Re: [PATCH 1/4] zbd: move zone finish operation to zbd_convert_to_write_zone()

On 4/11/25 8:12 PM, Shin'ichiro Kawasaki wrote:
> Currently, when a write target zone has fewer remainder sectors than
> the block size, fio finishes the zone to make the zone inactive (not
> open), so that another zone can be open and used as a write target zone.
> This zone finish operation is implemented in zbd_adjust_block().
> However, this placement is less ideal because zbd_adjust_block() manages
> not just write requests but also read and trim requests.
> 
> Since the zone finish operation is exclusively necessary for write
> requests, implement it into zbd_convert_to_write_zone(). While at it,
> improve the function comment.
> 
> Signed-off-by: Shin'ichiro Kawasaki <[email protected]>

A few nits below. But otherwise, nice cleanup !

Reviewed-by: Damien Le Moal <[email protected]>

> ---
>  zbd.c | 84 ++++++++++++++++++++++++++++++++---------------------------
>  1 file changed, 45 insertions(+), 39 deletions(-)
> 
> diff --git a/zbd.c b/zbd.c
> index ee095b1d..5e0f7f4e 100644
> --- a/zbd.c
> +++ b/zbd.c
> @@ -1396,16 +1396,23 @@ static bool any_io_in_flight(void)
>  	return false;
>  }
>  
> -/*
> +/**
> + * zbd_convert_to_write_zone - Convert the write request to writable zone

Maybe: Convert the target zone of an io_u to a writable zone
?

> + * @td: The fio thread data
> + * @io_u: The I/O unit write request to convert

We do not convert the io_u, but rather, the zones it targets, no ?

> + * @zb: The zone selected at the beginning of the function call. The caller must
> + *      hold zb->mutex.
> + *
>   * Modify the offset of an I/O unit that does not refer to a zone such that
> - * in write target zones array. Add a zone to or remove a zone from the lsit if
> + * in write target zones array. Add a zone to or remove a zone from the list if

Nit: s/list/array

>   * necessary. The write target zone is searched across sequential zones.
>   * This algorithm can only work correctly if all write pointers are
> - * a multiple of the fio block size. The caller must neither hold z->mutex
> - * nor f->zbd_info->mutex. Returns with z->mutex held upon success.
> + * a multiple of the fio block size. The caller must not hold
> + * f->zbd_info->mutex. Returns with z->mutex held upon success.
>   */


-- 
Damien Le Moal
Western Digital Research