From e1c42045203071c4634b89e696037357810d3083 Mon Sep 17 00:00:00 2001 From: arter97 Date: Wed, 6 Aug 2014 23:22:50 +0900 Subject: f2fs: fix typo Fix typo and some grammatical errors. The words "filesystem" and "readahead" are being used without the space treewide. Signed-off-by: Park Ju Hyung Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 657582fc760..633315acfef 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -663,7 +663,7 @@ restore_gc: if (need_restart_gc) { if (start_gc_thread(sbi)) f2fs_msg(sbi->sb, KERN_WARNING, - "background gc thread is stop"); + "background gc thread has stopped"); } else if (need_stop_gc) { stop_gc_thread(sbi); } -- cgit v1.2.3-70-g09d2 From 97c3c5cac2bba0ecc4b0de83d33a23aa427ef628 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 19 Aug 2014 09:13:01 -0700 Subject: f2fs: don't skip checkpoint if there is no dirty node pages This is the errorneous scenario. 1. write data 2. do checkpoint 3. produce some dirty node pages by the gc thread 4. write back dirty node pages 5. f2fs_put_super will skip the checkpoint, since dirty count for node pages is zero. This patch removes such the wrong condition check. Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 633315acfef..60e3554a6eb 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -432,7 +432,7 @@ static void f2fs_put_super(struct super_block *sb) stop_gc_thread(sbi); /* We don't need to do checkpoint when it's clean */ - if (sbi->s_dirty && get_pages(sbi, F2FS_DIRTY_NODES)) + if (sbi->s_dirty) write_checkpoint(sbi, true); iput(sbi->node_inode); -- cgit v1.2.3-70-g09d2 From 6f12ac25f0167adb5d9ad5547fd6838380261e5c Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 19 Aug 2014 09:48:22 -0700 Subject: f2fs: trigger release_dirty_inode in f2fs_put_super The generic_shutdown_super calls sync_filesystem, evict_inode, and then f2fs_put_super. In f2fs_evict_inode, we remain some dirty inode information so we should release them at f2fs_put_super. Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/f2fs.h | 1 + fs/f2fs/super.c | 3 +++ 3 files changed, 5 insertions(+), 1 deletion(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 7e1c13bfb99..f14af912a14 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -348,7 +348,7 @@ bool exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode) return e ? true : false; } -static void release_dirty_inode(struct f2fs_sb_info *sbi) +void release_dirty_inode(struct f2fs_sb_info *sbi) { struct ino_entry *e, *tmp; int i; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index cc5ead1bd37..cc89a7f490a 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1262,6 +1262,7 @@ int ra_meta_pages(struct f2fs_sb_info *, int, int, int); long sync_meta_pages(struct f2fs_sb_info *, enum page_type, long); void add_dirty_inode(struct f2fs_sb_info *, nid_t, int type); void remove_dirty_inode(struct f2fs_sb_info *, nid_t, int type); +void release_dirty_inode(struct f2fs_sb_info *); bool exist_written_data(struct f2fs_sb_info *, nid_t, int); int acquire_orphan_inode(struct f2fs_sb_info *); void release_orphan_inode(struct f2fs_sb_info *); diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 60e3554a6eb..7a5477915d9 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -435,6 +435,9 @@ static void f2fs_put_super(struct super_block *sb) if (sbi->s_dirty) write_checkpoint(sbi, true); + /* normally superblock is clean, so we need to release this */ + release_dirty_inode(sbi); + iput(sbi->node_inode); iput(sbi->meta_inode); -- cgit v1.2.3-70-g09d2 From ed2e621a95d704e6a4e904cc00524e8cbddda0c2 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 8 Aug 2014 15:37:41 -0700 Subject: f2fs: give a chance to mount again when encountering errors This patch gives another chance to try mount process when we encounter an error. This makes an effect on the roll-forward recovery failures as well. Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 7a5477915d9..e161e13958e 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -902,8 +902,10 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) struct buffer_head *raw_super_buf; struct inode *root; long err = -EINVAL; + bool retry = true; int i; +try_onemore: /* allocate memory for f2fs-specific super block info */ sbi = kzalloc(sizeof(struct f2fs_sb_info), GFP_KERNEL); if (!sbi) @@ -1083,9 +1085,11 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) /* recover fsynced data */ if (!test_opt(sbi, DISABLE_ROLL_FORWARD)) { err = recover_fsync_data(sbi); - if (err) + if (err) { f2fs_msg(sb, KERN_ERR, "Cannot recover all fsync data errno=%ld", err); + goto free_kobj; + } } /* @@ -1126,6 +1130,13 @@ free_sb_buf: brelse(raw_super_buf); free_sbi: kfree(sbi); + + /* give only one another chance */ + if (retry) { + retry = !retry; + shrink_dcache_sb(sb); + goto try_onemore; + } return err; } -- cgit v1.2.3-70-g09d2 From 1e968fdfe69e4060f05fa04059ecad93a0284e32 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 11 Aug 2014 16:49:25 -0700 Subject: f2fs: introduce f2fs_cp_error for readability This patch adds f2fs_cp_error for readability. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 4 ++-- fs/f2fs/f2fs.h | 5 +++++ fs/f2fs/gc.c | 2 +- fs/f2fs/super.c | 2 +- 4 files changed, 9 insertions(+), 4 deletions(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index f14af912a14..6f38aad8e65 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -162,7 +162,7 @@ static int f2fs_write_meta_page(struct page *page, goto redirty_out; /* Should not write any meta pages, if any IO error was occurred */ - if (unlikely(is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ERROR_FLAG))) + if (unlikely(f2fs_cp_error(sbi))) goto no_write; f2fs_wait_on_page_writeback(page, META); @@ -934,7 +934,7 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) /* Here, we only have one bio having CP pack */ sync_meta_pages(sbi, META_FLUSH, LONG_MAX); - if (!is_set_ckpt_flags(ckpt, CP_ERROR_FLAG)) { + if (!f2fs_cp_error(sbi)) { clear_prefree_segments(sbi); release_dirty_inode(sbi); F2FS_RESET_SB_DIRT(sbi); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index cc89a7f490a..2d009aed0c2 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1096,6 +1096,11 @@ static inline int f2fs_readonly(struct super_block *sb) return sb->s_flags & MS_RDONLY; } +static inline bool f2fs_cp_error(struct f2fs_sb_info *sbi) +{ + return is_set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG); +} + static inline void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi) { set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG); diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 87c50c50746..e8507b1c875 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -693,7 +693,7 @@ int f2fs_gc(struct f2fs_sb_info *sbi) gc_more: if (unlikely(!(sbi->sb->s_flags & MS_ACTIVE))) goto stop; - if (unlikely(is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ERROR_FLAG))) + if (unlikely(f2fs_cp_error(sbi))) goto stop; if (gc_type == BG_GC && has_not_enough_free_secs(sbi, nfree)) { diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index e161e13958e..8aabe3ef42f 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -815,7 +815,7 @@ static int sanity_check_ckpt(struct f2fs_sb_info *sbi) if (unlikely(fsmeta >= total)) return 1; - if (unlikely(is_set_ckpt_flags(ckpt, CP_ERROR_FLAG))) { + if (unlikely(f2fs_cp_error(sbi))) { f2fs_msg(sbi->sb, KERN_ERR, "A bug case: need to run fsck"); return 1; } -- cgit v1.2.3-70-g09d2 From 8501017e50fb7586ba522a2913ce664d6c2024f6 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 11 Aug 2014 18:37:46 -0700 Subject: f2fs: check s_dirty under cp_mutex It needs to check s_dirty under cp_mutex, since s_dirty is reset under that mutex. And previous condition was not correct, since we can omit doing checkpoint when checkpoint was done followed by all the node pages were written back. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 8 ++++++-- fs/f2fs/super.c | 3 --- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 6f38aad8e65..dc29b783768 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -952,6 +952,10 @@ void write_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) trace_f2fs_write_checkpoint(sbi->sb, is_umount, "start block_ops"); mutex_lock(&sbi->cp_mutex); + + if (!sbi->s_dirty) + goto out; + block_operations(sbi); trace_f2fs_write_checkpoint(sbi->sb, is_umount, "finish block_ops"); @@ -976,9 +980,9 @@ void write_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) do_checkpoint(sbi, is_umount); unblock_operations(sbi); - mutex_unlock(&sbi->cp_mutex); - stat_inc_cp_count(sbi->stat_info); +out: + mutex_unlock(&sbi->cp_mutex); trace_f2fs_write_checkpoint(sbi->sb, is_umount, "finish checkpoint"); } diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 8aabe3ef42f..e7a7b619ffd 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -460,9 +460,6 @@ int f2fs_sync_fs(struct super_block *sb, int sync) trace_f2fs_sync_fs(sb, sync); - if (!sbi->s_dirty && !get_pages(sbi, F2FS_DIRTY_NODES)) - return 0; - if (sync) { mutex_lock(&sbi->gc_mutex); write_checkpoint(sbi, false); -- cgit v1.2.3-70-g09d2 From cf779cab14d50a84b61399f758da269654b863db Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 11 Aug 2014 18:37:46 -0700 Subject: f2fs: handle EIO not to break fs consistency There are two rules when EIO is occurred. 1. don't write any checkpoint data to preserve the previous checkpoint 2. don't lose the cached dentry/node/meta pages So, at first, this patch adds set_page_dirty in f2fs_write_end_io's failure. Then, writing checkpoint/dentry/node blocks is not allowed. Note that, for the data pages, we can't just throw away by redirtying them. Otherwise, kworker can fall into infinite loop to flush them. (Ref. xfstests/019) Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 45 ++++++++++++++++++++++++++++++++------------- fs/f2fs/data.c | 11 ++++++++++- fs/f2fs/node.c | 2 ++ fs/f2fs/super.c | 5 ++++- 4 files changed, 48 insertions(+), 15 deletions(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index dc29b783768..c9c08d52ecf 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -160,14 +160,11 @@ static int f2fs_write_meta_page(struct page *page, goto redirty_out; if (wbc->for_reclaim) goto redirty_out; - - /* Should not write any meta pages, if any IO error was occurred */ if (unlikely(f2fs_cp_error(sbi))) - goto no_write; + goto redirty_out; f2fs_wait_on_page_writeback(page, META); write_meta_page(sbi, page); -no_write: dec_page_count(sbi, F2FS_DIRTY_META); unlock_page(page); return 0; @@ -737,7 +734,7 @@ retry: /* * Freeze all the FS-operations for checkpoint. */ -static void block_operations(struct f2fs_sb_info *sbi) +static int block_operations(struct f2fs_sb_info *sbi) { struct writeback_control wbc = { .sync_mode = WB_SYNC_ALL, @@ -745,6 +742,7 @@ static void block_operations(struct f2fs_sb_info *sbi) .for_reclaim = 0, }; struct blk_plug plug; + int err = 0; blk_start_plug(&plug); @@ -754,6 +752,10 @@ retry_flush_dents: if (get_pages(sbi, F2FS_DIRTY_DENTS)) { f2fs_unlock_all(sbi); sync_dirty_dir_inodes(sbi); + if (unlikely(f2fs_cp_error(sbi))) { + err = -EIO; + goto out; + } goto retry_flush_dents; } @@ -767,9 +769,16 @@ retry_flush_nodes: if (get_pages(sbi, F2FS_DIRTY_NODES)) { up_write(&sbi->node_write); sync_node_pages(sbi, 0, &wbc); + if (unlikely(f2fs_cp_error(sbi))) { + f2fs_unlock_all(sbi); + err = -EIO; + goto out; + } goto retry_flush_nodes; } +out: blk_finish_plug(&plug); + return err; } static void unblock_operations(struct f2fs_sb_info *sbi) @@ -813,8 +822,11 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) discard_next_dnode(sbi, NEXT_FREE_BLKADDR(sbi, curseg)); /* Flush all the NAT/SIT pages */ - while (get_pages(sbi, F2FS_DIRTY_META)) + while (get_pages(sbi, F2FS_DIRTY_META)) { sync_meta_pages(sbi, META, LONG_MAX); + if (unlikely(f2fs_cp_error(sbi))) + return; + } next_free_nid(sbi, &last_nid); @@ -924,6 +936,9 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) /* wait for previous submitted node/meta pages writeback */ wait_on_all_pages_writeback(sbi); + if (unlikely(f2fs_cp_error(sbi))) + return; + filemap_fdatawait_range(NODE_MAPPING(sbi), 0, LONG_MAX); filemap_fdatawait_range(META_MAPPING(sbi), 0, LONG_MAX); @@ -934,11 +949,13 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) /* Here, we only have one bio having CP pack */ sync_meta_pages(sbi, META_FLUSH, LONG_MAX); - if (!f2fs_cp_error(sbi)) { - clear_prefree_segments(sbi); - release_dirty_inode(sbi); - F2FS_RESET_SB_DIRT(sbi); - } + release_dirty_inode(sbi); + + if (unlikely(f2fs_cp_error(sbi))) + return; + + clear_prefree_segments(sbi); + F2FS_RESET_SB_DIRT(sbi); } /* @@ -955,8 +972,10 @@ void write_checkpoint(struct f2fs_sb_info *sbi, bool is_umount) if (!sbi->s_dirty) goto out; - - block_operations(sbi); + if (unlikely(f2fs_cp_error(sbi))) + goto out; + if (block_operations(sbi)) + goto out; trace_f2fs_write_checkpoint(sbi->sb, is_umount, "finish block_ops"); diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index ac3ccc25386..6ba52a39306 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -53,7 +53,7 @@ static void f2fs_write_end_io(struct bio *bio, int err) struct page *page = bvec->bv_page; if (unlikely(err)) { - SetPageError(page); + set_page_dirty(page); set_bit(AS_EIO, &page->mapping->flags); f2fs_stop_checkpoint(sbi); } @@ -836,10 +836,19 @@ write: /* Dentry blocks are controlled by checkpoint */ if (S_ISDIR(inode->i_mode)) { + if (unlikely(f2fs_cp_error(sbi))) + goto redirty_out; err = do_write_data_page(page, &fio); goto done; } + /* we should bypass data pages to proceed the kworkder jobs */ + if (unlikely(f2fs_cp_error(sbi))) { + SetPageError(page); + unlock_page(page); + return 0; + } + if (!wbc->for_reclaim) need_balance_fs = true; else if (has_not_enough_free_secs(sbi, 0)) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 9f126f80813..d2f78428342 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1215,6 +1215,8 @@ static int f2fs_write_node_page(struct page *page, if (unlikely(sbi->por_doing)) goto redirty_out; + if (unlikely(f2fs_cp_error(sbi))) + goto redirty_out; f2fs_wait_on_page_writeback(page, NODE); diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index e7a7b619ffd..ddb1e9d3636 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -435,7 +435,10 @@ static void f2fs_put_super(struct super_block *sb) if (sbi->s_dirty) write_checkpoint(sbi, true); - /* normally superblock is clean, so we need to release this */ + /* + * normally superblock is clean, so we need to release this. + * In addition, EIO will skip do checkpoint, we need this as well. + */ release_dirty_inode(sbi); iput(sbi->node_inode); -- cgit v1.2.3-70-g09d2 From c200b1aa6cb460ce8c3ecf6fdc690d3949c3cc5d Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 20 Aug 2014 18:36:46 +0800 Subject: f2fs: fix incorrect calculation with total/free inode num Theoretically, our total inodes number is the same as total node number, but there are three node ids are reserved in f2fs, they are 0, 1 (node nid), and 2 (meta nid), and they should never be used by user, so our total/free inode number calculated in ->statfs is wrong. This patch indroduces F2FS_RESERVED_NODE_NUM and then fixes this issue by recalculating total/free inode number with the macro. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 2 +- fs/f2fs/super.c | 4 ++-- include/linux/f2fs_fs.h | 3 +++ 3 files changed, 6 insertions(+), 3 deletions(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index b4d964029fc..044395c20ee 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1957,7 +1957,7 @@ static int init_node_manager(struct f2fs_sb_info *sbi) nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nat_blocks; /* not used nids: 0, node, meta, (and root counted as valid node) */ - nm_i->available_nids = nm_i->max_nid - 3; + nm_i->available_nids = nm_i->max_nid - F2FS_RESERVED_NODE_NUM; nm_i->fcnt = 0; nm_i->nat_cnt = 0; nm_i->ram_thresh = DEF_RAM_THRESHOLD; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index ddb1e9d3636..0f61f5aa587 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -508,8 +508,8 @@ static int f2fs_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_bfree = buf->f_blocks - valid_user_blocks(sbi) - ovp_count; buf->f_bavail = user_block_count - valid_user_blocks(sbi); - buf->f_files = sbi->total_node_count; - buf->f_ffree = sbi->total_node_count - valid_inode_count(sbi); + buf->f_files = sbi->total_node_count - F2FS_RESERVED_NODE_NUM; + buf->f_ffree = buf->f_files - valid_inode_count(sbi); buf->f_namelen = F2FS_NAME_LEN; buf->f_fsid.val[0] = (u32)id; diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index 6ff0b0b42d4..0ed77f32c9a 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -24,6 +24,9 @@ #define NULL_ADDR ((block_t)0) /* used as block_t addresses */ #define NEW_ADDR ((block_t)-1) /* used as block_t addresses */ +/* 0, 1(node nid), 2(meta nid) are reserved node id */ +#define F2FS_RESERVED_NODE_NUM 3 + #define F2FS_ROOT_INO(sbi) (sbi->root_ino_num) #define F2FS_NODE_INO(sbi) (sbi->node_ino_num) #define F2FS_META_INO(sbi) (sbi->meta_ino_num) -- cgit v1.2.3-70-g09d2 From 922cedbd00b68e506eb6f37cbe07cdf399a37b27 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 28 Aug 2014 16:13:21 +0300 Subject: f2fs: simplify by using a literal We can make the code a bit simpler because we know that "!retry" is zero. Signed-off-by: Dan Carpenter Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/f2fs/super.c') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 0f61f5aa587..41bdf511003 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1133,7 +1133,7 @@ free_sbi: /* give only one another chance */ if (retry) { - retry = !retry; + retry = 0; shrink_dcache_sb(sb); goto try_onemore; } -- cgit v1.2.3-70-g09d2