The current implementation replies on H_IOCTL call to issue a H_SESSION_ERR_DETECTED command to let the hypervisor to send a failover signal. However, it may not work while the vnic is already in error state, e.g., "ibmvnic 30000003 env3: rx buffer returned with rc 6". Add a last resort, that is to schedule a failover reset via CRQ command.
Signed-off-by: Lijun Pan <lijunp...@gmail.com> --- drivers/net/ethernet/ibm/ibmvnic.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index ee9bf18c597f..d44a7b5b8f67 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -5512,14 +5512,14 @@ static ssize_t failover_store(struct device *dev, struct device_attribute *attr, rc = plpar_hcall_norets(H_VIOCTL, adapter->vdev->unit_address, H_SESSION_ERR_DETECTED, session_token, 0, 0); if (rc) { - netdev_err(netdev, "Client initiated failover failed, rc %ld\n", + netdev_err(netdev, + "H_VIOCTL initiated failover failed, rc %ld, trying to send CRQ_CMD, the last resort\n", rc); - return -EINVAL; + ibmvnic_reset(adapter, VNIC_RESET_FAILOVER); } return count; } - static DEVICE_ATTR_WO(failover); static unsigned long ibmvnic_get_desired_dma(struct vio_dev *vdev) -- 2.23.0