Re: [f2fs-dev] [PATCH v3] f2fs: add reserved blocks for root user

[Jaegeuk Kim]

On 01/02, Chao Yu wrote:
> On 2018/1/1 9:29, Jaegeuk Kim wrote:
> > This patch allows root to reserve some blocks via mount option.
> > 
> > "-o reserve_root=N" means N x 4KB-sized blocks for root only.
> > 
> > Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
> > ---
> > 
> > Change log from v2:
> >  - wrong submission. :P
> > 
> >  fs/f2fs/f2fs.h  | 26 ++++++++++++++++++++++----
> >  fs/f2fs/super.c | 26 +++++++++++++++++++++++---
> >  fs/f2fs/sysfs.c |  3 ++-
> >  3 files changed, 47 insertions(+), 8 deletions(-)
> > 
> > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > index 5f7f42267221..123d875f7293 100644
> > --- a/fs/f2fs/f2fs.h
> > +++ b/fs/f2fs/f2fs.h
> > @@ -95,6 +95,7 @@ extern char *fault_name[FAULT_MAX];
> >  #define F2FS_MOUNT_PRJQUOTA		0x00200000
> >  #define F2FS_MOUNT_QUOTA		0x00400000
> >  #define F2FS_MOUNT_INLINE_XATTR_SIZE	0x00800000
> > +#define F2FS_MOUNT_RESERVE_ROOT		0x01000000
> >  
> >  #define clear_opt(sbi, option)	((sbi)->mount_opt.opt &= ~F2FS_MOUNT_##option)
> >  #define set_opt(sbi, option)	((sbi)->mount_opt.opt |= F2FS_MOUNT_##option)
> > @@ -1110,6 +1111,7 @@ struct f2fs_sb_info {
> >  	block_t last_valid_block_count;		/* for recovery */
> >  	block_t reserved_blocks;		/* configurable reserved blocks */
> >  	block_t current_reserved_blocks;	/* current reserved blocks */
> > +	block_t root_reserved_blocks;		/* root reserved blocks */
> >  
> >  	unsigned int nquota_files;		/* # of quota sysfile */
> >  
> > @@ -1562,6 +1564,12 @@ static inline bool f2fs_has_xattr_block(unsigned int ofs)
> >  	return ofs == XATTR_NODE_OFFSET;
> >  }
> >  
> > +static inline block_t reserve_root_limit(struct f2fs_sb_info *sbi)
> > +{
> > +	/* limit is 0.2% */
> 
> Should be 2% according to below calculation?

Oh, right. Fixed.

> 
> > +	return (sbi->user_block_count << 1) / 100;
> > +}
> > +
> >  static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool);
> >  static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
> >  				 struct inode *inode, blkcnt_t *count)
> > @@ -1591,11 +1599,17 @@ static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
> >  	sbi->total_valid_block_count += (block_t)(*count);
> >  	avail_user_block_count = sbi->user_block_count -
> >  					sbi->current_reserved_blocks;
> > +
> > +	if (!(test_opt(sbi, RESERVE_ROOT) && capable(CAP_SYS_RESOURCE)))
> 
> How about adding uid & gid verification also like ext4?

Again, that's another feature which requires a mount option. I think it'd be
better to add that, once we have a use-case.

> As this is a mount option, in ->remount_fs, we should consider to recover
> original reserved block number if we encounter some error during remount.

Yup, agreed.

Thanks,

> 
> Thanks,
> 
> > +		avail_user_block_count -= sbi->root_reserved_blocks;
> > +
> >  	if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) {
> >  		diff = sbi->total_valid_block_count - avail_user_block_count;
> > +		if (diff > *count)
> > +			diff = *count;
> >  		*count -= diff;
> >  		release = diff;
> > -		sbi->total_valid_block_count = avail_user_block_count;
> > +		sbi->total_valid_block_count -= diff;
> >  		if (!*count) {
> >  			spin_unlock(&sbi->stat_lock);
> >  			percpu_counter_sub(&sbi->alloc_valid_block_count, diff);
> > @@ -1784,9 +1798,13 @@ static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
> >  
> >  	spin_lock(&sbi->stat_lock);
> >  
> > -	valid_block_count = sbi->total_valid_block_count + 1;
> > -	if (unlikely(valid_block_count + sbi->current_reserved_blocks >
> > -						sbi->user_block_count)) {
> > +	valid_block_count = sbi->total_valid_block_count +
> > +					sbi->current_reserved_blocks + 1;
> > +
> > +	if (!(test_opt(sbi, RESERVE_ROOT) && capable(CAP_SYS_RESOURCE)))
> > +		valid_block_count += sbi->root_reserved_blocks;
> > +
> > +	if (unlikely(valid_block_count > sbi->user_block_count)) {
> >  		spin_unlock(&sbi->stat_lock);
> >  		goto enospc;
> >  	}
> > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > index cb876d905ca5..9221b013db98 100644
> > --- a/fs/f2fs/super.c
> > +++ b/fs/f2fs/super.c
> > @@ -107,6 +107,7 @@ enum {
> >  	Opt_noextent_cache,
> >  	Opt_noinline_data,
> >  	Opt_data_flush,
> > +	Opt_reserve_root,
> >  	Opt_mode,
> >  	Opt_io_size_bits,
> >  	Opt_fault_injection,
> > @@ -157,6 +158,7 @@ static match_table_t f2fs_tokens = {
> >  	{Opt_noextent_cache, "noextent_cache"},
> >  	{Opt_noinline_data, "noinline_data"},
> >  	{Opt_data_flush, "data_flush"},
> > +	{Opt_reserve_root, "reserve_root=%u"},
> >  	{Opt_mode, "mode=%s"},
> >  	{Opt_io_size_bits, "io_bits=%u"},
> >  	{Opt_fault_injection, "fault_injection=%u"},
> > @@ -488,6 +490,12 @@ static int parse_options(struct super_block *sb, char *options)
> >  		case Opt_data_flush:
> >  			set_opt(sbi, DATA_FLUSH);
> >  			break;
> > +		case Opt_reserve_root:
> > +			if (args->from && match_int(args, &arg))
> > +				return -EINVAL;
> > +			sbi->root_reserved_blocks = arg;
> > +			set_opt(sbi, RESERVE_ROOT);
> > +			break;
> >  		case Opt_mode:
> >  			name = match_strdup(&args[0]);
> >  
> > @@ -1005,10 +1013,12 @@ static int f2fs_statfs(struct dentry *dentry, struct kstatfs *buf)
> >  	buf->f_bsize = sbi->blocksize;
> >  
> >  	buf->f_blocks = total_count - start_count;
> > -	buf->f_bfree = user_block_count - valid_user_blocks(sbi) + ovp_count;
> > +	buf->f_bfree = user_block_count - valid_user_blocks(sbi) -
> > +						sbi->current_reserved_blocks +
> > +						sbi->root_reserved_blocks;
> >  	buf->f_bavail = user_block_count - valid_user_blocks(sbi) -
> > -						sbi->current_reserved_blocks;
> > -
> > +						sbi->current_reserved_blocks -
> > +						sbi->root_reserved_blocks;
> >  	avail_node_count = sbi->total_node_count - sbi->nquota_files -
> >  						F2FS_RESERVED_NODE_NUM;
> >  
> > @@ -1136,6 +1146,9 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
> >  	else if (test_opt(sbi, LFS))
> >  		seq_puts(seq, "lfs");
> >  	seq_printf(seq, ",active_logs=%u", sbi->active_logs);
> > +	if (test_opt(sbi, RESERVE_ROOT))
> > +		seq_printf(seq, ",reserve_root=%u",
> > +				sbi->root_reserved_blocks);
> >  	if (F2FS_IO_SIZE_BITS(sbi))
> >  		seq_printf(seq, ",io_size=%uKB", F2FS_IO_SIZE_KB(sbi));
> >  #ifdef CONFIG_F2FS_FAULT_INJECTION
> > @@ -2571,6 +2584,13 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> >  	sbi->reserved_blocks = 0;
> >  	sbi->current_reserved_blocks = 0;
> >  
> > +	if (test_opt(sbi, RESERVE_ROOT) &&
> > +		sbi->root_reserved_blocks > reserve_root_limit(sbi)) {
> > +		sbi->root_reserved_blocks = reserve_root_limit(sbi);
> > +		f2fs_msg(sb, KERN_INFO,
> > +			"Reduce reserved blocks for root = %u\n",
> > +				sbi->root_reserved_blocks);
> > +	}
> >  	for (i = 0; i < NR_INODE_TYPE; i++) {
> >  		INIT_LIST_HEAD(&sbi->inode_list[i]);
> >  		spin_lock_init(&sbi->inode_lock[i]);
> > diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
> > index 7e15cbc5b8e1..41887e6ec1b3 100644
> > --- a/fs/f2fs/sysfs.c
> > +++ b/fs/f2fs/sysfs.c
> > @@ -162,7 +162,8 @@ static ssize_t f2fs_sbi_store(struct f2fs_attr *a,
> >  #endif
> >  	if (a->struct_type == RESERVED_BLOCKS) {
> >  		spin_lock(&sbi->stat_lock);
> > -		if (t > (unsigned long)sbi->user_block_count) {
> > +		if (t > (unsigned long)(sbi->user_block_count -
> > +					sbi->root_reserved_blocks)) {
> >  			spin_unlock(&sbi->stat_lock);
> >  			return -EINVAL;
> >  		}
> >