blktests nvme 041,042 leak memory

[Maurizio Lombardi]

st 29. 5. 2024 v 14:51 odesílatel Sagi Grimberg <sagi at grimberg.me> napsal:
>
> --
> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> index f62fd49c1411..8a9d51e1ccd6 100644
> --- a/drivers/nvme/host/core.c
> +++ b/drivers/nvme/host/core.c
> @@ -4702,6 +4702,14 @@ int nvme_init_ctrl(struct nvme_ctrl *ctrl, struct
> device *dev,
>          nvme_fault_inject_fini(&ctrl->fault_inject);
>          dev_pm_qos_hide_latency_tolerance(ctrl->device);
>          cdev_device_del(&ctrl->cdev, ctrl->device);
> +       /*
> +        * this is nasty hack, but device_add allocated a device private
> +        * dev->p, which is freed assymmetricly in device_release, and
> +        * in order to cleanup after ourselves, and not rely on the device
> +        * .release handler (which also calls the nvme transport free
> handler)
> +        * we explicitly free the device private.
> +        */
> +       kfree(ctrl->device->p);
>   out_free_name:
>          nvme_put_ctrl(ctrl);
>          kfree_const(ctrl->device->kobj.name)
> --
>
> But this is ugly...
>

I like the fact that it's just a one-liner in nvme_init_ctrl, but
yes... it's really a hack that may stop working if device_add()'s
implementation changes.

Maurizio