ice: Add handler to configure SR-IOV

This patch implements parts of ice_sriov_configure and VF reset flow.

To create virtual functions (VFs), the user sets a value in num_vfs
through sysfs. This results in the kernel calling the handler for
.sriov_configure which is ice_sriov_configure.

VF setup first starts with a VF reset, followed by allocation of the VF
VSI using ice_vf_vsi_setup. Once the VF setup is complete a state bit
ICE_VF_STATE_INIT is set in the vf->states bitmap to indicate that
the VF is ready to go.

Also for VF reset to go into effect, it's necessary to issue a disable
queue command (ice_aqc_opc_dis_txqs). So this patch updates multiple
functions in the disable queue flow to take additional parameters that
distinguish if queues are being disabled due to VF reset.

Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

11 files changed, 1061 insertions, 9 deletions

diff --git a/drivers/net/ethernet/intel/ice/Makefile b/drivers/net/ethernet/intel/ice/Makefile
index 45125bd074d9..1999cd09239e 100644
--- a/drivers/net/ethernet/intel/ice/Makefile
+++ b/drivers/net/ethernet/intel/ice/Makefile
@@ -16,3 +16,4 @@ ice-y := ice_main.o	\
 	 ice_lib.o	\
 	 ice_txrx.o	\
 	 ice_ethtool.o
+ice-$(CONFIG_PCI_IOV) += ice_virtchnl_pf.o
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 639d45d1da49..f788cd63237a 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -28,6 +28,7 @@
 #include <linux/ip.h>
 #include <linux/ipv6.h>
 #include <linux/if_bridge.h>
+#include <linux/avf/virtchnl.h>
 #include <net/ipv6.h>
 #include "ice_devids.h"
 #include "ice_type.h"
@@ -35,6 +36,7 @@
 #include "ice_switch.h"
 #include "ice_common.h"
 #include "ice_sched.h"
+#include "ice_virtchnl_pf.h"
 
 extern const char ice_drv_ver[];
 #define ICE_BAR0		0
@@ -65,6 +67,12 @@ extern const char ice_drv_ver[];
 #define ICE_INVAL_Q_INDEX	0xffff
 #define ICE_INVAL_VFID		256
 #define ICE_MAX_VF_COUNT	256
+#define ICE_MAX_QS_PER_VF		256
+#define ICE_MIN_QS_PER_VF		1
+#define ICE_DFLT_QS_PER_VF		4
+#define ICE_MAX_INTR_PER_VF		65
+#define ICE_MIN_INTR_PER_VF		(ICE_MIN_QS_PER_VF + 1)
+#define ICE_DFLT_INTR_PER_VF		(ICE_DFLT_QS_PER_VF + 1)
 
 #define ICE_VSIQF_HKEY_ARRAY_SIZE	((VSIQF_HKEY_MAX_INDEX + 1) *	4)
 
@@ -135,10 +143,20 @@ enum ice_state {
 	__ICE_EMPR_RECV,		/* set by OICR handler */
 	__ICE_SUSPENDED,		/* set on module remove path */
 	__ICE_RESET_FAILED,		/* set by reset/rebuild */
+	/* When checking for the PF to be in a nominal operating state, the
+	 * bits that are grouped at the beginning of the list need to be
+	 * checked.  Bits occurring before __ICE_STATE_NOMINAL_CHECK_BITS will
+	 * be checked.  If you need to add a bit into consideration for nominal
+	 * operating state, it must be added before
+	 * __ICE_STATE_NOMINAL_CHECK_BITS.  Do not move this entry's position
+	 * without appropriate consideration.
+	 */
+	__ICE_STATE_NOMINAL_CHECK_BITS,
 	__ICE_ADMINQ_EVENT_PENDING,
 	__ICE_MAILBOXQ_EVENT_PENDING,
 	__ICE_MDD_EVENT_PENDING,
 	__ICE_FLTR_OVERFLOW_PROMISC,
+	__ICE_VF_DIS,
 	__ICE_CFG_BUSY,
 	__ICE_SERVICE_SCHED,
 	__ICE_SERVICE_DIS,
@@ -243,6 +261,7 @@ enum ice_pf_flags {
 	ICE_FLAG_MSIX_ENA,
 	ICE_FLAG_FLTR_SYNC,
 	ICE_FLAG_RSS_ENA,
+	ICE_FLAG_SRIOV_ENA,
 	ICE_FLAG_SRIOV_CAPABLE,
 	ICE_PF_FLAGS_NBITS		/* must be last */
 };
@@ -259,7 +278,12 @@ struct ice_pf {
 
 	struct ice_vsi **vsi;		/* VSIs created by the driver */
 	struct ice_sw *first_sw;	/* first switch created by firmware */
+	/* Virtchnl/SR-IOV config info */
+	struct ice_vf *vf;
+	int num_alloc_vfs;		/* actual number of VFs allocated */
 	u16 num_vfs_supported;		/* num VFs supported for this PF */
+	u16 num_vf_qps;			/* num queue pairs per VF */
+	u16 num_vf_msix;		/* num vectors per VF */
 	DECLARE_BITMAP(state, __ICE_STATE_NBITS);
 	DECLARE_BITMAP(avail_txqs, ICE_MAX_TXQS);
 	DECLARE_BITMAP(avail_rxqs, ICE_MAX_RXQS);
diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index 0fe054e4bfb8..c52f450f2c0d 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -2287,6 +2287,8 @@ ice_aq_add_lan_txq(struct ice_hw *hw, u8 num_qgrps,
  * @num_qgrps: number of groups in the list
  * @qg_list: the list of groups to disable
  * @buf_size: the total size of the qg_list buffer in bytes
+ * @rst_src: if called due to reset, specifies the RST source
+ * @vmvf_num: the relative VM or VF number that is undergoing the reset
  * @cd: pointer to command details structure or NULL
  *
  * Disable LAN Tx queue (0x0C31)
@@ -2294,6 +2296,7 @@ ice_aq_add_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 static enum ice_status
 ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 		   struct ice_aqc_dis_txq_item *qg_list, u16 buf_size,
+		   enum ice_disq_rst_src rst_src, u16 vmvf_num,
 		   struct ice_sq_cd *cd)
 {
 	struct ice_aqc_dis_txqs *cmd;
@@ -2303,14 +2306,45 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 	cmd = &desc.params.dis_txqs;
 	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_dis_txqs);
 
-	if (!qg_list)
+	/* qg_list can be NULL only in VM/VF reset flow */
+	if (!qg_list && !rst_src)
 		return ICE_ERR_PARAM;
 
 	if (num_qgrps > ICE_LAN_TXQ_MAX_QGRPS)
 		return ICE_ERR_PARAM;
-	desc.flags |= cpu_to_le16(ICE_AQ_FLAG_RD);
+
 	cmd->num_entries = num_qgrps;
 
+	cmd->vmvf_and_timeout = cpu_to_le16((5 << ICE_AQC_Q_DIS_TIMEOUT_S) &
+					    ICE_AQC_Q_DIS_TIMEOUT_M);
+
+	switch (rst_src) {
+	case ICE_VM_RESET:
+		cmd->cmd_type = ICE_AQC_Q_DIS_CMD_VM_RESET;
+		cmd->vmvf_and_timeout |=
+			cpu_to_le16(vmvf_num & ICE_AQC_Q_DIS_VMVF_NUM_M);
+		break;
+	case ICE_VF_RESET:
+		cmd->cmd_type = ICE_AQC_Q_DIS_CMD_VF_RESET;
+		/* In this case, FW expects vmvf_num to be absolute VF id */
+		cmd->vmvf_and_timeout |=
+			cpu_to_le16((vmvf_num + hw->func_caps.vf_base_id) &
+				    ICE_AQC_Q_DIS_VMVF_NUM_M);
+		break;
+	case ICE_NO_RESET:
+	default:
+		break;
+	}
+
+	/* If no queue group info, we are in a reset flow. Issue the AQ */
+	if (!qg_list)
+		goto do_aq;
+
+	/* set RD bit to indicate that command buffer is provided by the driver
+	 * and it needs to be read by the firmware
+	 */
+	desc.flags |= cpu_to_le16(ICE_AQ_FLAG_RD);
+
 	for (i = 0; i < num_qgrps; ++i) {
 		/* Calculate the size taken up by the queue IDs in this group */
 		sz += qg_list[i].num_qs * sizeof(qg_list[i].q_id);
@@ -2326,6 +2360,7 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 	if (buf_size != sz)
 		return ICE_ERR_PARAM;
 
+do_aq:
 	return ice_aq_send_cmd(hw, &desc, qg_list, buf_size, cd);
 }
 
@@ -2632,13 +2667,16 @@ ena_txq_exit:
  * @num_queues: number of queues
  * @q_ids: pointer to the q_id array
  * @q_teids: pointer to queue node teids
+ * @rst_src: if called due to reset, specifies the RST source
+ * @vmvf_num: the relative VM or VF number that is undergoing the reset
  * @cd: pointer to command details structure or NULL
  *
  * This function removes queues and their corresponding nodes in SW DB
  */
 enum ice_status
 ice_dis_vsi_txq(struct ice_port_info *pi, u8 num_queues, u16 *q_ids,
-		u32 *q_teids, struct ice_sq_cd *cd)
+		u32 *q_teids, enum ice_disq_rst_src rst_src, u16 vmvf_num,
+		struct ice_sq_cd *cd)
 {
 	enum ice_status status = ICE_ERR_DOES_NOT_EXIST;
 	struct ice_aqc_dis_txq_item qg_list;
@@ -2647,6 +2685,15 @@ ice_dis_vsi_txq(struct ice_port_info *pi, u8 num_queues, u16 *q_ids,
 	if (!pi || pi->port_state != ICE_SCHED_PORT_STATE_READY)
 		return ICE_ERR_CFG;
 
+	/* if queue is disabled already yet the disable queue command has to be
+	 * sent to complete the VF reset, then call ice_aq_dis_lan_txq without
+	 * any queue information
+	 */
+
+	if (!num_queues && rst_src)
+		return ice_aq_dis_lan_txq(pi->hw, 0, NULL, 0, rst_src, vmvf_num,
+					  NULL);
+
 	mutex_lock(&pi->sched_lock);
 
 	for (i = 0; i < num_queues; i++) {
@@ -2659,7 +2706,8 @@ ice_dis_vsi_txq(struct ice_port_info *pi, u8 num_queues, u16 *q_ids,
 		qg_list.num_qs = 1;
 		qg_list.q_id[0] = cpu_to_le16(q_ids[i]);
 		status = ice_aq_dis_lan_txq(pi->hw, 1, &qg_list,
-					    sizeof(qg_list), cd);
+					    sizeof(qg_list), rst_src, vmvf_num,
+					    cd);
 
 		if (status)
 			break;
diff --git a/drivers/net/ethernet/intel/ice/ice_common.h b/drivers/net/ethernet/intel/ice/ice_common.h
index 7b2a5bb2e550..1900681289a4 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.h
+++ b/drivers/net/ethernet/intel/ice/ice_common.h
@@ -7,6 +7,7 @@
 #include "ice.h"
 #include "ice_type.h"
 #include "ice_switch.h"
+#include <linux/avf/virtchnl.h>
 
 void ice_debug_cq(struct ice_hw *hw, u32 mask, void *desc, void *buf,
 		  u16 buf_len);
@@ -89,7 +90,8 @@ ice_aq_set_event_mask(struct ice_hw *hw, u8 port_num, u16 mask,
 		      struct ice_sq_cd *cd);
 enum ice_status
 ice_dis_vsi_txq(struct ice_port_info *pi, u8 num_queues, u16 *q_ids,
-		u32 *q_teids, struct ice_sq_cd *cmd_details);
+		u32 *q_teids, enum ice_disq_rst_src rst_src, u16 vmvf_num,
+		struct ice_sq_cd *cmd_details);
 enum ice_status
 ice_cfg_vsi_lan(struct ice_port_info *pi, u16 vsi_handle, u8 tc_bitmap,
 		u16 *max_lanqs);
diff --git a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
index c2d867b756ef..b676b3151d04 100644
--- a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
+++ b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
@@ -90,10 +90,16 @@
 #define GLGEN_RTRIG_CORER_M			BIT(0)
 #define GLGEN_RTRIG_GLOBR_M			BIT(1)
 #define GLGEN_STAT				0x000B612C
+#define GLGEN_VFLRSTAT(_i)			(0x00093A04 + ((_i) * 4))
 #define PFGEN_CTRL				0x00091000
 #define PFGEN_CTRL_PFSWR_M			BIT(0)
 #define PFGEN_STATE				0x00088000
 #define PRTGEN_STATUS				0x000B8100
+#define VFGEN_RSTAT(_VF)			(0x00074000 + ((_VF) * 4))
+#define VPGEN_VFRSTAT(_VF)			(0x00090800 + ((_VF) * 4))
+#define VPGEN_VFRSTAT_VFRD_M			BIT(0)
+#define VPGEN_VFRTRIG(_VF)			(0x00090000 + ((_VF) * 4))
+#define VPGEN_VFRTRIG_VFSWR_M			BIT(0)
 #define PFHMC_ERRORDATA				0x00520500
 #define PFHMC_ERRORINFO				0x00520400
 #define GLINT_DYN_CTL(_INT)			(0x00160000 + ((_INT) * 4))
@@ -106,6 +112,13 @@
 #define GLINT_ITR(_i, _INT)			(0x00154000 + ((_i) * 8192 + (_INT) * 4))
 #define GLINT_RATE(_INT)			(0x0015A000 + ((_INT) * 4))
 #define GLINT_RATE_INTRL_ENA_M			BIT(6)
+#define GLINT_VECT2FUNC(_INT)			(0x00162000 + ((_INT) * 4))
+#define GLINT_VECT2FUNC_VF_NUM_S		0
+#define GLINT_VECT2FUNC_VF_NUM_M		ICE_M(0xFF, 0)
+#define GLINT_VECT2FUNC_PF_NUM_S		12
+#define GLINT_VECT2FUNC_PF_NUM_M		ICE_M(0x7, 12)
+#define GLINT_VECT2FUNC_IS_PF_S			16
+#define GLINT_VECT2FUNC_IS_PF_M			BIT(16)
 #define PFINT_FW_CTL				0x0016C800
 #define PFINT_FW_CTL_MSIX_INDX_M		ICE_M(0x7FF, 0)
 #define PFINT_FW_CTL_ITR_INDX_S			11
@@ -137,6 +150,12 @@
 #define QINT_TQCTL_MSIX_INDX_S			0
 #define QINT_TQCTL_ITR_INDX_S			11
 #define QINT_TQCTL_CAUSE_ENA_M			BIT(30)
+#define VPINT_ALLOC(_VF)			(0x001D1000 + ((_VF) * 4))
+#define VPINT_ALLOC_FIRST_S			0
+#define VPINT_ALLOC_FIRST_M			ICE_M(0x7FF, 0)
+#define VPINT_ALLOC_LAST_S			12
+#define VPINT_ALLOC_LAST_M			ICE_M(0x7FF, 12)
+#define VPINT_ALLOC_VALID_M			BIT(31)
 #define QRX_CONTEXT(_i, _QRX)			(0x00280000 + ((_i) * 8192 + (_QRX) * 4))
 #define QRX_CTRL(_QRX)				(0x00120000 + ((_QRX) * 4))
 #define QRX_CTRL_MAX_INDEX			2047
@@ -149,6 +168,20 @@
 #define QRX_TAIL_MAX_INDEX			2047
 #define QRX_TAIL_TAIL_S				0
 #define QRX_TAIL_TAIL_M				ICE_M(0x1FFF, 0)
+#define VPLAN_RX_QBASE(_VF)			(0x00072000 + ((_VF) * 4))
+#define VPLAN_RX_QBASE_VFFIRSTQ_S		0
+#define VPLAN_RX_QBASE_VFFIRSTQ_M		ICE_M(0x7FF, 0)
+#define VPLAN_RX_QBASE_VFNUMQ_S			16
+#define VPLAN_RX_QBASE_VFNUMQ_M			ICE_M(0xFF, 16)
+#define VPLAN_RXQ_MAPENA(_VF)			(0x00073000 + ((_VF) * 4))
+#define VPLAN_RXQ_MAPENA_RX_ENA_M		BIT(0)
+#define VPLAN_TX_QBASE(_VF)			(0x001D1800 + ((_VF) * 4))
+#define VPLAN_TX_QBASE_VFFIRSTQ_S		0
+#define VPLAN_TX_QBASE_VFFIRSTQ_M		ICE_M(0x3FFF, 0)
+#define VPLAN_TX_QBASE_VFNUMQ_S			16
+#define VPLAN_TX_QBASE_VFNUMQ_M			ICE_M(0xFF, 16)
+#define VPLAN_TXQ_MAPENA(_VF)			(0x00073800 + ((_VF) * 4))
+#define VPLAN_TXQ_MAPENA_TX_ENA_M		BIT(0)
 #define GL_MDET_RX				0x00294C00
 #define GL_MDET_RX_QNUM_S			0
 #define GL_MDET_RX_QNUM_M			ICE_M(0x7FFF, 0)
@@ -196,6 +229,9 @@
 #define PF_FUNC_RID				0x0009E880
 #define PF_FUNC_RID_FUNC_NUM_S			0
 #define PF_FUNC_RID_FUNC_NUM_M			ICE_M(0x7, 0)
+#define PF_PCI_CIAA				0x0009E580
+#define PF_PCI_CIAA_VF_NUM_S			12
+#define PF_PCI_CIAD				0x0009E500
 #define GL_PWR_MODE_CTL				0x000B820C
 #define GL_PWR_MODE_CTL_CAR_MAX_BW_S		30
 #define GL_PWR_MODE_CTL_CAR_MAX_BW_M		ICE_M(0x3, 30)
@@ -276,5 +312,7 @@
 #define GLV_UPTCH(_i)				(0x0030A004 + ((_i) * 8))
 #define GLV_UPTCL(_i)				(0x0030A000 + ((_i) * 8))
 #define VSIQF_HKEY_MAX_INDEX			12
+#define VFINT_DYN_CTLN(_i)			(0x00003800 + ((_i) * 4))
+#define VFINT_DYN_CTLN_CLEARPBA_M		BIT(1)
 
 #endif /* _ICE_HW_AUTOGEN_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index acf3478a3f3b..4b26705a9ab5 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -1784,8 +1784,11 @@ int ice_vsi_stop_rx_rings(struct ice_vsi *vsi)
 /**
  * ice_vsi_stop_tx_rings - Disable Tx rings
  * @vsi: the VSI being configured
+ * @rst_src: reset source
+ * @rel_vmvf_num: Relative id of VF/VM
  */
-int ice_vsi_stop_tx_rings(struct ice_vsi *vsi)
+int ice_vsi_stop_tx_rings(struct ice_vsi *vsi, enum ice_disq_rst_src rst_src,
+			  u16 rel_vmvf_num)
 {
 	struct ice_pf *pf = vsi->back;
 	struct ice_hw *hw = &pf->hw;
@@ -1837,7 +1840,7 @@ int ice_vsi_stop_tx_rings(struct ice_vsi *vsi)
 		     GLINT_DYN_CTL_SWINT_TRIG_M | GLINT_DYN_CTL_INTENA_MSK_M);
 	}
 	status = ice_dis_vsi_txq(vsi->port_info, vsi->num_txq, q_ids, q_teids,
-				 NULL);
+				 rst_src, rel_vmvf_num, NULL);
 	/* if the disable queue command was exercised during an active reset
 	 * flow, ICE_ERR_RESET_ONGOING is returned. This is not an error as
 	 * the reset operation disables queues at the hardware level anyway.
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index 2617afe01c82..677db40338f5 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -31,7 +31,8 @@ int ice_vsi_start_rx_rings(struct ice_vsi *vsi);
 
 int ice_vsi_stop_rx_rings(struct ice_vsi *vsi);
 
-int ice_vsi_stop_tx_rings(struct ice_vsi *vsi);
+int ice_vsi_stop_tx_rings(struct ice_vsi *vsi, enum ice_disq_rst_src rst_src,
+			  u16 rel_vmvf_num);
 
 int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena);
 
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 3fd3bb783707..5b8c950d219a 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -2185,6 +2185,8 @@ static void ice_remove(struct pci_dev *pdev)
 	set_bit(__ICE_DOWN, pf->state);
 	ice_service_task_stop(pf);
 
+	if (test_bit(ICE_FLAG_SRIOV_ENA, pf->flags))
+		ice_free_vfs(pf);
 	ice_vsi_release_all(pf);
 	ice_free_irq_msix_misc(pf);
 	ice_for_each_vsi(pf, i) {
@@ -2220,6 +2222,7 @@ static struct pci_driver ice_driver = {
 	.id_table = ice_pci_tbl,
 	.probe = ice_probe,
 	.remove = ice_remove,
+	.sriov_configure = ice_sriov_configure,
 };
 
 /**
@@ -2955,7 +2958,7 @@ int ice_down(struct ice_vsi *vsi)
 	}
 
 	ice_vsi_dis_irq(vsi);
-	tx_err = ice_vsi_stop_tx_rings(vsi);
+	tx_err = ice_vsi_stop_tx_rings(vsi, ICE_NO_RESET, 0);
 	if (tx_err)
 		netdev_err(vsi->netdev,
 			   "Failed stop Tx rings, VSI %d error %d\n",
@@ -3357,6 +3360,7 @@ static void ice_rebuild(struct ice_pf *pf)
 		goto err_vsi_rebuild;
 	}
 
+	ice_reset_all_vfs(pf, true);
 	/* if we get here, reset flow is successful */
 	clear_bit(__ICE_RESET_FAILED, pf->state);
 	return;
diff --git a/drivers/net/ethernet/intel/ice/ice_type.h b/drivers/net/ethernet/intel/ice/ice_type.h
index 6d053fb5f941..15b3c999006a 100644
--- a/drivers/net/ethernet/intel/ice/ice_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_type.h
@@ -104,6 +104,15 @@ struct ice_link_status {
 	u8 module_type[ICE_MODULE_TYPE_TOTAL_BYTE];
 };
 
+/* Different reset sources for which a disable queue AQ call has to be made in
+ * order to clean the TX scheduler as a part of the reset
+ */
+enum ice_disq_rst_src {
+	ICE_NO_RESET = 0,
+	ICE_VM_RESET,
+	ICE_VF_RESET,
+};
+
 /* PHY info such as phy_type, etc... */
 struct ice_phy_info {
 	struct ice_link_status link_info;
@@ -130,6 +139,7 @@ struct ice_hw_common_caps {
 
 	/* Virtualization support */
 	u8 sr_iov_1_1;			/* SR-IOV enabled */
+
 	/* RSS related capabilities */
 	u16 rss_table_size;		/* 512 for PFs and 64 for VFs */
 	u8 rss_table_entry_width;	/* RSS Entry width in bits */
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
new file mode 100644
index 000000000000..7f041fd785d6
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -0,0 +1,847 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2018, Intel Corporation. */
+
+#include "ice.h"
+#include "ice_lib.h"
+
+/**
+ * ice_get_vf_vector - get VF interrupt vector register offset
+ * @vf_msix: number of MSIx vector per VF on a PF
+ * @vf_id: VF identifier
+ * @i: index of MSIx vector
+ */
+static u32 ice_get_vf_vector(int vf_msix, int vf_id, int i)
+{
+	return ((i == 0) ? VFINT_DYN_CTLN(vf_id) :
+		 VFINT_DYN_CTLN(((vf_msix - 1) * (vf_id)) + (i - 1)));
+}
+
+/**
+ * ice_free_vf_res - Free a VF's resources
+ * @vf: pointer to the VF info
+ */
+static void ice_free_vf_res(struct ice_vf *vf)
+{
+	struct ice_pf *pf = vf->pf;
+	int i, pf_vf_msix;
+
+	/* First, disable VF's configuration API to prevent OS from
+	 * accessing the VF's VSI after it's freed or invalidated.
+	 */
+	clear_bit(ICE_VF_STATE_INIT, vf->vf_states);
+
+	/* free vsi & disconnect it from the parent uplink */
+	if (vf->lan_vsi_idx) {
+		ice_vsi_release(pf->vsi[vf->lan_vsi_idx]);
+		vf->lan_vsi_idx = 0;
+		vf->lan_vsi_num = 0;
+		vf->num_mac = 0;
+	}
+
+	pf_vf_msix = pf->num_vf_msix;
+	/* Disable interrupts so that VF starts in a known state */
+	for (i = 0; i < pf_vf_msix; i++) {
+		u32 reg_idx;
+
+		reg_idx = ice_get_vf_vector(pf_vf_msix, vf->vf_id, i);
+		wr32(&pf->hw, reg_idx, VFINT_DYN_CTLN_CLEARPBA_M);
+		ice_flush(&pf->hw);
+	}
+	/* reset some of the state variables keeping track of the resources */
+	clear_bit(ICE_VF_STATE_MC_PROMISC, vf->vf_states);
+	clear_bit(ICE_VF_STATE_UC_PROMISC, vf->vf_states);
+}
+
+/***********************enable_vf routines*****************************/
+
+/**
+ * ice_dis_vf_mappings
+ * @vf: pointer to the VF structure
+ */
+static void ice_dis_vf_mappings(struct ice_vf *vf)
+{
+	struct ice_pf *pf = vf->pf;
+	struct ice_vsi *vsi;
+	int first, last, v;
+	struct ice_hw *hw;
+
+	hw = &pf->hw;
+	vsi = pf->vsi[vf->lan_vsi_idx];
+
+	wr32(hw, VPINT_ALLOC(vf->vf_id), 0);
+
+	first = vf->first_vector_idx;
+	last = first + pf->num_vf_msix - 1;
+	for (v = first; v <= last; v++) {
+		u32 reg;
+
+		reg = (((1 << GLINT_VECT2FUNC_IS_PF_S) &
+			GLINT_VECT2FUNC_IS_PF_M) |
+		       ((hw->pf_id << GLINT_VECT2FUNC_PF_NUM_S) &
+			GLINT_VECT2FUNC_PF_NUM_M));
+		wr32(hw, GLINT_VECT2FUNC(v), reg);
+	}
+
+	if (vsi->tx_mapping_mode == ICE_VSI_MAP_CONTIG)
+		wr32(hw, VPLAN_TX_QBASE(vf->vf_id), 0);
+	else
+		dev_err(&pf->pdev->dev,
+			"Scattered mode for VF Tx queues is not yet implemented\n");
+
+	if (vsi->rx_mapping_mode == ICE_VSI_MAP_CONTIG)
+		wr32(hw, VPLAN_RX_QBASE(vf->vf_id), 0);
+	else
+		dev_err(&pf->pdev->dev,
+			"Scattered mode for VF Rx queues is not yet implemented\n");
+}
+
+/**
+ * ice_free_vfs - Free all VFs
+ * @pf: pointer to the PF structure
+ */
+void ice_free_vfs(struct ice_pf *pf)
+{
+	struct ice_hw *hw = &pf->hw;
+	int tmp, i;
+
+	if (!pf->vf)
+		return;
+
+	while (test_and_set_bit(__ICE_VF_DIS, pf->state))
+		usleep_range(1000, 2000);
+
+	/* Avoid wait time by stopping all VFs at the same time */
+	for (i = 0; i < pf->num_alloc_vfs; i++) {
+		if (!test_bit(ICE_VF_STATE_ENA, pf->vf[i].vf_states))
+			continue;
+
+		/* stop rings without wait time */
+		ice_vsi_stop_tx_rings(pf->vsi[pf->vf[i].lan_vsi_idx],
+				      ICE_NO_RESET, i);
+		ice_vsi_stop_rx_rings(pf->vsi[pf->vf[i].lan_vsi_idx]);
+
+		clear_bit(ICE_VF_STATE_ENA, pf->vf[i].vf_states);
+	}
+
+	/* Disable IOV before freeing resources. This lets any VF drivers
+	 * running in the host get themselves cleaned up before we yank
+	 * the carpet out from underneath their feet.
+	 */
+	if (!pci_vfs_assigned(pf->pdev))
+		pci_disable_sriov(pf->pdev);
+	else
+		dev_warn(&pf->pdev->dev, "VFs are assigned - not disabling SR-IOV\n");
+
+	tmp = pf->num_alloc_vfs;
+	pf->num_vf_qps = 0;
+	pf->num_alloc_vfs = 0;
+	for (i = 0; i < tmp; i++) {
+		if (test_bit(ICE_VF_STATE_INIT, pf->vf[i].vf_states)) {
+			/* disable VF qp mappings */
+			ice_dis_vf_mappings(&pf->vf[i]);
+
+			/* Set this state so that assigned VF vectors can be
+			 * reclaimed by PF for reuse in ice_vsi_release(). No
+			 * need to clear this bit since pf->vf array is being
+			 * freed anyways after this for loop
+			 */
+			set_bit(ICE_VF_STATE_CFG_INTR, pf->vf[i].vf_states);
+			ice_free_vf_res(&pf->vf[i]);
+		}
+	}
+
+	devm_kfree(&pf->pdev->dev, pf->vf);
+	pf->vf = NULL;
+
+	/* This check is for when the driver is unloaded while VFs are
+	 * assigned. Setting the number of VFs to 0 through sysfs is caught
+	 * before this function ever gets called.
+	 */
+	if (!pci_vfs_assigned(pf->pdev)) {
+		int vf_id;
+
+		/* Acknowledge VFLR for all VFs. Without this, VFs will fail to
+		 * work correctly when SR-IOV gets re-enabled.
+		 */
+		for (vf_id = 0; vf_id < tmp; vf_id++) {
+			u32 reg_idx, bit_idx;
+
+			reg_idx = (hw->func_caps.vf_base_id + vf_id) / 32;
+			bit_idx = (hw->func_caps.vf_base_id + vf_id) % 32;
+			wr32(hw, GLGEN_VFLRSTAT(reg_idx), BIT(bit_idx));
+		}
+	}
+	clear_bit(__ICE_VF_DIS, pf->state);
+	clear_bit(ICE_FLAG_SRIOV_ENA, pf->flags);
+}
+
+/**
+ * ice_trigger_vf_reset - Reset a VF on HW
+ * @vf: pointer to the VF structure
+ * @is_vflr: true if VFLR was issued, false if not
+ *
+ * Trigger hardware to start a reset for a particular VF. Expects the caller
+ * to wait the proper amount of time to allow hardware to reset the VF before
+ * it cleans up and restores VF functionality.
+ */
+static void ice_trigger_vf_reset(struct ice_vf *vf, bool is_vflr)
+{
+	struct ice_pf *pf = vf->pf;
+	u32 reg, reg_idx, bit_idx;
+	struct ice_hw *hw;
+	int vf_abs_id, i;
+
+	hw = &pf->hw;
+	vf_abs_id = vf->vf_id + hw->func_caps.vf_base_id;
+
+	/* Inform VF that it is no longer active, as a warning */
+	clear_bit(ICE_VF_STATE_ACTIVE, vf->vf_states);
+
+	/* Disable VF's configuration API during reset. The flag is re-enabled
+	 * in ice_alloc_vf_res(), when it's safe again to access VF's VSI.
+	 * It's normally disabled in ice_free_vf_res(), but it's safer
+	 * to do it earlier to give some time to finish to any VF config
+	 * functions that may still be running at this point.
+	 */
+	clear_bit(ICE_VF_STATE_INIT, vf->vf_states);
+
+	/* In the case of a VFLR, the HW has already reset the VF and we
+	 * just need to clean up, so don't hit the VFRTRIG register.
+	 */
+	if (!is_vflr) {
+		/* reset VF using VPGEN_VFRTRIG reg */
+		reg = rd32(hw, VPGEN_VFRTRIG(vf->vf_id));
+		reg |= VPGEN_VFRTRIG_VFSWR_M;
+		wr32(hw, VPGEN_VFRTRIG(vf->vf_id), reg);
+	}
+	/* clear the VFLR bit in GLGEN_VFLRSTAT */
+	reg_idx = (vf_abs_id) / 32;
+	bit_idx = (vf_abs_id) % 32;
+	wr32(hw, GLGEN_VFLRSTAT(reg_idx), BIT(bit_idx));
+	ice_flush(hw);
+
+	wr32(hw, PF_PCI_CIAA,
+	     VF_DEVICE_STATUS | (vf_abs_id << PF_PCI_CIAA_VF_NUM_S));
+	for (i = 0; i < 100; i++) {
+		reg = rd32(hw, PF_PCI_CIAD);
+		if ((reg & VF_TRANS_PENDING_M) != 0)
+			dev_err(&pf->pdev->dev,
+				"VF %d PCI transactions stuck\n", vf->vf_id);
+		udelay(1);
+	}
+}
+
+/**
+ * ice_vsi_set_pvid - Set port VLAN id for the VSI
+ * @vsi: the VSI being changed
+ * @vid: the VLAN id to set as a PVID
+ */
+static int ice_vsi_set_pvid(struct ice_vsi *vsi, u16 vid)
+{
+	struct device *dev = &vsi->back->pdev->dev;
+	struct ice_hw *hw = &vsi->back->hw;
+	struct ice_vsi_ctx ctxt = { 0 };
+	enum ice_status status;
+
+	ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_MODE_TAGGED |
+			       ICE_AQ_VSI_PVLAN_INSERT_PVID |
+			       ICE_AQ_VSI_VLAN_EMOD_STR;
+	ctxt.info.pvid = cpu_to_le16(vid);
+	ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
+
+	status = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+	if (status) {
+		dev_info(dev, "update VSI for VLAN insert failed, err %d aq_err %d\n",
+			 status, hw->adminq.sq_last_status);
+		return -EIO;
+	}
+
+	vsi->info.pvid = ctxt.info.pvid;
+	vsi->info.vlan_flags = ctxt.info.vlan_flags;
+	return 0;
+}
+
+/**
+ * ice_vf_vsi_setup - Set up a VF VSI
+ * @pf: board private structure
+ * @pi: pointer to the port_info instance
+ * @vf_id: defines VF id to which this VSI connects.
+ *
+ * Returns pointer to the successfully allocated VSI struct on success,
+ * otherwise returns NULL on failure.
+ */
+static struct ice_vsi *
+ice_vf_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi, u16 vf_id)
+{
+	return ice_vsi_setup(pf, pi, ICE_VSI_VF, vf_id);
+}
+
+/**
+ * ice_alloc_vsi_res - Setup VF VSI and its resources
+ * @vf: pointer to the VF structure
+ *
+ * Returns 0 on success, negative value on failure
+ */
+static int ice_alloc_vsi_res(struct ice_vf *vf)
+{
+	struct ice_pf *pf = vf->pf;
+	LIST_HEAD(tmp_add_list);
+	u8 broadcast[ETH_ALEN];
+	struct ice_vsi *vsi;
+	int status = 0;
+
+	vsi = ice_vf_vsi_setup(pf, pf->hw.port_info, vf->vf_id);
+
+	if (!vsi) {
+		dev_err(&pf->pdev->dev, "Failed to create VF VSI\n");
+		return -ENOMEM;
+	}
+
+	vf->lan_vsi_idx = vsi->idx;
+	vf->lan_vsi_num = vsi->vsi_num;
+
+	/* first vector index is the VFs OICR index */
+	vf->first_vector_idx = vsi->hw_base_vector;
+	/* Since hw_base_vector holds the vector where data queue interrupts
+	 * starts, increment by 1 since VFs allocated vectors include OICR intr
+	 * as well.
+	 */
+	vsi->hw_base_vector += 1;
+
+	/* Check if port VLAN exist before, and restore it accordingly */
+	if (vf->port_vlan_id)
+		ice_vsi_set_pvid(vsi, vf->port_vlan_id);
+
+	eth_broadcast_addr(broadcast);
+
+	status = ice_add_mac_to_list(vsi, &tmp_add_list, broadcast);
+	if (status)
+		goto ice_alloc_vsi_res_exit;
+
+	if (is_valid_ether_addr(vf->dflt_lan_addr.addr)) {
+		status = ice_add_mac_to_list(vsi, &tmp_add_list,
+					     vf->dflt_lan_addr.addr);
+		if (status)
+			goto ice_alloc_vsi_res_exit;
+	}
+
+	status = ice_add_mac(&pf->hw, &tmp_add_list);
+	if (status)
+		dev_err(&pf->pdev->dev, "could not add mac filters\n");
+
+	/* Clear this bit after VF initialization since we shouldn't reclaim
+	 * and reassign interrupts for synchronous or asynchronous VFR events.
+	 * We don't want to reconfigure interrupts since AVF driver doesn't
+	 * expect vector assignment to be changed unless there is a request for
+	 * more vectors.
+	 */
+	clear_bit(ICE_VF_STATE_CFG_INTR, vf->vf_states);
+ice_alloc_vsi_res_exit:
+	ice_free_fltr_list(&pf->pdev->dev, &tmp_add_list);
+	return status;
+}
+
+/**
+ * ice_alloc_vf_res - Allocate VF resources
+ * @vf: pointer to the VF structure
+ */
+static int ice_alloc_vf_res(struct ice_vf *vf)
+{
+	int status;
+
+	/* setup VF VSI and necessary resources */
+	status = ice_alloc_vsi_res(vf);
+	if (status)
+		goto ice_alloc_vf_res_exit;
+
+	if (vf->trusted)
+		set_bit(ICE_VIRTCHNL_VF_CAP_PRIVILEGE, &vf->vf_caps);
+	else
+		clear_bit(ICE_VIRTCHNL_VF_CAP_PRIVILEGE, &vf->vf_caps);
+
+	/* VF is now completely initialized */
+	set_bit(ICE_VF_STATE_INIT, vf->vf_states);
+
+	return status;
+
+ice_alloc_vf_res_exit:
+	ice_free_vf_res(vf);
+	return status;
+}
+
+/**
+ * ice_ena_vf_mappings
+ * @vf: pointer to the VF structure
+ *
+ * Enable VF vectors and queues allocation by writing the details into
+ * respective registers.
+ */
+static void ice_ena_vf_mappings(struct ice_vf *vf)
+{
+	struct ice_pf *pf = vf->pf;
+	struct ice_vsi *vsi;
+	int first, last, v;
+	struct ice_hw *hw;
+	int abs_vf_id;
+	u32 reg;
+
+	hw = &pf->hw;
+	vsi = pf->vsi[vf->lan_vsi_idx];
+	first = vf->first_vector_idx;
+	last = (first + pf->num_vf_msix) - 1;
+	abs_vf_id = vf->vf_id + hw->func_caps.vf_base_id;
+
+	/* VF Vector allocation */
+	reg = (((first << VPINT_ALLOC_FIRST_S) & VPINT_ALLOC_FIRST_M) |
+	       ((last << VPINT_ALLOC_LAST_S) & VPINT_ALLOC_LAST_M) |
+	       VPINT_ALLOC_VALID_M);
+	wr32(hw, VPINT_ALLOC(vf->vf_id), reg);
+
+	/* map the interrupts to its functions */
+	for (v = first; v <= last; v++) {
+		reg = (((abs_vf_id << GLINT_VECT2FUNC_VF_NUM_S) &
+			GLINT_VECT2FUNC_VF_NUM_M) |
+		       ((hw->pf_id << GLINT_VECT2FUNC_PF_NUM_S) &
+			GLINT_VECT2FUNC_PF_NUM_M));
+		wr32(hw, GLINT_VECT2FUNC(v), reg);
+	}
+
+	/* VF Tx queues allocation */
+	if (vsi->tx_mapping_mode == ICE_VSI_MAP_CONTIG) {
+		wr32(hw, VPLAN_TXQ_MAPENA(vf->vf_id),
+		     VPLAN_TXQ_MAPENA_TX_ENA_M);
+		/* set the VF PF Tx queue range
+		 * VFNUMQ value should be set to (number of queues - 1). A value
+		 * of 0 means 1 queue and a value of 255 means 256 queues
+		 */
+		reg = (((vsi->txq_map[0] << VPLAN_TX_QBASE_VFFIRSTQ_S) &
+			VPLAN_TX_QBASE_VFFIRSTQ_M) |
+		       (((vsi->alloc_txq - 1) << VPLAN_TX_QBASE_VFNUMQ_S) &
+			VPLAN_TX_QBASE_VFNUMQ_M));
+		wr32(hw, VPLAN_TX_QBASE(vf->vf_id), reg);
+	} else {
+		dev_err(&pf->pdev->dev,
+			"Scattered mode for VF Tx queues is not yet implemented\n");
+	}
+
+	/* VF Rx queues allocation */
+	if (vsi->rx_mapping_mode == ICE_VSI_MAP_CONTIG) {
+		wr32(hw, VPLAN_RXQ_MAPENA(vf->vf_id),
+		     VPLAN_RXQ_MAPENA_RX_ENA_M);
+		/* set the VF PF Rx queue range
+		 * VFNUMQ value should be set to (number of queues - 1). A value
+		 * of 0 means 1 queue and a value of 255 means 256 queues
+		 */
+		reg = (((vsi->rxq_map[0] << VPLAN_RX_QBASE_VFFIRSTQ_S) &
+			VPLAN_RX_QBASE_VFFIRSTQ_M) |
+		       (((vsi->alloc_txq - 1) << VPLAN_RX_QBASE_VFNUMQ_S) &
+			VPLAN_RX_QBASE_VFNUMQ_M));
+		wr32(hw, VPLAN_RX_QBASE(vf->vf_id), reg);
+	} else {
+		dev_err(&pf->pdev->dev,
+			"Scattered mode for VF Rx queues is not yet implemented\n");
+	}
+}
+
+/**
+ * ice_determine_res
+ * @pf: pointer to the PF structure
+ * @avail_res: available resources in the PF structure
+ * @max_res: maximum resources that can be given per VF
+ * @min_res: minimum resources that can be given per VF
+ *
+ * Returns non-zero value if resources (queues/vectors) are available or
+ * returns zero if PF cannot accommodate for all num_alloc_vfs.
+ */
+static int
+ice_determine_res(struct ice_pf *pf, u16 avail_res, u16 max_res, u16 min_res)
+{
+	bool checked_min_res = false;
+	int res;
+
+	/* start by checking if PF can assign max number of resources for
+	 * all num_alloc_vfs.
+	 * if yes, return number per VF
+	 * If no, divide by 2 and roundup, check again
+	 * repeat the loop till we reach a point where even minimum resources
+	 * are not available, in that case return 0
+	 */
+	res = max_res;
+	while ((res >= min_res) && !checked_min_res) {
+		int num_all_res;
+
+		num_all_res = pf->num_alloc_vfs * res;
+		if (num_all_res <= avail_res)
+			return res;
+
+		if (res == min_res)
+			checked_min_res = true;
+
+		res = DIV_ROUND_UP(res, 2);
+	}
+	return 0;
+}
+
+/**
+ * ice_check_avail_res - check if vectors and queues are available
+ * @pf: pointer to the PF structure
+ *
+ * This function is where we calculate actual number of resources for VF VSIs,
+ * we don't reserve ahead of time during probe. Returns success if vectors and
+ * queues resources are available, otherwise returns error code
+ */
+static int ice_check_avail_res(struct ice_pf *pf)
+{
+	u16 num_msix, num_txq, num_rxq;
+
+	if (!pf->num_alloc_vfs)
+		return -EINVAL;
+
+	/* Grab from HW interrupts common pool
+	 * Note: By the time the user decides it needs more vectors in a VF
+	 * its already too late since one must decide this prior to creating the
+	 * VF interface. So the best we can do is take a guess as to what the
+	 * user might want.
+	 *
+	 * We have two policies for vector allocation:
+	 * 1. if num_alloc_vfs is from 1 to 16, then we consider this as small
+	 * number of NFV VFs used for NFV appliances, since this is a special
+	 * case, we try to assign maximum vectors per VF (65) as much as
+	 * possible, based on determine_resources algorithm.
+	 * 2. if num_alloc_vfs is from 17 to 256, then its large number of
+	 * regular VFs which are not used for any special purpose. Hence try to
+	 * grab default interrupt vectors (5 as supported by AVF driver).
+	 */
+	if (pf->num_alloc_vfs <= 16) {
+		num_msix = ice_determine_res(pf, pf->num_avail_hw_msix,
+					     ICE_MAX_INTR_PER_VF,
+					     ICE_MIN_INTR_PER_VF);
+	} else if (pf->num_alloc_vfs <= ICE_MAX_VF_COUNT) {
+		num_msix = ice_determine_res(pf, pf->num_avail_hw_msix,
+					     ICE_DFLT_INTR_PER_VF,
+					     ICE_MIN_INTR_PER_VF);
+	} else {
+		dev_err(&pf->pdev->dev,
+			"Number of VFs %d exceeds max VF count %d\n",
+			pf->num_alloc_vfs, ICE_MAX_VF_COUNT);
+		return -EIO;
+	}
+
+	if (!num_msix)
+		return -EIO;
+
+	/* Grab from the common pool
+	 * start by requesting Default queues (4 as supported by AVF driver),
+	 * Note that, the main difference between queues and vectors is, latter
+	 * can only be reserved at init time but queues can be requested by VF
+	 * at runtime through Virtchnl, that is the reason we start by reserving
+	 * few queues.
+	 */
+	num_txq = ice_determine_res(pf, pf->q_left_tx, ICE_DFLT_QS_PER_VF,
+				    ICE_MIN_QS_PER_VF);
+
+	num_rxq = ice_determine_res(pf, pf->q_left_rx, ICE_DFLT_QS_PER_VF,
+				    ICE_MIN_QS_PER_VF);
+
+	if (!num_txq || !num_rxq)
+		return -EIO;
+
+	/* since AVF driver works with only queue pairs which means, it expects
+	 * to have equal number of Rx and Tx queues, so take the minimum of
+	 * available Tx or Rx queues
+	 */
+	pf->num_vf_qps = min_t(int, num_txq, num_rxq);
+	pf->num_vf_msix = num_msix;
+
+	return 0;
+}
+
+/**
+ * ice_cleanup_and_realloc_vf - Clean up VF and reallocate resources after reset
+ * @vf: pointer to the VF structure
+ *
+ * Cleanup a VF after the hardware reset is finished. Expects the caller to
+ * have verified whether the reset is finished properly, and ensure the
+ * minimum amount of wait time has passed. Reallocate VF resources back to make
+ * VF state active
+ */
+static void ice_cleanup_and_realloc_vf(struct ice_vf *vf)
+{
+	struct ice_pf *pf = vf->pf;
+	struct ice_hw *hw;
+	u32 reg;
+
+	hw = &pf->hw;
+
+	/* PF software completes the flow by notifying VF that reset flow is
+	 * completed. This is done by enabling hardware by clearing the reset
+	 * bit in the VPGEN_VFRTRIG reg and setting VFR_STATE in the VFGEN_RSTAT
+	 * register to VFR completed (done at the end of this function)
+	 * By doing this we allow HW to access VF memory at any point. If we
+	 * did it any sooner, HW could access memory while it was being freed
+	 * in ice_free_vf_res(), causing an IOMMU fault.
+	 *
+	 * On the other hand, this needs to be done ASAP, because the VF driver
+	 * is waiting for this to happen and may report a timeout. It's
+	 * harmless, but it gets logged into Guest OS kernel log, so best avoid
+	 * it.
+	 */
+	reg = rd32(hw, VPGEN_VFRTRIG(vf->vf_id));
+	reg &= ~VPGEN_VFRTRIG_VFSWR_M;
+	wr32(hw, VPGEN_VFRTRIG(vf->vf_id), reg);
+
+	/* reallocate VF resources to finish resetting the VSI state */
+	if (!ice_alloc_vf_res(vf)) {
+		ice_ena_vf_mappings(vf);
+		set_bit(ICE_VF_STATE_ACTIVE, vf->vf_states);
+		clear_bit(ICE_VF_STATE_DIS, vf->vf_states);
+		vf->num_vlan = 0;
+	}
+
+	/* Tell the VF driver the reset is done. This needs to be done only
+	 * after VF has been fully initialized, because the VF driver may
+	 * request resources immediately after setting this flag.
+	 */
+	wr32(hw, VFGEN_RSTAT(vf->vf_id), VIRTCHNL_VFR_VFACTIVE);
+}
+
+/**
+ * ice_reset_all_vfs - reset all allocated VFs in one go
+ * @pf: pointer to the PF structure
+ * @is_vflr: true if VFLR was issued, false if not
+ *
+ * First, tell the hardware to reset each VF, then do all the waiting in one
+ * chunk, and finally finish restoring each VF after the wait. This is useful
+ * during PF routines which need to reset all VFs, as otherwise it must perform
+ * these resets in a serialized fashion.
+ *
+ * Returns true if any VFs were reset, and false otherwise.
+ */
+bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
+{
+	struct ice_hw *hw = &pf->hw;
+	int v, i;
+
+	/* If we don't have any VFs, then there is nothing to reset */
+	if (!pf->num_alloc_vfs)
+		return false;
+
+	/* If VFs have been disabled, there is no need to reset */
+	if (test_and_set_bit(__ICE_VF_DIS, pf->state))
+		return false;
+
+	/* Begin reset on all VFs at once */
+	for (v = 0; v < pf->num_alloc_vfs; v++)
+		ice_trigger_vf_reset(&pf->vf[v], is_vflr);
+
+	/* Call Disable LAN Tx queue AQ call with VFR bit set and 0
+	 * queues to inform Firmware about VF reset.
+	 */
+	for (v = 0; v < pf->num_alloc_vfs; v++)
+		ice_dis_vsi_txq(pf->vsi[0]->port_info, 0, NULL, NULL,
+				ICE_VF_RESET, v, NULL);
+
+	/* HW requires some time to make sure it can flush the FIFO for a VF
+	 * when it resets it. Poll the VPGEN_VFRSTAT register for each VF in
+	 * sequence to make sure that it has completed. We'll keep track of
+	 * the VFs using a simple iterator that increments once that VF has
+	 * finished resetting.
+	 */
+	for (i = 0, v = 0; i < 10 && v < pf->num_alloc_vfs; i++) {
+		usleep_range(10000, 20000);
+
+		/* Check each VF in sequence */
+		while (v < pf->num_alloc_vfs) {
+			struct ice_vf *vf = &pf->vf[v];
+			u32 reg;
+
+			reg = rd32(hw, VPGEN_VFRSTAT(vf->vf_id));
+			if (!(reg & VPGEN_VFRSTAT_VFRD_M))
+				break;
+
+			/* If the current VF has finished resetting, move on
+			 * to the next VF in sequence.
+			 */
+			v++;
+		}
+	}
+
+	/* Display a warning if at least one VF didn't manage to reset in
+	 * time, but continue on with the operation.
+	 */
+	if (v < pf->num_alloc_vfs)
+		dev_warn(&pf->pdev->dev, "VF reset check timeout\n");
+	usleep_range(10000, 20000);
+
+	/* free VF resources to begin resetting the VSI state */
+	for (v = 0; v < pf->num_alloc_vfs; v++)
+		ice_free_vf_res(&pf->vf[v]);
+
+	if (ice_check_avail_res(pf)) {
+		dev_err(&pf->pdev->dev,
+			"Cannot allocate VF resources, try with fewer number of VFs\n");
+		return false;
+	}
+
+	/* Finish the reset on each VF */
+	for (v = 0; v < pf->num_alloc_vfs; v++)
+		ice_cleanup_and_realloc_vf(&pf->vf[v]);
+
+	ice_flush(hw);
+	clear_bit(__ICE_VF_DIS, pf->state);
+
+	return true;
+}
+
+/**
+ * ice_alloc_vfs - Allocate and set up VFs resources
+ * @pf: pointer to the PF structure
+ * @num_alloc_vfs: number of VFs to allocate
+ */
+static int ice_alloc_vfs(struct ice_pf *pf, u16 num_alloc_vfs)
+{
+	struct ice_hw *hw = &pf->hw;
+	struct ice_vf *vfs;
+	int i, ret;
+
+	/* Disable global interrupt 0 so we don't try to handle the VFLR. */
+	wr32(hw, GLINT_DYN_CTL(pf->hw_oicr_idx),
+	     ICE_ITR_NONE << GLINT_DYN_CTL_ITR_INDX_S);
+
+	ice_flush(hw);
+
+	ret = pci_enable_sriov(pf->pdev, num_alloc_vfs);
+	if (ret) {
+		pf->num_alloc_vfs = 0;
+		goto err_unroll_intr;
+	}
+	/* allocate memory */
+	vfs = devm_kcalloc(&pf->pdev->dev, num_alloc_vfs, sizeof(*vfs),
+			   GFP_KERNEL);
+	if (!vfs) {
+		ret = -ENOMEM;
+		goto err_unroll_sriov;
+	}
+	pf->vf = vfs;
+
+	/* apply default profile */
+	for (i = 0; i < num_alloc_vfs; i++) {
+		vfs[i].pf = pf;
+		vfs[i].vf_sw_id = pf->first_sw;
+		vfs[i].vf_id = i;
+
+		/* assign default capabilities */
+		set_bit(ICE_VIRTCHNL_VF_CAP_L2, &vfs[i].vf_caps);
+		vfs[i].spoofchk = true;
+
+		/* Set this state so that PF driver does VF vector assignment */
+		set_bit(ICE_VF_STATE_CFG_INTR, vfs[i].vf_states);
+	}
+	pf->num_alloc_vfs = num_alloc_vfs;
+
+	/* VF resources get allocated during reset */
+	if (!ice_reset_all_vfs(pf, false))
+		goto err_unroll_sriov;
+
+	goto err_unroll_intr;
+
+err_unroll_sriov:
+	pci_disable_sriov(pf->pdev);
+err_unroll_intr:
+	/* rearm interrupts here */
+	ice_irq_dynamic_ena(hw, NULL, NULL);
+	return ret;
+}
+
+/**
+ * ice_pf_state_is_nominal - checks the pf for nominal state
+ * @pf: pointer to pf to check
+ *
+ * Check the PF's state for a collection of bits that would indicate
+ * the PF is in a state that would inhibit normal operation for
+ * driver functionality.
+ *
+ * Returns true if PF is in a nominal state.
+ * Returns false otherwise
+ */
+static bool ice_pf_state_is_nominal(struct ice_pf *pf)
+{
+	DECLARE_BITMAP(check_bits, __ICE_STATE_NBITS) = { 0 };
+
+	if (!pf)
+		return false;
+
+	bitmap_set(check_bits, 0, __ICE_STATE_NOMINAL_CHECK_BITS);
+	if (bitmap_intersects(pf->state, check_bits, __ICE_STATE_NBITS))
+		return false;
+
+	return true;
+}
+
+/**
+ * ice_pci_sriov_ena - Enable or change number of VFs
+ * @pf: pointer to the PF structure
+ * @num_vfs: number of VFs to allocate
+ */
+static int ice_pci_sriov_ena(struct ice_pf *pf, int num_vfs)
+{
+	int pre_existing_vfs = pci_num_vf(pf->pdev);
+	struct device *dev = &pf->pdev->dev;
+	int err;
+
+	if (!ice_pf_state_is_nominal(pf)) {
+		dev_err(dev, "Cannot enable SR-IOV, device not ready\n");
+		return -EBUSY;
+	}
+
+	if (!test_bit(ICE_FLAG_SRIOV_CAPABLE, pf->flags)) {
+		dev_err(dev, "This device is not capable of SR-IOV\n");
+		return -ENODEV;
+	}
+
+	if (pre_existing_vfs && pre_existing_vfs != num_vfs)
+		ice_free_vfs(pf);
+	else if (pre_existing_vfs && pre_existing_vfs == num_vfs)
+		return num_vfs;
+
+	if (num_vfs > pf->num_vfs_supported) {
+		dev_err(dev, "Can't enable %d VFs, max VFs supported is %d\n",
+			num_vfs, pf->num_vfs_supported);
+		return -ENOTSUPP;
+	}
+
+	dev_info(dev, "Allocating %d VFs\n", num_vfs);
+	err = ice_alloc_vfs(pf, num_vfs);
+	if (err) {
+		dev_err(dev, "Failed to enable SR-IOV: %d\n", err);
+		return err;
+	}
+
+	set_bit(ICE_FLAG_SRIOV_ENA, pf->flags);
+	return num_vfs;
+}
+
+/**
+ * ice_sriov_configure - Enable or change number of VFs via sysfs
+ * @pdev: pointer to a pci_dev structure
+ * @num_vfs: number of VFs to allocate
+ *
+ * This function is called when the user updates the number of VFs in sysfs.
+ */
+int ice_sriov_configure(struct pci_dev *pdev, int num_vfs)
+{
+	struct ice_pf *pf = pci_get_drvdata(pdev);
+
+	if (num_vfs)
+		return ice_pci_sriov_ena(pf, num_vfs);
+
+	if (!pci_vfs_assigned(pdev)) {
+		ice_free_vfs(pf);
+	} else {
+		dev_err(&pf->pdev->dev,
+			"can't free VFs because some are assigned to VMs.\n");
+		return -EBUSY;
+	}
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
new file mode 100644
index 000000000000..85c263a7494c
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
@@ -0,0 +1,74 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2018, Intel Corporation. */
+
+#ifndef _ICE_VIRTCHNL_PF_H_
+#define _ICE_VIRTCHNL_PF_H_
+#include "ice.h"
+
+/* Static VF transaction/status register def */
+#define VF_DEVICE_STATUS		0xAA
+#define VF_TRANS_PENDING_M		0x20
+
+/* Specific VF states */
+enum ice_vf_states {
+	ICE_VF_STATE_INIT = 0,
+	ICE_VF_STATE_ACTIVE,
+	ICE_VF_STATE_ENA,
+	ICE_VF_STATE_DIS,
+	ICE_VF_STATE_MC_PROMISC,
+	ICE_VF_STATE_UC_PROMISC,
+	/* state to indicate if PF needs to do vector assignment for VF.
+	 * This needs to be set during first time VF initialization or later
+	 * when VF asks for more Vectors through virtchnl OP.
+	 */
+	ICE_VF_STATE_CFG_INTR,
+	ICE_VF_STATES_NBITS
+};
+
+/* VF capabilities */
+enum ice_virtchnl_cap {
+	ICE_VIRTCHNL_VF_CAP_L2 = 0,
+	ICE_VIRTCHNL_VF_CAP_PRIVILEGE,
+};
+
+/* VF information structure */
+struct ice_vf {
+	struct ice_pf *pf;
+
+	s16 vf_id;			/* VF id in the PF space */
+	int first_vector_idx;		/* first vector index of this VF */
+	struct ice_sw *vf_sw_id;	/* switch id the VF VSIs connect to */
+	struct virtchnl_ether_addr dflt_lan_addr;
+	u16 port_vlan_id;
+	u8 trusted;
+	u16 lan_vsi_idx;		/* index into PF struct */
+	u16 lan_vsi_num;		/* ID as used by firmware */
+	unsigned long vf_caps;		/* vf's adv. capabilities */
+	DECLARE_BITMAP(vf_states, ICE_VF_STATES_NBITS);	/* VF runtime states */
+	u8 spoofchk;
+	u16 num_mac;
+	u16 num_vlan;
+};
+
+#ifdef CONFIG_PCI_IOV
+int ice_sriov_configure(struct pci_dev *pdev, int num_vfs);
+void ice_free_vfs(struct ice_pf *pf);
+bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr);
+#else /* CONFIG_PCI_IOV */
+#define ice_free_vfs(pf) do {} while (0)
+
+static inline bool
+ice_reset_all_vfs(struct ice_pf __always_unused *pf,
+		  bool __always_unused is_vflr)
+{
+	return true;
+}
+
+static inline int
+ice_sriov_configure(struct pci_dev __always_unused *pdev,
+		    int __always_unused num_vfs)
+{
+	return -EOPNOTSUPP;
+}
+#endif /* CONFIG_PCI_IOV */
+#endif /* _ICE_VIRTCHNL_PF_H_ */