Re: [PATCH] f2fs: fix to release compress file for F2FS_IOC_RESERVE_COMPRESS_BLOCKS when has no space

[Jaegeuk Kim]

On 02/10, Yangtao Li wrote:
> When the released file is reserving and the space is insufficient,
> the fsck flag is set incorrectly, and lead to the file cannot be
> reserved and released normally.
> 
> $ mount -t f2fs -o compress_extension=*,compress_mode=user /mnt/9p/f2fs.img
> 	/mnt/f2fs/
> $ dd if=/dev/zero of=/mnt/f2fs/800M bs=1M count=800
> $ dd if=/dev/zero of=/mnt/f2fs/60M bs=1M count=60
> $ f2fs_io compress /mnt/f2fs/60M
> $ ./mnt/9p/my_f2fs_io release_cblocks /mnt/f2fs/60M
> 11520
> $ dd if=/dev/zero of=/mnt/f2fs/30M bs=1M count=30
> $ f2fs_io reserve_cblocks /mnt/f2fs/60M
> [   56.399712] F2FS-fs (loop0): f2fs_reserve_compress_blocks: partial blocks
> 	were released i_ino=cf iblocks=76104, reserved=5220, compr_blocks=5655,
> 	run fsck to fix.
> F2FS_IOC_RESERVE_COMPRESS_BLOCKS failed: No space left on device
> 
> $ ./mnt/9p/my_f2fs_io reserve_cblocks /mnt/f2fs/60M
> 0
> $ ./mnt/9p/my_f2fs_io release_cblocks /mnt/f2fs/60M
> F2FS_IOC_RELEASE_COMPRESS_BLOCKS failed: Invalid argument
> 
> $ rm /mnt/f2fs/30M
> $ ./mnt/9p/my_f2fs_io reserve_cblocks /mnt/f2fs/60M
> 0
> $ ./mnt/9p/my_f2fs_io release_cblocks /mnt/f2fs/60M
> F2FS_IOC_RELEASE_COMPRESS_BLOCKS failed: Invalid argument
> 
> In this case, let's release back the reserved part of the compressed file.
> 
> Fixes: c75488fb4d82 ("f2fs: introduce F2FS_IOC_RESERVE_COMPRESS_BLOCKS")
> Signed-off-by: Qi Han <hanqi@xxxxxxxx>
> Signed-off-by: Yangtao Li <frank.li@xxxxxxxx>
> ---
>  fs/f2fs/file.c | 97 ++++++++++++++++++++++++++++++++------------------
>  1 file changed, 63 insertions(+), 34 deletions(-)
> 
> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> index 300eae8b5415..8f3f55ac153a 100644
> --- a/fs/f2fs/file.c
> +++ b/fs/f2fs/file.c
> @@ -3427,11 +3427,52 @@ static int release_compress_blocks(struct dnode_of_data *dn, pgoff_t count)
>  	return released_blocks;
>  }
>  
> +static int f2fs_do_release_compress_blocks(struct inode *inode,
> +			pgoff_t page_count, unsigned int *released_blocks)
> +{
> +	pgoff_t page_idx = 0;
> +	int ret;
> +
> +	while (page_idx < page_count) {
> +		struct dnode_of_data dn;
> +		pgoff_t end_offset, count;
> +
> +		set_new_dnode(&dn, inode, NULL, NULL, 0);
> +		ret = f2fs_get_dnode_of_data(&dn, page_idx, LOOKUP_NODE);
> +		if (ret) {
> +			if (ret == -ENOENT) {
> +				page_idx = f2fs_get_next_page_offset(&dn,
> +								page_idx);
> +				ret = 0;
> +				continue;
> +			}
> +			break;
> +		}
> +
> +		end_offset = ADDRS_PER_PAGE(dn.node_page, inode);
> +		count = min(end_offset - dn.ofs_in_node, page_count - page_idx);
> +		count = round_up(count, F2FS_I(inode)->i_cluster_size);
> +
> +		ret = release_compress_blocks(&dn, count);
> +
> +		f2fs_put_dnode(&dn);
> +
> +		if (ret < 0)
> +			break;
> +
> +		page_idx += count;
> +		if (released_blocks)
> +			*released_blocks += ret;
> +	}
> +
> +	return ret;
> +}
> +
>  static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
>  {
>  	struct inode *inode = file_inode(filp);
>  	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> -	pgoff_t page_idx = 0, last_idx;
> +	pgoff_t last_idx;
>  	unsigned int released_blocks = 0;
>  	int ret;
>  	int writecount;
> @@ -3481,36 +3522,7 @@ static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
>  
>  	last_idx = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE);
>  
> -	while (page_idx < last_idx) {
> -		struct dnode_of_data dn;
> -		pgoff_t end_offset, count;
> -
> -		set_new_dnode(&dn, inode, NULL, NULL, 0);
> -		ret = f2fs_get_dnode_of_data(&dn, page_idx, LOOKUP_NODE);
> -		if (ret) {
> -			if (ret == -ENOENT) {
> -				page_idx = f2fs_get_next_page_offset(&dn,
> -								page_idx);
> -				ret = 0;
> -				continue;
> -			}
> -			break;
> -		}
> -
> -		end_offset = ADDRS_PER_PAGE(dn.node_page, inode);
> -		count = min(end_offset - dn.ofs_in_node, last_idx - page_idx);
> -		count = round_up(count, F2FS_I(inode)->i_cluster_size);
> -
> -		ret = release_compress_blocks(&dn, count);
> -
> -		f2fs_put_dnode(&dn);
> -
> -		if (ret < 0)
> -			break;
> -
> -		page_idx += count;
> -		released_blocks += ret;
> -	}
> +	ret = f2fs_do_release_compress_blocks(inode, last_idx, &released_blocks);
>  
>  	filemap_invalidate_unlock(inode->i_mapping);
>  	f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
> @@ -3585,8 +3597,22 @@ static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count)
>  		if (ret)
>  			return ret;
>  
> -		if (reserved != cluster_size - compr_blocks)
> -			return -ENOSPC;
> +		if (reserved != cluster_size - compr_blocks) {
> +			dec_valid_block_count(sbi, dn->inode, reserved);

This looks breaking the consistency?

> +
> +			for (i = cluster_size - 1; i > 0; i--) {
> +				dn->ofs_in_node--;
> +				blkaddr = f2fs_data_blkaddr(dn);
> +
> +				if (__is_valid_data_blkaddr(blkaddr)) {
> +					dn->ofs_in_node -= i;
> +					return -ENOSPC;
> +				}
> +
> +				dn->data_blkaddr = NULL_ADDR;
> +				f2fs_set_data_blkaddr(dn);
> +			}
> +		}
>  
>  		f2fs_i_compr_blocks_update(dn->inode, compr_blocks, true);
>  
> @@ -3604,6 +3630,7 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
>  	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
>  	pgoff_t page_idx = 0, last_idx;
>  	unsigned int reserved_blocks = 0;
> +	struct dnode_of_data dn;
>  	int ret;
>  
>  	if (!f2fs_sb_has_compression(F2FS_I_SB(inode)))
> @@ -3637,7 +3664,6 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
>  	last_idx = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE);
>  
>  	while (page_idx < last_idx) {
> -		struct dnode_of_data dn;
>  		pgoff_t end_offset, count;
>  
>  		set_new_dnode(&dn, inode, NULL, NULL, 0);
> @@ -3667,6 +3693,9 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
>  		reserved_blocks += ret;
>  	}
>  
> +	if (ret == -ENOSPC)
> +		f2fs_do_release_compress_blocks(inode, page_idx + dn.ofs_in_node, NULL);
> +
>  	filemap_invalidate_unlock(inode->i_mapping);
>  	f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
>  
> -- 
> 2.25.1