Merge branch '10GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue

Tony Nguyen says:

====================
10GbE Intel Wired LAN Driver Updates 2021-11-17

Radoslaw Tyl says:

The change is a consequence of errors reported by the ixgbevf driver
while starting several virtual guests at the same time on ESX host.
During this, VF was not able to communicate correctly with the PF,
as a result reported "PF still in reset state. Is the PF interface up?"
and then goes to locked state. The only thing left was to reload
the VF driver on the guest OS.

The background of the problem is that the current PFU and VFU
semaphore locking mechanism between sender and receiver may cause
overriding Mailbox memory (VFMBMEM), in such scenario receiver of
the original message will read the invalid, corrupted or one (or more)
message may be lost.

This change is actually as a support for communication with PF ESX
driver and does not contains changes and support for ixgbe driver.
For maintain backward compatibility, previous communication method
has been preserved in the form of LEGACY functions.

In the future there is a plan to add a support for a 1.5 mailbox API
communication also to ixgbe driver.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/intel/ixgbevf/defines.h

@@ -281,6 +281,10 @@ struct ixgbe_adv_tx_context_desc {
 #define IXGBE_ERR_INVALID_MAC_ADDR	-1
 #define IXGBE_ERR_RESET_FAILED		-2
 #define IXGBE_ERR_INVALID_ARGUMENT	-3
+#define IXGBE_ERR_CONFIG		-4
+#define IXGBE_ERR_MBX			-5
+#define IXGBE_ERR_TIMEOUT		-6
+#define IXGBE_ERR_PARAM			-7
 
 /* Transmit Config masks */
 #define IXGBE_TXDCTL_ENABLE		0x02000000 /* Ena specific Tx Queue */

drivers/net/ethernet/intel/ixgbevf/ipsec.c

@@ -40,16 +40,16 @@ static int ixgbevf_ipsec_set_pf_sa(struct ixgbevf_adapter *adapter,
 
 	spin_lock_bh(&adapter->mbx_lock);
 
-	ret = hw->mbx.ops.write_posted(hw, msgbuf, IXGBE_VFMAILBOX_SIZE);
+	ret = ixgbevf_write_mbx(hw, msgbuf, IXGBE_VFMAILBOX_SIZE);
 	if (ret)
 		goto out;
 
-	ret = hw->mbx.ops.read_posted(hw, msgbuf, 2);
+	ret = ixgbevf_poll_mbx(hw, msgbuf, 2);
 	if (ret)
 		goto out;
 
 	ret = (int)msgbuf[1];
-	if (msgbuf[0] & IXGBE_VT_MSGTYPE_NACK && ret >= 0)
+	if (msgbuf[0] & IXGBE_VT_MSGTYPE_FAILURE && ret >= 0)
 		ret = -1;
 
 out:
@@ -77,11 +77,11 @@ static int ixgbevf_ipsec_del_pf_sa(struct ixgbevf_adapter *adapter, int pfsa)
 
 	spin_lock_bh(&adapter->mbx_lock);
 
-	err = hw->mbx.ops.write_posted(hw, msgbuf, 2);
+	err = ixgbevf_write_mbx(hw, msgbuf, 2);
 	if (err)
 		goto out;
 
-	err = hw->mbx.ops.read_posted(hw, msgbuf, 2);
+	err = ixgbevf_poll_mbx(hw, msgbuf, 2);
 	if (err)
 		goto out;
 
@@ -623,6 +623,7 @@ void ixgbevf_init_ipsec_offload(struct ixgbevf_adapter *adapter)
 
 	switch (adapter->hw.api_version) {
 	case ixgbe_mbox_api_14:
+	case ixgbe_mbox_api_15:
 		break;
 	default:
 		return;

drivers/net/ethernet/intel/ixgbevf/ixgbevf.h

@@ -430,6 +430,7 @@ extern const struct ixgbevf_info ixgbevf_X540_vf_info;
 extern const struct ixgbevf_info ixgbevf_X550_vf_info;
 extern const struct ixgbevf_info ixgbevf_X550EM_x_vf_info;
 extern const struct ixgbe_mbx_operations ixgbevf_mbx_ops;
+extern const struct ixgbe_mbx_operations ixgbevf_mbx_ops_legacy;
 extern const struct ixgbevf_info ixgbevf_x550em_a_vf_info;
 
 extern const struct ixgbevf_info ixgbevf_82599_vf_hv_info;
@@ -491,4 +492,8 @@ void ixgbe_napi_del_all(struct ixgbevf_adapter *adapter);
 
 #define hw_dbg(hw, format, arg...) \
 	netdev_dbg(ixgbevf_hw_to_netdev(hw), format, ## arg)
+
+s32 ixgbevf_poll_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size);
+s32 ixgbevf_write_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size);
+
 #endif /* _IXGBEVF_H_ */

drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c

@@ -2266,6 +2266,7 @@ static void ixgbevf_negotiate_api(struct ixgbevf_adapter *adapter)
 {
 	struct ixgbe_hw *hw = &adapter->hw;
 	static const int api[] = {
+		ixgbe_mbox_api_15,
 		ixgbe_mbox_api_14,
 		ixgbe_mbox_api_13,
 		ixgbe_mbox_api_12,
@@ -2284,6 +2285,12 @@ static void ixgbevf_negotiate_api(struct ixgbevf_adapter *adapter)
 		idx++;
 	}
 
+	if (hw->api_version >= ixgbe_mbox_api_15) {
+		hw->mbx.ops.init_params(hw);
+		memcpy(&hw->mbx.ops, &ixgbevf_mbx_ops,
+		       sizeof(struct ixgbe_mbx_operations));
+	}
+
 	spin_unlock_bh(&adapter->mbx_lock);
 }
 
@@ -2627,6 +2634,7 @@ static void ixgbevf_set_num_queues(struct ixgbevf_adapter *adapter)
 		case ixgbe_mbox_api_12:
 		case ixgbe_mbox_api_13:
 		case ixgbe_mbox_api_14:
+		case ixgbe_mbox_api_15:
 			if (adapter->xdp_prog &&
 			    hw->mac.max_tx_queues == rss)
 				rss = rss > 3 ? 2 : 1;
@@ -4565,7 +4573,7 @@ static int ixgbevf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	memcpy(&hw->mac.ops, ii->mac_ops, sizeof(hw->mac.ops));
 	hw->mac.type  = ii->mac;
 
-	memcpy(&hw->mbx.ops, &ixgbevf_mbx_ops,
+	memcpy(&hw->mbx.ops, &ixgbevf_mbx_ops_legacy,
 	       sizeof(struct ixgbe_mbx_operations));
 
 	/* setup the private structure */
@@ -4625,6 +4633,7 @@ static int ixgbevf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	case ixgbe_mbox_api_12:
 	case ixgbe_mbox_api_13:
 	case ixgbe_mbox_api_14:
+	case ixgbe_mbox_api_15:
 		netdev->max_mtu = IXGBE_MAX_JUMBO_FRAME_SIZE -
 				  (ETH_HLEN + ETH_FCS_LEN);
 		break;

drivers/net/ethernet/intel/ixgbevf/mbx.h

@@ -7,7 +7,6 @@
 #include "vf.h"
 
 #define IXGBE_VFMAILBOX_SIZE	16 /* 16 32 bit words - 64 bytes */
-#define IXGBE_ERR_MBX		-100
 
 #define IXGBE_VFMAILBOX		0x002FC
 #define IXGBE_VFMBMEM		0x00200
@@ -39,14 +38,17 @@
 
 /* If it's a IXGBE_VF_* msg then it originates in the VF and is sent to the
  * PF.  The reverse is true if it is IXGBE_PF_*.
- * Message ACK's are the value or'd with 0xF0000000
+ * Message results are the value or'd with 0xF0000000
  */
-/* Messages below or'd with this are the ACK */
-#define IXGBE_VT_MSGTYPE_ACK	0x80000000
-/* Messages below or'd with this are the NACK */
-#define IXGBE_VT_MSGTYPE_NACK	0x40000000
-/* Indicates that VF is still clear to send requests */
-#define IXGBE_VT_MSGTYPE_CTS	0x20000000
+#define IXGBE_VT_MSGTYPE_SUCCESS	0x80000000 /* Messages or'd with this
+						    * have succeeded
+						    */
+#define IXGBE_VT_MSGTYPE_FAILURE	0x40000000 /* Messages or'd with this
+						    * have failed
+						    */
+#define IXGBE_VT_MSGTYPE_CTS		0x20000000 /* Indicates that VF is still
+						    * clear to send requests
+						    */
 #define IXGBE_VT_MSGINFO_SHIFT	16
 /* bits 23:16 are used for exra info for certain messages */
 #define IXGBE_VT_MSGINFO_MASK	(0xFF << IXGBE_VT_MSGINFO_SHIFT)
@@ -63,6 +65,7 @@ enum ixgbe_pfvf_api_rev {
 	ixgbe_mbox_api_12,	/* API version 1.2, linux/freebsd VF driver */
 	ixgbe_mbox_api_13,	/* API version 1.3, linux/freebsd VF driver */
 	ixgbe_mbox_api_14,	/* API version 1.4, linux/freebsd VF driver */
+	ixgbe_mbox_api_15,	/* API version 1.5, linux/freebsd VF driver */
 	/* This value should always be last */
 	ixgbe_mbox_api_unknown,	/* indicates that API version is not known */
 };

drivers/net/ethernet/intel/ixgbevf/vf.c

@@ -13,13 +13,12 @@
 static inline s32 ixgbevf_write_msg_read_ack(struct ixgbe_hw *hw, u32 *msg,
 					     u32 *retmsg, u16 size)
 {
-	struct ixgbe_mbx_info *mbx = &hw->mbx;
-	s32 retval = mbx->ops.write_posted(hw, msg, size);
+	s32 retval = ixgbevf_write_mbx(hw, msg, size);
 
 	if (retval)
 		return retval;
 
-	return mbx->ops.read_posted(hw, retmsg, size);
+	return ixgbevf_poll_mbx(hw, retmsg, size);
 }
 
 /**
@@ -75,6 +74,9 @@ static s32 ixgbevf_reset_hw_vf(struct ixgbe_hw *hw)
 
 	/* reset the api version */
 	hw->api_version = ixgbe_mbox_api_10;
+	hw->mbx.ops.init_params(hw);
+	memcpy(&hw->mbx.ops, &ixgbevf_mbx_ops_legacy,
+	       sizeof(struct ixgbe_mbx_operations));
 
 	IXGBE_WRITE_REG(hw, IXGBE_VFCTRL, IXGBE_CTRL_RST);
 	IXGBE_WRITE_FLUSH(hw);
@@ -92,7 +94,7 @@ static s32 ixgbevf_reset_hw_vf(struct ixgbe_hw *hw)
 	mbx->timeout = IXGBE_VF_MBX_INIT_TIMEOUT;
 
 	msgbuf[0] = IXGBE_VF_RESET;
-	mbx->ops.write_posted(hw, msgbuf, 1);
+	ixgbevf_write_mbx(hw, msgbuf, 1);
 
 	mdelay(10);
 
@@ -100,7 +102,7 @@ static s32 ixgbevf_reset_hw_vf(struct ixgbe_hw *hw)
 	 * also set up the mc_filter_type which is piggy backed
 	 * on the mac address in word 3
 	 */
-	ret_val = mbx->ops.read_posted(hw, msgbuf, IXGBE_VF_PERMADDR_MSG_LEN);
+	ret_val = ixgbevf_poll_mbx(hw, msgbuf, IXGBE_VF_PERMADDR_MSG_LEN);
 	if (ret_val)
 		return ret_val;
 
@@ -108,11 +110,11 @@ static s32 ixgbevf_reset_hw_vf(struct ixgbe_hw *hw)
 	 * to indicate that no MAC address has yet been assigned for
 	 * the VF.
 	 */
-	if (msgbuf[0] != (IXGBE_VF_RESET | IXGBE_VT_MSGTYPE_ACK) &&
-	    msgbuf[0] != (IXGBE_VF_RESET | IXGBE_VT_MSGTYPE_NACK))
+	if (msgbuf[0] != (IXGBE_VF_RESET | IXGBE_VT_MSGTYPE_SUCCESS) &&
+	    msgbuf[0] != (IXGBE_VF_RESET | IXGBE_VT_MSGTYPE_FAILURE))
 		return IXGBE_ERR_INVALID_MAC_ADDR;
 
-	if (msgbuf[0] == (IXGBE_VF_RESET | IXGBE_VT_MSGTYPE_ACK))
+	if (msgbuf[0] == (IXGBE_VF_RESET | IXGBE_VT_MSGTYPE_SUCCESS))
 		ether_addr_copy(hw->mac.perm_addr, addr);
 
 	hw->mac.mc_filter_type = msgbuf[IXGBE_VF_MC_TYPE_WORD];
@@ -269,7 +271,7 @@ static s32 ixgbevf_set_uc_addr_vf(struct ixgbe_hw *hw, u32 index, u8 *addr)
 	if (!ret_val) {
 		msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
-		if (msgbuf[0] == (msgbuf_chk | IXGBE_VT_MSGTYPE_NACK))
+		if (msgbuf[0] == (msgbuf_chk | IXGBE_VT_MSGTYPE_FAILURE))
 			return -ENOMEM;
 	}
 
@@ -311,6 +313,7 @@ int ixgbevf_get_reta_locked(struct ixgbe_hw *hw, u32 *reta, int num_rx_queues)
 	 * is not supported for this device type.
 	 */
 	switch (hw->api_version) {
+	case ixgbe_mbox_api_15:
 	case ixgbe_mbox_api_14:
 	case ixgbe_mbox_api_13:
 	case ixgbe_mbox_api_12:
@@ -323,12 +326,12 @@ int ixgbevf_get_reta_locked(struct ixgbe_hw *hw, u32 *reta, int num_rx_queues)
 
 	msgbuf[0] = IXGBE_VF_GET_RETA;
 
-	err = hw->mbx.ops.write_posted(hw, msgbuf, 1);
+	err = ixgbevf_write_mbx(hw, msgbuf, 1);
 
 	if (err)
 		return err;
 
-	err = hw->mbx.ops.read_posted(hw, msgbuf, dwords + 1);
+	err = ixgbevf_poll_mbx(hw, msgbuf, dwords + 1);
 
 	if (err)
 		return err;
@@ -336,14 +339,14 @@ int ixgbevf_get_reta_locked(struct ixgbe_hw *hw, u32 *reta, int num_rx_queues)
 	msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
 	/* If the operation has been refused by a PF return -EPERM */
-	if (msgbuf[0] == (IXGBE_VF_GET_RETA | IXGBE_VT_MSGTYPE_NACK))
+	if (msgbuf[0] == (IXGBE_VF_GET_RETA | IXGBE_VT_MSGTYPE_FAILURE))
 		return -EPERM;
 
 	/* If we didn't get an ACK there must have been
 	 * some sort of mailbox error so we should treat it
 	 * as such.
 	 */
-	if (msgbuf[0] != (IXGBE_VF_GET_RETA | IXGBE_VT_MSGTYPE_ACK))
+	if (msgbuf[0] != (IXGBE_VF_GET_RETA | IXGBE_VT_MSGTYPE_SUCCESS))
 		return IXGBE_ERR_MBX;
 
 	/* ixgbevf doesn't support more than 2 queues at the moment */
@@ -379,6 +382,7 @@ int ixgbevf_get_rss_key_locked(struct ixgbe_hw *hw, u8 *rss_key)
 	 * or if the operation is not supported for this device type.
 	 */
 	switch (hw->api_version) {
+	case ixgbe_mbox_api_15:
 	case ixgbe_mbox_api_14:
 	case ixgbe_mbox_api_13:
 	case ixgbe_mbox_api_12:
@@ -390,12 +394,12 @@ int ixgbevf_get_rss_key_locked(struct ixgbe_hw *hw, u8 *rss_key)
 	}
 
 	msgbuf[0] = IXGBE_VF_GET_RSS_KEY;
-	err = hw->mbx.ops.write_posted(hw, msgbuf, 1);
+	err = ixgbevf_write_mbx(hw, msgbuf, 1);
 
 	if (err)
 		return err;
 
-	err = hw->mbx.ops.read_posted(hw, msgbuf, 11);
+	err = ixgbevf_poll_mbx(hw, msgbuf, 11);
 
 	if (err)
 		return err;
@@ -403,14 +407,14 @@ int ixgbevf_get_rss_key_locked(struct ixgbe_hw *hw, u8 *rss_key)
 	msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
 	/* If the operation has been refused by a PF return -EPERM */
-	if (msgbuf[0] == (IXGBE_VF_GET_RSS_KEY | IXGBE_VT_MSGTYPE_NACK))
+	if (msgbuf[0] == (IXGBE_VF_GET_RSS_KEY | IXGBE_VT_MSGTYPE_FAILURE))
 		return -EPERM;
 
 	/* If we didn't get an ACK there must have been
 	 * some sort of mailbox error so we should treat it
 	 * as such.
 	 */
-	if (msgbuf[0] != (IXGBE_VF_GET_RSS_KEY | IXGBE_VT_MSGTYPE_ACK))
+	if (msgbuf[0] != (IXGBE_VF_GET_RSS_KEY | IXGBE_VT_MSGTYPE_SUCCESS))
 		return IXGBE_ERR_MBX;
 
 	memcpy(rss_key, msgbuf + 1, IXGBEVF_RSS_HASH_KEY_SIZE);
@@ -442,7 +446,7 @@ static s32 ixgbevf_set_rar_vf(struct ixgbe_hw *hw, u32 index, u8 *addr,
 
 	/* if nacked the address was rejected, use "perm_addr" */
 	if (!ret_val &&
-	    (msgbuf[0] == (IXGBE_VF_SET_MAC_ADDR | IXGBE_VT_MSGTYPE_NACK))) {
+	    (msgbuf[0] == (IXGBE_VF_SET_MAC_ADDR | IXGBE_VT_MSGTYPE_FAILURE))) {
 		ixgbevf_get_mac_addr_vf(hw, hw->mac.addr);
 		return IXGBE_ERR_MBX;
 	}
@@ -545,8 +549,9 @@ static s32 ixgbevf_update_xcast_mode(struct ixgbe_hw *hw, int xcast_mode)
 		if (xcast_mode == IXGBEVF_XCAST_MODE_PROMISC)
 			return -EOPNOTSUPP;
 		fallthrough;
-	case ixgbe_mbox_api_14:
 	case ixgbe_mbox_api_13:
+	case ixgbe_mbox_api_14:
+	case ixgbe_mbox_api_15:
 		break;
 	default:
 		return -EOPNOTSUPP;
@@ -561,7 +566,7 @@ static s32 ixgbevf_update_xcast_mode(struct ixgbe_hw *hw, int xcast_mode)
 		return err;
 
 	msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
-	if (msgbuf[0] == (IXGBE_VF_UPDATE_XCAST_MODE | IXGBE_VT_MSGTYPE_NACK))
+	if (msgbuf[0] == (IXGBE_VF_UPDATE_XCAST_MODE | IXGBE_VT_MSGTYPE_FAILURE))
 		return -EPERM;
 
 	return 0;
@@ -606,7 +611,7 @@ static s32 ixgbevf_set_vfta_vf(struct ixgbe_hw *hw, u32 vlan, u32 vind,
 	msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 	msgbuf[0] &= ~(0xFF << IXGBE_VT_MSGINFO_SHIFT);
 
-	if (msgbuf[0] != (IXGBE_VF_SET_VLAN | IXGBE_VT_MSGTYPE_ACK))
+	if (msgbuf[0] != (IXGBE_VF_SET_VLAN | IXGBE_VT_MSGTYPE_SUCCESS))
 		err = IXGBE_ERR_INVALID_ARGUMENT;
 
 mbx_err:
@@ -705,12 +710,15 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw,
 	/* if the read failed it could just be a mailbox collision, best wait
 	 * until we are called again and don't report an error
 	 */
-	if (mbx->ops.read(hw, &in_msg, 1))
+	if (mbx->ops.read(hw, &in_msg, 1)) {
+		if (hw->api_version >= ixgbe_mbox_api_15)
+			mac->get_link_status = false;
 		goto out;
+	}
 
 	if (!(in_msg & IXGBE_VT_MSGTYPE_CTS)) {
 		/* msg is not CTS and is NACK we must have lost CTS status */
-		if (in_msg & IXGBE_VT_MSGTYPE_NACK)
+		if (in_msg & IXGBE_VT_MSGTYPE_FAILURE)
 			ret_val = -1;
 		goto out;
 	}
@@ -816,7 +824,7 @@ static s32 ixgbevf_set_rlpml_vf(struct ixgbe_hw *hw, u16 max_size)
 	if (ret_val)
 		return ret_val;
 	if ((msgbuf[0] & IXGBE_VF_SET_LPE) &&
-	    (msgbuf[0] & IXGBE_VT_MSGTYPE_NACK))
+	    (msgbuf[0] & IXGBE_VT_MSGTYPE_FAILURE))
 		return IXGBE_ERR_MBX;
 
 	return 0;
@@ -863,7 +871,8 @@ static int ixgbevf_negotiate_api_version_vf(struct ixgbe_hw *hw, int api)
 		msg[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
 		/* Store value and return 0 on success */
-		if (msg[0] == (IXGBE_VF_API_NEGOTIATE | IXGBE_VT_MSGTYPE_ACK)) {
+		if (msg[0] == (IXGBE_VF_API_NEGOTIATE |
+			      IXGBE_VT_MSGTYPE_SUCCESS)) {
 			hw->api_version = api;
 			return 0;
 		}
@@ -901,6 +910,7 @@ int ixgbevf_get_queues(struct ixgbe_hw *hw, unsigned int *num_tcs,
 	case ixgbe_mbox_api_12:
 	case ixgbe_mbox_api_13:
 	case ixgbe_mbox_api_14:
+	case ixgbe_mbox_api_15:
 		break;
 	default:
 		return 0;
@@ -918,7 +928,7 @@ int ixgbevf_get_queues(struct ixgbe_hw *hw, unsigned int *num_tcs,
 		 * some sort of mailbox error so we should treat it
 		 * as such
 		 */
-		if (msg[0] != (IXGBE_VF_GET_QUEUE | IXGBE_VT_MSGTYPE_ACK))
+		if (msg[0] != (IXGBE_VF_GET_QUEUE | IXGBE_VT_MSGTYPE_SUCCESS))
 			return IXGBE_ERR_MBX;
 
 		/* record and validate values from message */

drivers/net/ethernet/intel/ixgbevf/vf.h

@@ -73,10 +73,9 @@ struct ixgbe_mac_info {
 
 struct ixgbe_mbx_operations {
 	s32 (*init_params)(struct ixgbe_hw *hw);
+	void (*release)(struct ixgbe_hw *hw);
 	s32 (*read)(struct ixgbe_hw *, u32 *, u16);
 	s32 (*write)(struct ixgbe_hw *, u32 *, u16);
-	s32 (*read_posted)(struct ixgbe_hw *, u32 *, u16);
-	s32 (*write_posted)(struct ixgbe_hw *, u32 *, u16);
 	s32 (*check_for_msg)(struct ixgbe_hw *);
 	s32 (*check_for_ack)(struct ixgbe_hw *);
 	s32 (*check_for_rst)(struct ixgbe_hw *);
@@ -96,7 +95,7 @@ struct ixgbe_mbx_info {
 	struct ixgbe_mbx_stats stats;
 	u32 timeout;
 	u32 udelay;
-	u32 v2p_mailbox;
+	u32 vf_mailbox;
 	u16 size;
 };