Re: [f2fs-dev] [PATCH v5 1/2] f2fs: correct i_size change for atomic writes

[Chao Yu <chao@kernel.org>]

On 2022/10/19 1:00, Daeho Jeong wrote:
> From: Daeho Jeong <daehojeong@google.com>
> 
> We need to make sure i_size doesn't change until atomic write commit is
> successful and restore it when commit is failed.
> 
> Signed-off-by: Daeho Jeong <daehojeong@google.com>
> ---
> v5: prevent only i_size update for atomic files when dirtying inode
> v4: move i_size update after clearing atomic file flag in
>      f2fs_abort_atomic_write()
> v3: make sure inode is clean while atomic writing
> ---
>   fs/f2fs/f2fs.h    |  8 ++++++++
>   fs/f2fs/file.c    | 18 +++++++++++-------
>   fs/f2fs/inode.c   |  5 ++++-
>   fs/f2fs/segment.c | 14 ++++++++++----
>   4 files changed, 33 insertions(+), 12 deletions(-)
> 
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index e7e750e6b332..68d2a3383c38 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -764,6 +764,7 @@ enum {
>   	FI_COMPRESS_RELEASED,	/* compressed blocks were released */
>   	FI_ALIGNED_WRITE,	/* enable aligned write */
>   	FI_COW_FILE,		/* indicate COW file */
> +	FI_ATOMIC_COMMIT,	/* indicate atomic commit completed except disk sync */

How about FI_ATOMIC_WRITTEN or FI_ATOMIC_COMMITTED? which may be
more readable?

>   	FI_MAX,			/* max flag, never be used */
>   };
>   
> @@ -822,6 +823,7 @@ struct f2fs_inode_info {
>   	unsigned int i_cluster_size;		/* cluster size */
>   
>   	unsigned int atomic_write_cnt;
> +	loff_t original_i_size;		/* original i_size before atomic write */
>   };
>   
>   static inline void get_extent_info(struct extent_info *ext,
> @@ -3072,6 +3074,8 @@ static inline void f2fs_i_blocks_write(struct inode *inode,
>   		set_inode_flag(inode, FI_AUTO_RECOVER);
>   }
>   
> +static inline bool f2fs_is_atomic_file(struct inode *inode);
> +
>   static inline void f2fs_i_size_write(struct inode *inode, loff_t i_size)
>   {
>   	bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
> @@ -3081,6 +3085,10 @@ static inline void f2fs_i_size_write(struct inode *inode, loff_t i_size)
>   		return;
>   
>   	i_size_write(inode, i_size);
> +
> +	if (f2fs_is_atomic_file(inode))
> +		return;
> +
>   	f2fs_mark_inode_dirty_sync(inode, true);
>   	if (clean || recover)
>   		set_inode_flag(inode, FI_AUTO_RECOVER);
> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> index ec9ee0f6d502..7ce629c95f4a 100644
> --- a/fs/f2fs/file.c
> +++ b/fs/f2fs/file.c
> @@ -1991,6 +1991,7 @@ static int f2fs_ioc_start_atomic_write(struct file *filp)
>   	struct f2fs_inode_info *fi = F2FS_I(inode);
>   	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
>   	struct inode *pinode;
> +	loff_t isize;
>   	int ret;
>   
>   	if (!inode_owner_or_capable(mnt_userns, inode))
> @@ -2049,7 +2050,12 @@ static int f2fs_ioc_start_atomic_write(struct file *filp)
>   		f2fs_up_write(&fi->i_gc_rwsem[WRITE]);
>   		goto out;
>   	}
> -	f2fs_i_size_write(fi->cow_inode, i_size_read(inode));
> +
> +	f2fs_write_inode(inode, NULL);
> +
> +	isize = i_size_read(inode);
> +	fi->original_i_size = isize;
> +	f2fs_i_size_write(fi->cow_inode, isize);
>   
>   	stat_inc_atomic_inode(inode);
>   
> @@ -2087,16 +2093,14 @@ static int f2fs_ioc_commit_atomic_write(struct file *filp)
>   
>   	if (f2fs_is_atomic_file(inode)) {
>   		ret = f2fs_commit_atomic_write(inode);
> -		if (ret)
> -			goto unlock_out;
> -
> -		ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 0, true);
>   		if (!ret)
> -			f2fs_abort_atomic_write(inode, false);
> +			ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 0, true);
> +
> +		f2fs_abort_atomic_write(inode, ret);
>   	} else {
>   		ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 1, false);
>   	}
> -unlock_out:
> +
>   	inode_unlock(inode);
>   	mnt_drop_write_file(filp);
>   	return ret;
> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> index 9f0d3864d9f1..044f132b87bc 100644
> --- a/fs/f2fs/inode.c
> +++ b/fs/f2fs/inode.c
> @@ -621,9 +621,12 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
>   	ri->i_uid = cpu_to_le32(i_uid_read(inode));
>   	ri->i_gid = cpu_to_le32(i_gid_read(inode));
>   	ri->i_links = cpu_to_le32(inode->i_nlink);
> -	ri->i_size = cpu_to_le64(i_size_read(inode));
>   	ri->i_blocks = cpu_to_le64(SECTOR_TO_BLOCK(inode->i_blocks) + 1);
>   
> +	if (!f2fs_is_atomic_file(inode) ||
> +			is_inode_flag_set(inode, FI_ATOMIC_COMMIT))
> +		ri->i_size = cpu_to_le64(i_size_read(inode));

So we expect that below flow can update i_size correctly, right?

- f2fs_ioc_commit_atomic_write
  - f2fs_do_sync_file
   - f2fs_skip_inode_update
   - f2fs_write_inode
    - f2fs_update_inode_page
     - f2fs_update_inode

Thanks,

> +
>   	if (et) {
>   		read_lock(&et->lock);
>   		set_raw_extent(&et->largest, &ri->i_ext);
> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> index 289bcb7ca300..264b4e352319 100644
> --- a/fs/f2fs/segment.c
> +++ b/fs/f2fs/segment.c
> @@ -192,14 +192,18 @@ void f2fs_abort_atomic_write(struct inode *inode, bool clean)
>   	if (!f2fs_is_atomic_file(inode))
>   		return;
>   
> -	if (clean)
> -		truncate_inode_pages_final(inode->i_mapping);
>   	clear_inode_flag(fi->cow_inode, FI_COW_FILE);
>   	iput(fi->cow_inode);
>   	fi->cow_inode = NULL;
>   	release_atomic_write_cnt(inode);
> +	clear_inode_flag(inode, FI_ATOMIC_COMMIT);
>   	clear_inode_flag(inode, FI_ATOMIC_FILE);
>   	stat_dec_atomic_inode(inode);
> +
> +	if (clean) {
> +		truncate_inode_pages_final(inode->i_mapping);
> +		f2fs_i_size_write(inode, fi->original_i_size);
> +	}
>   }
>   
>   static int __replace_atomic_write_block(struct inode *inode, pgoff_t index,
> @@ -335,10 +339,12 @@ static int __f2fs_commit_atomic_write(struct inode *inode)
>   	}
>   
>   out:
> -	if (ret)
> +	if (ret) {
>   		sbi->revoked_atomic_block += fi->atomic_write_cnt;
> -	else
> +	} else {
>   		sbi->committed_atomic_block += fi->atomic_write_cnt;
> +		set_inode_flag(inode, FI_ATOMIC_COMMIT);
> +	}
>   
>   	__complete_revoke_list(inode, &revoke_list, ret ? true : false);
>