[PATCH blktests 00/15] nvme: prepare for ANA support

[Daniel Wagner]

On Thu, Oct 24, 2024 at 10:00:10AM GMT, Shin'ichiro Kawasaki wrote:
> The 3rd patch caused the failure of the test case nvme/037 with fc transport.
> I'm not sure if it unveils a kernel bug, or if it should be fixed in blktests.
> I found the following change for the test case avoids the failure, which
> adds short waits after nvme disconnect and target cleanup. Comments on this
> change by nvme-fc experts will be appreciated.
> 
> diff --git a/tests/nvme/037 b/tests/nvme/037
> index 33a6857..e00f632 100755
> --- a/tests/nvme/037
> +++ b/tests/nvme/037
> @@ -36,7 +36,13 @@ test_device() {
> 
>                 _nvme_disconnect_subsys \
>                         --subsysnqn "${subsys}${i}" >>"${FULL}" 2>&1
> +               if [[ ${nvme_trtype} == fc ]]; then
> +                       sleep .1
> +               fi
>                 _nvmet_passthru_target_cleanup --subsysnqn "${subsys}${i}"
> +               if [[ ${nvme_trtype} == fc ]]; then
> +                       sleep .1
> +               fi

I think this is a workaround for a kernel bug which not really
surprising. I'll have a look.