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

Tony Nguyen says:

====================
iwl-next Intel Wired LAN Driver Updates 2021-06-01

This pull request is targeting net-next and rdma-next branches.
These patches have been reviewed by netdev and rdma mailing lists[1].

This series adds RDMA support to the ice driver for E810 devices and
converts the i40e driver to use the auxiliary bus infrastructure
for X722 devices. The PCI netdev drivers register auxiliary RDMA devices
that will bind to auxiliary drivers registered by the new irdma module.

[1] https://lore.kernel.org/netdev/20210520143809.819-1-shiraz.saleem@intel.com/
---
v3:
- ice_aq_add_rdma_qsets(), ice_cfg_vsi_rdma(), ice_[ena|dis]_vsi_rdma_qset(),
and ice_cfg_rdma_fltr() no longer return ice_status
- Remove null check from ice_aq_add_rdma_qsets()

v2:
- Added patch 'i40e: Replace one-element array with flexible-array
member'

Changes since linked review (v6):
- Removed unnecessary checks in i40e_client_device_register() and
i40e_client_device_unregister()
- Simplified the i40e_client_device_register() API
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

MAINTAINERS

@@ -9137,6 +9137,7 @@ F:	Documentation/networking/device_drivers/ethernet/intel/
 F:	drivers/net/ethernet/intel/
 F:	drivers/net/ethernet/intel/*/
 F:	include/linux/avf/virtchnl.h
+F:	include/linux/net/intel/iidc.h
 
 INTEL FRAMEBUFFER DRIVER (excluding 810 and 815)
 M:	Maik Broemme <mbroemme@libmpq.org>

drivers/infiniband/hw/i40iw/i40iw_main.c

@@ -1423,7 +1423,7 @@ static enum i40iw_status_code i40iw_save_msix_info(struct i40iw_device *iwdev,
 	struct i40e_qv_info *iw_qvinfo;
 	u32 ceq_idx;
 	u32 i;
-	u32 size;
+	size_t size;
 
 	if (!ldev->msix_count) {
 		i40iw_pr_err("No MSI-X vectors\n");
@@ -1433,8 +1433,7 @@ static enum i40iw_status_code i40iw_save_msix_info(struct i40iw_device *iwdev,
 	iwdev->msix_count = ldev->msix_count;
 
 	size = sizeof(struct i40iw_msix_vector) * iwdev->msix_count;
-	size += sizeof(struct i40e_qvlist_info);
-	size +=  sizeof(struct i40e_qv_info) * iwdev->msix_count - 1;
+	size += struct_size(iw_qvlist, qv_info, iwdev->msix_count);
 	iwdev->iw_msixtbl = kzalloc(size, GFP_KERNEL);
 
 	if (!iwdev->iw_msixtbl)

drivers/net/ethernet/intel/Kconfig

@@ -241,6 +241,7 @@ config I40E
 	tristate "Intel(R) Ethernet Controller XL710 Family support"
 	imply PTP_1588_CLOCK
 	depends on PCI
+	select AUXILIARY_BUS
 	help
 	  This driver supports Intel(R) Ethernet Controller XL710 Family of
 	  devices.  For more information on how to identify your adapter, go
@@ -294,6 +295,7 @@ config ICE
 	tristate "Intel(R) Ethernet Connection E800 Series Support"
 	default n
 	depends on PCI_MSI
+	select AUXILIARY_BUS
 	select DIMLIB
 	select NET_DEVLINK
 	select PLDMFW

drivers/net/ethernet/intel/i40e/i40e.h

@@ -870,6 +870,8 @@ struct i40e_netdev_priv {
 	struct i40e_vsi *vsi;
 };
 
+extern struct ida i40e_client_ida;
+
 /* struct that defines an interrupt vector */
 struct i40e_q_vector {
 	struct i40e_vsi *vsi;

drivers/net/ethernet/intel/i40e/i40e_client.c

@@ -12,6 +12,7 @@ static const char i40e_client_interface_version_str[] = I40E_CLIENT_VERSION_STR;
 static struct i40e_client *registered_client;
 static LIST_HEAD(i40e_devices);
 static DEFINE_MUTEX(i40e_device_mutex);
+DEFINE_IDA(i40e_client_ida);
 
 static int i40e_client_virtchnl_send(struct i40e_info *ldev,
 				     struct i40e_client *client,
@@ -275,6 +276,57 @@ void i40e_client_update_msix_info(struct i40e_pf *pf)
 	cdev->lan_info.msix_entries = &pf->msix_entries[pf->iwarp_base_vector];
 }
 
+static void i40e_auxiliary_dev_release(struct device *dev)
+{
+	struct i40e_auxiliary_device *i40e_aux_dev =
+			container_of(dev, struct i40e_auxiliary_device, aux_dev.dev);
+
+	ida_free(&i40e_client_ida, i40e_aux_dev->aux_dev.id);
+	kfree(i40e_aux_dev);
+}
+
+static int i40e_register_auxiliary_dev(struct i40e_info *ldev, const char *name)
+{
+	struct i40e_auxiliary_device *i40e_aux_dev;
+	struct pci_dev *pdev = ldev->pcidev;
+	struct auxiliary_device *aux_dev;
+	int ret;
+
+	i40e_aux_dev = kzalloc(sizeof(*i40e_aux_dev), GFP_KERNEL);
+	if (!i40e_aux_dev)
+		return -ENOMEM;
+
+	i40e_aux_dev->ldev = ldev;
+
+	aux_dev = &i40e_aux_dev->aux_dev;
+	aux_dev->name = name;
+	aux_dev->dev.parent = &pdev->dev;
+	aux_dev->dev.release = i40e_auxiliary_dev_release;
+	ldev->aux_dev = aux_dev;
+
+	ret = ida_alloc(&i40e_client_ida, GFP_KERNEL);
+	if (ret < 0) {
+		kfree(i40e_aux_dev);
+		return ret;
+	}
+	aux_dev->id = ret;
+
+	ret = auxiliary_device_init(aux_dev);
+	if (ret < 0) {
+		ida_free(&i40e_client_ida, aux_dev->id);
+		kfree(i40e_aux_dev);
+		return ret;
+	}
+
+	ret = auxiliary_device_add(aux_dev);
+	if (ret) {
+		auxiliary_device_uninit(aux_dev);
+		return ret;
+	}
+
+	return ret;
+}
+
 /**
  * i40e_client_add_instance - add a client instance struct to the instance list
  * @pf: pointer to the board struct
@@ -286,9 +338,6 @@ static void i40e_client_add_instance(struct i40e_pf *pf)
 	struct netdev_hw_addr *mac = NULL;
 	struct i40e_vsi *vsi = pf->vsi[pf->lan_vsi];
 
-	if (!registered_client || pf->cinst)
-		return;
-
 	cdev = kzalloc(sizeof(*cdev), GFP_KERNEL);
 	if (!cdev)
 		return;
@@ -308,11 +357,8 @@ static void i40e_client_add_instance(struct i40e_pf *pf)
 	cdev->lan_info.fw_build = pf->hw.aq.fw_build;
 	set_bit(__I40E_CLIENT_INSTANCE_NONE, &cdev->state);
 
-	if (i40e_client_get_params(vsi, &cdev->lan_info.params)) {
-		kfree(cdev);
-		cdev = NULL;
-		return;
-	}
+	if (i40e_client_get_params(vsi, &cdev->lan_info.params))
+		goto free_cdev;
 
 	mac = list_first_entry(&cdev->lan_info.netdev->dev_addrs.list,
 			       struct netdev_hw_addr, list);
@@ -324,7 +370,17 @@ static void i40e_client_add_instance(struct i40e_pf *pf)
 	cdev->client = registered_client;
 	pf->cinst = cdev;
 
-	i40e_client_update_msix_info(pf);
+	cdev->lan_info.msix_count = pf->num_iwarp_msix;
+	cdev->lan_info.msix_entries = &pf->msix_entries[pf->iwarp_base_vector];
+
+	if (i40e_register_auxiliary_dev(&cdev->lan_info, "iwarp"))
+		goto free_cdev;
+
+	return;
+
+free_cdev:
+	kfree(cdev);
+	pf->cinst = NULL;
 }
 
 /**
@@ -345,7 +401,7 @@ void i40e_client_del_instance(struct i40e_pf *pf)
  **/
 void i40e_client_subtask(struct i40e_pf *pf)
 {
-	struct i40e_client *client = registered_client;
+	struct i40e_client *client;
 	struct i40e_client_instance *cdev;
 	struct i40e_vsi *vsi = pf->vsi[pf->lan_vsi];
 	int ret = 0;
@@ -359,9 +415,11 @@ void i40e_client_subtask(struct i40e_pf *pf)
 	    test_bit(__I40E_CONFIG_BUSY, pf->state))
 		return;
 
-	if (!client || !cdev)
+	if (!cdev || !cdev->client)
 		return;
 
+	client = cdev->client;
+
 	/* Here we handle client opens. If the client is down, and
 	 * the netdev is registered, then open the client.
 	 */
@@ -423,16 +481,8 @@ int i40e_lan_add_device(struct i40e_pf *pf)
 		 pf->hw.pf_id, pf->hw.bus.bus_id,
 		 pf->hw.bus.device, pf->hw.bus.func);
 
-	/* If a client has already been registered, we need to add an instance
-	 * of it to our new LAN device.
-	 */
-	if (registered_client)
-		i40e_client_add_instance(pf);
+	i40e_client_add_instance(pf);
 
-	/* Since in some cases register may have happened before a device gets
-	 * added, we can schedule a subtask to go initiate the clients if
-	 * they can be launched at probe time.
-	 */
 	set_bit(__I40E_CLIENT_SERVICE_REQUESTED, pf->state);
 	i40e_service_event_schedule(pf);
 
@@ -449,9 +499,13 @@ int i40e_lan_add_device(struct i40e_pf *pf)
  **/
 int i40e_lan_del_device(struct i40e_pf *pf)
 {
+	struct auxiliary_device *aux_dev = pf->cinst->lan_info.aux_dev;
 	struct i40e_device *ldev, *tmp;
 	int ret = -ENODEV;
 
+	auxiliary_device_delete(aux_dev);
+	auxiliary_device_uninit(aux_dev);
+
 	/* First, remove any client instance. */
 	i40e_client_del_instance(pf);
 
@@ -579,7 +633,7 @@ static int i40e_client_setup_qvlist(struct i40e_info *ldev,
 	u32 v_idx, i, reg_idx, reg;
 
 	ldev->qvlist_info = kzalloc(struct_size(ldev->qvlist_info, qv_info,
-				    qvlist_info->num_vectors - 1), GFP_KERNEL);
+				    qvlist_info->num_vectors), GFP_KERNEL);
 	if (!ldev->qvlist_info)
 		return -ENOMEM;
 	ldev->qvlist_info->num_vectors = qvlist_info->num_vectors;
@@ -732,6 +786,42 @@ static int i40e_client_update_vsi_ctxt(struct i40e_info *ldev,
 	return err;
 }
 
+void i40e_client_device_register(struct i40e_info *ldev, struct i40e_client *client)
+{
+	struct i40e_pf *pf = ldev->pf;
+
+	pf->cinst->client = client;
+	set_bit(__I40E_CLIENT_SERVICE_REQUESTED, pf->state);
+	i40e_service_event_schedule(pf);
+}
+EXPORT_SYMBOL_GPL(i40e_client_device_register);
+
+void i40e_client_device_unregister(struct i40e_info *ldev)
+{
+	struct i40e_pf *pf = ldev->pf;
+	struct i40e_client_instance *cdev = pf->cinst;
+
+	if (!cdev)
+		return;
+
+	while (test_and_set_bit(__I40E_SERVICE_SCHED, pf->state))
+		usleep_range(500, 1000);
+
+	if (test_bit(__I40E_CLIENT_INSTANCE_OPENED, &cdev->state)) {
+		cdev->client->ops->close(&cdev->lan_info, cdev->client, false);
+		clear_bit(__I40E_CLIENT_INSTANCE_OPENED, &cdev->state);
+		i40e_client_release_qvlist(&cdev->lan_info);
+	}
+
+	pf->cinst->client = NULL;
+	clear_bit(__I40E_SERVICE_SCHED, pf->state);
+}
+EXPORT_SYMBOL_GPL(i40e_client_device_unregister);
+
+/* Retain these legacy global registration/unregistration calls till i40iw is
+ * removed from the kernel. The irdma unified driver does not use these
+ * exported symbols.
+ */
 /**
  * i40e_register_client - Register a i40e client driver with the L2 driver
  * @client: pointer to the i40e_client struct

drivers/net/ethernet/intel/i40e/i40e_main.c

@@ -16270,6 +16270,7 @@ static void __exit i40e_exit_module(void)
 {
 	pci_unregister_driver(&i40e_driver);
 	destroy_workqueue(i40e_wq);
+	ida_destroy(&i40e_client_ida);
 	i40e_dbg_exit();
 }
 module_exit(i40e_exit_module);

drivers/net/ethernet/intel/ice/Makefile

@@ -22,6 +22,7 @@ ice-y := ice_main.o	\
 	 ice_ethtool_fdir.o \
 	 ice_flex_pipe.o \
 	 ice_flow.o	\
+	 ice_idc.o	\
 	 ice_devlink.o	\
 	 ice_fw_update.o \
 	 ice_lag.o	\

drivers/net/ethernet/intel/ice/ice.h

@@ -34,6 +34,7 @@
 #include <linux/if_bridge.h>
 #include <linux/ctype.h>
 #include <linux/bpf.h>
+#include <linux/auxiliary_bus.h>
 #include <linux/avf/virtchnl.h>
 #include <linux/cpu_rmap.h>
 #include <linux/dim.h>
@@ -55,6 +56,7 @@
 #include "ice_switch.h"
 #include "ice_common.h"
 #include "ice_sched.h"
+#include "ice_idc_int.h"
 #include "ice_virtchnl_pf.h"
 #include "ice_sriov.h"
 #include "ice_fdir.h"
@@ -78,6 +80,8 @@
 #define ICE_MIN_LAN_OICR_MSIX	1
 #define ICE_MIN_MSIX		(ICE_MIN_LAN_TXRX_MSIX + ICE_MIN_LAN_OICR_MSIX)
 #define ICE_FDIR_MSIX		2
+#define ICE_RDMA_NUM_AEQ_MSIX	4
+#define ICE_MIN_RDMA_MSIX	2
 #define ICE_NO_VSI		0xffff
 #define ICE_VSI_MAP_CONTIG	0
 #define ICE_VSI_MAP_SCATTER	1
@@ -88,8 +92,9 @@
 #define ICE_MAX_LG_RSS_QS	256
 #define ICE_RES_VALID_BIT	0x8000
 #define ICE_RES_MISC_VEC_ID	(ICE_RES_VALID_BIT - 1)
+#define ICE_RES_RDMA_VEC_ID	(ICE_RES_MISC_VEC_ID - 1)
 /* All VF control VSIs share the same IRQ, so assign a unique ID for them */
-#define ICE_RES_VF_CTRL_VEC_ID	(ICE_RES_MISC_VEC_ID - 1)
+#define ICE_RES_VF_CTRL_VEC_ID	(ICE_RES_RDMA_VEC_ID - 1)
 #define ICE_INVAL_Q_INDEX	0xffff
 #define ICE_INVAL_VFID		256
 
@@ -203,9 +208,9 @@ enum ice_pf_state {
 	ICE_NEEDS_RESTART,
 	ICE_PREPARED_FOR_RESET,	/* set by driver when prepared */
 	ICE_RESET_OICR_RECV,		/* set by driver after rcv reset OICR */
-	ICE_PFR_REQ,			/* set by driver and peers */
-	ICE_CORER_REQ,		/* set by driver and peers */
-	ICE_GLOBR_REQ,		/* set by driver and peers */
+	ICE_PFR_REQ,		/* set by driver */
+	ICE_CORER_REQ,		/* set by driver */
+	ICE_GLOBR_REQ,		/* set by driver */
 	ICE_CORER_RECV,		/* set by OICR handler */
 	ICE_GLOBR_RECV,		/* set by OICR handler */
 	ICE_EMPR_RECV,		/* set by OICR handler */
@@ -332,6 +337,7 @@ struct ice_vsi {
 	u16 req_rxq;			 /* User requested Rx queues */
 	u16 num_rx_desc;
 	u16 num_tx_desc;
+	u16 qset_handle[ICE_MAX_TRAFFIC_CLASS];
 	struct ice_tc_cfg tc_cfg;
 	struct bpf_prog *xdp_prog;
 	struct ice_ring **xdp_rings;	 /* XDP ring array */
@@ -373,12 +379,14 @@ struct ice_q_vector {
 
 enum ice_pf_flags {
 	ICE_FLAG_FLTR_SYNC,
+	ICE_FLAG_RDMA_ENA,
 	ICE_FLAG_RSS_ENA,
 	ICE_FLAG_SRIOV_ENA,
 	ICE_FLAG_SRIOV_CAPABLE,
 	ICE_FLAG_DCB_CAPABLE,
 	ICE_FLAG_DCB_ENA,
 	ICE_FLAG_FD_ENA,
+	ICE_FLAG_AUX_ENA,
 	ICE_FLAG_ADV_FEATURES,
 	ICE_FLAG_LINK_DOWN_ON_CLOSE_ENA,
 	ICE_FLAG_TOTAL_PORT_SHUTDOWN_ENA,
@@ -439,6 +447,8 @@ struct ice_pf {
 	struct mutex sw_mutex;		/* lock for protecting VSI alloc flow */
 	struct mutex tc_mutex;		/* lock to protect TC changes */
 	u32 msg_enable;
+	u16 num_rdma_msix;		/* Total MSIX vectors for RDMA driver */
+	u16 rdma_base_vector;
 
 	/* spinlock to protect the AdminQ wait list */
 	spinlock_t aq_wait_lock;
@@ -471,6 +481,8 @@ struct ice_pf {
 	unsigned long tx_timeout_last_recovery;
 	u32 tx_timeout_recovery_level;
 	char int_name[ICE_INT_NAME_STR_LEN];
+	struct auxiliary_device *adev;
+	int aux_idx;
 	u32 sw_int_count;
 
 	__le64 nvm_phy_type_lo; /* NVM PHY type low */
@@ -636,6 +648,9 @@ int ice_get_rss_key(struct ice_vsi *vsi, u8 *seed);
 void ice_fill_rss_lut(u8 *lut, u16 rss_table_size, u16 rss_size);
 int ice_schedule_reset(struct ice_pf *pf, enum ice_reset_req reset);
 void ice_print_link_msg(struct ice_vsi *vsi, bool isup);
+int ice_plug_aux_dev(struct ice_pf *pf);
+void ice_unplug_aux_dev(struct ice_pf *pf);
+int ice_init_rdma(struct ice_pf *pf);
 const char *ice_stat_str(enum ice_status stat_err);
 const char *ice_aq_str(enum ice_aq_err aq_err);
 bool ice_is_wol_supported(struct ice_hw *hw);
@@ -660,4 +675,25 @@ int ice_open_internal(struct net_device *netdev);
 int ice_stop(struct net_device *netdev);
 void ice_service_task_schedule(struct ice_pf *pf);
 
+/**
+ * ice_set_rdma_cap - enable RDMA support
+ * @pf: PF struct
+ */
+static inline void ice_set_rdma_cap(struct ice_pf *pf)
+{
+	if (pf->hw.func_caps.common_cap.rdma && pf->num_rdma_msix) {
+		set_bit(ICE_FLAG_RDMA_ENA, pf->flags);
+		ice_plug_aux_dev(pf);
+	}
+}
+
+/**
+ * ice_clear_rdma_cap - disable RDMA support
+ * @pf: PF struct
+ */
+static inline void ice_clear_rdma_cap(struct ice_pf *pf)
+{
+	ice_unplug_aux_dev(pf);
+	clear_bit(ICE_FLAG_RDMA_ENA, pf->flags);
+}
 #endif /* _ICE_H_ */

drivers/net/ethernet/intel/ice/ice_adminq_cmd.h

@@ -115,6 +115,7 @@ struct ice_aqc_list_caps_elem {
 #define ICE_AQC_CAPS_PENDING_OROM_VER			0x004B
 #define ICE_AQC_CAPS_NET_VER				0x004C
 #define ICE_AQC_CAPS_PENDING_NET_VER			0x004D
+#define ICE_AQC_CAPS_RDMA				0x0051
 #define ICE_AQC_CAPS_NVM_MGMT				0x0080
 
 	u8 major_ver;
@@ -1684,6 +1685,36 @@ struct ice_aqc_dis_txq_item {
 	__le16 q_id[];
 } __packed;
 
+/* Add Tx RDMA Queue Set (indirect 0x0C33) */
+struct ice_aqc_add_rdma_qset {
+	u8 num_qset_grps;
+	u8 reserved[7];
+	__le32 addr_high;
+	__le32 addr_low;
+};
+
+/* This is the descriptor of each Qset entry for the Add Tx RDMA Queue Set
+ * command (0x0C33). Only used within struct ice_aqc_add_rdma_qset.
+ */
+struct ice_aqc_add_tx_rdma_qset_entry {
+	__le16 tx_qset_id;
+	u8 rsvd[2];
+	__le32 qset_teid;
+	struct ice_aqc_txsched_elem info;
+};
+
+/* The format of the command buffer for Add Tx RDMA Queue Set(0x0C33)
+ * is an array of the following structs. Please note that the length of
+ * each struct ice_aqc_add_rdma_qset is variable due to the variable
+ * number of queues in each group!
+ */
+struct ice_aqc_add_rdma_qset_data {
+	__le32 parent_teid;
+	__le16 num_qsets;
+	u8 rsvd[2];
+	struct ice_aqc_add_tx_rdma_qset_entry rdma_qsets[];
+};
+
 /* Configure Firmware Logging Command (indirect 0xFF09)
  * Logging Information Read Response (indirect 0xFF10)
  * Note: The 0xFF10 command has no input parameters.
@@ -1880,6 +1911,7 @@ struct ice_aq_desc {
 		struct ice_aqc_get_set_rss_key get_set_rss_key;
 		struct ice_aqc_add_txqs add_txqs;
 		struct ice_aqc_dis_txqs dis_txqs;
+		struct ice_aqc_add_rdma_qset add_rdma_qset;
 		struct ice_aqc_add_get_update_free_vsi vsi_cmd;
 		struct ice_aqc_add_update_free_vsi_resp add_update_free_vsi_res;
 		struct ice_aqc_fw_logging fw_logging;
@@ -2028,6 +2060,7 @@ enum ice_adminq_opc {
 	/* Tx queue handling commands/events */
 	ice_aqc_opc_add_txqs				= 0x0C30,
 	ice_aqc_opc_dis_txqs				= 0x0C31,
+	ice_aqc_opc_add_rdma_qset			= 0x0C33,
 
 	/* package commands */
 	ice_aqc_opc_download_pkg			= 0x0C40,

drivers/net/ethernet/intel/ice/ice_common.c

@@ -2,6 +2,7 @@
 /* Copyright (c) 2018, Intel Corporation. */
 
 #include "ice_common.h"
+#include "ice_lib.h"
 #include "ice_sched.h"
 #include "ice_adminq_cmd.h"
 #include "ice_flow.h"
@@ -1062,7 +1063,8 @@ enum ice_status ice_check_reset(struct ice_hw *hw)
 				 GLNVM_ULD_POR_DONE_1_M |\
 				 GLNVM_ULD_PCIER_DONE_2_M)
 
-	uld_mask = ICE_RESET_DONE_MASK;
+	uld_mask = ICE_RESET_DONE_MASK | (hw->func_caps.common_cap.rdma ?
+					  GLNVM_ULD_PE_DONE_M : 0);
 
 	/* Device is Active; check Global Reset processes are done */
 	for (cnt = 0; cnt < ICE_PF_RESET_WAIT_COUNT; cnt++) {
@@ -1938,6 +1940,10 @@ ice_parse_common_caps(struct ice_hw *hw, struct ice_hw_common_caps *caps,
 		ice_debug(hw, ICE_DBG_INIT, "%s: nvm_unified_update = %d\n", prefix,
 			  caps->nvm_unified_update);
 		break;
+	case ICE_AQC_CAPS_RDMA:
+		caps->rdma = (number == 1);
+		ice_debug(hw, ICE_DBG_INIT, "%s: rdma = %d\n", prefix, caps->rdma);
+		break;
 	case ICE_AQC_CAPS_MAX_MTU:
 		caps->max_mtu = number;
 		ice_debug(hw, ICE_DBG_INIT, "%s: max_mtu = %d\n",
@@ -1971,6 +1977,16 @@ ice_recalc_port_limited_caps(struct ice_hw *hw, struct ice_hw_common_caps *caps)
 		caps->maxtc = 4;
 		ice_debug(hw, ICE_DBG_INIT, "reducing maxtc to %d (based on #ports)\n",
 			  caps->maxtc);
+		if (caps->rdma) {
+			ice_debug(hw, ICE_DBG_INIT, "forcing RDMA off\n");
+			caps->rdma = 0;
+		}
+
+		/* print message only when processing device capabilities
+		 * during initialization.
+		 */
+		if (caps == &hw->dev_caps.common_cap)
+			dev_info(ice_hw_to_dev(hw), "RDMA functionality is not available with the current device configuration.\n");
 	}
 }
 
@@ -3635,6 +3651,52 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 	return status;
 }
 
+/**
+ * ice_aq_add_rdma_qsets
+ * @hw: pointer to the hardware structure
+ * @num_qset_grps: Number of RDMA Qset groups
+ * @qset_list: list of Qset groups to be added
+ * @buf_size: size of buffer for indirect command
+ * @cd: pointer to command details structure or NULL
+ *
+ * Add Tx RDMA Qsets (0x0C33)
+ */
+static int
+ice_aq_add_rdma_qsets(struct ice_hw *hw, u8 num_qset_grps,
+		      struct ice_aqc_add_rdma_qset_data *qset_list,
+		      u16 buf_size, struct ice_sq_cd *cd)
+{
+	struct ice_aqc_add_rdma_qset_data *list;
+	struct ice_aqc_add_rdma_qset *cmd;
+	struct ice_aq_desc desc;
+	u16 i, sum_size = 0;
+
+	cmd = &desc.params.add_rdma_qset;
+
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_add_rdma_qset);
+
+	if (num_qset_grps > ICE_LAN_TXQ_MAX_QGRPS)
+		return -EINVAL;
+
+	for (i = 0, list = qset_list; i < num_qset_grps; i++) {
+		u16 num_qsets = le16_to_cpu(list->num_qsets);
+
+		sum_size += struct_size(list, rdma_qsets, num_qsets);
+		list = (struct ice_aqc_add_rdma_qset_data *)(list->rdma_qsets +
+							     num_qsets);
+	}
+
+	if (buf_size != sum_size)
+		return -EINVAL;
+
+	desc.flags |= cpu_to_le16(ICE_AQ_FLAG_RD);
+
+	cmd->num_qset_grps = num_qset_grps;
+
+	return ice_status_to_errno(ice_aq_send_cmd(hw, &desc, qset_list,
+						   buf_size, cd));
+}
+
 /* End of FW Admin Queue command wrappers */
 
 /**
@@ -4132,6 +4194,162 @@ ice_cfg_vsi_lan(struct ice_port_info *pi, u16 vsi_handle, u8 tc_bitmap,
 			      ICE_SCHED_NODE_OWNER_LAN);
 }
 
+/**
+ * ice_cfg_vsi_rdma - configure the VSI RDMA queues
+ * @pi: port information structure
+ * @vsi_handle: software VSI handle
+ * @tc_bitmap: TC bitmap
+ * @max_rdmaqs: max RDMA queues array per TC
+ *
+ * This function adds/updates the VSI RDMA queues per TC.
+ */
+int
+ice_cfg_vsi_rdma(struct ice_port_info *pi, u16 vsi_handle, u16 tc_bitmap,
+		 u16 *max_rdmaqs)
+{
+	return ice_status_to_errno(ice_cfg_vsi_qs(pi, vsi_handle, tc_bitmap,
+						  max_rdmaqs,
+						  ICE_SCHED_NODE_OWNER_RDMA));
+}
+
+/**
+ * ice_ena_vsi_rdma_qset
+ * @pi: port information structure
+ * @vsi_handle: software VSI handle
+ * @tc: TC number
+ * @rdma_qset: pointer to RDMA Qset
+ * @num_qsets: number of RDMA Qsets
+ * @qset_teid: pointer to Qset node TEIDs
+ *
+ * This function adds RDMA Qset
+ */
+int
+ice_ena_vsi_rdma_qset(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
+		      u16 *rdma_qset, u16 num_qsets, u32 *qset_teid)
+{
+	struct ice_aqc_txsched_elem_data node = { 0 };
+	struct ice_aqc_add_rdma_qset_data *buf;
+	struct ice_sched_node *parent;
+	enum ice_status status;
+	struct ice_hw *hw;
+	u16 i, buf_size;
+	int ret;
+
+	if (!pi || pi->port_state != ICE_SCHED_PORT_STATE_READY)
+		return -EIO;
+	hw = pi->hw;
+
+	if (!ice_is_vsi_valid(hw, vsi_handle))
+		return -EINVAL;
+
+	buf_size = struct_size(buf, rdma_qsets, num_qsets);
+	buf = kzalloc(buf_size, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+	mutex_lock(&pi->sched_lock);
+
+	parent = ice_sched_get_free_qparent(pi, vsi_handle, tc,
+					    ICE_SCHED_NODE_OWNER_RDMA);
+	if (!parent) {
+		ret = -EINVAL;
+		goto rdma_error_exit;
+	}
+	buf->parent_teid = parent->info.node_teid;
+	node.parent_teid = parent->info.node_teid;
+
+	buf->num_qsets = cpu_to_le16(num_qsets);
+	for (i = 0; i < num_qsets; i++) {
+		buf->rdma_qsets[i].tx_qset_id = cpu_to_le16(rdma_qset[i]);
+		buf->rdma_qsets[i].info.valid_sections =
+			ICE_AQC_ELEM_VALID_GENERIC | ICE_AQC_ELEM_VALID_CIR |
+			ICE_AQC_ELEM_VALID_EIR;
+		buf->rdma_qsets[i].info.generic = 0;
+		buf->rdma_qsets[i].info.cir_bw.bw_profile_idx =
+			cpu_to_le16(ICE_SCHED_DFLT_RL_PROF_ID);
+		buf->rdma_qsets[i].info.cir_bw.bw_alloc =
+			cpu_to_le16(ICE_SCHED_DFLT_BW_WT);
+		buf->rdma_qsets[i].info.eir_bw.bw_profile_idx =
+			cpu_to_le16(ICE_SCHED_DFLT_RL_PROF_ID);
+		buf->rdma_qsets[i].info.eir_bw.bw_alloc =
+			cpu_to_le16(ICE_SCHED_DFLT_BW_WT);
+	}
+	ret = ice_aq_add_rdma_qsets(hw, 1, buf, buf_size, NULL);
+	if (ret) {
+		ice_debug(hw, ICE_DBG_RDMA, "add RDMA qset failed\n");
+		goto rdma_error_exit;
+	}
+	node.data.elem_type = ICE_AQC_ELEM_TYPE_LEAF;
+	for (i = 0; i < num_qsets; i++) {
+		node.node_teid = buf->rdma_qsets[i].qset_teid;
+		status = ice_sched_add_node(pi, hw->num_tx_sched_layers - 1,
+					    &node);
+		if (status) {
+			ret = ice_status_to_errno(status);
+			break;
+		}
+		qset_teid[i] = le32_to_cpu(node.node_teid);
+	}
+rdma_error_exit:
+	mutex_unlock(&pi->sched_lock);
+	kfree(buf);
+	return ret;
+}
+
+/**
+ * ice_dis_vsi_rdma_qset - free RDMA resources
+ * @pi: port_info struct
+ * @count: number of RDMA Qsets to free
+ * @qset_teid: TEID of Qset node
+ * @q_id: list of queue IDs being disabled
+ */
+int
+ice_dis_vsi_rdma_qset(struct ice_port_info *pi, u16 count, u32 *qset_teid,
+		      u16 *q_id)
+{
+	struct ice_aqc_dis_txq_item *qg_list;
+	enum ice_status status = 0;
+	struct ice_hw *hw;
+	u16 qg_size;
+	int i;
+
+	if (!pi || pi->port_state != ICE_SCHED_PORT_STATE_READY)
+		return -EIO;
+
+	hw = pi->hw;
+
+	qg_size = struct_size(qg_list, q_id, 1);
+	qg_list = kzalloc(qg_size, GFP_KERNEL);
+	if (!qg_list)
+		return -ENOMEM;
+
+	mutex_lock(&pi->sched_lock);
+
+	for (i = 0; i < count; i++) {
+		struct ice_sched_node *node;
+
+		node = ice_sched_find_node_by_teid(pi->root, qset_teid[i]);
+		if (!node)
+			continue;
+
+		qg_list->parent_teid = node->info.parent_teid;
+		qg_list->num_qs = 1;
+		qg_list->q_id[0] =
+			cpu_to_le16(q_id[i] |
+				    ICE_AQC_Q_DIS_BUF_ELEM_TYPE_RDMA_QSET);
+
+		status = ice_aq_dis_lan_txq(hw, 1, qg_list, qg_size,
+					    ICE_NO_RESET, 0, NULL);
+		if (status)
+			break;
+
+		ice_free_sched_node(pi, node);
+	}
+
+	mutex_unlock(&pi->sched_lock);
+	kfree(qg_list);
+	return ice_status_to_errno(status);
+}
+
 /**
  * ice_replay_pre_init - replay pre initialization
  * @hw: pointer to the HW struct

drivers/net/ethernet/intel/ice/ice_common.h

@@ -147,6 +147,15 @@ ice_aq_sff_eeprom(struct ice_hw *hw, u16 lport, u8 bus_addr,
 		  u16 mem_addr, u8 page, u8 set_page, u8 *data, u8 length,
 		  bool write, struct ice_sq_cd *cd);
 
+int
+ice_cfg_vsi_rdma(struct ice_port_info *pi, u16 vsi_handle, u16 tc_bitmap,
+		 u16 *max_rdmaqs);
+int
+ice_ena_vsi_rdma_qset(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
+		      u16 *rdma_qset, u16 num_qsets, u32 *qset_teid);
+int
+ice_dis_vsi_rdma_qset(struct ice_port_info *pi, u16 count, u32 *qset_teid,
+		      u16 *q_id);
 enum ice_status
 ice_dis_vsi_txq(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u8 num_queues,
 		u16 *q_handle, u16 *q_ids, u32 *q_teids,

drivers/net/ethernet/intel/ice/ice_dcb_lib.c

@@ -275,6 +275,7 @@ int ice_pf_dcb_cfg(struct ice_pf *pf, struct ice_dcbx_cfg *new_cfg, bool locked)
 	struct ice_dcbx_cfg *old_cfg, *curr_cfg;
 	struct device *dev = ice_pf_to_dev(pf);
 	int ret = ICE_DCB_NO_HW_CHG;
+	struct iidc_event *event;
 	struct ice_vsi *pf_vsi;
 
 	curr_cfg = &pf->hw.port_info->qos_cfg.local_dcbx_cfg;
@@ -313,6 +314,15 @@ int ice_pf_dcb_cfg(struct ice_pf *pf, struct ice_dcbx_cfg *new_cfg, bool locked)
 		goto free_cfg;
 	}
 
+	/* Notify AUX drivers about impending change to TCs */
+	event = kzalloc(sizeof(*event), GFP_KERNEL);
+	if (!event)
+		return -ENOMEM;
+
+	set_bit(IIDC_EVENT_BEFORE_TC_CHANGE, event->type);
+	ice_send_event_to_aux(pf, event);
+	kfree(event);
+
 	/* avoid race conditions by holding the lock while disabling and
 	 * re-enabling the VSI
 	 */
@@ -640,6 +650,7 @@ static int ice_dcb_noncontig_cfg(struct ice_pf *pf)
 void ice_pf_dcb_recfg(struct ice_pf *pf)
 {
 	struct ice_dcbx_cfg *dcbcfg = &pf->hw.port_info->qos_cfg.local_dcbx_cfg;
+	struct iidc_event *event;
 	u8 tc_map = 0;
 	int v, ret;
 
@@ -675,6 +686,14 @@ void ice_pf_dcb_recfg(struct ice_pf *pf)
 		if (vsi->type == ICE_VSI_PF)
 			ice_dcbnl_set_all(vsi);
 	}
+	/* Notify the AUX drivers that TC change is finished */
+	event = kzalloc(sizeof(*event), GFP_KERNEL);
+	if (!event)
+		return;
+
+	set_bit(IIDC_EVENT_AFTER_TC_CHANGE, event->type);
+	ice_send_event_to_aux(pf, event);
+	kfree(event);
 }
 
 /**

drivers/net/ethernet/intel/ice/ice_hw_autogen.h

@@ -110,8 +110,6 @@
 #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_CTL				0x0016CC54
 #define GLINT_CTL_DIS_AUTOMASK_M		BIT(0)
 #define GLINT_CTL_ITR_GRAN_200_S		16
@@ -160,6 +158,7 @@
 #define PFINT_OICR_GRST_M			BIT(20)
 #define PFINT_OICR_PCI_EXCEPTION_M		BIT(21)
 #define PFINT_OICR_HMC_ERR_M			BIT(26)
+#define PFINT_OICR_PE_PUSH_M			BIT(27)
 #define PFINT_OICR_PE_CRITERR_M			BIT(28)
 #define PFINT_OICR_VFLR_M			BIT(29)
 #define PFINT_OICR_SWINT_M			BIT(31)

drivers/net/ethernet/intel/ice/ice_idc.c

+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (C) 2021, Intel Corporation. */
+
+/* Inter-Driver Communication */
+#include "ice.h"
+#include "ice_lib.h"
+#include "ice_dcb_lib.h"
+
+/**
+ * ice_get_auxiliary_drv - retrieve iidc_auxiliary_drv struct
+ * @pf: pointer to PF struct
+ *
+ * This function has to be called with a device_lock on the
+ * pf->adev.dev to avoid race conditions.
+ */
+static struct iidc_auxiliary_drv *ice_get_auxiliary_drv(struct ice_pf *pf)
+{
+	struct auxiliary_device *adev;
+
+	adev = pf->adev;
+	if (!adev || !adev->dev.driver)
+		return NULL;
+
+	return container_of(adev->dev.driver, struct iidc_auxiliary_drv,
+			    adrv.driver);
+}
+
+/**
+ * ice_send_event_to_aux - send event to RDMA AUX driver
+ * @pf: pointer to PF struct
+ * @event: event struct
+ */
+void ice_send_event_to_aux(struct ice_pf *pf, struct iidc_event *event)
+{
+	struct iidc_auxiliary_drv *iadrv;
+
+	if (!pf->adev)
+		return;
+
+	device_lock(&pf->adev->dev);
+	iadrv = ice_get_auxiliary_drv(pf);
+	if (iadrv && iadrv->event_handler)
+		iadrv->event_handler(pf, event);
+	device_unlock(&pf->adev->dev);
+}
+
+/**
+ * ice_find_vsi - Find the VSI from VSI ID
+ * @pf: The PF pointer to search in
+ * @vsi_num: The VSI ID to search for
+ */
+static struct ice_vsi *ice_find_vsi(struct ice_pf *pf, u16 vsi_num)
+{
+	int i;
+
+	ice_for_each_vsi(pf, i)
+		if (pf->vsi[i] && pf->vsi[i]->vsi_num == vsi_num)
+			return  pf->vsi[i];
+	return NULL;
+}
+
+/**
+ * ice_add_rdma_qset - Add Leaf Node for RDMA Qset
+ * @pf: PF struct
+ * @qset: Resource to be allocated
+ */
+int ice_add_rdma_qset(struct ice_pf *pf, struct iidc_rdma_qset_params *qset)
+{
+	u16 max_rdmaqs[ICE_MAX_TRAFFIC_CLASS];
+	struct ice_vsi *vsi;
+	struct device *dev;
+	u32 qset_teid;
+	u16 qs_handle;
+	int status;
+	int i;
+
+	if (WARN_ON(!pf || !qset))
+		return -EINVAL;
+
+	dev = ice_pf_to_dev(pf);
+
+	if (!test_bit(ICE_FLAG_RDMA_ENA, pf->flags))
+		return -EINVAL;
+
+	vsi = ice_get_main_vsi(pf);
+	if (!vsi) {
+		dev_err(dev, "RDMA QSet invalid VSI\n");
+		return -EINVAL;
+	}
+
+	ice_for_each_traffic_class(i)
+		max_rdmaqs[i] = 0;
+
+	max_rdmaqs[qset->tc]++;
+	qs_handle = qset->qs_handle;
+
+	status = ice_cfg_vsi_rdma(vsi->port_info, vsi->idx, vsi->tc_cfg.ena_tc,
+				  max_rdmaqs);
+	if (status) {
+		dev_err(dev, "Failed VSI RDMA Qset config\n");
+		return status;
+	}
+
+	status = ice_ena_vsi_rdma_qset(vsi->port_info, vsi->idx, qset->tc,
+				       &qs_handle, 1, &qset_teid);
+	if (status) {
+		dev_err(dev, "Failed VSI RDMA Qset enable\n");
+		return status;
+	}
+	vsi->qset_handle[qset->tc] = qset->qs_handle;
+	qset->teid = qset_teid;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ice_add_rdma_qset);
+
+/**
+ * ice_del_rdma_qset - Delete leaf node for RDMA Qset
+ * @pf: PF struct
+ * @qset: Resource to be freed
+ */
+int ice_del_rdma_qset(struct ice_pf *pf, struct iidc_rdma_qset_params *qset)
+{
+	struct ice_vsi *vsi;
+	u32 teid;
+	u16 q_id;
+
+	if (WARN_ON(!pf || !qset))
+		return -EINVAL;
+
+	vsi = ice_find_vsi(pf, qset->vport_id);
+	if (!vsi) {
+		dev_err(ice_pf_to_dev(pf), "RDMA Invalid VSI\n");
+		return -EINVAL;
+	}
+
+	q_id = qset->qs_handle;
+	teid = qset->teid;
+
+	vsi->qset_handle[qset->tc] = 0;
+
+	return ice_dis_vsi_rdma_qset(vsi->port_info, 1, &teid, &q_id);
+}
+EXPORT_SYMBOL_GPL(ice_del_rdma_qset);
+
+/**
+ * ice_rdma_request_reset - accept request from RDMA to perform a reset
+ * @pf: struct for PF
+ * @reset_type: type of reset
+ */
+int ice_rdma_request_reset(struct ice_pf *pf, enum iidc_reset_type reset_type)
+{
+	enum ice_reset_req reset;
+
+	if (WARN_ON(!pf))
+		return -EINVAL;
+
+	switch (reset_type) {
+	case IIDC_PFR:
+		reset = ICE_RESET_PFR;
+		break;
+	case IIDC_CORER:
+		reset = ICE_RESET_CORER;
+		break;
+	case IIDC_GLOBR:
+		reset = ICE_RESET_GLOBR;
+		break;
+	default:
+		dev_err(ice_pf_to_dev(pf), "incorrect reset request\n");
+		return -EINVAL;
+	}
+
+	return ice_schedule_reset(pf, reset);
+}
+EXPORT_SYMBOL_GPL(ice_rdma_request_reset);
+
+/**
+ * ice_rdma_update_vsi_filter - update main VSI filters for RDMA
+ * @pf: pointer to struct for PF
+ * @vsi_id: VSI HW idx to update filter on
+ * @enable: bool whether to enable or disable filters
+ */
+int ice_rdma_update_vsi_filter(struct ice_pf *pf, u16 vsi_id, bool enable)
+{
+	struct ice_vsi *vsi;
+	int status;
+
+	if (WARN_ON(!pf))
+		return -EINVAL;
+
+	vsi = ice_find_vsi(pf, vsi_id);
+	if (!vsi)
+		return -EINVAL;
+
+	status = ice_cfg_rdma_fltr(&pf->hw, vsi->idx, enable);
+	if (status) {
+		dev_err(ice_pf_to_dev(pf), "Failed to  %sable RDMA filtering\n",
+			enable ? "en" : "dis");
+	} else {
+		if (enable)
+			vsi->info.q_opt_flags |= ICE_AQ_VSI_Q_OPT_PE_FLTR_EN;
+		else
+			vsi->info.q_opt_flags &= ~ICE_AQ_VSI_Q_OPT_PE_FLTR_EN;
+	}
+
+	return status;
+}
+EXPORT_SYMBOL_GPL(ice_rdma_update_vsi_filter);
+
+/**
+ * ice_get_qos_params - parse QoS params for RDMA consumption
+ * @pf: pointer to PF struct
+ * @qos: set of QoS values
+ */
+void ice_get_qos_params(struct ice_pf *pf, struct iidc_qos_params *qos)
+{
+	struct ice_dcbx_cfg *dcbx_cfg;
+	unsigned int i;
+	u32 up2tc;
+
+	dcbx_cfg = &pf->hw.port_info->qos_cfg.local_dcbx_cfg;
+	up2tc = rd32(&pf->hw, PRTDCB_TUP2TC);
+
+	qos->num_tc = ice_dcb_get_num_tc(dcbx_cfg);
+	for (i = 0; i < IIDC_MAX_USER_PRIORITY; i++)
+		qos->up2tc[i] = (up2tc >> (i * 3)) & 0x7;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++)
+		qos->tc_info[i].rel_bw = dcbx_cfg->etscfg.tcbwtable[i];
+}
+EXPORT_SYMBOL_GPL(ice_get_qos_params);
+
+/**
+ * ice_reserve_rdma_qvector - Reserve vector resources for RDMA driver
+ * @pf: board private structure to initialize
+ */
+static int ice_reserve_rdma_qvector(struct ice_pf *pf)
+{
+	if (test_bit(ICE_FLAG_RDMA_ENA, pf->flags)) {
+		int index;
+
+		index = ice_get_res(pf, pf->irq_tracker, pf->num_rdma_msix,
+				    ICE_RES_RDMA_VEC_ID);
+		if (index < 0)
+			return index;
+		pf->num_avail_sw_msix -= pf->num_rdma_msix;
+		pf->rdma_base_vector = (u16)index;
+	}
+	return 0;
+}
+
+/**
+ * ice_adev_release - function to be mapped to AUX dev's release op
+ * @dev: pointer to device to free
+ */
+static void ice_adev_release(struct device *dev)
+{
+	struct iidc_auxiliary_dev *iadev;
+
+	iadev = container_of(dev, struct iidc_auxiliary_dev, adev.dev);
+	kfree(iadev);
+}
+
+/**
+ * ice_plug_aux_dev - allocate and register AUX device
+ * @pf: pointer to pf struct
+ */
+int ice_plug_aux_dev(struct ice_pf *pf)
+{
+	struct iidc_auxiliary_dev *iadev;
+	struct auxiliary_device *adev;
+	int ret;
+
+	iadev = kzalloc(sizeof(*iadev), GFP_KERNEL);
+	if (!iadev)
+		return -ENOMEM;
+
+	adev = &iadev->adev;
+	pf->adev = adev;
+	iadev->pf = pf;
+
+	adev->id = pf->aux_idx;
+	adev->dev.release = ice_adev_release;
+	adev->dev.parent = &pf->pdev->dev;
+	adev->name = IIDC_RDMA_ROCE_NAME;
+
+	ret = auxiliary_device_init(adev);
+	if (ret) {
+		pf->adev = NULL;
+		kfree(iadev);
+		return ret;
+	}
+
+	ret = auxiliary_device_add(adev);
+	if (ret) {
+		pf->adev = NULL;
+		auxiliary_device_uninit(adev);
+		return ret;
+	}
+
+	return 0;
+}
+
+/* ice_unplug_aux_dev - unregister and free AUX device
+ * @pf: pointer to pf struct
+ */
+void ice_unplug_aux_dev(struct ice_pf *pf)
+{
+	if (!pf->adev)
+		return;
+
+	auxiliary_device_delete(pf->adev);
+	auxiliary_device_uninit(pf->adev);
+	pf->adev = NULL;
+}
+
+/**
+ * ice_init_rdma - initializes PF for RDMA use
+ * @pf: ptr to ice_pf
+ */
+int ice_init_rdma(struct ice_pf *pf)
+{
+	struct device *dev = &pf->pdev->dev;
+	int ret;
+
+	/* Reserve vector resources */
+	ret = ice_reserve_rdma_qvector(pf);
+	if (ret < 0) {
+		dev_err(dev, "failed to reserve vectors for RDMA\n");
+		return ret;
+	}
+
+	return ice_plug_aux_dev(pf);
+}

drivers/net/ethernet/intel/ice/ice_idc_int.h

+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2021, Intel Corporation. */
+
+#ifndef _ICE_IDC_INT_H_
+#define _ICE_IDC_INT_H_
+
+#include <linux/net/intel/iidc.h>
+#include "ice.h"
+
+struct ice_pf;
+
+void ice_send_event_to_aux(struct ice_pf *pf, struct iidc_event *event);
+
+#endif /* !_ICE_IDC_INT_H_ */

drivers/net/ethernet/intel/ice/ice_lag.c

@@ -172,6 +172,7 @@ ice_lag_link(struct ice_lag *lag, struct netdev_notifier_changeupper_info *info)
 	}
 
 	ice_clear_sriov_cap(pf);
+	ice_clear_rdma_cap(pf);
 
 	lag->bonded = true;
 	lag->role = ICE_LAG_UNSET;
@@ -222,6 +223,7 @@ ice_lag_unlink(struct ice_lag *lag,
 	}
 
 	ice_set_sriov_cap(pf);
+	ice_set_rdma_cap(pf);
 	lag->bonded = false;
 	lag->role = ICE_LAG_NONE;
 }

drivers/net/ethernet/intel/ice/ice_lib.c

@@ -616,6 +616,17 @@ bool ice_is_safe_mode(struct ice_pf *pf)
 	return !test_bit(ICE_FLAG_ADV_FEATURES, pf->flags);
 }
 
+/**
+ * ice_is_aux_ena
+ * @pf: pointer to the PF struct
+ *
+ * returns true if AUX devices/drivers are supported, false otherwise
+ */
+bool ice_is_aux_ena(struct ice_pf *pf)
+{
+	return test_bit(ICE_FLAG_AUX_ENA, pf->flags);
+}
+
 /**
  * ice_vsi_clean_rss_flow_fld - Delete RSS configuration
  * @vsi: the VSI being cleaned up

drivers/net/ethernet/intel/ice/ice_lib.h

@@ -102,7 +102,7 @@ enum ice_status
 ice_vsi_cfg_mac_fltr(struct ice_vsi *vsi, const u8 *macaddr, bool set);
 
 bool ice_is_safe_mode(struct ice_pf *pf);
-
+bool ice_is_aux_ena(struct ice_pf *pf);
 bool ice_is_dflt_vsi_in_use(struct ice_sw *sw);
 
 bool ice_is_vsi_dflt_vsi(struct ice_sw *sw, struct ice_vsi *vsi);

drivers/net/ethernet/intel/ice/ice_main.c

@@ -35,6 +35,8 @@ MODULE_PARM_DESC(debug, "netif level (0=none,...,16=all), hw debug_mask (0x8XXXX
 MODULE_PARM_DESC(debug, "netif level (0=none,...,16=all)");
 #endif /* !CONFIG_DYNAMIC_DEBUG */
 
+static DEFINE_IDA(ice_aux_ida);
+
 static struct workqueue_struct *ice_wq;
 static const struct net_device_ops ice_netdev_safe_mode_ops;
 static const struct net_device_ops ice_netdev_ops;
@@ -454,6 +456,8 @@ ice_prepare_for_reset(struct ice_pf *pf)
 	if (test_bit(ICE_PREPARED_FOR_RESET, pf->state))
 		return;
 
+	ice_unplug_aux_dev(pf);
+
 	/* Notify VFs of impending reset */
 	if (ice_check_sq_alive(hw, &hw->mailboxq))
 		ice_vc_notify_reset(pf);
@@ -2118,6 +2122,8 @@ int ice_schedule_reset(struct ice_pf *pf, enum ice_reset_req reset)
 		return -EBUSY;
 	}
 
+	ice_unplug_aux_dev(pf);
+
 	switch (reset) {
 	case ICE_RESET_PFR:
 		set_bit(ICE_PFR_REQ, pf->state);
@@ -2608,6 +2614,7 @@ static void ice_ena_misc_vector(struct ice_pf *pf)
 	       PFINT_OICR_PCI_EXCEPTION_M |
 	       PFINT_OICR_VFLR_M |
 	       PFINT_OICR_HMC_ERR_M |
+	       PFINT_OICR_PE_PUSH_M |
 	       PFINT_OICR_PE_CRITERR_M);
 
 	wr32(hw, PFINT_OICR_ENA, val);
@@ -2678,8 +2685,6 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
 
 		/* If a reset cycle isn't already in progress, we set a bit in
 		 * pf->state so that the service task can start a reset/rebuild.
-		 * We also make note of which reset happened so that peer
-		 * devices/drivers can be informed.
 		 */
 		if (!test_and_set_bit(ICE_RESET_OICR_RECV, pf->state)) {
 			if (reset == ICE_RESET_CORER)
@@ -2706,11 +2711,19 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
 		}
 	}
 
-	if (oicr & PFINT_OICR_HMC_ERR_M) {
-		ena_mask &= ~PFINT_OICR_HMC_ERR_M;
-		dev_dbg(dev, "HMC Error interrupt - info 0x%x, data 0x%x\n",
-			rd32(hw, PFHMC_ERRORINFO),
-			rd32(hw, PFHMC_ERRORDATA));
+#define ICE_AUX_CRIT_ERR (PFINT_OICR_PE_CRITERR_M | PFINT_OICR_HMC_ERR_M | PFINT_OICR_PE_PUSH_M)
+	if (oicr & ICE_AUX_CRIT_ERR) {
+		struct iidc_event *event;
+
+		ena_mask &= ~ICE_AUX_CRIT_ERR;
+		event = kzalloc(sizeof(*event), GFP_KERNEL);
+		if (event) {
+			set_bit(IIDC_EVENT_CRIT_ERR, event->type);
+			/* report the entire OICR value to AUX driver */
+			event->reg = oicr;
+			ice_send_event_to_aux(pf, event);
+			kfree(event);
+		}
 	}
 
 	/* Report any remaining unexpected interrupts */
@@ -2720,8 +2733,7 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
 		/* If a critical error is pending there is no choice but to
 		 * reset the device.
 		 */
-		if (oicr & (PFINT_OICR_PE_CRITERR_M |
-			    PFINT_OICR_PCI_EXCEPTION_M |
+		if (oicr & (PFINT_OICR_PCI_EXCEPTION_M |
 			    PFINT_OICR_ECC_ERR_M)) {
 			set_bit(ICE_PFR_REQ, pf->state);
 			ice_service_task_schedule(pf);
@@ -3276,6 +3288,12 @@ static void ice_set_pf_caps(struct ice_pf *pf)
 {
 	struct ice_hw_func_caps *func_caps = &pf->hw.func_caps;
 
+	clear_bit(ICE_FLAG_RDMA_ENA, pf->flags);
+	clear_bit(ICE_FLAG_AUX_ENA, pf->flags);
+	if (func_caps->common_cap.rdma) {
+		set_bit(ICE_FLAG_RDMA_ENA, pf->flags);
+		set_bit(ICE_FLAG_AUX_ENA, pf->flags);
+	}
 	clear_bit(ICE_FLAG_DCB_CAPABLE, pf->flags);
 	if (func_caps->common_cap.dcb)
 		set_bit(ICE_FLAG_DCB_CAPABLE, pf->flags);
@@ -3355,11 +3373,12 @@ static int ice_init_pf(struct ice_pf *pf)
  */
 static int ice_ena_msix_range(struct ice_pf *pf)
 {
-	int v_left, v_actual, v_other, v_budget = 0;
+	int num_cpus, v_left, v_actual, v_other, v_budget = 0;
 	struct device *dev = ice_pf_to_dev(pf);
 	int needed, err, i;
 
 	v_left = pf->hw.func_caps.common_cap.num_msix_vectors;
+	num_cpus = num_online_cpus();
 
 	/* reserve for LAN miscellaneous handler */
 	needed = ICE_MIN_LAN_OICR_MSIX;
@@ -3381,13 +3400,23 @@ static int ice_ena_msix_range(struct ice_pf *pf)
 	v_other = v_budget;
 
 	/* reserve vectors for LAN traffic */
-	needed = min_t(int, num_online_cpus(), v_left);
+	needed = num_cpus;
 	if (v_left < needed)
 		goto no_hw_vecs_left_err;
 	pf->num_lan_msix = needed;
 	v_budget += needed;
 	v_left -= needed;
 
+	/* reserve vectors for RDMA auxiliary driver */
+	if (test_bit(ICE_FLAG_RDMA_ENA, pf->flags)) {
+		needed = num_cpus + ICE_RDMA_NUM_AEQ_MSIX;
+		if (v_left < needed)
+			goto no_hw_vecs_left_err;
+		pf->num_rdma_msix = needed;
+		v_budget += needed;
+		v_left -= needed;
+	}
+
 	pf->msix_entries = devm_kcalloc(dev, v_budget,
 					sizeof(*pf->msix_entries), GFP_KERNEL);
 	if (!pf->msix_entries) {
@@ -3417,16 +3446,46 @@ static int ice_ena_msix_range(struct ice_pf *pf)
 			err = -ERANGE;
 			goto msix_err;
 		} else {
-			int v_traffic = v_actual - v_other;
+			int v_remain = v_actual - v_other;
+			int v_rdma = 0, v_min_rdma = 0;
+
+			if (test_bit(ICE_FLAG_RDMA_ENA, pf->flags)) {
+				/* Need at least 1 interrupt in addition to
+				 * AEQ MSIX
+				 */
+				v_rdma = ICE_RDMA_NUM_AEQ_MSIX + 1;
+				v_min_rdma = ICE_MIN_RDMA_MSIX;
+			}
 
 			if (v_actual == ICE_MIN_MSIX ||
-			    v_traffic < ICE_MIN_LAN_TXRX_MSIX)
+			    v_remain < ICE_MIN_LAN_TXRX_MSIX + v_min_rdma) {
+				dev_warn(dev, "Not enough MSI-X vectors to support RDMA.\n");
+				clear_bit(ICE_FLAG_RDMA_ENA, pf->flags);
+
+				pf->num_rdma_msix = 0;
 				pf->num_lan_msix = ICE_MIN_LAN_TXRX_MSIX;
-			else
-				pf->num_lan_msix = v_traffic;
+			} else if ((v_remain < ICE_MIN_LAN_TXRX_MSIX + v_rdma) ||
+				   (v_remain - v_rdma < v_rdma)) {
+				/* Support minimum RDMA and give remaining
+				 * vectors to LAN MSIX
+				 */
+				pf->num_rdma_msix = v_min_rdma;
+				pf->num_lan_msix = v_remain - v_min_rdma;
+			} else {
+				/* Split remaining MSIX with RDMA after
+				 * accounting for AEQ MSIX
+				 */
+				pf->num_rdma_msix = (v_remain - ICE_RDMA_NUM_AEQ_MSIX) / 2 +
+						    ICE_RDMA_NUM_AEQ_MSIX;
+				pf->num_lan_msix = v_remain - pf->num_rdma_msix;
+			}
 
 			dev_notice(dev, "Enabled %d MSI-X vectors for LAN traffic.\n",
 				   pf->num_lan_msix);
+
+			if (test_bit(ICE_FLAG_RDMA_ENA, pf->flags))
+				dev_notice(dev, "Enabled %d MSI-X vectors for RDMA.\n",
+					   pf->num_rdma_msix);
 		}
 	}
 
@@ -3441,6 +3500,7 @@ static int ice_ena_msix_range(struct ice_pf *pf)
 		needed, v_left);
 	err = -ERANGE;
 exit_err:
+	pf->num_rdma_msix = 0;
 	pf->num_lan_msix = 0;
 	return err;
 }
@@ -4268,8 +4328,29 @@ ice_probe(struct pci_dev *pdev, const struct pci_device_id __always_unused *ent)
 
 	/* ready to go, so clear down state bit */
 	clear_bit(ICE_DOWN, pf->state);
+	if (ice_is_aux_ena(pf)) {
+		pf->aux_idx = ida_alloc(&ice_aux_ida, GFP_KERNEL);
+		if (pf->aux_idx < 0) {
+			dev_err(dev, "Failed to allocate device ID for AUX driver\n");
+			err = -ENOMEM;
+			goto err_netdev_reg;
+		}
+
+		err = ice_init_rdma(pf);
+		if (err) {
+			dev_err(dev, "Failed to initialize RDMA: %d\n", err);
+			err = -EIO;
+			goto err_init_aux_unroll;
+		}
+	} else {
+		dev_warn(dev, "RDMA is not supported on this device\n");
+	}
+
 	return 0;
 
+err_init_aux_unroll:
+	pf->adev = NULL;
+	ida_free(&ice_aux_ida, pf->aux_idx);
 err_netdev_reg:
 err_send_version_unroll:
 	ice_vsi_release_all(pf);
@@ -4379,10 +4460,12 @@ static void ice_remove(struct pci_dev *pdev)
 		ice_free_vfs(pf);
 	}
 
-	set_bit(ICE_DOWN, pf->state);
 	ice_service_task_stop(pf);
 
 	ice_aq_cancel_waiting_tasks(pf);
+	ice_unplug_aux_dev(pf);
+	ida_free(&ice_aux_ida, pf->aux_idx);
+	set_bit(ICE_DOWN, pf->state);
 
 	mutex_destroy(&(&pf->hw)->fdir_fltr_lock);
 	ice_deinit_lag(pf);
@@ -4538,6 +4621,8 @@ static int __maybe_unused ice_suspend(struct device *dev)
 	 */
 	disabled = ice_service_task_stop(pf);
 
+	ice_unplug_aux_dev(pf);
+
 	/* Already suspended?, then there is nothing to do */
 	if (test_and_set_bit(ICE_SUSPENDED, pf->state)) {
 		if (!disabled)
@@ -6208,6 +6293,8 @@ static void ice_rebuild(struct ice_pf *pf, enum ice_reset_req reset_type)
 
 	/* if we get here, reset flow is successful */
 	clear_bit(ICE_RESET_FAILED, pf->state);
+
+	ice_plug_aux_dev(pf);
 	return;
 
 err_vsi_rebuild:
@@ -6246,7 +6333,9 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu)
 	struct ice_netdev_priv *np = netdev_priv(netdev);
 	struct ice_vsi *vsi = np->vsi;
 	struct ice_pf *pf = vsi->back;
+	struct iidc_event *event;
 	u8 count = 0;
+	int err = 0;
 
 	if (new_mtu == (int)netdev->mtu) {
 		netdev_warn(netdev, "MTU is already %u\n", netdev->mtu);
@@ -6279,27 +6368,38 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu)
 		return -EBUSY;
 	}
 
+	event = kzalloc(sizeof(*event), GFP_KERNEL);
+	if (!event)
+		return -ENOMEM;
+
+	set_bit(IIDC_EVENT_BEFORE_MTU_CHANGE, event->type);
+	ice_send_event_to_aux(pf, event);
+	clear_bit(IIDC_EVENT_BEFORE_MTU_CHANGE, event->type);
+
 	netdev->mtu = (unsigned int)new_mtu;
 
 	/* if VSI is up, bring it down and then back up */
 	if (!test_and_set_bit(ICE_VSI_DOWN, vsi->state)) {
-		int err;
-
 		err = ice_down(vsi);
 		if (err) {
 			netdev_err(netdev, "change MTU if_down err %d\n", err);
-			return err;
+			goto event_after;
 		}
 
 		err = ice_up(vsi);
 		if (err) {
 			netdev_err(netdev, "change MTU if_up err %d\n", err);
-			return err;
+			goto event_after;
 		}
 	}
 
 	netdev_dbg(netdev, "changed MTU to %d\n", new_mtu);
-	return 0;
+event_after:
+	set_bit(IIDC_EVENT_AFTER_MTU_CHANGE, event->type);
+	ice_send_event_to_aux(pf, event);
+	kfree(event);
+
+	return err;
 }
 
 /**

drivers/net/ethernet/intel/ice/ice_sched.c

@@ -595,6 +595,50 @@ ice_alloc_lan_q_ctx(struct ice_hw *hw, u16 vsi_handle, u8 tc, u16 new_numqs)
 	return 0;
 }
 
+/**
+ * ice_alloc_rdma_q_ctx - allocate RDMA queue contexts for the given VSI and TC
+ * @hw: pointer to the HW struct
+ * @vsi_handle: VSI handle
+ * @tc: TC number
+ * @new_numqs: number of queues
+ */
+static enum ice_status
+ice_alloc_rdma_q_ctx(struct ice_hw *hw, u16 vsi_handle, u8 tc, u16 new_numqs)
+{
+	struct ice_vsi_ctx *vsi_ctx;
+	struct ice_q_ctx *q_ctx;
+
+	vsi_ctx = ice_get_vsi_ctx(hw, vsi_handle);
+	if (!vsi_ctx)
+		return ICE_ERR_PARAM;
+	/* allocate RDMA queue contexts */
+	if (!vsi_ctx->rdma_q_ctx[tc]) {
+		vsi_ctx->rdma_q_ctx[tc] = devm_kcalloc(ice_hw_to_dev(hw),
+						       new_numqs,
+						       sizeof(*q_ctx),
+						       GFP_KERNEL);
+		if (!vsi_ctx->rdma_q_ctx[tc])
+			return ICE_ERR_NO_MEMORY;
+		vsi_ctx->num_rdma_q_entries[tc] = new_numqs;
+		return 0;
+	}
+	/* num queues are increased, update the queue contexts */
+	if (new_numqs > vsi_ctx->num_rdma_q_entries[tc]) {
+		u16 prev_num = vsi_ctx->num_rdma_q_entries[tc];
+
+		q_ctx = devm_kcalloc(ice_hw_to_dev(hw), new_numqs,
+				     sizeof(*q_ctx), GFP_KERNEL);
+		if (!q_ctx)
+			return ICE_ERR_NO_MEMORY;
+		memcpy(q_ctx, vsi_ctx->rdma_q_ctx[tc],
+		       prev_num * sizeof(*q_ctx));
+		devm_kfree(ice_hw_to_dev(hw), vsi_ctx->rdma_q_ctx[tc]);
+		vsi_ctx->rdma_q_ctx[tc] = q_ctx;
+		vsi_ctx->num_rdma_q_entries[tc] = new_numqs;
+	}
+	return 0;
+}
+
 /**
  * ice_aq_rl_profile - performs a rate limiting task
  * @hw: pointer to the HW struct
@@ -1774,13 +1818,22 @@ ice_sched_update_vsi_child_nodes(struct ice_port_info *pi, u16 vsi_handle,
 	if (!vsi_ctx)
 		return ICE_ERR_PARAM;
 
-	prev_numqs = vsi_ctx->sched.max_lanq[tc];
+	if (owner == ICE_SCHED_NODE_OWNER_LAN)
+		prev_numqs = vsi_ctx->sched.max_lanq[tc];
+	else
+		prev_numqs = vsi_ctx->sched.max_rdmaq[tc];
 	/* num queues are not changed or less than the previous number */
 	if (new_numqs <= prev_numqs)
 		return status;
-	status = ice_alloc_lan_q_ctx(hw, vsi_handle, tc, new_numqs);
-	if (status)
-		return status;
+	if (owner == ICE_SCHED_NODE_OWNER_LAN) {
+		status = ice_alloc_lan_q_ctx(hw, vsi_handle, tc, new_numqs);
+		if (status)
+			return status;
+	} else {
+		status = ice_alloc_rdma_q_ctx(hw, vsi_handle, tc, new_numqs);
+		if (status)
+			return status;
+	}
 
 	if (new_numqs)
 		ice_sched_calc_vsi_child_nodes(hw, new_numqs, new_num_nodes);
@@ -1795,7 +1848,10 @@ ice_sched_update_vsi_child_nodes(struct ice_port_info *pi, u16 vsi_handle,
 					       new_num_nodes, owner);
 	if (status)
 		return status;
-	vsi_ctx->sched.max_lanq[tc] = new_numqs;
+	if (owner == ICE_SCHED_NODE_OWNER_LAN)
+		vsi_ctx->sched.max_lanq[tc] = new_numqs;
+	else
+		vsi_ctx->sched.max_rdmaq[tc] = new_numqs;
 
 	return 0;
 }
@@ -1861,6 +1917,7 @@ ice_sched_cfg_vsi(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u16 maxqs,
 		 * recreate the child nodes all the time in these cases.
 		 */
 		vsi_ctx->sched.max_lanq[tc] = 0;
+		vsi_ctx->sched.max_rdmaq[tc] = 0;
 	}
 
 	/* update the VSI child nodes */
@@ -1990,6 +2047,8 @@ ice_sched_rm_vsi_cfg(struct ice_port_info *pi, u16 vsi_handle, u8 owner)
 		}
 		if (owner == ICE_SCHED_NODE_OWNER_LAN)
 			vsi_ctx->sched.max_lanq[i] = 0;
+		else
+			vsi_ctx->sched.max_rdmaq[i] = 0;
 	}
 	status = 0;
 

drivers/net/ethernet/intel/ice/ice_switch.c

 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2018, Intel Corporation. */
 
+#include "ice_lib.h"
 #include "ice_switch.h"
 
 #define ICE_ETH_DA_OFFSET		0
@@ -302,6 +303,10 @@ static void ice_clear_vsi_q_ctx(struct ice_hw *hw, u16 vsi_handle)
 			devm_kfree(ice_hw_to_dev(hw), vsi->lan_q_ctx[i]);
 			vsi->lan_q_ctx[i] = NULL;
 		}
+		if (vsi->rdma_q_ctx[i]) {
+			devm_kfree(ice_hw_to_dev(hw), vsi->rdma_q_ctx[i]);
+			vsi->rdma_q_ctx[i] = NULL;
+		}
 	}
 }
 
@@ -422,6 +427,29 @@ ice_update_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
 	return ice_aq_update_vsi(hw, vsi_ctx, cd);
 }
 
+/**
+ * ice_cfg_rdma_fltr - enable/disable RDMA filtering on VSI
+ * @hw: pointer to HW struct
+ * @vsi_handle: VSI SW index
+ * @enable: boolean for enable/disable
+ */
+int
+ice_cfg_rdma_fltr(struct ice_hw *hw, u16 vsi_handle, bool enable)
+{
+	struct ice_vsi_ctx *ctx;
+
+	ctx = ice_get_vsi_ctx(hw, vsi_handle);
+	if (!ctx)
+		return -EIO;
+
+	if (enable)
+		ctx->info.q_opt_flags |= ICE_AQ_VSI_Q_OPT_PE_FLTR_EN;
+	else
+		ctx->info.q_opt_flags &= ~ICE_AQ_VSI_Q_OPT_PE_FLTR_EN;
+
+	return ice_status_to_errno(ice_update_vsi(hw, vsi_handle, ctx, NULL));
+}
+
 /**
  * ice_aq_alloc_free_vsi_list
  * @hw: pointer to the HW struct

drivers/net/ethernet/intel/ice/ice_switch.h

@@ -26,6 +26,8 @@ struct ice_vsi_ctx {
 	u8 vf_num;
 	u16 num_lan_q_entries[ICE_MAX_TRAFFIC_CLASS];
 	struct ice_q_ctx *lan_q_ctx[ICE_MAX_TRAFFIC_CLASS];
+	u16 num_rdma_q_entries[ICE_MAX_TRAFFIC_CLASS];
+	struct ice_q_ctx *rdma_q_ctx[ICE_MAX_TRAFFIC_CLASS];
 };
 
 enum ice_sw_fwd_act_type {
@@ -223,6 +225,8 @@ enum ice_status
 ice_add_eth_mac(struct ice_hw *hw, struct list_head *em_list);
 enum ice_status
 ice_remove_eth_mac(struct ice_hw *hw, struct list_head *em_list);
+int
+ice_cfg_rdma_fltr(struct ice_hw *hw, u16 vsi_handle, bool enable);
 void ice_remove_vsi_fltr(struct ice_hw *hw, u16 vsi_handle);
 enum ice_status
 ice_add_vlan(struct ice_hw *hw, struct list_head *m_list);

drivers/net/ethernet/intel/ice/ice_type.h

@@ -45,6 +45,7 @@ static inline u32 ice_round_to_num(u32 N, u32 R)
 #define ICE_DBG_FLOW		BIT_ULL(9)
 #define ICE_DBG_SW		BIT_ULL(13)
 #define ICE_DBG_SCHED		BIT_ULL(14)
+#define ICE_DBG_RDMA		BIT_ULL(15)
 #define ICE_DBG_PKG		BIT_ULL(16)
 #define ICE_DBG_RES		BIT_ULL(17)
 #define ICE_DBG_AQ_MSG		BIT_ULL(24)
@@ -262,6 +263,7 @@ struct ice_hw_common_caps {
 	u8 rss_table_entry_width;	/* RSS Entry width in bits */
 
 	u8 dcb;
+	u8 rdma;
 
 	bool nvm_update_pending_nvm;
 	bool nvm_update_pending_orom;
@@ -440,6 +442,7 @@ struct ice_sched_node {
 	u8 tc_num;
 	u8 owner;
 #define ICE_SCHED_NODE_OWNER_LAN	0
+#define ICE_SCHED_NODE_OWNER_RDMA	2
 };
 
 /* Access Macros for Tx Sched Elements data */
@@ -511,6 +514,7 @@ struct ice_sched_vsi_info {
 	struct ice_sched_node *ag_node[ICE_MAX_TRAFFIC_CLASS];
 	struct list_head list_entry;
 	u16 max_lanq[ICE_MAX_TRAFFIC_CLASS];
+	u16 max_rdmaq[ICE_MAX_TRAFFIC_CLASS];
 };
 
 /* driver defines the policy */

include/linux/net/intel/i40e_client.h

@@ -4,6 +4,8 @@
 #ifndef _I40E_CLIENT_H_
 #define _I40E_CLIENT_H_
 
+#include <linux/auxiliary_bus.h>
+
 #define I40E_CLIENT_STR_LENGTH 10
 
 /* Client interface version should be updated anytime there is a change in the
@@ -48,7 +50,7 @@ struct i40e_qv_info {
 
 struct i40e_qvlist_info {
 	u32 num_vectors;
-	struct i40e_qv_info qv_info[1];
+	struct i40e_qv_info qv_info[];
 };
 
 
@@ -78,6 +80,7 @@ struct i40e_info {
 	u8 lanmac[6];
 	struct net_device *netdev;
 	struct pci_dev *pcidev;
+	struct auxiliary_device *aux_dev;
 	u8 __iomem *hw_addr;
 	u8 fid;	/* function id, PF id or VF id */
 #define I40E_CLIENT_FTYPE_PF 0
@@ -100,6 +103,11 @@ struct i40e_info {
 	u32 fw_build;                   /* firmware build number */
 };
 
+struct i40e_auxiliary_device {
+	struct auxiliary_device aux_dev;
+	struct i40e_info *ldev;
+};
+
 #define I40E_CLIENT_RESET_LEVEL_PF   1
 #define I40E_CLIENT_RESET_LEVEL_CORE 2
 #define I40E_CLIENT_VSI_FLAG_TCP_ENABLE  BIT(1)
@@ -187,6 +195,8 @@ static inline bool i40e_client_is_registered(struct i40e_client *client)
 	return test_bit(__I40E_CLIENT_REGISTERED, &client->state);
 }
 
+void i40e_client_device_register(struct i40e_info *ldev, struct i40e_client *client);
+void i40e_client_device_unregister(struct i40e_info *ldev);
 /* used by clients */
 int i40e_register_client(struct i40e_client *client);
 int i40e_unregister_client(struct i40e_client *client);

include/linux/net/intel/iidc.h

+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (C) 2021, Intel Corporation. */
+
+#ifndef _IIDC_H_
+#define _IIDC_H_
+
+#include <linux/auxiliary_bus.h>
+#include <linux/dcbnl.h>
+#include <linux/device.h>
+#include <linux/if_ether.h>
+#include <linux/kernel.h>
+#include <linux/netdevice.h>
+
+enum iidc_event_type {
+	IIDC_EVENT_BEFORE_MTU_CHANGE,
+	IIDC_EVENT_AFTER_MTU_CHANGE,
+	IIDC_EVENT_BEFORE_TC_CHANGE,
+	IIDC_EVENT_AFTER_TC_CHANGE,
+	IIDC_EVENT_CRIT_ERR,
+	IIDC_EVENT_NBITS		/* must be last */
+};
+
+enum iidc_reset_type {
+	IIDC_PFR,
+	IIDC_CORER,
+	IIDC_GLOBR,
+};
+
+#define IIDC_MAX_USER_PRIORITY		8
+
+/* Struct to hold per RDMA Qset info */
+struct iidc_rdma_qset_params {
+	/* Qset TEID returned to the RDMA driver in
+	 * ice_add_rdma_qset and used by RDMA driver
+	 * for calls to ice_del_rdma_qset
+	 */
+	u32 teid;	/* Qset TEID */
+	u16 qs_handle; /* RDMA driver provides this */
+	u16 vport_id; /* VSI index */
+	u8 tc; /* TC branch the Qset should belong to */
+};
+
+struct iidc_qos_info {
+	u64 tc_ctx;
+	u8 rel_bw;
+	u8 prio_type;
+	u8 egress_virt_up;
+	u8 ingress_virt_up;
+};
+
+/* Struct to pass QoS info */
+struct iidc_qos_params {
+	struct iidc_qos_info tc_info[IEEE_8021QAZ_MAX_TCS];
+	u8 up2tc[IIDC_MAX_USER_PRIORITY];
+	u8 vport_relative_bw;
+	u8 vport_priority_type;
+	u8 num_tc;
+};
+
+struct iidc_event {
+	DECLARE_BITMAP(type, IIDC_EVENT_NBITS);
+	u32 reg;
+};
+
+struct ice_pf;
+
+int ice_add_rdma_qset(struct ice_pf *pf, struct iidc_rdma_qset_params *qset);
+int ice_del_rdma_qset(struct ice_pf *pf, struct iidc_rdma_qset_params *qset);
+int ice_rdma_request_reset(struct ice_pf *pf, enum iidc_reset_type reset_type);
+int ice_rdma_update_vsi_filter(struct ice_pf *pf, u16 vsi_id, bool enable);
+void ice_get_qos_params(struct ice_pf *pf, struct iidc_qos_params *qos);
+
+#define IIDC_RDMA_ROCE_NAME	"roce"
+
+/* Structure representing auxiliary driver tailored information about the core
+ * PCI dev, each auxiliary driver using the IIDC interface will have an
+ * instance of this struct dedicated to it.
+ */
+
+struct iidc_auxiliary_dev {
+	struct auxiliary_device adev;
+	struct ice_pf *pf;
+};
+
+/* structure representing the auxiliary driver. This struct is to be
+ * allocated and populated by the auxiliary driver's owner. The core PCI
+ * driver will access these ops by performing a container_of on the
+ * auxiliary_device->dev.driver.
+ */
+struct iidc_auxiliary_drv {
+	struct auxiliary_driver adrv;
+	/* This event_handler is meant to be a blocking call.  For instance,
+	 * when a BEFORE_MTU_CHANGE event comes in, the event_handler will not
+	 * return until the auxiliary driver is ready for the MTU change to
+	 * happen.
+	 */
+	void (*event_handler)(struct ice_pf *pf, struct iidc_event *event);
+};
+
+#endif /* _IIDC_H_*/