On Fri, Jun 04, 2021 at 02:33:49PM +0900, Shin'ichiro Kawasaki wrote:
> Recent commit d2f442bc0bd5 ("ioengines: add get_max_open_zones zoned
> block device operation") modified fio to compare --max_open_zones option
> value and max_open_zones reported by the device. The device limit is
> fetched through sysfs or through an ioengine specific implementation.
>
> The test script currently try to fetch the max open zones limit using
> libzbc tools or sg_inq. If either of these fail, default value 128 is
> supplied. This default value can be too high when the test script is
> run for certain zoned block devices, and can therefore result in fio
> error and test case failure.
>
> To avoid the failure, modify the default value used in the test script
> from 128 to 0. With this, --max_open_zones=0 is passed to fio, and it
> makes fio use the max_open_zones reported by the device. Also add
> comments to describe why the test script gets max_open_zones with tools.
>
> Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@xxxxxxx>
> ---
> t/zbd/functions | 15 ++++++++++++---
> 1 file changed, 12 insertions(+), 3 deletions(-)
>
> diff --git a/t/zbd/functions b/t/zbd/functions
> index 40ffe1de..06973f79 100644
> --- a/t/zbd/functions
> +++ b/t/zbd/functions
> @@ -173,15 +173,24 @@ last_online_zone() {
> fi
> }
>
> +# Get max_open_zones of SMR drives using sg_inq or libzbc tools. Two test cases
> +# 31 and 32 use this max_open_zones value. The test case 31 uses max_open_zones
> +# to decide number of write target zones. The test case 32 passes max_open_zones
> +# value to fio with --max_open_zones option. Of note is that fio itself as the
> +# feature to get max_open_zones from the device through sysfs or ioengine
has the feature?
> +# specific implementation. This max_open_zones fetch by test script is required
> +# in case 1) ioengine lacks max_open_zones fetch feature, 2) Linux kernel lacks
> +# zoned block device support, or 3) Linux kernel supports zoned block devices
> +# but lacks max_open_zones in sysfs.
1) ioengine lacks max_open_zones
makes it sound like that is something that is needed,
it is not, there is the oslib implementation.
I think that you can replace 1, 2, 3 with something simpler like:
... is required in case fio is running on an old Linux kernel version which
lacks max_open_zones in sysfs.
Or if you want something more specific:
... is required in case fio is running on an old Linux kernel version which
lacks max_oper_zones in sysfs, or which lacks zbd support completely.
> max_open_zones() {
> local dev=$1
>
> if [ -n "${sg_inq}" ] && [ ! -n "${use_libzbc}" ]; then
> if ! ${sg_inq} -e --page=0xB6 --len=20 --hex "$dev" \
> > /dev/null 2>&1; then
> - # Non scsi device such as null_blk can not return max open zones.
> - # Use default value.
> - echo 128
> + # When sg_inq can not get max open zones, specify 0 which indicates
> + # fio to get max open zones limit from the device.
> + echo 0
> else
> ${sg_inq} -e --page=0xB6 --len=20 --hex "$dev" | tail -1 |
> {
> --
> 2.31.1
>
