On Thu, Aug 30, 2018 at 05:01:29PM +0300, Yuval Shaia wrote:
> When Ethernet function goes up the driver updates the device.
> Add support to device quiesce operation as a result of netdev down
> event.
>
> Since this operation might not be supported by the device yet a warning
> is printed and not error.
>
> While there fix the error message in the netdev up event.
>
> Signed-off-by: Yuval Shaia <yuval.shaia@xxxxxxxxxx>
> drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api.h | 1 +
> drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c | 11 ++++++++++-
> 2 files changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api.h b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api.h
> index 6fd5a8f4e2f6..fe966393991c 100644
> +++ b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api.h
> @@ -167,6 +167,7 @@ enum pvrdma_device_ctl {
> PVRDMA_DEVICE_CTL_ACTIVATE, /* Activate device. */
> PVRDMA_DEVICE_CTL_UNQUIESCE, /* Unquiesce device. */
> PVRDMA_DEVICE_CTL_RESET, /* Reset device. */
> + PVRDMA_DEVICE_CTL_QUIESCE, /* Quiesce device. */
> };
>
> enum pvrdma_intr_vector {
> diff --git a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c
> index a5719899f49a..0183faa2f54d 100644
> +++ b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c
> @@ -701,6 +701,15 @@ static void pvrdma_netdevice_event_handle(struct pvrdma_dev *dev,
> switch (event) {
> case NETDEV_REBOOT:
> case NETDEV_DOWN:
> + pvrdma_write_reg(dev, PVRDMA_REG_CTL,
> + PVRDMA_DEVICE_CTL_QUIESCE);
> +
> + mb();
Don't add mb()'s without a huge comment exmapling what it is supposed
to be doing.. kinda looks wrong really.
Jason
