Re: [PATCH v2 1/2] zonefs: add zone-capacity support

[Damien Le Moal <Damien.LeMoal@wdc.com>]

On 2020/07/20 17:52, Johannes Thumshirn wrote:
> In the zoned storage model, the sectors within a zone are typically all
> writeable. With the introduction of the Zoned Namespace (ZNS) Command
> Set in the NVM Express organization, the model was extended to have a
> specific writeable capacity.
> 
> This zone capacity can be less than the overall zone size for a NVMe ZNS
> device. For other zoned block devices like ZBC or null_blk in zoned-mode
> the zone capacity is always equal to the zone size.

null_blk has the zone_capacity option now to emulate ZNS smaller zone
capacities. But that option applies to sequential zones only. null_blk
conventional zones always have a capacity equal to zone size. Is it what you
meant to say here ?

> 
> Use the zone capacity field instead from blk_zone for determining the
> maximum inode size and inode blocks in zonefs.
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
> ---
>  fs/zonefs/super.c  | 11 +++++++----
>  fs/zonefs/zonefs.h |  3 +++
>  2 files changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
> index b13c332a3513..337249f98cae 100644
> --- a/fs/zonefs/super.c
> +++ b/fs/zonefs/super.c
> @@ -335,7 +335,7 @@ static void zonefs_io_error(struct inode *inode, bool write)
>  	struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
>  	unsigned int noio_flag;
>  	unsigned int nr_zones =
> -		zi->i_max_size >> (sbi->s_zone_sectors_shift + SECTOR_SHIFT);
> +		zi->i_zone_size >> (sbi->s_zone_sectors_shift + SECTOR_SHIFT);
>  	struct zonefs_ioerr_data err = {
>  		.inode = inode,
>  		.write = write,
> @@ -398,7 +398,7 @@ static int zonefs_file_truncate(struct inode *inode, loff_t isize)
>  		goto unlock;
>  
>  	ret = blkdev_zone_mgmt(inode->i_sb->s_bdev, op, zi->i_zsector,
> -			       zi->i_max_size >> SECTOR_SHIFT, GFP_NOFS);
> +			       zi->i_zone_size >> SECTOR_SHIFT, GFP_NOFS);
>  	if (ret) {
>  		zonefs_err(inode->i_sb,
>  			   "Zone management operation at %llu failed %d",
> @@ -1051,14 +1051,16 @@ static void zonefs_init_file_inode(struct inode *inode, struct blk_zone *zone,
>  
>  	zi->i_ztype = type;
>  	zi->i_zsector = zone->start;
> +	zi->i_zone_size = zone->len << SECTOR_SHIFT;
> +
>  	zi->i_max_size = min_t(loff_t, MAX_LFS_FILESIZE,
> -			       zone->len << SECTOR_SHIFT);
> +			       zone->capacity << SECTOR_SHIFT);
>  	zi->i_wpoffset = zonefs_check_zone_condition(inode, zone, true, true);
>  
>  	inode->i_uid = sbi->s_uid;
>  	inode->i_gid = sbi->s_gid;
>  	inode->i_size = zi->i_wpoffset;
> -	inode->i_blocks = zone->len;
> +	inode->i_blocks = zi->i_max_size >> SECTOR_SHIFT;
>  
>  	inode->i_op = &zonefs_file_inode_operations;
>  	inode->i_fop = &zonefs_file_operations;
> @@ -1169,6 +1171,7 @@ static int zonefs_create_zgroup(struct zonefs_zone_data *zd,
>  				else if (next->cond == BLK_ZONE_COND_OFFLINE)
>  					zone->cond = BLK_ZONE_COND_OFFLINE;
>  			}
> +			zone->capacity = zone->len;
>  		}

Normally, conventional zones on all known zoned devices will always have a zone
capacity equal to the zone size. But I would rather check that this is the case
here as the AGGRCNV option can only work if zone capacity is equal to the zone
size. So something like:


diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index abfb17f88f9a..db4853c7ec75 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -1164,12 +1164,17 @@ static int zonefs_create_zgroup(struct zonefs_zone_data *zd,
                                if (zonefs_zone_type(next) != type)
                                        break;
                                zone->len += next->len;
+                               zone->capacity += next->capacity;
                                if (next->cond == BLK_ZONE_COND_READONLY &&
                                    zone->cond != BLK_ZONE_COND_OFFLINE)
                                        zone->cond = BLK_ZONE_COND_READONLY;
                                else if (next->cond == BLK_ZONE_COND_OFFLINE)
                                        zone->cond = BLK_ZONE_COND_OFFLINE;
                        }
+                       if (zone->capacity != zone->len) {
+                               zonefs_err(sb, "Invalid conventional zone
capacity\n");
+                               ret = -EINVAL;
+                       }
                }

would be better.


>  
>  		/*
> diff --git a/fs/zonefs/zonefs.h b/fs/zonefs/zonefs.h
> index ad17fef7ce91..55b39970acb2 100644
> --- a/fs/zonefs/zonefs.h
> +++ b/fs/zonefs/zonefs.h
> @@ -56,6 +56,9 @@ struct zonefs_inode_info {
>  	/* File maximum size */
>  	loff_t			i_max_size;
>  
> +	/* File zone size */
> +	loff_t			i_zone_size;
> +
>  	/*
>  	 * To serialise fully against both syscall and mmap based IO and
>  	 * sequential file truncation, two locks are used. For serializing
> 
-- 
Damien Le Moal
Western Digital Research