[PATCH] btrfs: zoned: fix full zone SB reading on ZNS

[PATCH] btrfs: zoned: fix full zone SB reading on ZNS

[Naohiro Aota]

When both of the superblock zones are full, we need to check which
superblock is newer. The calculation of last superblock position is wrong
as it does not consider zone_capacity. Fix it.

Fixes: 9658b72ef300 ("btrfs: zoned: locate superblock position using zone capacity")
CC: stable@vger.kernel.org # 6.1+
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
---
 fs/btrfs/zoned.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index e3fe02aae641..cd1fee22998c 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -122,10 +122,9 @@ static int sb_write_pointer(struct block_device *bdev, struct blk_zone *zones,
 		int i;
 
 		for (i = 0; i < BTRFS_NR_SB_LOG_ZONES; i++) {
-			u64 bytenr;
-
-			bytenr = ((zones[i].start + zones[i].len)
-				   << SECTOR_SHIFT) - BTRFS_SUPER_INFO_SIZE;
+			u64 zone_end = (zones[i].start + zones[i].capacity) << SECTOR_SHIFT;
+			u64 bytenr = ALIGN_DOWN(zone_end, BTRFS_SUPER_INFO_SIZE) -
+				BTRFS_SUPER_INFO_SIZE;
 
 			page[i] = read_cache_page_gfp(mapping,
 					bytenr >> PAGE_SHIFT, GFP_NOFS);
-- 
2.40.1

Re: [PATCH] btrfs: zoned: fix full zone SB reading on ZNS

[Johannes Thumshirn]

Looks good,
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Re: [PATCH] btrfs: zoned: fix full zone SB reading on ZNS

[David Sterba]

On Tue, May 09, 2023 at 06:29:15PM +0000, Naohiro Aota wrote:
> When both of the superblock zones are full, we need to check which
> superblock is newer. The calculation of last superblock position is wrong
> as it does not consider zone_capacity. Fix it.
> 
> Fixes: 9658b72ef300 ("btrfs: zoned: locate superblock position using zone capacity")
> CC: stable@vger.kernel.org # 6.1+
> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>

Added to misc-next, thanks.

> ---
>  fs/btrfs/zoned.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
> index e3fe02aae641..cd1fee22998c 100644
> --- a/fs/btrfs/zoned.c
> +++ b/fs/btrfs/zoned.c
> @@ -122,10 +122,9 @@ static int sb_write_pointer(struct block_device *bdev, struct blk_zone *zones,
>  		int i;
>  
>  		for (i = 0; i < BTRFS_NR_SB_LOG_ZONES; i++) {
> -			u64 bytenr;
> -
> -			bytenr = ((zones[i].start + zones[i].len)
> -				   << SECTOR_SHIFT) - BTRFS_SUPER_INFO_SIZE;
> +			u64 zone_end = (zones[i].start + zones[i].capacity) << SECTOR_SHIFT;
> +			u64 bytenr = ALIGN_DOWN(zone_end, BTRFS_SUPER_INFO_SIZE) -
> +				BTRFS_SUPER_INFO_SIZE;

I did a quick grep for 'zone.*len' as a wide search for potential
candidates of the same pattern but haven't spotted anything obvious.
With the ZNS support it may be useful to have a helper to read the zone
end in case the real blocks need to be considered, like here.