(dn); } diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 507e604..73ca511 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1568,12 +1568,17 @@ void f2fs_submit_page_mbio(struct f2fs_sb_info *, struct page *, struct f2fs_io_info *); int reserve_new_block(struct dnode_of_data *); int f2fs_reserve_block(struct dnode_of_data *, pgoff_t); +void f2fs_shrink_extent_tree(struct f2fs_sb_info *, int); +void f2fs_destroy_extent_tree(struct inode *); void f2fs_update_extent_cache(struct dnode_of_data *); struct page *find_data_page(struct inode *, pgoff_t, bool); struct page *get_lock_data_page(struct inode *, pgoff_t); struct page *get_new_data_page(struct inode *, struct page *, pgoff_t, bool); int do_write_data_page(struct page *, struct f2fs_io_info *); int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *, u64, u64); +void init_extent_cache_info(struct f2fs_sb_info *); +int __init create_extent_cache(void); +void destroy_extent_cache(void); /* * gc.c diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index 28dd26a..b508744 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -335,6 +335,7 @@ void f2fs_evict_inode(struct inode *inode) no_delete: stat_dec_inline_dir(inode); stat_dec_inline_inode(inode); + f2fs_destroy_extent_tree(inode); invalidate_mapping_pages(NODE_MAPPING(sbi), inode->i_ino, inode->i_ino); if (xnid) invalidate_mapping_pages(NODE_MAPPING(sbi), xnid, xnid); diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index b85bb97..f3f2d6f 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -277,6 +277,9 @@ void f2fs_balance_fs(struct f2fs_sb_info *sbi) void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi) { + /* try to shrink extent cache when there is no enough memory */ + f2fs_shrink_extent_tree(sbi, EXTENT_CACHE_SHRINK_NUMBER); + /* check the # of cached NAT entries and prefree segments */ if (try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK) || excess_prefree_segs(sbi) || diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 07a5564..3aaf2b5 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1077,6 +1077,8 @@ try_onemore: INIT_LIST_HEAD(&sbi->dir_inode_list); spin_lock_init(&sbi->dir_inode_lock); + init_extent_cache_info(sbi); + init_ino_entry_info(sbi); /* setup f2fs internal modules */ @@ -1283,10 +1285,13 @@ static int __init init_f2fs_fs(void) err = create_checkpoint_caches(); if (err) goto free_segment_manager_caches; + err = create_extent_cache(); + if (err) + goto free_checkpoint_caches; f2fs_kset = kset_create_and_add("f2fs", NULL, fs_kobj); if (!f2fs_kset) { err = -ENOMEM; - goto free_checkpoint_caches; + goto free_extent_cache; } err = register_filesystem(&f2fs_fs_type); if (err) @@ -1297,6 +1302,8 @@ static int __init init_f2fs_fs(void) free_kset: kset_unregister(f2fs_kset); +free_extent_cache: + destroy_extent_cache(); free_checkpoint_caches: destroy_checkpoint_caches(); free_segment_manager_caches: -- 2.2.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/