Re: [dpdk-dev] [PATCH] net/ixgbe: add or remove MAC address

[Ye Xiaolong <xiaolong.ye@intel.com>]

Hi, guinan

For the title, better to use

Add support for vf MAC address add and remove

or something like so.

On 12/03, Guinan Sun wrote:
>Ixgbe PMD pf host code needs to support ixgbevf mac address
>add and remove. For this purpose, a response was added
>between pf and vf to update the mac address.

Does this mean each one vf can have multiple MAC addresses after this patch,
or this `add` actually means to replace the old mac address with the new one?

>
>Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
>---
> drivers/net/ixgbe/ixgbe_ethdev.h |  1 +
> drivers/net/ixgbe/ixgbe_pf.c     | 35 ++++++++++++++++++++++++++++++++
> 2 files changed, 36 insertions(+)
>
>diff --git a/drivers/net/ixgbe/ixgbe_ethdev.h b/drivers/net/ixgbe/ixgbe_ethdev.h
>index 76a1b9d18..e1cd8fd16 100644
>--- a/drivers/net/ixgbe/ixgbe_ethdev.h
>+++ b/drivers/net/ixgbe/ixgbe_ethdev.h
>@@ -270,6 +270,7 @@ struct ixgbe_vf_info {
> 	uint8_t api_version;
> 	uint16_t switch_domain_id;
> 	uint16_t xcast_mode;
>+	uint16_t mac_count;

How is this mac_count initialized? 

> };
> 
> /*
>diff --git a/drivers/net/ixgbe/ixgbe_pf.c b/drivers/net/ixgbe/ixgbe_pf.c
>index d0d85e138..76dbed2ab 100644
>--- a/drivers/net/ixgbe/ixgbe_pf.c
>+++ b/drivers/net/ixgbe/ixgbe_pf.c
>@@ -748,6 +748,37 @@ ixgbe_set_vf_mc_promisc(struct rte_eth_dev *dev, uint32_t vf, uint32_t *msgbuf)
> 	return 0;
> }
> 
>+static int
>+ixgbe_set_vf_macvlan_msg(struct rte_eth_dev *dev, uint32_t vf, uint32_t *msgbuf)
>+{
>+	struct ixgbe_hw *hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
>+	struct ixgbe_vf_info *vf_info =
>+		*(IXGBE_DEV_PRIVATE_TO_P_VFDATA(dev->data->dev_private));
>+	uint8_t *new_mac = (uint8_t *)(&msgbuf[1]);
>+	int index = (msgbuf[0] & IXGBE_VT_MSGINFO_MASK) >>
>+		    IXGBE_VT_MSGINFO_SHIFT;
>+
>+	if (index) {
>+		if (!rte_is_valid_assigned_ether_addr(
>+			(struct rte_ether_addr *)new_mac)) {
>+			RTE_LOG(ERR, PMD, "set invalid mac vf:%d\n", vf);
>+			return -1;
>+		}
>+
>+		if (new_mac == NULL)
>+			return -1;

I feel the null check should be in front of valid ether addr check, otherwise
there might be null pointer reference issue.

Thanks,
Xiaolong

>+
>+		vf_info[vf].mac_count++;
>+
>+		hw->mac.ops.set_rar(hw, vf_info[vf].mac_count,
>+				new_mac, vf, IXGBE_RAH_AV);
>+	} else {
>+		hw->mac.ops.clear_rar(hw, vf_info[vf].mac_count);
>+		vf_info[vf].mac_count = 0;
>+	}
>+	return 0;
>+}
>+
> static int
> ixgbe_rcv_msg_from_vf(struct rte_eth_dev *dev, uint16_t vf)
> {
>@@ -835,6 +866,10 @@ ixgbe_rcv_msg_from_vf(struct rte_eth_dev *dev, uint16_t vf)
> 		if (retval == RTE_PMD_IXGBE_MB_EVENT_PROCEED)
> 			retval = ixgbe_set_vf_mc_promisc(dev, vf, msgbuf);
> 		break;
>+	case IXGBE_VF_SET_MACVLAN:
>+		if (retval == RTE_PMD_IXGBE_MB_EVENT_PROCEED)
>+			retval = ixgbe_set_vf_macvlan_msg(dev, vf, msgbuf);
>+		break;
> 	default:
> 		PMD_DRV_LOG(DEBUG, "Unhandled Msg %8.8x", (unsigned)msgbuf[0]);
> 		retval = IXGBE_ERR_MBX;
>-- 
>2.17.1
>