[PATCH net-next 00/12][pull request] 100GbE Intel Wired LAN Driver Updates 2021-04-22

[PATCH net-next 00/12][pull request] 100GbE Intel Wired LAN Driver Updates 2021-04-22

[Tony Nguyen]

This series contains updates to virtchnl header file, ice, and iavf
drivers.

Vignesh adds support to warn about potentially malicious VFs; those that
are overflowing the mailbox for the ice driver.

Michal adds support for an allowlist/denylist of VF commands based on
supported capabilities for the ice driver.

Brett adds support for iavf UDP segmentation offload by adding the
capability bit to virtchnl, advertising support in the ice driver, and
enabling it in the iavf driver. He also adds a helper function for
getting the VF VSI for ice.

Colin Ian King removes an unneeded pointer assignment.

Qi enables support in the ice driver to support virtchnl requests from
the iavf to configure its own RSS input set. This includes adding new
capability bits, structures, and commands to virtchnl header file.

Haiyue enables configuring RSS flow hash via ethtool to support TCP, UDP
and SCTP protocols in iavf.

The following are changes since commit 5d869070569a23aa909c6e7e9d010fc438a492ef:
  net: phy: marvell: don't use empty switch default case
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue 100GbE

Brett Creeley (3):
  ice: Advertise virtchnl UDP segmentation offload capability
  iavf: add support for UDP Segmentation Offload
  ice: Add helper function to get the VF's VSI

Colin Ian King (1):
  ice: remove redundant assignment to pointer vsi

Haiyue Wang (4):
  iavf: Add framework to enable ethtool RSS config
  iavf: Support for modifying TCP RSS flow hashing
  iavf: Support for modifying UDP RSS flow hashing
  iavf: Support for modifying SCTP RSS flow hashing

Michal Swiatkowski (1):
  ice: Allow ignoring opcodes on specific VF

Qi Zhang (2):
  ice: Enable RSS configure for AVF
  ice: Support RSS configure removal for AVF

Vignesh Sridhar (1):
  ice: warn about potentially malicious VFs

 drivers/net/ethernet/intel/iavf/Makefile      |   1 +
 drivers/net/ethernet/intel/iavf/iavf.h        |  10 +
 .../net/ethernet/intel/iavf/iavf_adv_rss.c    | 218 ++++++
 .../net/ethernet/intel/iavf/iavf_adv_rss.h    |  95 +++
 .../net/ethernet/intel/iavf/iavf_ethtool.c    | 252 ++++++-
 drivers/net/ethernet/intel/iavf/iavf_main.c   |  29 +
 drivers/net/ethernet/intel/iavf/iavf_txrx.c   |  15 +-
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   | 164 +++++
 drivers/net/ethernet/intel/ice/Makefile       |   1 +
 drivers/net/ethernet/intel/ice/ice.h          |   1 +
 drivers/net/ethernet/intel/ice/ice_flow.c     |  88 +++
 drivers/net/ethernet/intel/ice/ice_flow.h     |   6 +
 drivers/net/ethernet/intel/ice/ice_main.c     |   7 +-
 drivers/net/ethernet/intel/ice/ice_sriov.c    | 400 ++++++++++-
 drivers/net/ethernet/intel/ice/ice_sriov.h    |  20 +-
 drivers/net/ethernet/intel/ice/ice_type.h     |  75 ++
 .../intel/ice/ice_virtchnl_allowlist.c        | 171 +++++
 .../intel/ice/ice_virtchnl_allowlist.h        |  13 +
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  | 663 ++++++++++++++++--
 .../net/ethernet/intel/ice/ice_virtchnl_pf.h  |  13 +
 include/linux/avf/virtchnl.h                  |  27 +-
 21 files changed, 2214 insertions(+), 55 deletions(-)
 create mode 100644 drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
 create mode 100644 drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
 create mode 100644 drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
 create mode 100644 drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.h

-- 
2.26.2

[PATCH net-next 01/12] ice: warn about potentially malicious VFs

[Tony Nguyen]

From: Vignesh Sridhar <vignesh.sridhar@intel.com>

Attempt to detect malicious VFs and, if suspected, log the information but
keep going to allow the user to take any desired actions.

Potentially malicious VFs are identified by checking if the VFs are
transmitting too many messages via the PF-VF mailbox which could cause an
overflow of this channel resulting in denial of service. This is done by
creating a snapshot or static capture of the mailbox buffer which can be
traversed and in which the messages sent by VFs are tracked.

Co-developed-by: Yashaswini Raghuram Prathivadi Bhayankaram <yashaswini.raghuram.prathivadi.bhayankaram@intel.com>
Signed-off-by: Yashaswini Raghuram Prathivadi Bhayankaram <yashaswini.raghuram.prathivadi.bhayankaram@intel.com>
Co-developed-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Co-developed-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Vignesh Sridhar <vignesh.sridhar@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/ice/ice.h          |   1 +
 drivers/net/ethernet/intel/ice/ice_main.c     |   7 +-
 drivers/net/ethernet/intel/ice/ice_sriov.c    | 400 +++++++++++++++++-
 drivers/net/ethernet/intel/ice/ice_sriov.h    |  20 +-
 drivers/net/ethernet/intel/ice/ice_type.h     |  75 ++++
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  |  94 +++-
 .../net/ethernet/intel/ice/ice_virtchnl_pf.h  |  12 +
 7 files changed, 605 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 7ae10fd87265..e35db3ff583b 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -426,6 +426,7 @@ struct ice_pf {
 	u16 num_msix_per_vf;
 	/* used to ratelimit the MDD event logging */
 	unsigned long last_printed_mdd_jiffies;
+	DECLARE_BITMAP(malvfs, ICE_MAX_VF_COUNT);
 	DECLARE_BITMAP(state, ICE_STATE_NBITS);
 	DECLARE_BITMAP(flags, ICE_PF_FLAGS_NBITS);
 	unsigned long *avail_txqs;	/* bitmap to track PF Tx queue usage */
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 6dbaa9099fdf..4ee85a217c6f 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1193,6 +1193,10 @@ static int __ice_clean_ctrlq(struct ice_pf *pf, enum ice_ctl_q q_type)
 	case ICE_CTL_Q_MAILBOX:
 		cq = &hw->mailboxq;
 		qtype = "Mailbox";
+		/* we are going to try to detect a malicious VF, so set the
+		 * state to begin detection
+		 */
+		hw->mbx_snapshot.mbx_buf.state = ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT;
 		break;
 	default:
 		dev_warn(dev, "Unknown control queue type 0x%x\n", q_type);
@@ -1274,7 +1278,8 @@ static int __ice_clean_ctrlq(struct ice_pf *pf, enum ice_ctl_q q_type)
 			ice_vf_lan_overflow_event(pf, &event);
 			break;
 		case ice_mbx_opc_send_msg_to_pf:
-			ice_vc_process_vf_msg(pf, &event);
+			if (!ice_is_malicious_vf(pf, &event, i, pending))
+				ice_vc_process_vf_msg(pf, &event);
 			break;
 		case ice_aqc_opc_fw_logging:
 			ice_output_fw_log(hw, &event.desc, event.msg_buf);
diff --git a/drivers/net/ethernet/intel/ice/ice_sriov.c b/drivers/net/ethernet/intel/ice/ice_sriov.c
index 554f567476f3..aa11d07793d4 100644
--- a/drivers/net/ethernet/intel/ice/ice_sriov.c
+++ b/drivers/net/ethernet/intel/ice/ice_sriov.c
@@ -2,7 +2,6 @@
 /* Copyright (c) 2018, Intel Corporation. */
 
 #include "ice_common.h"
-#include "ice_adminq_cmd.h"
 #include "ice_sriov.h"
 
 /**
@@ -132,3 +131,402 @@ u32 ice_conv_link_speed_to_virtchnl(bool adv_link_support, u16 link_speed)
 
 	return speed;
 }
+
+/* The mailbox overflow detection algorithm helps to check if there
+ * is a possibility of a malicious VF transmitting too many MBX messages to the
+ * PF.
+ * 1. The mailbox snapshot structure, ice_mbx_snapshot, is initialized during
+ * driver initialization in ice_init_hw() using ice_mbx_init_snapshot().
+ * The struct ice_mbx_snapshot helps to track and traverse a static window of
+ * messages within the mailbox queue while looking for a malicious VF.
+ *
+ * 2. When the caller starts processing its mailbox queue in response to an
+ * interrupt, the structure ice_mbx_snapshot is expected to be cleared before
+ * the algorithm can be run for the first time for that interrupt. This can be
+ * done via ice_mbx_reset_snapshot().
+ *
+ * 3. For every message read by the caller from the MBX Queue, the caller must
+ * call the detection algorithm's entry function ice_mbx_vf_state_handler().
+ * Before every call to ice_mbx_vf_state_handler() the struct ice_mbx_data is
+ * filled as it is required to be passed to the algorithm.
+ *
+ * 4. Every time a message is read from the MBX queue, a VFId is received which
+ * is passed to the state handler. The boolean output is_malvf of the state
+ * handler ice_mbx_vf_state_handler() serves as an indicator to the caller
+ * whether this VF is malicious or not.
+ *
+ * 5. When a VF is identified to be malicious, the caller can send a message
+ * to the system administrator. The caller can invoke ice_mbx_report_malvf()
+ * to help determine if a malicious VF is to be reported or not. This function
+ * requires the caller to maintain a global bitmap to track all malicious VFs
+ * and pass that to ice_mbx_report_malvf() along with the VFID which was identified
+ * to be malicious by ice_mbx_vf_state_handler().
+ *
+ * 6. The global bitmap maintained by PF can be cleared completely if PF is in
+ * reset or the bit corresponding to a VF can be cleared if that VF is in reset.
+ * When a VF is shut down and brought back up, we assume that the new VF
+ * brought up is not malicious and hence report it if found malicious.
+ *
+ * 7. The function ice_mbx_reset_snapshot() is called to reset the information
+ * in ice_mbx_snapshot for every new mailbox interrupt handled.
+ *
+ * 8. The memory allocated for variables in ice_mbx_snapshot is de-allocated
+ * when driver is unloaded.
+ */
+#define ICE_RQ_DATA_MASK(rq_data) ((rq_data) & PF_MBX_ARQH_ARQH_M)
+/* Using the highest value for an unsigned 16-bit value 0xFFFF to indicate that
+ * the max messages check must be ignored in the algorithm
+ */
+#define ICE_IGNORE_MAX_MSG_CNT	0xFFFF
+
+/**
+ * ice_mbx_traverse - Pass through mailbox snapshot
+ * @hw: pointer to the HW struct
+ * @new_state: new algorithm state
+ *
+ * Traversing the mailbox static snapshot without checking
+ * for malicious VFs.
+ */
+static void
+ice_mbx_traverse(struct ice_hw *hw,
+		 enum ice_mbx_snapshot_state *new_state)
+{
+	struct ice_mbx_snap_buffer_data *snap_buf;
+	u32 num_iterations;
+
+	snap_buf = &hw->mbx_snapshot.mbx_buf;
+
+	/* As mailbox buffer is circular, applying a mask
+	 * on the incremented iteration count.
+	 */
+	num_iterations = ICE_RQ_DATA_MASK(++snap_buf->num_iterations);
+
+	/* Checking either of the below conditions to exit snapshot traversal:
+	 * Condition-1: If the number of iterations in the mailbox is equal to
+	 * the mailbox head which would indicate that we have reached the end
+	 * of the static snapshot.
+	 * Condition-2: If the maximum messages serviced in the mailbox for a
+	 * given interrupt is the highest possible value then there is no need
+	 * to check if the number of messages processed is equal to it. If not
+	 * check if the number of messages processed is greater than or equal
+	 * to the maximum number of mailbox entries serviced in current work item.
+	 */
+	if (num_iterations == snap_buf->head ||
+	    (snap_buf->max_num_msgs_mbx < ICE_IGNORE_MAX_MSG_CNT &&
+	     ++snap_buf->num_msg_proc >= snap_buf->max_num_msgs_mbx))
+		*new_state = ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT;
+}
+
+/**
+ * ice_mbx_detect_malvf - Detect malicious VF in snapshot
+ * @hw: pointer to the HW struct
+ * @vf_id: relative virtual function ID
+ * @new_state: new algorithm state
+ * @is_malvf: boolean output to indicate if VF is malicious
+ *
+ * This function tracks the number of asynchronous messages
+ * sent per VF and marks the VF as malicious if it exceeds
+ * the permissible number of messages to send.
+ */
+static enum ice_status
+ice_mbx_detect_malvf(struct ice_hw *hw, u16 vf_id,
+		     enum ice_mbx_snapshot_state *new_state,
+		     bool *is_malvf)
+{
+	struct ice_mbx_snapshot *snap = &hw->mbx_snapshot;
+
+	if (vf_id >= snap->mbx_vf.vfcntr_len)
+		return ICE_ERR_OUT_OF_RANGE;
+
+	/* increment the message count in the VF array */
+	snap->mbx_vf.vf_cntr[vf_id]++;
+
+	if (snap->mbx_vf.vf_cntr[vf_id] >= ICE_ASYNC_VF_MSG_THRESHOLD)
+		*is_malvf = true;
+
+	/* continue to iterate through the mailbox snapshot */
+	ice_mbx_traverse(hw, new_state);
+
+	return 0;
+}
+
+/**
+ * ice_mbx_reset_snapshot - Reset mailbox snapshot structure
+ * @snap: pointer to mailbox snapshot structure in the ice_hw struct
+ *
+ * Reset the mailbox snapshot structure and clear VF counter array.
+ */
+static void ice_mbx_reset_snapshot(struct ice_mbx_snapshot *snap)
+{
+	u32 vfcntr_len;
+
+	if (!snap || !snap->mbx_vf.vf_cntr)
+		return;
+
+	/* Clear VF counters. */
+	vfcntr_len = snap->mbx_vf.vfcntr_len;
+	if (vfcntr_len)
+		memset(snap->mbx_vf.vf_cntr, 0,
+		       (vfcntr_len * sizeof(*snap->mbx_vf.vf_cntr)));
+
+	/* Reset mailbox snapshot for a new capture. */
+	memset(&snap->mbx_buf, 0, sizeof(snap->mbx_buf));
+	snap->mbx_buf.state = ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT;
+}
+
+/**
+ * ice_mbx_vf_state_handler - Handle states of the overflow algorithm
+ * @hw: pointer to the HW struct
+ * @mbx_data: pointer to structure containing mailbox data
+ * @vf_id: relative virtual function (VF) ID
+ * @is_malvf: boolean output to indicate if VF is malicious
+ *
+ * The function serves as an entry point for the malicious VF
+ * detection algorithm by handling the different states and state
+ * transitions of the algorithm:
+ * New snapshot: This state is entered when creating a new static
+ * snapshot. The data from any previous mailbox snapshot is
+ * cleared and a new capture of the mailbox head and tail is
+ * logged. This will be the new static snapshot to detect
+ * asynchronous messages sent by VFs. On capturing the snapshot
+ * and depending on whether the number of pending messages in that
+ * snapshot exceed the watermark value, the state machine enters
+ * traverse or detect states.
+ * Traverse: If pending message count is below watermark then iterate
+ * through the snapshot without any action on VF.
+ * Detect: If pending message count exceeds watermark traverse
+ * the static snapshot and look for a malicious VF.
+ */
+enum ice_status
+ice_mbx_vf_state_handler(struct ice_hw *hw,
+			 struct ice_mbx_data *mbx_data, u16 vf_id,
+			 bool *is_malvf)
+{
+	struct ice_mbx_snapshot *snap = &hw->mbx_snapshot;
+	struct ice_mbx_snap_buffer_data *snap_buf;
+	struct ice_ctl_q_info *cq = &hw->mailboxq;
+	enum ice_mbx_snapshot_state new_state;
+	enum ice_status status = 0;
+
+	if (!is_malvf || !mbx_data)
+		return ICE_ERR_BAD_PTR;
+
+	/* When entering the mailbox state machine assume that the VF
+	 * is not malicious until detected.
+	 */
+	*is_malvf = false;
+
+	 /* Checking if max messages allowed to be processed while servicing current
+	  * interrupt is not less than the defined AVF message threshold.
+	  */
+	if (mbx_data->max_num_msgs_mbx <= ICE_ASYNC_VF_MSG_THRESHOLD)
+		return ICE_ERR_INVAL_SIZE;
+
+	/* The watermark value should not be lesser than the threshold limit
+	 * set for the number of asynchronous messages a VF can send to mailbox
+	 * nor should it be greater than the maximum number of messages in the
+	 * mailbox serviced in current interrupt.
+	 */
+	if (mbx_data->async_watermark_val < ICE_ASYNC_VF_MSG_THRESHOLD ||
+	    mbx_data->async_watermark_val > mbx_data->max_num_msgs_mbx)
+		return ICE_ERR_PARAM;
+
+	new_state = ICE_MAL_VF_DETECT_STATE_INVALID;
+	snap_buf = &snap->mbx_buf;
+
+	switch (snap_buf->state) {
+	case ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT:
+		/* Clear any previously held data in mailbox snapshot structure. */
+		ice_mbx_reset_snapshot(snap);
+
+		/* Collect the pending ARQ count, number of messages processed and
+		 * the maximum number of messages allowed to be processed from the
+		 * Mailbox for current interrupt.
+		 */
+		snap_buf->num_pending_arq = mbx_data->num_pending_arq;
+		snap_buf->num_msg_proc = mbx_data->num_msg_proc;
+		snap_buf->max_num_msgs_mbx = mbx_data->max_num_msgs_mbx;
+
+		/* Capture a new static snapshot of the mailbox by logging the
+		 * head and tail of snapshot and set num_iterations to the tail
+		 * value to mark the start of the iteration through the snapshot.
+		 */
+		snap_buf->head = ICE_RQ_DATA_MASK(cq->rq.next_to_clean +
+						  mbx_data->num_pending_arq);
+		snap_buf->tail = ICE_RQ_DATA_MASK(cq->rq.next_to_clean - 1);
+		snap_buf->num_iterations = snap_buf->tail;
+
+		/* Pending ARQ messages returned by ice_clean_rq_elem
+		 * is the difference between the head and tail of the
+		 * mailbox queue. Comparing this value against the watermark
+		 * helps to check if we potentially have malicious VFs.
+		 */
+		if (snap_buf->num_pending_arq >=
+		    mbx_data->async_watermark_val) {
+			new_state = ICE_MAL_VF_DETECT_STATE_DETECT;
+			status = ice_mbx_detect_malvf(hw, vf_id, &new_state, is_malvf);
+		} else {
+			new_state = ICE_MAL_VF_DETECT_STATE_TRAVERSE;
+			ice_mbx_traverse(hw, &new_state);
+		}
+		break;
+
+	case ICE_MAL_VF_DETECT_STATE_TRAVERSE:
+		new_state = ICE_MAL_VF_DETECT_STATE_TRAVERSE;
+		ice_mbx_traverse(hw, &new_state);
+		break;
+
+	case ICE_MAL_VF_DETECT_STATE_DETECT:
+		new_state = ICE_MAL_VF_DETECT_STATE_DETECT;
+		status = ice_mbx_detect_malvf(hw, vf_id, &new_state, is_malvf);
+		break;
+
+	default:
+		new_state = ICE_MAL_VF_DETECT_STATE_INVALID;
+		status = ICE_ERR_CFG;
+	}
+
+	snap_buf->state = new_state;
+
+	return status;
+}
+
+/**
+ * ice_mbx_report_malvf - Track and note malicious VF
+ * @hw: pointer to the HW struct
+ * @all_malvfs: all malicious VFs tracked by PF
+ * @bitmap_len: length of bitmap in bits
+ * @vf_id: relative virtual function ID of the malicious VF
+ * @report_malvf: boolean to indicate if malicious VF must be reported
+ *
+ * This function will update a bitmap that keeps track of the malicious
+ * VFs attached to the PF. A malicious VF must be reported only once if
+ * discovered between VF resets or loading so the function checks
+ * the input vf_id against the bitmap to verify if the VF has been
+ * detected in any previous mailbox iterations.
+ */
+enum ice_status
+ice_mbx_report_malvf(struct ice_hw *hw, unsigned long *all_malvfs,
+		     u16 bitmap_len, u16 vf_id, bool *report_malvf)
+{
+	if (!all_malvfs || !report_malvf)
+		return ICE_ERR_PARAM;
+
+	*report_malvf = false;
+
+	if (bitmap_len < hw->mbx_snapshot.mbx_vf.vfcntr_len)
+		return ICE_ERR_INVAL_SIZE;
+
+	if (vf_id >= bitmap_len)
+		return ICE_ERR_OUT_OF_RANGE;
+
+	/* If the vf_id is found in the bitmap set bit and boolean to true */
+	if (!test_and_set_bit(vf_id, all_malvfs))
+		*report_malvf = true;
+
+	return 0;
+}
+
+/**
+ * ice_mbx_clear_malvf - Clear VF bitmap and counter for VF ID
+ * @snap: pointer to the mailbox snapshot structure
+ * @all_malvfs: all malicious VFs tracked by PF
+ * @bitmap_len: length of bitmap in bits
+ * @vf_id: relative virtual function ID of the malicious VF
+ *
+ * In case of a VF reset, this function can be called to clear
+ * the bit corresponding to the VF ID in the bitmap tracking all
+ * malicious VFs attached to the PF. The function also clears the
+ * VF counter array at the index of the VF ID. This is to ensure
+ * that the new VF loaded is not considered malicious before going
+ * through the overflow detection algorithm.
+ */
+enum ice_status
+ice_mbx_clear_malvf(struct ice_mbx_snapshot *snap, unsigned long *all_malvfs,
+		    u16 bitmap_len, u16 vf_id)
+{
+	if (!snap || !all_malvfs)
+		return ICE_ERR_PARAM;
+
+	if (bitmap_len < snap->mbx_vf.vfcntr_len)
+		return ICE_ERR_INVAL_SIZE;
+
+	/* Ensure VF ID value is not larger than bitmap or VF counter length */
+	if (vf_id >= bitmap_len || vf_id >= snap->mbx_vf.vfcntr_len)
+		return ICE_ERR_OUT_OF_RANGE;
+
+	/* Clear VF ID bit in the bitmap tracking malicious VFs attached to PF */
+	clear_bit(vf_id, all_malvfs);
+
+	/* Clear the VF counter in the mailbox snapshot structure for that VF ID.
+	 * This is to ensure that if a VF is unloaded and a new one brought back
+	 * up with the same VF ID for a snapshot currently in traversal or detect
+	 * state the counter for that VF ID does not increment on top of existing
+	 * values in the mailbox overflow detection algorithm.
+	 */
+	snap->mbx_vf.vf_cntr[vf_id] = 0;
+
+	return 0;
+}
+
+/**
+ * ice_mbx_init_snapshot - Initialize mailbox snapshot structure
+ * @hw: pointer to the hardware structure
+ * @vf_count: number of VFs allocated on a PF
+ *
+ * Clear the mailbox snapshot structure and allocate memory
+ * for the VF counter array based on the number of VFs allocated
+ * on that PF.
+ *
+ * Assumption: This function will assume ice_get_caps() has already been
+ * called to ensure that the vf_count can be compared against the number
+ * of VFs supported as defined in the functional capabilities of the device.
+ */
+enum ice_status ice_mbx_init_snapshot(struct ice_hw *hw, u16 vf_count)
+{
+	struct ice_mbx_snapshot *snap = &hw->mbx_snapshot;
+
+	/* Ensure that the number of VFs allocated is non-zero and
+	 * is not greater than the number of supported VFs defined in
+	 * the functional capabilities of the PF.
+	 */
+	if (!vf_count || vf_count > hw->func_caps.num_allocd_vfs)
+		return ICE_ERR_INVAL_SIZE;
+
+	snap->mbx_vf.vf_cntr = devm_kcalloc(ice_hw_to_dev(hw), vf_count,
+					    sizeof(*snap->mbx_vf.vf_cntr),
+					    GFP_KERNEL);
+	if (!snap->mbx_vf.vf_cntr)
+		return ICE_ERR_NO_MEMORY;
+
+	/* Setting the VF counter length to the number of allocated
+	 * VFs for given PF's functional capabilities.
+	 */
+	snap->mbx_vf.vfcntr_len = vf_count;
+
+	/* Clear mbx_buf in the mailbox snaphot structure and setting the
+	 * mailbox snapshot state to a new capture.
+	 */
+	memset(&snap->mbx_buf, 0, sizeof(snap->mbx_buf));
+	snap->mbx_buf.state = ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT;
+
+	return 0;
+}
+
+/**
+ * ice_mbx_deinit_snapshot - Free mailbox snapshot structure
+ * @hw: pointer to the hardware structure
+ *
+ * Clear the mailbox snapshot structure and free the VF counter array.
+ */
+void ice_mbx_deinit_snapshot(struct ice_hw *hw)
+{
+	struct ice_mbx_snapshot *snap = &hw->mbx_snapshot;
+
+	/* Free VF counter array and reset VF counter length */
+	devm_kfree(ice_hw_to_dev(hw), snap->mbx_vf.vf_cntr);
+	snap->mbx_vf.vfcntr_len = 0;
+
+	/* Clear mbx_buf in the mailbox snaphot structure */
+	memset(&snap->mbx_buf, 0, sizeof(snap->mbx_buf));
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_sriov.h b/drivers/net/ethernet/intel/ice/ice_sriov.h
index 3d78a0795138..161dc55d9e9c 100644
--- a/drivers/net/ethernet/intel/ice/ice_sriov.h
+++ b/drivers/net/ethernet/intel/ice/ice_sriov.h
@@ -4,7 +4,14 @@
 #ifndef _ICE_SRIOV_H_
 #define _ICE_SRIOV_H_
 
-#include "ice_common.h"
+#include "ice_type.h"
+#include "ice_controlq.h"
+
+/* Defining the mailbox message threshold as 63 asynchronous
+ * pending messages. Normal VF functionality does not require
+ * sending more than 63 asynchronous pending message.
+ */
+#define ICE_ASYNC_VF_MSG_THRESHOLD	63
 
 #ifdef CONFIG_PCI_IOV
 enum ice_status
@@ -12,6 +19,17 @@ ice_aq_send_msg_to_vf(struct ice_hw *hw, u16 vfid, u32 v_opcode, u32 v_retval,
 		      u8 *msg, u16 msglen, struct ice_sq_cd *cd);
 
 u32 ice_conv_link_speed_to_virtchnl(bool adv_link_support, u16 link_speed);
+enum ice_status
+ice_mbx_vf_state_handler(struct ice_hw *hw, struct ice_mbx_data *mbx_data,
+			 u16 vf_id, bool *is_mal_vf);
+enum ice_status
+ice_mbx_clear_malvf(struct ice_mbx_snapshot *snap, unsigned long *all_malvfs,
+		    u16 bitmap_len, u16 vf_id);
+enum ice_status ice_mbx_init_snapshot(struct ice_hw *hw, u16 vf_count);
+void ice_mbx_deinit_snapshot(struct ice_hw *hw);
+enum ice_status
+ice_mbx_report_malvf(struct ice_hw *hw, unsigned long *all_malvfs,
+		     u16 bitmap_len, u16 vf_id, bool *report_malvf);
 #else /* CONFIG_PCI_IOV */
 static inline enum ice_status
 ice_aq_send_msg_to_vf(struct ice_hw __always_unused *hw,
diff --git a/drivers/net/ethernet/intel/ice/ice_type.h b/drivers/net/ethernet/intel/ice/ice_type.h
index 9b80962ff92f..4474dd6a7ba1 100644
--- a/drivers/net/ethernet/intel/ice/ice_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_type.h
@@ -630,6 +630,80 @@ struct ice_fw_log_cfg {
 	struct ice_fw_log_evnt evnts[ICE_AQC_FW_LOG_ID_MAX];
 };
 
+/* Enum defining the different states of the mailbox snapshot in the
+ * PF-VF mailbox overflow detection algorithm. The snapshot can be in
+ * states:
+ * 1. ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT - generate a new static snapshot
+ * within the mailbox buffer.
+ * 2. ICE_MAL_VF_DETECT_STATE_TRAVERSE - iterate through the mailbox snaphot
+ * 3. ICE_MAL_VF_DETECT_STATE_DETECT - track the messages sent per VF via the
+ * mailbox and mark any VFs sending more messages than the threshold limit set.
+ * 4. ICE_MAL_VF_DETECT_STATE_INVALID - Invalid mailbox state set to 0xFFFFFFFF.
+ */
+enum ice_mbx_snapshot_state {
+	ICE_MAL_VF_DETECT_STATE_NEW_SNAPSHOT = 0,
+	ICE_MAL_VF_DETECT_STATE_TRAVERSE,
+	ICE_MAL_VF_DETECT_STATE_DETECT,
+	ICE_MAL_VF_DETECT_STATE_INVALID = 0xFFFFFFFF,
+};
+
+/* Structure to hold information of the static snapshot and the mailbox
+ * buffer data used to generate and track the snapshot.
+ * 1. state: the state of the mailbox snapshot in the malicious VF
+ * detection state handler ice_mbx_vf_state_handler()
+ * 2. head: head of the mailbox snapshot in a circular mailbox buffer
+ * 3. tail: tail of the mailbox snapshot in a circular mailbox buffer
+ * 4. num_iterations: number of messages traversed in circular mailbox buffer
+ * 5. num_msg_proc: number of messages processed in mailbox
+ * 6. num_pending_arq: number of pending asynchronous messages
+ * 7. max_num_msgs_mbx: maximum messages in mailbox for currently
+ * serviced work item or interrupt.
+ */
+struct ice_mbx_snap_buffer_data {
+	enum ice_mbx_snapshot_state state;
+	u32 head;
+	u32 tail;
+	u32 num_iterations;
+	u16 num_msg_proc;
+	u16 num_pending_arq;
+	u16 max_num_msgs_mbx;
+};
+
+/* Structure to track messages sent by VFs on mailbox:
+ * 1. vf_cntr: a counter array of VFs to track the number of
+ * asynchronous messages sent by each VF
+ * 2. vfcntr_len: number of entries in VF counter array
+ */
+struct ice_mbx_vf_counter {
+	u32 *vf_cntr;
+	u32 vfcntr_len;
+};
+
+/* Structure to hold data relevant to the captured static snapshot
+ * of the PF-VF mailbox.
+ */
+struct ice_mbx_snapshot {
+	struct ice_mbx_snap_buffer_data mbx_buf;
+	struct ice_mbx_vf_counter mbx_vf;
+};
+
+/* Structure to hold data to be used for capturing or updating a
+ * static snapshot.
+ * 1. num_msg_proc: number of messages processed in mailbox
+ * 2. num_pending_arq: number of pending asynchronous messages
+ * 3. max_num_msgs_mbx: maximum messages in mailbox for currently
+ * serviced work item or interrupt.
+ * 4. async_watermark_val: An upper threshold set by caller to determine
+ * if the pending arq count is large enough to assume that there is
+ * the possibility of a mailicious VF.
+ */
+struct ice_mbx_data {
+	u16 num_msg_proc;
+	u16 num_pending_arq;
+	u16 max_num_msgs_mbx;
+	u16 async_watermark_val;
+};
+
 /* Port hardware description */
 struct ice_hw {
 	u8 __iomem *hw_addr;
@@ -761,6 +835,7 @@ struct ice_hw {
 	DECLARE_BITMAP(fdir_perfect_fltr, ICE_FLTR_PTYPE_MAX);
 	struct mutex rss_locks;	/* protect RSS configuration */
 	struct list_head rss_list_head;
+	struct ice_mbx_snapshot mbx_snapshot;
 };
 
 /* Statistics collected by each port, VSI, VEB, and S-channel */
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index e38d4adc5b8d..a3ed4b84bba6 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -424,6 +424,14 @@ void ice_free_vfs(struct ice_pf *pf)
 			wr32(hw, GLGEN_VFLRSTAT(reg_idx), BIT(bit_idx));
 		}
 	}
+
+	/* clear malicious info if the VFs are getting released */
+	for (i = 0; i < tmp; i++)
+		if (ice_mbx_clear_malvf(&hw->mbx_snapshot, pf->malvfs,
+					ICE_MAX_VF_COUNT, i))
+			dev_dbg(dev, "failed to clear malicious VF state for VF %u\n",
+				i);
+
 	clear_bit(ICE_VF_DIS, pf->state);
 	clear_bit(ICE_FLAG_SRIOV_ENA, pf->flags);
 }
@@ -1257,6 +1265,11 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
 	if (!pf->num_alloc_vfs)
 		return false;
 
+	/* clear all malicious info if the VFs are getting reset */
+	ice_for_each_vf(pf, i)
+		if (ice_mbx_clear_malvf(&hw->mbx_snapshot, pf->malvfs, ICE_MAX_VF_COUNT, i))
+			dev_dbg(dev, "failed to clear malicious VF state for VF %u\n", i);
+
 	/* If VFs have been disabled, there is no need to reset */
 	if (test_and_set_bit(ICE_VF_DIS, pf->state))
 		return false;
@@ -1437,6 +1450,10 @@ bool ice_reset_vf(struct ice_vf *vf, bool is_vflr)
 	ice_vf_rebuild_vsi_with_release(vf);
 	ice_vf_post_vsi_rebuild(vf);
 
+	/* if the VF has been reset allow it to come up again */
+	if (ice_mbx_clear_malvf(&hw->mbx_snapshot, pf->malvfs, ICE_MAX_VF_COUNT, vf->vf_id))
+		dev_dbg(dev, "failed to clear malicious VF state for VF %u\n", i);
+
 	return true;
 }
 
@@ -1769,6 +1786,7 @@ int ice_sriov_configure(struct pci_dev *pdev, int num_vfs)
 {
 	struct ice_pf *pf = pci_get_drvdata(pdev);
 	struct device *dev = ice_pf_to_dev(pf);
+	enum ice_status status;
 	int err;
 
 	err = ice_check_sriov_allowed(pf);
@@ -1777,6 +1795,7 @@ int ice_sriov_configure(struct pci_dev *pdev, int num_vfs)
 
 	if (!num_vfs) {
 		if (!pci_vfs_assigned(pdev)) {
+			ice_mbx_deinit_snapshot(&pf->hw);
 			ice_free_vfs(pf);
 			if (pf->lag)
 				ice_enable_lag(pf->lag);
@@ -1787,9 +1806,15 @@ int ice_sriov_configure(struct pci_dev *pdev, int num_vfs)
 		return -EBUSY;
 	}
 
+	status = ice_mbx_init_snapshot(&pf->hw, num_vfs);
+	if (status)
+		return ice_status_to_errno(status);
+
 	err = ice_pci_sriov_ena(pf, num_vfs);
-	if (err)
+	if (err) {
+		ice_mbx_deinit_snapshot(&pf->hw);
 		return err;
+	}
 
 	if (pf->lag)
 		ice_disable_lag(pf->lag);
@@ -4255,3 +4280,70 @@ void ice_restore_all_vfs_msi_state(struct pci_dev *pdev)
 		}
 	}
 }
+
+/**
+ * ice_is_malicious_vf - helper function to detect a malicious VF
+ * @pf: ptr to struct ice_pf
+ * @event: pointer to the AQ event
+ * @num_msg_proc: the number of messages processed so far
+ * @num_msg_pending: the number of messages peinding in admin queue
+ */
+bool
+ice_is_malicious_vf(struct ice_pf *pf, struct ice_rq_event_info *event,
+		    u16 num_msg_proc, u16 num_msg_pending)
+{
+	s16 vf_id = le16_to_cpu(event->desc.retval);
+	struct device *dev = ice_pf_to_dev(pf);
+	struct ice_mbx_data mbxdata;
+	enum ice_status status;
+	bool malvf = false;
+	struct ice_vf *vf;
+
+	if (ice_validate_vf_id(pf, vf_id))
+		return false;
+
+	vf = &pf->vf[vf_id];
+	/* Check if VF is disabled. */
+	if (test_bit(ICE_VF_STATE_DIS, vf->vf_states))
+		return false;
+
+	mbxdata.num_msg_proc = num_msg_proc;
+	mbxdata.num_pending_arq = num_msg_pending;
+	mbxdata.max_num_msgs_mbx = pf->hw.mailboxq.num_rq_entries;
+#define ICE_MBX_OVERFLOW_WATERMARK 64
+	mbxdata.async_watermark_val = ICE_MBX_OVERFLOW_WATERMARK;
+
+	/* check to see if we have a malicious VF */
+	status = ice_mbx_vf_state_handler(&pf->hw, &mbxdata, vf_id, &malvf);
+	if (status)
+		return false;
+
+	if (malvf) {
+		bool report_vf = false;
+
+		/* if the VF is malicious and we haven't let the user
+		 * know about it, then let them know now
+		 */
+		status = ice_mbx_report_malvf(&pf->hw, pf->malvfs,
+					      ICE_MAX_VF_COUNT, vf_id,
+					      &report_vf);
+		if (status)
+			dev_dbg(dev, "Error reporting malicious VF\n");
+
+		if (report_vf) {
+			struct ice_vsi *pf_vsi = ice_get_main_vsi(pf);
+
+			if (pf_vsi)
+				dev_warn(dev, "VF MAC %pM on PF MAC %pM is generating asynchronous messages and may be overflowing the PF message queue. Please see the Adapter User Guide for more information\n",
+					 &vf->dflt_lan_addr.addr[0],
+					 pf_vsi->netdev->dev_addr);
+		}
+
+		return true;
+	}
+
+	/* if there was an error in detection or the VF is not malicious then
+	 * return false
+	 */
+	return false;
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
index 46abc5388fc7..bcc2890c930a 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
@@ -119,6 +119,9 @@ void ice_vc_notify_reset(struct ice_pf *pf);
 bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr);
 bool ice_reset_vf(struct ice_vf *vf, bool is_vflr);
 void ice_restore_all_vfs_msi_state(struct pci_dev *pdev);
+bool
+ice_is_malicious_vf(struct ice_pf *pf, struct ice_rq_event_info *event,
+		    u16 num_msg_proc, u16 num_msg_pending);
 
 int
 ice_set_vf_port_vlan(struct net_device *netdev, int vf_id, u16 vlan_id, u8 qos,
@@ -158,6 +161,15 @@ bool ice_vc_isvalid_vsi_id(struct ice_vf *vf, u16 vsi_id);
 #define ice_print_vf_rx_mdd_event(vf) do {} while (0)
 #define ice_restore_all_vfs_msi_state(pdev) do {} while (0)
 
+static inline bool
+ice_is_malicious_vf(struct ice_pf __always_unused *pf,
+		    struct ice_rq_event_info __always_unused *event,
+		    u16 __always_unused num_msg_proc,
+		    u16 __always_unused num_msg_pending)
+{
+	return false;
+}
+
 static inline bool
 ice_reset_all_vfs(struct ice_pf __always_unused *pf,
 		  bool __always_unused is_vflr)
-- 
2.26.2

[PATCH net-next 02/12] ice: Allow ignoring opcodes on specific VF

[Tony Nguyen]

From: Michal Swiatkowski <michal.swiatkowski@intel.com>

Declare bitmap of allowed commands on VF. Initialize default
opcodes list that should be always supported. Declare array of
supported opcodes for each caps used in virtchnl code.

Change allowed bitmap by setting or clearing corresponding
bit to allowlist (bit set) or denylist (bit clear).

Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/ice/Makefile       |   1 +
 .../intel/ice/ice_virtchnl_allowlist.c        | 165 ++++++++++++++++++
 .../intel/ice/ice_virtchnl_allowlist.h        |  13 ++
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  |  18 ++
 .../net/ethernet/intel/ice/ice_virtchnl_pf.h  |   1 +
 include/linux/avf/virtchnl.h                  |   1 +
 6 files changed, 199 insertions(+)
 create mode 100644 drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
 create mode 100644 drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.h

diff --git a/drivers/net/ethernet/intel/ice/Makefile b/drivers/net/ethernet/intel/ice/Makefile
index f391691e2c7e..07fe857e9e3a 100644
--- a/drivers/net/ethernet/intel/ice/Makefile
+++ b/drivers/net/ethernet/intel/ice/Makefile
@@ -26,6 +26,7 @@ ice-y := ice_main.o	\
 	 ice_fw_update.o \
 	 ice_lag.o	\
 	 ice_ethtool.o
+ice-$(CONFIG_PCI_IOV) += ice_virtchnl_allowlist.o
 ice-$(CONFIG_PCI_IOV) += ice_virtchnl_pf.o ice_sriov.o ice_virtchnl_fdir.o
 ice-$(CONFIG_DCB) += ice_dcb.o ice_dcb_nl.o ice_dcb_lib.o
 ice-$(CONFIG_RFS_ACCEL) += ice_arfs.o
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
new file mode 100644
index 000000000000..5a0fbb47346f
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
@@ -0,0 +1,165 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (C) 2021, Intel Corporation. */
+
+#include "ice_virtchnl_allowlist.h"
+
+/* Purpose of this file is to share functionality to allowlist or denylist
+ * opcodes used in PF <-> VF communication. Group of opcodes:
+ * - default -> should be always allowed after creating VF,
+ *   default_allowlist_opcodes
+ * - opcodes needed by VF to work correctly, but not associated with caps ->
+ *   should be allowed after successful VF resources allocation,
+ *   working_allowlist_opcodes
+ * - opcodes needed by VF when caps are activated
+ *
+ * Caps that don't use new opcodes (no opcodes should be allowed):
+ * - VIRTCHNL_VF_OFFLOAD_RSS_AQ
+ * - VIRTCHNL_VF_OFFLOAD_RSS_REG
+ * - VIRTCHNL_VF_OFFLOAD_WB_ON_ITR
+ * - VIRTCHNL_VF_OFFLOAD_CRC
+ * - VIRTCHNL_VF_OFFLOAD_RX_POLLING
+ * - VIRTCHNL_VF_OFFLOAD_RSS_PCTYPE_V2
+ * - VIRTCHNL_VF_OFFLOAD_ENCAP
+ * - VIRTCHNL_VF_OFFLOAD_ENCAP_CSUM
+ * - VIRTCHNL_VF_OFFLOAD_RX_ENCAP_CSUM
+ * - VIRTCHNL_VF_OFFLOAD_USO
+ */
+
+/* default opcodes to communicate with VF */
+static const u32 default_allowlist_opcodes[] = {
+	VIRTCHNL_OP_GET_VF_RESOURCES, VIRTCHNL_OP_VERSION, VIRTCHNL_OP_RESET_VF,
+};
+
+/* opcodes supported after successful VIRTCHNL_OP_GET_VF_RESOURCES */
+static const u32 working_allowlist_opcodes[] = {
+	VIRTCHNL_OP_CONFIG_TX_QUEUE, VIRTCHNL_OP_CONFIG_RX_QUEUE,
+	VIRTCHNL_OP_CONFIG_VSI_QUEUES, VIRTCHNL_OP_CONFIG_IRQ_MAP,
+	VIRTCHNL_OP_ENABLE_QUEUES, VIRTCHNL_OP_DISABLE_QUEUES,
+	VIRTCHNL_OP_GET_STATS, VIRTCHNL_OP_EVENT,
+};
+
+/* VIRTCHNL_VF_OFFLOAD_L2 */
+static const u32 l2_allowlist_opcodes[] = {
+	VIRTCHNL_OP_ADD_ETH_ADDR, VIRTCHNL_OP_DEL_ETH_ADDR,
+	VIRTCHNL_OP_CONFIG_PROMISCUOUS_MODE,
+};
+
+/* VIRTCHNL_VF_OFFLOAD_REQ_QUEUES */
+static const u32 req_queues_allowlist_opcodes[] = {
+	VIRTCHNL_OP_REQUEST_QUEUES,
+};
+
+/* VIRTCHNL_VF_OFFLOAD_VLAN */
+static const u32 vlan_allowlist_opcodes[] = {
+	VIRTCHNL_OP_ADD_VLAN, VIRTCHNL_OP_DEL_VLAN,
+	VIRTCHNL_OP_ENABLE_VLAN_STRIPPING, VIRTCHNL_OP_DISABLE_VLAN_STRIPPING,
+};
+
+/* VIRTCHNL_VF_OFFLOAD_RSS_PF */
+static const u32 rss_pf_allowlist_opcodes[] = {
+	VIRTCHNL_OP_CONFIG_RSS_KEY, VIRTCHNL_OP_CONFIG_RSS_LUT,
+	VIRTCHNL_OP_GET_RSS_HENA_CAPS, VIRTCHNL_OP_SET_RSS_HENA,
+};
+
+/* VIRTCHNL_VF_OFFLOAD_FDIR_PF */
+static const u32 fdir_pf_allowlist_opcodes[] = {
+	VIRTCHNL_OP_ADD_FDIR_FILTER, VIRTCHNL_OP_DEL_FDIR_FILTER,
+};
+
+struct allowlist_opcode_info {
+	const u32 *opcodes;
+	size_t size;
+};
+
+#define BIT_INDEX(caps) (HWEIGHT((caps) - 1))
+#define ALLOW_ITEM(caps, list) \
+	[BIT_INDEX(caps)] = { \
+		.opcodes = list, \
+		.size = ARRAY_SIZE(list) \
+	}
+static const struct allowlist_opcode_info allowlist_opcodes[] = {
+	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_L2, l2_allowlist_opcodes),
+	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_REQ_QUEUES, req_queues_allowlist_opcodes),
+	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_VLAN, vlan_allowlist_opcodes),
+	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_RSS_PF, rss_pf_allowlist_opcodes),
+	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_FDIR_PF, fdir_pf_allowlist_opcodes),
+};
+
+/**
+ * ice_vc_is_opcode_allowed - check if this opcode is allowed on this VF
+ * @vf: pointer to VF structure
+ * @opcode: virtchnl opcode
+ *
+ * Return true if message is allowed on this VF
+ */
+bool ice_vc_is_opcode_allowed(struct ice_vf *vf, u32 opcode)
+{
+	if (opcode >= VIRTCHNL_OP_MAX)
+		return false;
+
+	return test_bit(opcode, vf->opcodes_allowlist);
+}
+
+/**
+ * ice_vc_allowlist_opcodes - allowlist selected opcodes
+ * @vf: pointer to VF structure
+ * @opcodes: array of opocodes to allowlist
+ * @size: size of opcodes array
+ *
+ * Function should be called to allowlist opcodes on VF.
+ */
+static void
+ice_vc_allowlist_opcodes(struct ice_vf *vf, const u32 *opcodes, size_t size)
+{
+	unsigned int i;
+
+	for (i = 0; i < size; i++)
+		set_bit(opcodes[i], vf->opcodes_allowlist);
+}
+
+/**
+ * ice_vc_clear_allowlist - clear all allowlist opcodes
+ * @vf: pointer to VF structure
+ */
+static void ice_vc_clear_allowlist(struct ice_vf *vf)
+{
+	bitmap_zero(vf->opcodes_allowlist, VIRTCHNL_OP_MAX);
+}
+
+/**
+ * ice_vc_set_default_allowlist - allowlist default opcodes for VF
+ * @vf: pointer to VF structure
+ */
+void ice_vc_set_default_allowlist(struct ice_vf *vf)
+{
+	ice_vc_clear_allowlist(vf);
+	ice_vc_allowlist_opcodes(vf, default_allowlist_opcodes,
+				 ARRAY_SIZE(default_allowlist_opcodes));
+}
+
+/**
+ * ice_vc_set_working_allowlist - allowlist opcodes needed to by VF to work
+ * @vf: pointer to VF structure
+ *
+ * allowlist opcodes that aren't associated with specific caps, but
+ * are needed by VF to work.
+ */
+void ice_vc_set_working_allowlist(struct ice_vf *vf)
+{
+	ice_vc_allowlist_opcodes(vf, working_allowlist_opcodes,
+				 ARRAY_SIZE(working_allowlist_opcodes));
+}
+
+/**
+ * ice_vc_set_caps_allowlist - allowlist VF opcodes according caps
+ * @vf: pointer to VF structure
+ */
+void ice_vc_set_caps_allowlist(struct ice_vf *vf)
+{
+	unsigned long caps = vf->driver_caps;
+	unsigned int i;
+
+	for_each_set_bit(i, &caps, ARRAY_SIZE(allowlist_opcodes))
+		ice_vc_allowlist_opcodes(vf, allowlist_opcodes[i].opcodes,
+					 allowlist_opcodes[i].size);
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.h
new file mode 100644
index 000000000000..d3ae86ded219
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (C) 2021, Intel Corporation. */
+
+#ifndef _ICE_VIRTCHNL_ALLOWLIST_H_
+#define _ICE_VIRTCHNL_ALLOWLIST_H_
+#include "ice.h"
+
+bool ice_vc_is_opcode_allowed(struct ice_vf *vf, u32 opcode);
+
+void ice_vc_set_default_allowlist(struct ice_vf *vf);
+void ice_vc_set_working_allowlist(struct ice_vf *vf);
+void ice_vc_set_caps_allowlist(struct ice_vf *vf);
+#endif /* _ICE_VIRTCHNL_ALLOWLIST_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index a3ed4b84bba6..ccd6b3e8a5a9 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -5,6 +5,7 @@
 #include "ice_base.h"
 #include "ice_lib.h"
 #include "ice_fltr.h"
+#include "ice_virtchnl_allowlist.h"
 
 /**
  * ice_validate_vf_id - helper to check if VF ID is valid
@@ -1314,6 +1315,9 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
 	ice_for_each_vf(pf, v) {
 		vf = &pf->vf[v];
 
+		vf->driver_caps = 0;
+		ice_vc_set_default_allowlist(vf);
+
 		ice_vf_fdir_exit(vf);
 		/* clean VF control VSI when resetting VFs since it should be
 		 * setup only when VF creates its first FDIR rule.
@@ -1418,6 +1422,9 @@ bool ice_reset_vf(struct ice_vf *vf, bool is_vflr)
 		usleep_range(10, 20);
 	}
 
+	vf->driver_caps = 0;
+	ice_vc_set_default_allowlist(vf);
+
 	/* Display a warning if VF didn't manage to reset in time, but need to
 	 * continue on with the operation.
 	 */
@@ -1625,6 +1632,7 @@ static void ice_set_dflt_settings_vfs(struct ice_pf *pf)
 		set_bit(ICE_VIRTCHNL_VF_CAP_L2, &vf->vf_caps);
 		vf->spoofchk = true;
 		vf->num_vf_qs = pf->num_qps_per_vf;
+		ice_vc_set_default_allowlist(vf);
 
 		/* ctrl_vsi_idx will be set to a valid value only when VF
 		 * creates its first fdir rule.
@@ -2127,6 +2135,9 @@ static int ice_vc_get_vf_res_msg(struct ice_vf *vf, u8 *msg)
 	/* match guest capabilities */
 	vf->driver_caps = vfres->vf_cap_flags;
 
+	ice_vc_set_caps_allowlist(vf);
+	ice_vc_set_working_allowlist(vf);
+
 	set_bit(ICE_VF_STATE_ACTIVE, vf->vf_states);
 
 err:
@@ -3840,6 +3851,13 @@ void ice_vc_process_vf_msg(struct ice_pf *pf, struct ice_rq_event_info *event)
 			err = -EINVAL;
 	}
 
+	if (!ice_vc_is_opcode_allowed(vf, v_opcode)) {
+		ice_vc_send_msg_to_vf(vf, v_opcode,
+				      VIRTCHNL_STATUS_ERR_NOT_SUPPORTED, NULL,
+				      0);
+		return;
+	}
+
 error_handler:
 	if (err) {
 		ice_vc_send_msg_to_vf(vf, v_opcode, VIRTCHNL_STATUS_ERR_PARAM,
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
index bcc2890c930a..d800ed83d6c3 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
@@ -103,6 +103,7 @@ struct ice_vf {
 	u16 num_vf_qs;			/* num of queue configured per VF */
 	struct ice_mdd_vf_events mdd_rx_events;
 	struct ice_mdd_vf_events mdd_tx_events;
+	DECLARE_BITMAP(opcodes_allowlist, VIRTCHNL_OP_MAX);
 };
 
 #ifdef CONFIG_PCI_IOV
diff --git a/include/linux/avf/virtchnl.h b/include/linux/avf/virtchnl.h
index 40dd6afbfd81..debdd196773b 100644
--- a/include/linux/avf/virtchnl.h
+++ b/include/linux/avf/virtchnl.h
@@ -139,6 +139,7 @@ enum virtchnl_ops {
 	/* opcode 34 - 46 are reserved */
 	VIRTCHNL_OP_ADD_FDIR_FILTER = 47,
 	VIRTCHNL_OP_DEL_FDIR_FILTER = 48,
+	VIRTCHNL_OP_MAX,
 };
 
 /* These macros are used to generate compilation errors if a structure/union
-- 
2.26.2

[PATCH net-next 03/12] ice: Advertise virtchnl UDP segmentation offload capability

[Tony Nguyen]

From: Brett Creeley <brett.creeley@intel.com>

As the hardware is capable of supporting UDP segmentation offload, add a
capability bit to virtchnl.h to communicate this and have the driver
advertise its support.

Suggested-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 3 +++
 include/linux/avf/virtchnl.h                     | 1 +
 2 files changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index ccd6b3e8a5a9..1292a0b06eb5 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -2118,6 +2118,9 @@ static int ice_vc_get_vf_res_msg(struct ice_vf *vf, u8 *msg)
 	if (vf->driver_caps & VIRTCHNL_VF_CAP_ADV_LINK_SPEED)
 		vfres->vf_cap_flags |= VIRTCHNL_VF_CAP_ADV_LINK_SPEED;
 
+	if (vf->driver_caps & VIRTCHNL_VF_OFFLOAD_USO)
+		vfres->vf_cap_flags |= VIRTCHNL_VF_OFFLOAD_USO;
+
 	vfres->num_vsis = 1;
 	/* Tx and Rx queue are equal for VF */
 	vfres->num_queue_pairs = vsi->num_txq;
diff --git a/include/linux/avf/virtchnl.h b/include/linux/avf/virtchnl.h
index debdd196773b..9e0341cf2c36 100644
--- a/include/linux/avf/virtchnl.h
+++ b/include/linux/avf/virtchnl.h
@@ -251,6 +251,7 @@ VIRTCHNL_CHECK_STRUCT_LEN(16, virtchnl_vsi_resource);
 #define VIRTCHNL_VF_OFFLOAD_ENCAP_CSUM		0X00200000
 #define VIRTCHNL_VF_OFFLOAD_RX_ENCAP_CSUM	0X00400000
 #define VIRTCHNL_VF_OFFLOAD_ADQ			0X00800000
+#define VIRTCHNL_VF_OFFLOAD_USO			0X02000000
 #define VIRTCHNL_VF_OFFLOAD_FDIR_PF		0X10000000
 
 /* Define below the capability flags that are not offloads */
-- 
2.26.2

[PATCH net-next 04/12] iavf: add support for UDP Segmentation Offload

[Tony Nguyen]

From: Brett Creeley <brett.creeley@intel.com>

Add code to support UDP segmentation offload (USO) for
hardware that supports it.

Suggested-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/iavf/iavf_main.c     |  2 ++
 drivers/net/ethernet/intel/iavf/iavf_txrx.c     | 15 +++++++++++----
 drivers/net/ethernet/intel/iavf/iavf_virtchnl.c |  1 +
 3 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c
index a3268c894d85..6cd6b9dfe5d4 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_main.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_main.c
@@ -3542,6 +3542,8 @@ int iavf_process_config(struct iavf_adapter *adapter)
 	/* Enable cloud filter if ADQ is supported */
 	if (vfres->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_ADQ)
 		hw_features |= NETIF_F_HW_TC;
+	if (vfres->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_USO)
+		hw_features |= NETIF_F_GSO_UDP_L4;
 
 	netdev->hw_features |= hw_features;
 
diff --git a/drivers/net/ethernet/intel/iavf/iavf_txrx.c b/drivers/net/ethernet/intel/iavf/iavf_txrx.c
index d6cba53a3a21..3525eab8e9f9 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_txrx.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_txrx.c
@@ -1905,13 +1905,20 @@ static int iavf_tso(struct iavf_tx_buffer *first, u8 *hdr_len,
 
 	/* determine offset of inner transport header */
 	l4_offset = l4.hdr - skb->data;
-
 	/* remove payload length from inner checksum */
 	paylen = skb->len - l4_offset;
-	csum_replace_by_diff(&l4.tcp->check, (__force __wsum)htonl(paylen));
 
-	/* compute length of segmentation header */
-	*hdr_len = (l4.tcp->doff * 4) + l4_offset;
+	if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP_L4) {
+		csum_replace_by_diff(&l4.udp->check,
+				     (__force __wsum)htonl(paylen));
+		/* compute length of UDP segmentation header */
+		*hdr_len = (u8)sizeof(l4.udp) + l4_offset;
+	} else {
+		csum_replace_by_diff(&l4.tcp->check,
+				     (__force __wsum)htonl(paylen));
+		/* compute length of TCP segmentation header */
+		*hdr_len = (u8)((l4.tcp->doff * 4) + l4_offset);
+	}
 
 	/* pull values out of skb_shinfo */
 	gso_size = skb_shinfo(skb)->gso_size;
diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
index 3069092468b2..9aaade0aae4c 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
@@ -140,6 +140,7 @@ int iavf_send_vf_config_msg(struct iavf_adapter *adapter)
 	       VIRTCHNL_VF_OFFLOAD_ENCAP_CSUM |
 	       VIRTCHNL_VF_OFFLOAD_REQ_QUEUES |
 	       VIRTCHNL_VF_OFFLOAD_ADQ |
+	       VIRTCHNL_VF_OFFLOAD_USO |
 	       VIRTCHNL_VF_OFFLOAD_FDIR_PF |
 	       VIRTCHNL_VF_CAP_ADV_LINK_SPEED;
 
-- 
2.26.2

[PATCH net-next 05/12] ice: remove redundant assignment to pointer vsi

[Tony Nguyen]

From: Colin Ian King <colin.king@canonical.com>

Pointer vsi is being re-assigned a value that is never read,
the assignment is redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 1292a0b06eb5..c40560a9443b 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -2792,7 +2792,6 @@ static int ice_vc_ena_qs_msg(struct ice_vf *vf, u8 *msg)
 		set_bit(vf_q_id, vf->rxq_ena);
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
 	q_map = vqs->tx_queues;
 	for_each_set_bit(vf_q_id, &q_map, ICE_MAX_RSS_QS_PER_VF) {
 		if (!ice_vc_isvalid_q_id(vf, vqs->vsi_id, vf_q_id)) {
-- 
2.26.2

[PATCH net-next 06/12] ice: Add helper function to get the VF's VSI

[Tony Nguyen]

From: Brett Creeley <brett.creeley@intel.com>

Currently, the driver gets the VF's VSI by using a long string of
dereferences (i.e. vf->pf->vsi[vf->lan_vsi_idx]). If the method to get
the VF's VSI were to change the driver would have to change it in every
location. Fix this by adding the helper ice_get_vf_vsi().

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  | 82 +++++++++----------
 1 file changed, 39 insertions(+), 43 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index c40560a9443b..baada80c98ab 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -7,6 +7,15 @@
 #include "ice_fltr.h"
 #include "ice_virtchnl_allowlist.h"
 
+/**
+ * ice_get_vf_vsi - get VF's VSI based on the stored index
+ * @vf: VF used to get VSI
+ */
+static struct ice_vsi *ice_get_vf_vsi(struct ice_vf *vf)
+{
+	return vf->pf->vsi[vf->lan_vsi_idx];
+}
+
 /**
  * ice_validate_vf_id - helper to check if VF ID is valid
  * @pf: pointer to the PF structure
@@ -198,7 +207,7 @@ static void ice_vf_invalidate_vsi(struct ice_vf *vf)
  */
 static void ice_vf_vsi_release(struct ice_vf *vf)
 {
-	ice_vsi_release(vf->pf->vsi[vf->lan_vsi_idx]);
+	ice_vsi_release(ice_get_vf_vsi(vf));
 	ice_vf_invalidate_vsi(vf);
 }
 
@@ -274,7 +283,7 @@ static void ice_dis_vf_mappings(struct ice_vf *vf)
 	struct ice_hw *hw;
 
 	hw = &pf->hw;
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 
 	dev = ice_pf_to_dev(pf);
 	wr32(hw, VPINT_ALLOC(vf->vf_id), 0);
@@ -349,10 +358,7 @@ void ice_set_vf_state_qs_dis(struct ice_vf *vf)
  */
 static void ice_dis_vf_qs(struct ice_vf *vf)
 {
-	struct ice_pf *pf = vf->pf;
-	struct ice_vsi *vsi;
-
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 
 	ice_vsi_stop_lan_tx_rings(vsi, ICE_NO_RESET, vf->vf_id);
 	ice_vsi_stop_all_rx_rings(vsi);
@@ -639,8 +645,8 @@ static int ice_calc_vf_first_vector_idx(struct ice_pf *pf, struct ice_vf *vf)
  */
 static int ice_vf_rebuild_host_vlan_cfg(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
 	struct device *dev = ice_pf_to_dev(vf->pf);
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 	u16 vlan_id = 0;
 	int err;
 
@@ -676,8 +682,8 @@ static int ice_vf_rebuild_host_vlan_cfg(struct ice_vf *vf)
  */
 static int ice_vf_rebuild_host_mac_cfg(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
 	struct device *dev = ice_pf_to_dev(vf->pf);
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 	enum ice_status status;
 	u8 broadcast[ETH_ALEN];
 
@@ -778,8 +784,8 @@ static void ice_ena_vf_msix_mappings(struct ice_vf *vf)
  */
 static void ice_ena_vf_q_mappings(struct ice_vf *vf, u16 max_txq, u16 max_rxq)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
 	struct device *dev = ice_pf_to_dev(vf->pf);
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 	struct ice_hw *hw = &vf->pf->hw;
 	u32 reg;
 
@@ -826,7 +832,7 @@ static void ice_ena_vf_q_mappings(struct ice_vf *vf, u16 max_txq, u16 max_rxq)
  */
 static void ice_ena_vf_mappings(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 
 	ice_ena_vf_msix_mappings(vf);
 	ice_ena_vf_q_mappings(vf, vsi->alloc_txq, vsi->alloc_rxq);
@@ -1089,7 +1095,7 @@ ice_vf_set_vsi_promisc(struct ice_vf *vf, struct ice_vsi *vsi, u8 promisc_m,
 
 static void ice_vf_clear_counters(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 
 	vf->num_mac = 0;
 	vsi->num_vlan = 0;
@@ -1149,8 +1155,8 @@ static void ice_vf_rebuild_aggregator_node_cfg(struct ice_vsi *vsi)
  */
 static void ice_vf_rebuild_host_cfg(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
 	struct device *dev = ice_pf_to_dev(vf->pf);
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 
 	ice_vf_set_host_trust_cfg(vf);
 
@@ -1190,10 +1196,8 @@ static int ice_vf_rebuild_vsi_with_release(struct ice_vf *vf)
  */
 static int ice_vf_rebuild_vsi(struct ice_vf *vf)
 {
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 	struct ice_pf *pf = vf->pf;
-	struct ice_vsi *vsi;
-
-	vsi = pf->vsi[vf->lan_vsi_idx];
 
 	if (ice_vsi_rebuild(vsi, true)) {
 		dev_err(ice_pf_to_dev(pf), "failed to rebuild VF %d VSI\n",
@@ -1392,7 +1396,7 @@ bool ice_reset_vf(struct ice_vf *vf, bool is_vflr)
 	set_bit(ICE_VF_STATE_DIS, vf->vf_states);
 	ice_trigger_vf_reset(vf, is_vflr, false);
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 
 	if (test_bit(ICE_VF_STATE_QS_ENA, vf->vf_states))
 		ice_dis_vf_qs(vf);
@@ -1441,7 +1445,7 @@ bool ice_reset_vf(struct ice_vf *vf, bool is_vflr)
 		else
 			promisc_m = ICE_UCAST_PROMISC_BITS;
 
-		vsi = pf->vsi[vf->lan_vsi_idx];
+		vsi = ice_get_vf_vsi(vf);
 		if (ice_vf_set_vsi_promisc(vf, vsi, promisc_m, true))
 			dev_err(dev, "disabling promiscuous mode failed\n");
 	}
@@ -1887,7 +1891,7 @@ static struct ice_vf *ice_get_vf_from_pfq(struct ice_pf *pf, u16 pfq)
 		struct ice_vsi *vsi;
 		u16 rxq_idx;
 
-		vsi = pf->vsi[vf->lan_vsi_idx];
+		vsi = ice_get_vf_vsi(vf);
 
 		ice_for_each_rxq(vsi, rxq_idx)
 			if (vsi->rxq_map[rxq_idx] == pfq)
@@ -2027,8 +2031,7 @@ static int ice_vc_get_ver_msg(struct ice_vf *vf, u8 *msg)
  */
 static u16 ice_vc_get_max_frame_size(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
-	struct ice_port_info *pi = vsi->port_info;
+	struct ice_port_info *pi = ice_vf_get_port_info(vf);
 	u16 max_frame_size;
 
 	max_frame_size = pi->phy.link_info.max_frame_size;
@@ -2076,7 +2079,7 @@ static int ice_vc_get_vf_res_msg(struct ice_vf *vf, u8 *msg)
 				  VIRTCHNL_VF_OFFLOAD_VLAN;
 
 	vfres->vf_cap_flags = VIRTCHNL_VF_OFFLOAD_L2;
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto err;
@@ -2243,7 +2246,6 @@ static int ice_vc_config_rss_key(struct ice_vf *vf, u8 *msg)
 	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
 	struct virtchnl_rss_key *vrk =
 		(struct virtchnl_rss_key *)msg;
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 
 	if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
@@ -2266,7 +2268,7 @@ static int ice_vc_config_rss_key(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -2290,7 +2292,6 @@ static int ice_vc_config_rss_lut(struct ice_vf *vf, u8 *msg)
 {
 	struct virtchnl_rss_lut *vrl = (struct virtchnl_rss_lut *)msg;
 	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 
 	if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
@@ -2313,7 +2314,7 @@ static int ice_vc_config_rss_lut(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -2396,7 +2397,7 @@ int ice_set_vf_spoofchk(struct net_device *netdev, int vf_id, bool ena)
 	if (ret)
 		return ret;
 
-	vf_vsi = pf->vsi[vf->lan_vsi_idx];
+	vf_vsi = ice_get_vf_vsi(vf);
 	if (!vf_vsi) {
 		netdev_err(netdev, "VSI %d for VF %d is null\n",
 			   vf->lan_vsi_idx, vf->vf_id);
@@ -2501,7 +2502,7 @@ static int ice_vc_cfg_promiscuous_mode_msg(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -2637,7 +2638,6 @@ static int ice_vc_get_stats_msg(struct ice_vf *vf, u8 *msg)
 	struct virtchnl_queue_select *vqs =
 		(struct virtchnl_queue_select *)msg;
 	struct ice_eth_stats stats = { 0 };
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 
 	if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
@@ -2650,7 +2650,7 @@ static int ice_vc_get_stats_msg(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -2740,7 +2740,6 @@ static int ice_vc_ena_qs_msg(struct ice_vf *vf, u8 *msg)
 	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
 	struct virtchnl_queue_select *vqs =
 	    (struct virtchnl_queue_select *)msg;
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 	unsigned long q_map;
 	u16 vf_q_id;
@@ -2760,7 +2759,7 @@ static int ice_vc_ena_qs_msg(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -2830,7 +2829,6 @@ static int ice_vc_dis_qs_msg(struct ice_vf *vf, u8 *msg)
 	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
 	struct virtchnl_queue_select *vqs =
 	    (struct virtchnl_queue_select *)msg;
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 	unsigned long q_map;
 	u16 vf_q_id;
@@ -2851,7 +2849,7 @@ static int ice_vc_dis_qs_msg(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -3016,7 +3014,7 @@ static int ice_vc_cfg_irq_map_msg(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -3093,7 +3091,7 @@ static int ice_vc_cfg_qs_msg(struct ice_vf *vf, u8 *msg)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -3328,7 +3326,7 @@ ice_vc_handle_mac_addr_msg(struct ice_vf *vf, u8 *msg, bool set)
 		goto handle_mac_exit;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto handle_mac_exit;
@@ -3560,7 +3558,7 @@ static int ice_vc_process_vlan_msg(struct ice_vf *vf, u8 *msg, bool add_v)
 	}
 
 	hw = &pf->hw;
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -3727,7 +3725,6 @@ static int ice_vc_remove_vlan_msg(struct ice_vf *vf, u8 *msg)
 static int ice_vc_ena_vlan_stripping(struct ice_vf *vf)
 {
 	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 
 	if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
@@ -3740,7 +3737,7 @@ static int ice_vc_ena_vlan_stripping(struct ice_vf *vf)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (ice_vsi_manage_vlan_stripping(vsi, true))
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 
@@ -3758,7 +3755,6 @@ static int ice_vc_ena_vlan_stripping(struct ice_vf *vf)
 static int ice_vc_dis_vlan_stripping(struct ice_vf *vf)
 {
 	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
-	struct ice_pf *pf = vf->pf;
 	struct ice_vsi *vsi;
 
 	if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
@@ -3771,7 +3767,7 @@ static int ice_vc_dis_vlan_stripping(struct ice_vf *vf)
 		goto error_param;
 	}
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi) {
 		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
 		goto error_param;
@@ -3797,7 +3793,7 @@ static int ice_vc_dis_vlan_stripping(struct ice_vf *vf)
  */
 static int ice_vf_init_vlan_stripping(struct ice_vf *vf)
 {
-	struct ice_vsi *vsi = vf->pf->vsi[vf->lan_vsi_idx];
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
 
 	if (!vsi)
 		return -EINVAL;
@@ -4185,7 +4181,7 @@ int ice_get_vf_stats(struct net_device *netdev, int vf_id,
 	if (ret)
 		return ret;
 
-	vsi = pf->vsi[vf->lan_vsi_idx];
+	vsi = ice_get_vf_vsi(vf);
 	if (!vsi)
 		return -EINVAL;
 
-- 
2.26.2

[PATCH net-next 07/12] ice: Enable RSS configure for AVF

[Tony Nguyen]

From: Qi Zhang <qi.z.zhang@intel.com>

Currently, RSS hash input is not available to AVF by ethtool, it is set
by the PF directly.

Add the RSS configure support for AVF through new virtchnl message, and
define the capability flag VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF to query this
new RSS offload support.

Signed-off-by: Jia Guo <jia.guo@intel.com>
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Tested-by: Bo Chen <BoX.C.Chen@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_flow.h     |   3 +
 .../intel/ice/ice_virtchnl_allowlist.c        |   6 +
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  | 453 ++++++++++++++++++
 include/linux/avf/virtchnl.h                  |  25 +-
 4 files changed, 486 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_flow.h b/drivers/net/ethernet/intel/ice/ice_flow.h
index eec9def8ffca..2f68b59ace7e 100644
--- a/drivers/net/ethernet/intel/ice/ice_flow.h
+++ b/drivers/net/ethernet/intel/ice/ice_flow.h
@@ -8,6 +8,9 @@
 #define ICE_FLOW_FLD_OFF_INVAL		0xffff
 
 /* Generate flow hash field from flow field type(s) */
+#define ICE_FLOW_HASH_ETH	\
+	(BIT_ULL(ICE_FLOW_FIELD_IDX_ETH_DA) | \
+	 BIT_ULL(ICE_FLOW_FIELD_IDX_ETH_SA))
 #define ICE_FLOW_HASH_IPV4	\
 	(BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_SA) | \
 	 BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_DA))
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
index 5a0fbb47346f..9feebe5f556c 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_allowlist.c
@@ -61,6 +61,11 @@ static const u32 rss_pf_allowlist_opcodes[] = {
 	VIRTCHNL_OP_GET_RSS_HENA_CAPS, VIRTCHNL_OP_SET_RSS_HENA,
 };
 
+/* VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF */
+static const u32 adv_rss_pf_allowlist_opcodes[] = {
+	VIRTCHNL_OP_ADD_RSS_CFG, VIRTCHNL_OP_DEL_RSS_CFG,
+};
+
 /* VIRTCHNL_VF_OFFLOAD_FDIR_PF */
 static const u32 fdir_pf_allowlist_opcodes[] = {
 	VIRTCHNL_OP_ADD_FDIR_FILTER, VIRTCHNL_OP_DEL_FDIR_FILTER,
@@ -82,6 +87,7 @@ static const struct allowlist_opcode_info allowlist_opcodes[] = {
 	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_REQ_QUEUES, req_queues_allowlist_opcodes),
 	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_VLAN, vlan_allowlist_opcodes),
 	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_RSS_PF, rss_pf_allowlist_opcodes),
+	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF, adv_rss_pf_allowlist_opcodes),
 	ALLOW_ITEM(VIRTCHNL_VF_OFFLOAD_FDIR_PF, fdir_pf_allowlist_opcodes),
 };
 
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index baada80c98ab..ca778a80d363 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -5,8 +5,248 @@
 #include "ice_base.h"
 #include "ice_lib.h"
 #include "ice_fltr.h"
+#include "ice_flow.h"
 #include "ice_virtchnl_allowlist.h"
 
+#define FIELD_SELECTOR(proto_hdr_field) \
+		BIT((proto_hdr_field) & PROTO_HDR_FIELD_MASK)
+
+struct ice_vc_hdr_match_type {
+	u32 vc_hdr;	/* virtchnl headers (VIRTCHNL_PROTO_HDR_XXX) */
+	u32 ice_hdr;	/* ice headers (ICE_FLOW_SEG_HDR_XXX) */
+};
+
+static const struct ice_vc_hdr_match_type ice_vc_hdr_list_os[] = {
+	{VIRTCHNL_PROTO_HDR_NONE,	ICE_FLOW_SEG_HDR_NONE},
+	{VIRTCHNL_PROTO_HDR_IPV4,	ICE_FLOW_SEG_HDR_IPV4 |
+					ICE_FLOW_SEG_HDR_IPV_OTHER},
+	{VIRTCHNL_PROTO_HDR_IPV6,	ICE_FLOW_SEG_HDR_IPV6 |
+					ICE_FLOW_SEG_HDR_IPV_OTHER},
+	{VIRTCHNL_PROTO_HDR_TCP,	ICE_FLOW_SEG_HDR_TCP},
+	{VIRTCHNL_PROTO_HDR_UDP,	ICE_FLOW_SEG_HDR_UDP},
+	{VIRTCHNL_PROTO_HDR_SCTP,	ICE_FLOW_SEG_HDR_SCTP},
+};
+
+static const struct ice_vc_hdr_match_type ice_vc_hdr_list_comms[] = {
+	{VIRTCHNL_PROTO_HDR_NONE,	ICE_FLOW_SEG_HDR_NONE},
+	{VIRTCHNL_PROTO_HDR_ETH,	ICE_FLOW_SEG_HDR_ETH},
+	{VIRTCHNL_PROTO_HDR_S_VLAN,	ICE_FLOW_SEG_HDR_VLAN},
+	{VIRTCHNL_PROTO_HDR_C_VLAN,	ICE_FLOW_SEG_HDR_VLAN},
+	{VIRTCHNL_PROTO_HDR_IPV4,	ICE_FLOW_SEG_HDR_IPV4 |
+					ICE_FLOW_SEG_HDR_IPV_OTHER},
+	{VIRTCHNL_PROTO_HDR_IPV6,	ICE_FLOW_SEG_HDR_IPV6 |
+					ICE_FLOW_SEG_HDR_IPV_OTHER},
+	{VIRTCHNL_PROTO_HDR_TCP,	ICE_FLOW_SEG_HDR_TCP},
+	{VIRTCHNL_PROTO_HDR_UDP,	ICE_FLOW_SEG_HDR_UDP},
+	{VIRTCHNL_PROTO_HDR_SCTP,	ICE_FLOW_SEG_HDR_SCTP},
+	{VIRTCHNL_PROTO_HDR_PPPOE,	ICE_FLOW_SEG_HDR_PPPOE},
+	{VIRTCHNL_PROTO_HDR_GTPU_IP,	ICE_FLOW_SEG_HDR_GTPU_IP},
+	{VIRTCHNL_PROTO_HDR_GTPU_EH,	ICE_FLOW_SEG_HDR_GTPU_EH},
+	{VIRTCHNL_PROTO_HDR_GTPU_EH_PDU_DWN,
+					ICE_FLOW_SEG_HDR_GTPU_DWN},
+	{VIRTCHNL_PROTO_HDR_GTPU_EH_PDU_UP,
+					ICE_FLOW_SEG_HDR_GTPU_UP},
+	{VIRTCHNL_PROTO_HDR_L2TPV3,	ICE_FLOW_SEG_HDR_L2TPV3},
+	{VIRTCHNL_PROTO_HDR_ESP,	ICE_FLOW_SEG_HDR_ESP},
+	{VIRTCHNL_PROTO_HDR_AH,		ICE_FLOW_SEG_HDR_AH},
+	{VIRTCHNL_PROTO_HDR_PFCP,	ICE_FLOW_SEG_HDR_PFCP_SESSION},
+};
+
+struct ice_vc_hash_field_match_type {
+	u32 vc_hdr;		/* virtchnl headers
+				 * (VIRTCHNL_PROTO_HDR_XXX)
+				 */
+	u32 vc_hash_field;	/* virtchnl hash fields selector
+				 * FIELD_SELECTOR((VIRTCHNL_PROTO_HDR_ETH_XXX))
+				 */
+	u64 ice_hash_field;	/* ice hash fields
+				 * (BIT_ULL(ICE_FLOW_FIELD_IDX_XXX))
+				 */
+};
+
+static const struct
+ice_vc_hash_field_match_type ice_vc_hash_field_list_os[] = {
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_SA)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_DA)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST),
+		ICE_FLOW_HASH_IPV4},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_SA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_DA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		ICE_FLOW_HASH_IPV4 | BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_SA)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_DA)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST),
+		ICE_FLOW_HASH_IPV6},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_SA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_DA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		ICE_FLOW_HASH_IPV6 | BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_TCP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_SRC_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_TCP_SRC_PORT)},
+	{VIRTCHNL_PROTO_HDR_TCP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_DST_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_TCP_DST_PORT)},
+	{VIRTCHNL_PROTO_HDR_TCP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_SRC_PORT) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_DST_PORT),
+		ICE_FLOW_HASH_TCP_PORT},
+	{VIRTCHNL_PROTO_HDR_UDP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_SRC_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_UDP_SRC_PORT)},
+	{VIRTCHNL_PROTO_HDR_UDP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_DST_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_UDP_DST_PORT)},
+	{VIRTCHNL_PROTO_HDR_UDP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_SRC_PORT) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_DST_PORT),
+		ICE_FLOW_HASH_UDP_PORT},
+	{VIRTCHNL_PROTO_HDR_SCTP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_SRC_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_SCTP_SRC_PORT)},
+	{VIRTCHNL_PROTO_HDR_SCTP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_DST_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_SCTP_DST_PORT)},
+	{VIRTCHNL_PROTO_HDR_SCTP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_SRC_PORT) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_DST_PORT),
+		ICE_FLOW_HASH_SCTP_PORT},
+};
+
+static const struct
+ice_vc_hash_field_match_type ice_vc_hash_field_list_comms[] = {
+	{VIRTCHNL_PROTO_HDR_ETH, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_ETH_SRC),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_ETH_SA)},
+	{VIRTCHNL_PROTO_HDR_ETH, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_ETH_DST),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_ETH_DA)},
+	{VIRTCHNL_PROTO_HDR_ETH, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_ETH_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_ETH_DST),
+		ICE_FLOW_HASH_ETH},
+	{VIRTCHNL_PROTO_HDR_ETH,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_ETH_ETHERTYPE),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_ETH_TYPE)},
+	{VIRTCHNL_PROTO_HDR_S_VLAN,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_S_VLAN_ID),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_S_VLAN)},
+	{VIRTCHNL_PROTO_HDR_C_VLAN,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_C_VLAN_ID),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_C_VLAN)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_SA)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_DA)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST),
+		ICE_FLOW_HASH_IPV4},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_SA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_DA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		ICE_FLOW_HASH_IPV4 | BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV4, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV4_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV4_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_SA)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_DA)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST),
+		ICE_FLOW_HASH_IPV6},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_SA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_DA) |
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_SRC) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_DST) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		ICE_FLOW_HASH_IPV6 | BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_IPV6, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_IPV6_PROT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_IPV6_PROT)},
+	{VIRTCHNL_PROTO_HDR_TCP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_SRC_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_TCP_SRC_PORT)},
+	{VIRTCHNL_PROTO_HDR_TCP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_DST_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_TCP_DST_PORT)},
+	{VIRTCHNL_PROTO_HDR_TCP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_SRC_PORT) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_TCP_DST_PORT),
+		ICE_FLOW_HASH_TCP_PORT},
+	{VIRTCHNL_PROTO_HDR_UDP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_SRC_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_UDP_SRC_PORT)},
+	{VIRTCHNL_PROTO_HDR_UDP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_DST_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_UDP_DST_PORT)},
+	{VIRTCHNL_PROTO_HDR_UDP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_SRC_PORT) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_UDP_DST_PORT),
+		ICE_FLOW_HASH_UDP_PORT},
+	{VIRTCHNL_PROTO_HDR_SCTP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_SRC_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_SCTP_SRC_PORT)},
+	{VIRTCHNL_PROTO_HDR_SCTP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_DST_PORT),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_SCTP_DST_PORT)},
+	{VIRTCHNL_PROTO_HDR_SCTP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_SRC_PORT) |
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_SCTP_DST_PORT),
+		ICE_FLOW_HASH_SCTP_PORT},
+	{VIRTCHNL_PROTO_HDR_PPPOE,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_PPPOE_SESS_ID),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_PPPOE_SESS_ID)},
+	{VIRTCHNL_PROTO_HDR_GTPU_IP,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_GTPU_IP_TEID),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_GTPU_IP_TEID)},
+	{VIRTCHNL_PROTO_HDR_L2TPV3,
+		FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_L2TPV3_SESS_ID),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_L2TPV3_SESS_ID)},
+	{VIRTCHNL_PROTO_HDR_ESP, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_ESP_SPI),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_ESP_SPI)},
+	{VIRTCHNL_PROTO_HDR_AH, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_AH_SPI),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_AH_SPI)},
+	{VIRTCHNL_PROTO_HDR_PFCP, FIELD_SELECTOR(VIRTCHNL_PROTO_HDR_PFCP_SEID),
+		BIT_ULL(ICE_FLOW_FIELD_IDX_PFCP_SEID)},
+};
+
 /**
  * ice_get_vf_vsi - get VF's VSI based on the stored index
  * @vf: VF used to get VSI
@@ -2121,6 +2361,9 @@ static int ice_vc_get_vf_res_msg(struct ice_vf *vf, u8 *msg)
 	if (vf->driver_caps & VIRTCHNL_VF_CAP_ADV_LINK_SPEED)
 		vfres->vf_cap_flags |= VIRTCHNL_VF_CAP_ADV_LINK_SPEED;
 
+	if (vf->driver_caps & VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF)
+		vfres->vf_cap_flags |= VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF;
+
 	if (vf->driver_caps & VIRTCHNL_VF_OFFLOAD_USO)
 		vfres->vf_cap_flags |= VIRTCHNL_VF_OFFLOAD_USO;
 
@@ -2234,6 +2477,210 @@ static bool ice_vc_isvalid_ring_len(u16 ring_len)
 		!(ring_len % ICE_REQ_DESC_MULTIPLE));
 }
 
+/**
+ * ice_vc_parse_rss_cfg - parses hash fields and headers from
+ * a specific virtchnl RSS cfg
+ * @hw: pointer to the hardware
+ * @rss_cfg: pointer to the virtchnl RSS cfg
+ * @addl_hdrs: pointer to the protocol header fields (ICE_FLOW_SEG_HDR_*)
+ * to configure
+ * @hash_flds: pointer to the hash bit fields (ICE_FLOW_HASH_*) to configure
+ *
+ * Return true if all the protocol header and hash fields in the RSS cfg could
+ * be parsed, else return false
+ *
+ * This function parses the virtchnl RSS cfg to be the intended
+ * hash fields and the intended header for RSS configuration
+ */
+static bool
+ice_vc_parse_rss_cfg(struct ice_hw *hw, struct virtchnl_rss_cfg *rss_cfg,
+		     u32 *addl_hdrs, u64 *hash_flds)
+{
+	const struct ice_vc_hash_field_match_type *hf_list;
+	const struct ice_vc_hdr_match_type *hdr_list;
+	int i, hf_list_len, hdr_list_len;
+
+	if (!strncmp(hw->active_pkg_name, "ICE COMMS Package",
+		     sizeof(hw->active_pkg_name))) {
+		hf_list = ice_vc_hash_field_list_comms;
+		hf_list_len = ARRAY_SIZE(ice_vc_hash_field_list_comms);
+		hdr_list = ice_vc_hdr_list_comms;
+		hdr_list_len = ARRAY_SIZE(ice_vc_hdr_list_comms);
+	} else {
+		hf_list = ice_vc_hash_field_list_os;
+		hf_list_len = ARRAY_SIZE(ice_vc_hash_field_list_os);
+		hdr_list = ice_vc_hdr_list_os;
+		hdr_list_len = ARRAY_SIZE(ice_vc_hdr_list_os);
+	}
+
+	for (i = 0; i < rss_cfg->proto_hdrs.count; i++) {
+		struct virtchnl_proto_hdr *proto_hdr =
+					&rss_cfg->proto_hdrs.proto_hdr[i];
+		bool hdr_found = false;
+		int j;
+
+		/* Find matched ice headers according to virtchnl headers. */
+		for (j = 0; j < hdr_list_len; j++) {
+			struct ice_vc_hdr_match_type hdr_map = hdr_list[j];
+
+			if (proto_hdr->type == hdr_map.vc_hdr) {
+				*addl_hdrs |= hdr_map.ice_hdr;
+				hdr_found = true;
+			}
+		}
+
+		if (!hdr_found)
+			return false;
+
+		/* Find matched ice hash fields according to
+		 * virtchnl hash fields.
+		 */
+		for (j = 0; j < hf_list_len; j++) {
+			struct ice_vc_hash_field_match_type hf_map = hf_list[j];
+
+			if (proto_hdr->type == hf_map.vc_hdr &&
+			    proto_hdr->field_selector == hf_map.vc_hash_field) {
+				*hash_flds |= hf_map.ice_hash_field;
+				break;
+			}
+		}
+	}
+
+	return true;
+}
+
+/**
+ * ice_vf_adv_rss_offload_ena - determine if capabilities support advanced
+ * RSS offloads
+ * @caps: VF driver negotiated capabilities
+ *
+ * Return true if VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF capability is set,
+ * else return false
+ */
+static bool ice_vf_adv_rss_offload_ena(u32 caps)
+{
+	return !!(caps & VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF);
+}
+
+/**
+ * ice_vc_handle_rss_cfg
+ * @vf: pointer to the VF info
+ * @msg: pointer to the message buffer
+ * @add: add a RSS config if true, otherwise delete a RSS config
+ *
+ * This function adds/deletes a RSS config
+ */
+static int ice_vc_handle_rss_cfg(struct ice_vf *vf, u8 *msg, bool add)
+{
+	u32 v_opcode = add ? VIRTCHNL_OP_ADD_RSS_CFG : VIRTCHNL_OP_DEL_RSS_CFG;
+	struct virtchnl_rss_cfg *rss_cfg = (struct virtchnl_rss_cfg *)msg;
+	enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
+	struct device *dev = ice_pf_to_dev(vf->pf);
+	struct ice_hw *hw = &vf->pf->hw;
+	struct ice_vsi *vsi;
+
+	if (!test_bit(ICE_FLAG_RSS_ENA, vf->pf->flags)) {
+		dev_dbg(dev, "VF %d attempting to configure RSS, but RSS is not supported by the PF\n",
+			vf->vf_id);
+		v_ret = VIRTCHNL_STATUS_ERR_NOT_SUPPORTED;
+		goto error_param;
+	}
+
+	if (!ice_vf_adv_rss_offload_ena(vf->driver_caps)) {
+		dev_dbg(dev, "VF %d attempting to configure RSS, but Advanced RSS offload is not supported\n",
+			vf->vf_id);
+		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+		goto error_param;
+	}
+
+	if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
+		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+		goto error_param;
+	}
+
+	if (rss_cfg->proto_hdrs.count > VIRTCHNL_MAX_NUM_PROTO_HDRS ||
+	    rss_cfg->rss_algorithm < VIRTCHNL_RSS_ALG_TOEPLITZ_ASYMMETRIC ||
+	    rss_cfg->rss_algorithm > VIRTCHNL_RSS_ALG_XOR_SYMMETRIC) {
+		dev_dbg(dev, "VF %d attempting to configure RSS, but RSS configuration is not valid\n",
+			vf->vf_id);
+		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+		goto error_param;
+	}
+
+	vsi = ice_get_vf_vsi(vf);
+	if (!vsi) {
+		v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+		goto error_param;
+	}
+
+	if (rss_cfg->rss_algorithm == VIRTCHNL_RSS_ALG_R_ASYMMETRIC) {
+		struct ice_vsi_ctx *ctx;
+		enum ice_status status;
+		u8 lut_type, hash_type;
+
+		lut_type = ICE_AQ_VSI_Q_OPT_RSS_LUT_VSI;
+		hash_type = add ? ICE_AQ_VSI_Q_OPT_RSS_XOR :
+				ICE_AQ_VSI_Q_OPT_RSS_TPLZ;
+
+		ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+		if (!ctx) {
+			v_ret = VIRTCHNL_STATUS_ERR_NO_MEMORY;
+			goto error_param;
+		}
+
+		ctx->info.q_opt_rss = ((lut_type <<
+					ICE_AQ_VSI_Q_OPT_RSS_LUT_S) &
+				       ICE_AQ_VSI_Q_OPT_RSS_LUT_M) |
+				       (hash_type &
+					ICE_AQ_VSI_Q_OPT_RSS_HASH_M);
+
+		/* Preserve existing queueing option setting */
+		ctx->info.q_opt_rss |= (vsi->info.q_opt_rss &
+					  ICE_AQ_VSI_Q_OPT_RSS_GBL_LUT_M);
+		ctx->info.q_opt_tc = vsi->info.q_opt_tc;
+		ctx->info.q_opt_flags = vsi->info.q_opt_rss;
+
+		ctx->info.valid_sections =
+				cpu_to_le16(ICE_AQ_VSI_PROP_Q_OPT_VALID);
+
+		status = ice_update_vsi(hw, vsi->idx, ctx, NULL);
+		if (status) {
+			dev_err(dev, "update VSI for RSS failed, err %s aq_err %s\n",
+				ice_stat_str(status),
+				ice_aq_str(hw->adminq.sq_last_status));
+			v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+		} else {
+			vsi->info.q_opt_rss = ctx->info.q_opt_rss;
+		}
+
+		kfree(ctx);
+	} else {
+		u32 addl_hdrs = ICE_FLOW_SEG_HDR_NONE;
+		u64 hash_flds = ICE_HASH_INVALID;
+
+		if (!ice_vc_parse_rss_cfg(hw, rss_cfg, &addl_hdrs,
+					  &hash_flds)) {
+			v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+			goto error_param;
+		}
+
+		if (add) {
+			if (ice_add_rss_cfg(hw, vsi->idx, hash_flds,
+					    addl_hdrs)) {
+				v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+				dev_err(dev, "ice_add_rss_cfg failed for vsi = %d, v_ret = %d\n",
+					vsi->vsi_num, v_ret);
+			}
+		} else {
+			v_ret = VIRTCHNL_STATUS_ERR_NOT_SUPPORTED;
+			dev_err(dev, "RSS removal not supported\n");
+		}
+	}
+
+error_param:
+	return ice_vc_send_msg_to_vf(vf, v_opcode, v_ret, NULL, 0);
+}
+
 /**
  * ice_vc_config_rss_key
  * @vf: pointer to the VF info
@@ -3931,6 +4378,12 @@ void ice_vc_process_vf_msg(struct ice_pf *pf, struct ice_rq_event_info *event)
 	case VIRTCHNL_OP_DEL_FDIR_FILTER:
 		err = ice_vc_del_fdir_fltr(vf, msg);
 		break;
+	case VIRTCHNL_OP_ADD_RSS_CFG:
+		err = ice_vc_handle_rss_cfg(vf, msg, true);
+		break;
+	case VIRTCHNL_OP_DEL_RSS_CFG:
+		err = ice_vc_handle_rss_cfg(vf, msg, false);
+		break;
 	case VIRTCHNL_OP_UNKNOWN:
 	default:
 		dev_err(dev, "Unsupported opcode %d from VF %d\n", v_opcode,
diff --git a/include/linux/avf/virtchnl.h b/include/linux/avf/virtchnl.h
index 9e0341cf2c36..565deea6ffe8 100644
--- a/include/linux/avf/virtchnl.h
+++ b/include/linux/avf/virtchnl.h
@@ -136,7 +136,9 @@ enum virtchnl_ops {
 	VIRTCHNL_OP_DISABLE_CHANNELS = 31,
 	VIRTCHNL_OP_ADD_CLOUD_FILTER = 32,
 	VIRTCHNL_OP_DEL_CLOUD_FILTER = 33,
-	/* opcode 34 - 46 are reserved */
+	/* opcode 34 - 44 are reserved */
+	VIRTCHNL_OP_ADD_RSS_CFG = 45,
+	VIRTCHNL_OP_DEL_RSS_CFG = 46,
 	VIRTCHNL_OP_ADD_FDIR_FILTER = 47,
 	VIRTCHNL_OP_DEL_FDIR_FILTER = 48,
 	VIRTCHNL_OP_MAX,
@@ -252,6 +254,7 @@ VIRTCHNL_CHECK_STRUCT_LEN(16, virtchnl_vsi_resource);
 #define VIRTCHNL_VF_OFFLOAD_RX_ENCAP_CSUM	0X00400000
 #define VIRTCHNL_VF_OFFLOAD_ADQ			0X00800000
 #define VIRTCHNL_VF_OFFLOAD_USO			0X02000000
+#define VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF		0X08000000
 #define VIRTCHNL_VF_OFFLOAD_FDIR_PF		0X10000000
 
 /* Define below the capability flags that are not offloads */
@@ -677,6 +680,14 @@ enum virtchnl_vfr_states {
 	VIRTCHNL_VFR_VFACTIVE,
 };
 
+/* Type of RSS algorithm */
+enum virtchnl_rss_algorithm {
+	VIRTCHNL_RSS_ALG_TOEPLITZ_ASYMMETRIC	= 0,
+	VIRTCHNL_RSS_ALG_R_ASYMMETRIC		= 1,
+	VIRTCHNL_RSS_ALG_TOEPLITZ_SYMMETRIC	= 2,
+	VIRTCHNL_RSS_ALG_XOR_SYMMETRIC		= 3,
+};
+
 #define VIRTCHNL_MAX_NUM_PROTO_HDRS	32
 #define PROTO_HDR_SHIFT			5
 #define PROTO_HDR_FIELD_START(proto_hdr_type) ((proto_hdr_type) << PROTO_HDR_SHIFT)
@@ -832,6 +843,14 @@ struct virtchnl_proto_hdrs {
 
 VIRTCHNL_CHECK_STRUCT_LEN(2312, virtchnl_proto_hdrs);
 
+struct virtchnl_rss_cfg {
+	struct virtchnl_proto_hdrs proto_hdrs;	   /* protocol headers */
+	enum virtchnl_rss_algorithm rss_algorithm; /* RSS algorithm type */
+	u8 reserved[128];			   /* reserve for future */
+};
+
+VIRTCHNL_CHECK_STRUCT_LEN(2444, virtchnl_rss_cfg);
+
 /* action configuration for FDIR */
 struct virtchnl_filter_action {
 	enum virtchnl_action type;
@@ -1100,6 +1119,10 @@ virtchnl_vc_validate_vf_msg(struct virtchnl_version_info *ver, u32 v_opcode,
 	case VIRTCHNL_OP_DEL_CLOUD_FILTER:
 		valid_len = sizeof(struct virtchnl_filter);
 		break;
+	case VIRTCHNL_OP_ADD_RSS_CFG:
+	case VIRTCHNL_OP_DEL_RSS_CFG:
+		valid_len = sizeof(struct virtchnl_rss_cfg);
+		break;
 	case VIRTCHNL_OP_ADD_FDIR_FILTER:
 		valid_len = sizeof(struct virtchnl_fdir_add);
 		break;
-- 
2.26.2

[PATCH net-next 08/12] ice: Support RSS configure removal for AVF

[Tony Nguyen]

From: Qi Zhang <qi.z.zhang@intel.com>

Add the handler for virtchnl message VIRTCHNL_OP_DEL_RSS_CFG to remove
an existing RSS configuration with matching hashed fields.

Signed-off-by: Vignesh Sridhar <vignesh.sridhar@intel.com>
Co-developed-by: Jia Guo <jia.guo@intel.com>
Signed-off-by: Jia Guo <jia.guo@intel.com>
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Tested-by: Bo Chen <BoX.C.Chen@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_flow.c     | 88 +++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_flow.h     |  3 +
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  | 16 +++-
 3 files changed, 105 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_flow.c b/drivers/net/ethernet/intel/ice/ice_flow.c
index 4d59eb96383b..f160672448a0 100644
--- a/drivers/net/ethernet/intel/ice/ice_flow.c
+++ b/drivers/net/ethernet/intel/ice/ice_flow.c
@@ -2149,6 +2149,94 @@ ice_add_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds,
 	return status;
 }
 
+/**
+ * ice_rem_rss_cfg_sync - remove an existing RSS configuration
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: software VSI handle
+ * @hashed_flds: Packet hash types (ICE_FLOW_HASH_*) to remove
+ * @addl_hdrs: Protocol header fields within a packet segment
+ * @segs_cnt: packet segment count
+ *
+ * Assumption: lock has already been acquired for RSS list
+ */
+static enum ice_status
+ice_rem_rss_cfg_sync(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds,
+		     u32 addl_hdrs, u8 segs_cnt)
+{
+	const enum ice_block blk = ICE_BLK_RSS;
+	struct ice_flow_seg_info *segs;
+	struct ice_flow_prof *prof;
+	enum ice_status status;
+
+	segs = kcalloc(segs_cnt, sizeof(*segs), GFP_KERNEL);
+	if (!segs)
+		return ICE_ERR_NO_MEMORY;
+
+	/* Construct the packet segment info from the hashed fields */
+	status = ice_flow_set_rss_seg_info(&segs[segs_cnt - 1], hashed_flds,
+					   addl_hdrs);
+	if (status)
+		goto out;
+
+	prof = ice_flow_find_prof_conds(hw, blk, ICE_FLOW_RX, segs, segs_cnt,
+					vsi_handle,
+					ICE_FLOW_FIND_PROF_CHK_FLDS);
+	if (!prof) {
+		status = ICE_ERR_DOES_NOT_EXIST;
+		goto out;
+	}
+
+	status = ice_flow_disassoc_prof(hw, blk, prof, vsi_handle);
+	if (status)
+		goto out;
+
+	/* Remove RSS configuration from VSI context before deleting
+	 * the flow profile.
+	 */
+	ice_rem_rss_list(hw, vsi_handle, prof);
+
+	if (bitmap_empty(prof->vsis, ICE_MAX_VSI))
+		status = ice_flow_rem_prof(hw, blk, prof->id);
+
+out:
+	kfree(segs);
+	return status;
+}
+
+/**
+ * ice_rem_rss_cfg - remove an existing RSS config with matching hashed fields
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: software VSI handle
+ * @hashed_flds: Packet hash types (ICE_FLOW_HASH_*) to remove
+ * @addl_hdrs: Protocol header fields within a packet segment
+ *
+ * This function will lookup the flow profile based on the input
+ * hash field bitmap, iterate through the profile entry list of
+ * that profile and find entry associated with input VSI to be
+ * removed. Calls are made to underlying flow s which will APIs
+ * turn build or update buffers for RSS XLT1 section.
+ */
+enum ice_status __maybe_unused
+ice_rem_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds,
+		u32 addl_hdrs)
+{
+	enum ice_status status;
+
+	if (hashed_flds == ICE_HASH_INVALID ||
+	    !ice_is_vsi_valid(hw, vsi_handle))
+		return ICE_ERR_PARAM;
+
+	mutex_lock(&hw->rss_locks);
+	status = ice_rem_rss_cfg_sync(hw, vsi_handle, hashed_flds, addl_hdrs,
+				      ICE_RSS_OUTER_HEADERS);
+	if (!status)
+		status = ice_rem_rss_cfg_sync(hw, vsi_handle, hashed_flds,
+					      addl_hdrs, ICE_RSS_INNER_HEADERS);
+	mutex_unlock(&hw->rss_locks);
+
+	return status;
+}
+
 /* Mapping of AVF hash bit fields to an L3-L4 hash combination.
  * As the ice_flow_avf_hdr_field represent individual bit shifts in a hash,
  * convert its values to their appropriate flow L3, L4 values.
diff --git a/drivers/net/ethernet/intel/ice/ice_flow.h b/drivers/net/ethernet/intel/ice/ice_flow.h
index 2f68b59ace7e..2a2d8c1536cb 100644
--- a/drivers/net/ethernet/intel/ice/ice_flow.h
+++ b/drivers/net/ethernet/intel/ice/ice_flow.h
@@ -409,5 +409,8 @@ enum ice_status ice_rem_vsi_rss_cfg(struct ice_hw *hw, u16 vsi_handle);
 enum ice_status
 ice_add_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds,
 		u32 addl_hdrs);
+enum ice_status
+ice_rem_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds,
+		u32 addl_hdrs);
 u64 ice_get_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u32 hdrs);
 #endif /* _ICE_FLOW_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index ca778a80d363..a1d22d2aa0bd 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -2672,8 +2672,20 @@ static int ice_vc_handle_rss_cfg(struct ice_vf *vf, u8 *msg, bool add)
 					vsi->vsi_num, v_ret);
 			}
 		} else {
-			v_ret = VIRTCHNL_STATUS_ERR_NOT_SUPPORTED;
-			dev_err(dev, "RSS removal not supported\n");
+			enum ice_status status;
+
+			status = ice_rem_rss_cfg(hw, vsi->idx, hash_flds,
+						 addl_hdrs);
+			/* We just ignore ICE_ERR_DOES_NOT_EXIST, because
+			 * if two configurations share the same profile remove
+			 * one of them actually removes both, since the
+			 * profile is deleted.
+			 */
+			if (status && status != ICE_ERR_DOES_NOT_EXIST) {
+				v_ret = VIRTCHNL_STATUS_ERR_PARAM;
+				dev_err(dev, "ice_rem_rss_cfg failed for VF ID:%d, error:%s\n",
+					vf->vf_id, ice_stat_str(status));
+			}
 		}
 	}
 
-- 
2.26.2

[PATCH net-next 09/12] iavf: Add framework to enable ethtool RSS config

[Tony Nguyen]

From: Haiyue Wang <haiyue.wang@intel.com>

Add the virtchnl message interface to VF, so that VF can request RSS
input set(s) based on PF's capability.

This framework allows ethtool RSS config support on the VF driver.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/iavf/iavf.h        |  10 ++
 .../net/ethernet/intel/iavf/iavf_adv_rss.h    |  25 +++
 drivers/net/ethernet/intel/iavf/iavf_main.c   |  27 ++++
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   | 152 ++++++++++++++++++
 4 files changed, 214 insertions(+)
 create mode 100644 drivers/net/ethernet/intel/iavf/iavf_adv_rss.h

diff --git a/drivers/net/ethernet/intel/iavf/iavf.h b/drivers/net/ethernet/intel/iavf/iavf.h
index bda2a900df8e..e8bd04100ecd 100644
--- a/drivers/net/ethernet/intel/iavf/iavf.h
+++ b/drivers/net/ethernet/intel/iavf/iavf.h
@@ -38,6 +38,7 @@
 #include <linux/avf/virtchnl.h>
 #include "iavf_txrx.h"
 #include "iavf_fdir.h"
+#include "iavf_adv_rss.h"
 
 #define DEFAULT_DEBUG_LEVEL_SHIFT 3
 #define PFX "iavf: "
@@ -303,6 +304,8 @@ struct iavf_adapter {
 #define IAVF_FLAG_AQ_DEL_CLOUD_FILTER		BIT(24)
 #define IAVF_FLAG_AQ_ADD_FDIR_FILTER		BIT(25)
 #define IAVF_FLAG_AQ_DEL_FDIR_FILTER		BIT(26)
+#define IAVF_FLAG_AQ_ADD_ADV_RSS_CFG		BIT(27)
+#define IAVF_FLAG_AQ_DEL_ADV_RSS_CFG		BIT(28)
 
 	/* OS defined structs */
 	struct net_device *netdev;
@@ -345,6 +348,8 @@ struct iavf_adapter {
 			      VIRTCHNL_VF_CAP_ADV_LINK_SPEED)
 #define FDIR_FLTR_SUPPORT(_a) ((_a)->vf_res->vf_cap_flags & \
 			       VIRTCHNL_VF_OFFLOAD_FDIR_PF)
+#define ADV_RSS_SUPPORT(_a) ((_a)->vf_res->vf_cap_flags & \
+			     VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF)
 	struct virtchnl_vf_resource *vf_res; /* incl. all VSIs */
 	struct virtchnl_vsi_resource *vsi_res; /* our LAN VSI */
 	struct virtchnl_version_info pf_version;
@@ -372,6 +377,9 @@ struct iavf_adapter {
 	u16 fdir_active_fltr;
 	struct list_head fdir_list_head;
 	spinlock_t fdir_fltr_lock;	/* protect the Flow Director filter list */
+
+	struct list_head adv_rss_list_head;
+	spinlock_t adv_rss_lock;	/* protect the RSS management list */
 };
 
 
@@ -444,6 +452,8 @@ void iavf_add_cloud_filter(struct iavf_adapter *adapter);
 void iavf_del_cloud_filter(struct iavf_adapter *adapter);
 void iavf_add_fdir_filter(struct iavf_adapter *adapter);
 void iavf_del_fdir_filter(struct iavf_adapter *adapter);
+void iavf_add_adv_rss_cfg(struct iavf_adapter *adapter);
+void iavf_del_adv_rss_cfg(struct iavf_adapter *adapter);
 struct iavf_mac_filter *iavf_add_filter(struct iavf_adapter *adapter,
 					const u8 *macaddr);
 #endif /* _IAVF_H_ */
diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
new file mode 100644
index 000000000000..66262090e697
--- /dev/null
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2021, Intel Corporation. */
+
+#ifndef _IAVF_ADV_RSS_H_
+#define _IAVF_ADV_RSS_H_
+
+struct iavf_adapter;
+
+/* State of advanced RSS configuration */
+enum iavf_adv_rss_state_t {
+	IAVF_ADV_RSS_ADD_REQUEST,	/* User requests to add RSS */
+	IAVF_ADV_RSS_ADD_PENDING,	/* RSS pending add by the PF */
+	IAVF_ADV_RSS_DEL_REQUEST,	/* Driver requests to delete RSS */
+	IAVF_ADV_RSS_DEL_PENDING,	/* RSS pending delete by the PF */
+	IAVF_ADV_RSS_ACTIVE,		/* RSS configuration is active */
+};
+
+/* bookkeeping of advanced RSS configuration */
+struct iavf_adv_rss {
+	enum iavf_adv_rss_state_t state;
+	struct list_head list;
+
+	struct virtchnl_rss_cfg cfg_msg;
+};
+#endif /* _IAVF_ADV_RSS_H_ */
diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c
index 6cd6b9dfe5d4..7a81e7ceea65 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_main.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_main.c
@@ -962,6 +962,7 @@ void iavf_down(struct iavf_adapter *adapter)
 	struct iavf_cloud_filter *cf;
 	struct iavf_fdir_fltr *fdir;
 	struct iavf_mac_filter *f;
+	struct iavf_adv_rss *rss;
 
 	if (adapter->state <= __IAVF_DOWN_PENDING)
 		return;
@@ -1004,6 +1005,12 @@ void iavf_down(struct iavf_adapter *adapter)
 	}
 	spin_unlock_bh(&adapter->fdir_fltr_lock);
 
+	/* remove all advance RSS configuration */
+	spin_lock_bh(&adapter->adv_rss_lock);
+	list_for_each_entry(rss, &adapter->adv_rss_list_head, list)
+		rss->state = IAVF_ADV_RSS_DEL_REQUEST;
+	spin_unlock_bh(&adapter->adv_rss_lock);
+
 	if (!(adapter->flags & IAVF_FLAG_PF_COMMS_FAILED) &&
 	    adapter->state != __IAVF_RESETTING) {
 		/* cancel any current operation */
@@ -1016,6 +1023,7 @@ void iavf_down(struct iavf_adapter *adapter)
 		adapter->aq_required |= IAVF_FLAG_AQ_DEL_VLAN_FILTER;
 		adapter->aq_required |= IAVF_FLAG_AQ_DEL_CLOUD_FILTER;
 		adapter->aq_required |= IAVF_FLAG_AQ_DEL_FDIR_FILTER;
+		adapter->aq_required |= IAVF_FLAG_AQ_DEL_ADV_RSS_CFG;
 		adapter->aq_required |= IAVF_FLAG_AQ_DISABLE_QUEUES;
 	}
 
@@ -1646,6 +1654,14 @@ static int iavf_process_aq_command(struct iavf_adapter *adapter)
 		iavf_del_fdir_filter(adapter);
 		return IAVF_SUCCESS;
 	}
+	if (adapter->aq_required & IAVF_FLAG_AQ_ADD_ADV_RSS_CFG) {
+		iavf_add_adv_rss_cfg(adapter);
+		return 0;
+	}
+	if (adapter->aq_required & IAVF_FLAG_AQ_DEL_ADV_RSS_CFG) {
+		iavf_del_adv_rss_cfg(adapter);
+		return 0;
+	}
 	return -EAGAIN;
 }
 
@@ -3758,11 +3774,13 @@ static int iavf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	spin_lock_init(&adapter->mac_vlan_list_lock);
 	spin_lock_init(&adapter->cloud_filter_list_lock);
 	spin_lock_init(&adapter->fdir_fltr_lock);
+	spin_lock_init(&adapter->adv_rss_lock);
 
 	INIT_LIST_HEAD(&adapter->mac_filter_list);
 	INIT_LIST_HEAD(&adapter->vlan_filter_list);
 	INIT_LIST_HEAD(&adapter->cloud_filter_list);
 	INIT_LIST_HEAD(&adapter->fdir_list_head);
+	INIT_LIST_HEAD(&adapter->adv_rss_list_head);
 
 	INIT_WORK(&adapter->reset_task, iavf_reset_task);
 	INIT_WORK(&adapter->adminq_task, iavf_adminq_task);
@@ -3868,6 +3886,7 @@ static void iavf_remove(struct pci_dev *pdev)
 	struct iavf_adapter *adapter = netdev_priv(netdev);
 	struct iavf_fdir_fltr *fdir, *fdirtmp;
 	struct iavf_vlan_filter *vlf, *vlftmp;
+	struct iavf_adv_rss *rss, *rsstmp;
 	struct iavf_mac_filter *f, *ftmp;
 	struct iavf_cloud_filter *cf, *cftmp;
 	struct iavf_hw *hw = &adapter->hw;
@@ -3955,6 +3974,14 @@ static void iavf_remove(struct pci_dev *pdev)
 	}
 	spin_unlock_bh(&adapter->fdir_fltr_lock);
 
+	spin_lock_bh(&adapter->adv_rss_lock);
+	list_for_each_entry_safe(rss, rsstmp, &adapter->adv_rss_list_head,
+				 list) {
+		list_del(&rss->list);
+		kfree(rss);
+	}
+	spin_unlock_bh(&adapter->adv_rss_lock);
+
 	free_netdev(netdev);
 
 	pci_disable_pcie_error_reporting(pdev);
diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
index 9aaade0aae4c..54d2efe1732d 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
@@ -142,6 +142,7 @@ int iavf_send_vf_config_msg(struct iavf_adapter *adapter)
 	       VIRTCHNL_VF_OFFLOAD_ADQ |
 	       VIRTCHNL_VF_OFFLOAD_USO |
 	       VIRTCHNL_VF_OFFLOAD_FDIR_PF |
+	       VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF |
 	       VIRTCHNL_VF_CAP_ADV_LINK_SPEED;
 
 	adapter->current_op = VIRTCHNL_OP_GET_VF_RESOURCES;
@@ -1294,6 +1295,102 @@ void iavf_del_fdir_filter(struct iavf_adapter *adapter)
 	iavf_send_pf_msg(adapter, VIRTCHNL_OP_DEL_FDIR_FILTER, (u8 *)&f, len);
 }
 
+/**
+ * iavf_add_adv_rss_cfg
+ * @adapter: the VF adapter structure
+ *
+ * Request that the PF add RSS configuration as specified
+ * by the user via ethtool.
+ **/
+void iavf_add_adv_rss_cfg(struct iavf_adapter *adapter)
+{
+	struct virtchnl_rss_cfg *rss_cfg;
+	struct iavf_adv_rss *rss;
+	bool process_rss = false;
+	int len;
+
+	if (adapter->current_op != VIRTCHNL_OP_UNKNOWN) {
+		/* bail because we already have a command pending */
+		dev_err(&adapter->pdev->dev, "Cannot add RSS configuration, command %d pending\n",
+			adapter->current_op);
+		return;
+	}
+
+	len = sizeof(struct virtchnl_rss_cfg);
+	rss_cfg = kzalloc(len, GFP_KERNEL);
+	if (!rss_cfg)
+		return;
+
+	spin_lock_bh(&adapter->adv_rss_lock);
+	list_for_each_entry(rss, &adapter->adv_rss_list_head, list) {
+		if (rss->state == IAVF_ADV_RSS_ADD_REQUEST) {
+			process_rss = true;
+			rss->state = IAVF_ADV_RSS_ADD_PENDING;
+			memcpy(rss_cfg, &rss->cfg_msg, len);
+			break;
+		}
+	}
+	spin_unlock_bh(&adapter->adv_rss_lock);
+
+	if (process_rss) {
+		adapter->current_op = VIRTCHNL_OP_ADD_RSS_CFG;
+		iavf_send_pf_msg(adapter, VIRTCHNL_OP_ADD_RSS_CFG,
+				 (u8 *)rss_cfg, len);
+	} else {
+		adapter->aq_required &= ~IAVF_FLAG_AQ_ADD_ADV_RSS_CFG;
+	}
+
+	kfree(rss_cfg);
+}
+
+/**
+ * iavf_del_adv_rss_cfg
+ * @adapter: the VF adapter structure
+ *
+ * Request that the PF delete RSS configuration as specified
+ * by the user via ethtool.
+ **/
+void iavf_del_adv_rss_cfg(struct iavf_adapter *adapter)
+{
+	struct virtchnl_rss_cfg *rss_cfg;
+	struct iavf_adv_rss *rss;
+	bool process_rss = false;
+	int len;
+
+	if (adapter->current_op != VIRTCHNL_OP_UNKNOWN) {
+		/* bail because we already have a command pending */
+		dev_err(&adapter->pdev->dev, "Cannot remove RSS configuration, command %d pending\n",
+			adapter->current_op);
+		return;
+	}
+
+	len = sizeof(struct virtchnl_rss_cfg);
+	rss_cfg = kzalloc(len, GFP_KERNEL);
+	if (!rss_cfg)
+		return;
+
+	spin_lock_bh(&adapter->adv_rss_lock);
+	list_for_each_entry(rss, &adapter->adv_rss_list_head, list) {
+		if (rss->state == IAVF_ADV_RSS_DEL_REQUEST) {
+			process_rss = true;
+			rss->state = IAVF_ADV_RSS_DEL_PENDING;
+			memcpy(rss_cfg, &rss->cfg_msg, len);
+			break;
+		}
+	}
+	spin_unlock_bh(&adapter->adv_rss_lock);
+
+	if (process_rss) {
+		adapter->current_op = VIRTCHNL_OP_DEL_RSS_CFG;
+		iavf_send_pf_msg(adapter, VIRTCHNL_OP_DEL_RSS_CFG,
+				 (u8 *)rss_cfg, len);
+	} else {
+		adapter->aq_required &= ~IAVF_FLAG_AQ_DEL_ADV_RSS_CFG;
+	}
+
+	kfree(rss_cfg);
+}
+
 /**
  * iavf_request_reset
  * @adapter: adapter structure
@@ -1494,6 +1591,37 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter,
 			spin_unlock_bh(&adapter->fdir_fltr_lock);
 			}
 			break;
+		case VIRTCHNL_OP_ADD_RSS_CFG: {
+			struct iavf_adv_rss *rss, *rss_tmp;
+
+			spin_lock_bh(&adapter->adv_rss_lock);
+			list_for_each_entry_safe(rss, rss_tmp,
+						 &adapter->adv_rss_list_head,
+						 list) {
+				if (rss->state == IAVF_ADV_RSS_ADD_PENDING) {
+					list_del(&rss->list);
+					kfree(rss);
+				}
+			}
+			spin_unlock_bh(&adapter->adv_rss_lock);
+			}
+			break;
+		case VIRTCHNL_OP_DEL_RSS_CFG: {
+			struct iavf_adv_rss *rss;
+
+			spin_lock_bh(&adapter->adv_rss_lock);
+			list_for_each_entry(rss, &adapter->adv_rss_list_head,
+					    list) {
+				if (rss->state == IAVF_ADV_RSS_DEL_PENDING) {
+					rss->state = IAVF_ADV_RSS_ACTIVE;
+					dev_err(&adapter->pdev->dev, "Failed to delete RSS configuration, error %s\n",
+						iavf_stat_str(&adapter->hw,
+							      v_retval));
+				}
+			}
+			spin_unlock_bh(&adapter->adv_rss_lock);
+			}
+			break;
 		case VIRTCHNL_OP_ENABLE_VLAN_STRIPPING:
 		case VIRTCHNL_OP_DISABLE_VLAN_STRIPPING:
 			dev_warn(&adapter->pdev->dev, "Changing VLAN Stripping is not allowed when Port VLAN is configured\n");
@@ -1683,6 +1811,30 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter,
 		spin_unlock_bh(&adapter->fdir_fltr_lock);
 		}
 		break;
+	case VIRTCHNL_OP_ADD_RSS_CFG: {
+		struct iavf_adv_rss *rss;
+
+		spin_lock_bh(&adapter->adv_rss_lock);
+		list_for_each_entry(rss, &adapter->adv_rss_list_head, list)
+			if (rss->state == IAVF_ADV_RSS_ADD_PENDING)
+				rss->state = IAVF_ADV_RSS_ACTIVE;
+		spin_unlock_bh(&adapter->adv_rss_lock);
+		}
+		break;
+	case VIRTCHNL_OP_DEL_RSS_CFG: {
+		struct iavf_adv_rss *rss, *rss_tmp;
+
+		spin_lock_bh(&adapter->adv_rss_lock);
+		list_for_each_entry_safe(rss, rss_tmp,
+					 &adapter->adv_rss_list_head, list) {
+			if (rss->state == IAVF_ADV_RSS_DEL_PENDING) {
+				list_del(&rss->list);
+				kfree(rss);
+			}
+		}
+		spin_unlock_bh(&adapter->adv_rss_lock);
+		}
+		break;
 	default:
 		if (adapter->current_op && (v_opcode != adapter->current_op))
 			dev_warn(&adapter->pdev->dev, "Expected response %d from PF, received %d\n",
-- 
2.26.2

[PATCH net-next 10/12] iavf: Support for modifying TCP RSS flow hashing

[Tony Nguyen]

From: Haiyue Wang <haiyue.wang@intel.com>

Provides the ability to enable TCP RSS hashing by ethtool.

It gives users option of generating RSS hash based on the TCP source
and destination ports numbers, IPv4 or IPv6 source and destination
addresses.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/iavf/Makefile      |   1 +
 .../net/ethernet/intel/iavf/iavf_adv_rss.c    | 170 ++++++++++++++
 .../net/ethernet/intel/iavf/iavf_adv_rss.h    |  54 +++++
 .../net/ethernet/intel/iavf/iavf_ethtool.c    | 214 +++++++++++++++++-
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  15 +-
 5 files changed, 450 insertions(+), 4 deletions(-)
 create mode 100644 drivers/net/ethernet/intel/iavf/iavf_adv_rss.c

diff --git a/drivers/net/ethernet/intel/iavf/Makefile b/drivers/net/ethernet/intel/iavf/Makefile
index 121e194ee734..9c3e45c54d01 100644
--- a/drivers/net/ethernet/intel/iavf/Makefile
+++ b/drivers/net/ethernet/intel/iavf/Makefile
@@ -12,4 +12,5 @@ subdir-ccflags-y += -I$(src)
 obj-$(CONFIG_IAVF) += iavf.o
 
 iavf-objs := iavf_main.o iavf_ethtool.o iavf_virtchnl.o iavf_fdir.o \
+	     iavf_adv_rss.o \
 	     iavf_txrx.o iavf_common.o iavf_adminq.o iavf_client.o
diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
new file mode 100644
index 000000000000..4c5771cdc445
--- /dev/null
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
@@ -0,0 +1,170 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2021, Intel Corporation. */
+
+/* advanced RSS configuration ethtool support for iavf */
+
+#include "iavf.h"
+
+/**
+ * iavf_fill_adv_rss_ip4_hdr - fill the IPv4 RSS protocol header
+ * @hdr: the virtchnl message protocol header data structure
+ * @hash_flds: the RSS configuration protocol hash fields
+ */
+static void
+iavf_fill_adv_rss_ip4_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
+{
+	VIRTCHNL_SET_PROTO_HDR_TYPE(hdr, IPV4);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_IPV4_SA)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, IPV4, SRC);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_IPV4_DA)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, IPV4, DST);
+}
+
+/**
+ * iavf_fill_adv_rss_ip6_hdr - fill the IPv6 RSS protocol header
+ * @hdr: the virtchnl message protocol header data structure
+ * @hash_flds: the RSS configuration protocol hash fields
+ */
+static void
+iavf_fill_adv_rss_ip6_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
+{
+	VIRTCHNL_SET_PROTO_HDR_TYPE(hdr, IPV6);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_IPV6_SA)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, IPV6, SRC);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_IPV6_DA)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, IPV6, DST);
+}
+
+/**
+ * iavf_fill_adv_rss_tcp_hdr - fill the TCP RSS protocol header
+ * @hdr: the virtchnl message protocol header data structure
+ * @hash_flds: the RSS configuration protocol hash fields
+ */
+static void
+iavf_fill_adv_rss_tcp_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
+{
+	VIRTCHNL_SET_PROTO_HDR_TYPE(hdr, TCP);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, TCP, SRC_PORT);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, TCP, DST_PORT);
+}
+
+/**
+ * iavf_fill_adv_rss_cfg_msg - fill the RSS configuration into virtchnl message
+ * @rss_cfg: the virtchnl message to be filled with RSS configuration setting
+ * @packet_hdrs: the RSS configuration protocol header types
+ * @hash_flds: the RSS configuration protocol hash fields
+ *
+ * Returns 0 if the RSS configuration virtchnl message is filled successfully
+ */
+int
+iavf_fill_adv_rss_cfg_msg(struct virtchnl_rss_cfg *rss_cfg,
+			  u32 packet_hdrs, u64 hash_flds)
+{
+	struct virtchnl_proto_hdrs *proto_hdrs = &rss_cfg->proto_hdrs;
+	struct virtchnl_proto_hdr *hdr;
+
+	rss_cfg->rss_algorithm = VIRTCHNL_RSS_ALG_TOEPLITZ_ASYMMETRIC;
+
+	proto_hdrs->tunnel_level = 0;	/* always outer layer */
+
+	hdr = &proto_hdrs->proto_hdr[proto_hdrs->count++];
+	switch (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_L3) {
+	case IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4:
+		iavf_fill_adv_rss_ip4_hdr(hdr, hash_flds);
+		break;
+	case IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6:
+		iavf_fill_adv_rss_ip6_hdr(hdr, hash_flds);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	hdr = &proto_hdrs->proto_hdr[proto_hdrs->count++];
+	switch (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_L4) {
+	case IAVF_ADV_RSS_FLOW_SEG_HDR_TCP:
+		iavf_fill_adv_rss_tcp_hdr(hdr, hash_flds);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * iavf_find_adv_rss_cfg_by_hdrs - find RSS configuration with header type
+ * @adapter: pointer to the VF adapter structure
+ * @packet_hdrs: protocol header type to find.
+ *
+ * Returns pointer to advance RSS configuration if found or null
+ */
+struct iavf_adv_rss *
+iavf_find_adv_rss_cfg_by_hdrs(struct iavf_adapter *adapter, u32 packet_hdrs)
+{
+	struct iavf_adv_rss *rss;
+
+	list_for_each_entry(rss, &adapter->adv_rss_list_head, list)
+		if (rss->packet_hdrs == packet_hdrs)
+			return rss;
+
+	return NULL;
+}
+
+/**
+ * iavf_print_adv_rss_cfg
+ * @adapter: pointer to the VF adapter structure
+ * @rss: pointer to the advance RSS configuration to print
+ * @action: the string description about how to handle the RSS
+ * @result: the string description about the virtchnl result
+ *
+ * Print the advance RSS configuration
+ **/
+void
+iavf_print_adv_rss_cfg(struct iavf_adapter *adapter, struct iavf_adv_rss *rss,
+		       const char *action, const char *result)
+{
+	u32 packet_hdrs = rss->packet_hdrs;
+	u64 hash_flds = rss->hash_flds;
+	static char hash_opt[300];
+	const char *proto;
+
+	if (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_TCP)
+		proto = "TCP";
+	else
+		return;
+
+	memset(hash_opt, 0, sizeof(hash_opt));
+
+	strcat(hash_opt, proto);
+	if (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4)
+		strcat(hash_opt, "v4 ");
+	else
+		strcat(hash_opt, "v6 ");
+
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_IPV4_SA |
+			 IAVF_ADV_RSS_HASH_FLD_IPV6_SA))
+		strcat(hash_opt, "IP SA,");
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_IPV4_DA |
+			 IAVF_ADV_RSS_HASH_FLD_IPV6_DA))
+		strcat(hash_opt, "IP DA,");
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT)
+		strcat(hash_opt, "src port,");
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT)
+		strcat(hash_opt, "dst port,");
+
+	if (!action)
+		action = "";
+
+	if (!result)
+		result = "";
+
+	dev_info(&adapter->pdev->dev, "%s %s %s\n", action, hash_opt, result);
+}
diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
index 66262090e697..339ecb42938b 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
@@ -15,11 +15,65 @@ enum iavf_adv_rss_state_t {
 	IAVF_ADV_RSS_ACTIVE,		/* RSS configuration is active */
 };
 
+enum iavf_adv_rss_flow_seg_hdr {
+	IAVF_ADV_RSS_FLOW_SEG_HDR_NONE	= 0x00000000,
+	IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4	= 0x00000001,
+	IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6	= 0x00000002,
+	IAVF_ADV_RSS_FLOW_SEG_HDR_TCP	= 0x00000004,
+};
+
+#define IAVF_ADV_RSS_FLOW_SEG_HDR_L3		\
+	(IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4	|	\
+	 IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6)
+
+#define IAVF_ADV_RSS_FLOW_SEG_HDR_L4		\
+	(IAVF_ADV_RSS_FLOW_SEG_HDR_TCP)
+
+enum iavf_adv_rss_flow_field {
+	/* L3 */
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV4_SA,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV4_DA,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV6_SA,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV6_DA,
+	/* L4 */
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_SRC_PORT,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_DST_PORT,
+
+	/* The total number of enums must not exceed 64 */
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_MAX
+};
+
+#define IAVF_ADV_RSS_HASH_INVALID	0
+#define IAVF_ADV_RSS_HASH_FLD_IPV4_SA	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV4_SA)
+#define IAVF_ADV_RSS_HASH_FLD_IPV6_SA	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV6_SA)
+#define IAVF_ADV_RSS_HASH_FLD_IPV4_DA	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV4_DA)
+#define IAVF_ADV_RSS_HASH_FLD_IPV6_DA	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_IPV6_DA)
+#define IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_SRC_PORT)
+#define IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_DST_PORT)
+
 /* bookkeeping of advanced RSS configuration */
 struct iavf_adv_rss {
 	enum iavf_adv_rss_state_t state;
 	struct list_head list;
 
+	u32 packet_hdrs;
+	u64 hash_flds;
+
 	struct virtchnl_rss_cfg cfg_msg;
 };
+
+int
+iavf_fill_adv_rss_cfg_msg(struct virtchnl_rss_cfg *rss_cfg,
+			  u32 packet_hdrs, u64 hash_flds);
+struct iavf_adv_rss *
+iavf_find_adv_rss_cfg_by_hdrs(struct iavf_adapter *adapter, u32 packet_hdrs);
+void
+iavf_print_adv_rss_cfg(struct iavf_adapter *adapter, struct iavf_adv_rss *rss,
+		       const char *action, const char *result);
 #endif /* _IAVF_ADV_RSS_H_ */
diff --git a/drivers/net/ethernet/intel/iavf/iavf_ethtool.c b/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
index 3ebfef737f5c..d8dca5645c21 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
@@ -1418,6 +1418,214 @@ static int iavf_del_fdir_ethtool(struct iavf_adapter *adapter, struct ethtool_rx
 	return err;
 }
 
+/**
+ * iavf_adv_rss_parse_hdrs - parses headers from RSS hash input
+ * @cmd: ethtool rxnfc command
+ *
+ * This function parses the rxnfc command and returns intended
+ * header types for RSS configuration
+ */
+static u32 iavf_adv_rss_parse_hdrs(struct ethtool_rxnfc *cmd)
+{
+	u32 hdrs = IAVF_ADV_RSS_FLOW_SEG_HDR_NONE;
+
+	switch (cmd->flow_type) {
+	case TCP_V4_FLOW:
+		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |
+			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4;
+		break;
+	case TCP_V6_FLOW:
+		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |
+			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6;
+		break;
+	default:
+		break;
+	}
+
+	return hdrs;
+}
+
+/**
+ * iavf_adv_rss_parse_hash_flds - parses hash fields from RSS hash input
+ * @cmd: ethtool rxnfc command
+ *
+ * This function parses the rxnfc command and returns intended hash fields for
+ * RSS configuration
+ */
+static u64 iavf_adv_rss_parse_hash_flds(struct ethtool_rxnfc *cmd)
+{
+	u64 hfld = IAVF_ADV_RSS_HASH_INVALID;
+
+	if (cmd->data & RXH_IP_SRC || cmd->data & RXH_IP_DST) {
+		switch (cmd->flow_type) {
+		case TCP_V4_FLOW:
+			if (cmd->data & RXH_IP_SRC)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV4_SA;
+			if (cmd->data & RXH_IP_DST)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV4_DA;
+			break;
+		case TCP_V6_FLOW:
+			if (cmd->data & RXH_IP_SRC)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV6_SA;
+			if (cmd->data & RXH_IP_DST)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV6_DA;
+			break;
+		default:
+			break;
+		}
+	}
+
+	if (cmd->data & RXH_L4_B_0_1 || cmd->data & RXH_L4_B_2_3) {
+		switch (cmd->flow_type) {
+		case TCP_V4_FLOW:
+		case TCP_V6_FLOW:
+			if (cmd->data & RXH_L4_B_0_1)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT;
+			if (cmd->data & RXH_L4_B_2_3)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT;
+			break;
+		default:
+			break;
+		}
+	}
+
+	return hfld;
+}
+
+/**
+ * iavf_set_adv_rss_hash_opt - Enable/Disable flow types for RSS hash
+ * @adapter: pointer to the VF adapter structure
+ * @cmd: ethtool rxnfc command
+ *
+ * Returns Success if the flow input set is supported.
+ */
+static int
+iavf_set_adv_rss_hash_opt(struct iavf_adapter *adapter,
+			  struct ethtool_rxnfc *cmd)
+{
+	struct iavf_adv_rss *rss_old, *rss_new;
+	bool rss_new_add = false;
+	int count = 50, err = 0;
+	u64 hash_flds;
+	u32 hdrs;
+
+	if (!ADV_RSS_SUPPORT(adapter))
+		return -EOPNOTSUPP;
+
+	hdrs = iavf_adv_rss_parse_hdrs(cmd);
+	if (hdrs == IAVF_ADV_RSS_FLOW_SEG_HDR_NONE)
+		return -EINVAL;
+
+	hash_flds = iavf_adv_rss_parse_hash_flds(cmd);
+	if (hash_flds == IAVF_ADV_RSS_HASH_INVALID)
+		return -EINVAL;
+
+	rss_new = kzalloc(sizeof(*rss_new), GFP_KERNEL);
+	if (!rss_new)
+		return -ENOMEM;
+
+	if (iavf_fill_adv_rss_cfg_msg(&rss_new->cfg_msg, hdrs, hash_flds)) {
+		kfree(rss_new);
+		return -EINVAL;
+	}
+
+	while (test_and_set_bit(__IAVF_IN_CRITICAL_TASK,
+				&adapter->crit_section)) {
+		if (--count == 0) {
+			kfree(rss_new);
+			return -EINVAL;
+		}
+
+		udelay(1);
+	}
+
+	spin_lock_bh(&adapter->adv_rss_lock);
+	rss_old = iavf_find_adv_rss_cfg_by_hdrs(adapter, hdrs);
+	if (rss_old) {
+		if (rss_old->state != IAVF_ADV_RSS_ACTIVE) {
+			err = -EBUSY;
+		} else if (rss_old->hash_flds != hash_flds) {
+			rss_old->state = IAVF_ADV_RSS_ADD_REQUEST;
+			rss_old->hash_flds = hash_flds;
+			memcpy(&rss_old->cfg_msg, &rss_new->cfg_msg,
+			       sizeof(rss_new->cfg_msg));
+			adapter->aq_required |= IAVF_FLAG_AQ_ADD_ADV_RSS_CFG;
+		} else {
+			err = -EEXIST;
+		}
+	} else {
+		rss_new_add = true;
+		rss_new->state = IAVF_ADV_RSS_ADD_REQUEST;
+		rss_new->packet_hdrs = hdrs;
+		rss_new->hash_flds = hash_flds;
+		list_add_tail(&rss_new->list, &adapter->adv_rss_list_head);
+		adapter->aq_required |= IAVF_FLAG_AQ_ADD_ADV_RSS_CFG;
+	}
+	spin_unlock_bh(&adapter->adv_rss_lock);
+
+	if (!err)
+		mod_delayed_work(iavf_wq, &adapter->watchdog_task, 0);
+
+	clear_bit(__IAVF_IN_CRITICAL_TASK, &adapter->crit_section);
+
+	if (!rss_new_add)
+		kfree(rss_new);
+
+	return err;
+}
+
+/**
+ * iavf_get_adv_rss_hash_opt - Retrieve hash fields for a given flow-type
+ * @adapter: pointer to the VF adapter structure
+ * @cmd: ethtool rxnfc command
+ *
+ * Returns Success if the flow input set is supported.
+ */
+static int
+iavf_get_adv_rss_hash_opt(struct iavf_adapter *adapter,
+			  struct ethtool_rxnfc *cmd)
+{
+	struct iavf_adv_rss *rss;
+	u64 hash_flds;
+	u32 hdrs;
+
+	if (!ADV_RSS_SUPPORT(adapter))
+		return -EOPNOTSUPP;
+
+	cmd->data = 0;
+
+	hdrs = iavf_adv_rss_parse_hdrs(cmd);
+	if (hdrs == IAVF_ADV_RSS_FLOW_SEG_HDR_NONE)
+		return -EINVAL;
+
+	spin_lock_bh(&adapter->adv_rss_lock);
+	rss = iavf_find_adv_rss_cfg_by_hdrs(adapter, hdrs);
+	if (rss)
+		hash_flds = rss->hash_flds;
+	else
+		hash_flds = IAVF_ADV_RSS_HASH_INVALID;
+	spin_unlock_bh(&adapter->adv_rss_lock);
+
+	if (hash_flds == IAVF_ADV_RSS_HASH_INVALID)
+		return -EINVAL;
+
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_IPV4_SA |
+			 IAVF_ADV_RSS_HASH_FLD_IPV6_SA))
+		cmd->data |= (u64)RXH_IP_SRC;
+
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_IPV4_DA |
+			 IAVF_ADV_RSS_HASH_FLD_IPV6_DA))
+		cmd->data |= (u64)RXH_IP_DST;
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT)
+		cmd->data |= (u64)RXH_L4_B_0_1;
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT)
+		cmd->data |= (u64)RXH_L4_B_2_3;
+
+	return 0;
+}
+
 /**
  * iavf_set_rxnfc - command to set Rx flow rules.
  * @netdev: network interface device structure
@@ -1437,6 +1645,9 @@ static int iavf_set_rxnfc(struct net_device *netdev, struct ethtool_rxnfc *cmd)
 	case ETHTOOL_SRXCLSRLDEL:
 		ret = iavf_del_fdir_ethtool(adapter, cmd);
 		break;
+	case ETHTOOL_SRXFH:
+		ret = iavf_set_adv_rss_hash_opt(adapter, cmd);
+		break;
 	default:
 		break;
 	}
@@ -1477,8 +1688,7 @@ static int iavf_get_rxnfc(struct net_device *netdev, struct ethtool_rxnfc *cmd,
 		ret = iavf_get_fdir_fltr_ids(adapter, cmd, (u32 *)rule_locs);
 		break;
 	case ETHTOOL_GRXFH:
-		netdev_info(netdev,
-			    "RSS hash info is not available to vf, use pf.\n");
+		ret = iavf_get_adv_rss_hash_opt(adapter, cmd);
 		break;
 	default:
 		break;
diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
index 54d2efe1732d..0eab3c43bdc5 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c
@@ -1327,6 +1327,9 @@ void iavf_add_adv_rss_cfg(struct iavf_adapter *adapter)
 			process_rss = true;
 			rss->state = IAVF_ADV_RSS_ADD_PENDING;
 			memcpy(rss_cfg, &rss->cfg_msg, len);
+			iavf_print_adv_rss_cfg(adapter, rss,
+					       "Input set change for",
+					       "is pending");
 			break;
 		}
 	}
@@ -1599,6 +1602,9 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter,
 						 &adapter->adv_rss_list_head,
 						 list) {
 				if (rss->state == IAVF_ADV_RSS_ADD_PENDING) {
+					iavf_print_adv_rss_cfg(adapter, rss,
+							       "Failed to change the input set for",
+							       NULL);
 					list_del(&rss->list);
 					kfree(rss);
 				}
@@ -1815,9 +1821,14 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter,
 		struct iavf_adv_rss *rss;
 
 		spin_lock_bh(&adapter->adv_rss_lock);
-		list_for_each_entry(rss, &adapter->adv_rss_list_head, list)
-			if (rss->state == IAVF_ADV_RSS_ADD_PENDING)
+		list_for_each_entry(rss, &adapter->adv_rss_list_head, list) {
+			if (rss->state == IAVF_ADV_RSS_ADD_PENDING) {
+				iavf_print_adv_rss_cfg(adapter, rss,
+						       "Input set change for",
+						       "successful");
 				rss->state = IAVF_ADV_RSS_ACTIVE;
+			}
+		}
 		spin_unlock_bh(&adapter->adv_rss_lock);
 		}
 		break;
-- 
2.26.2

[PATCH net-next 11/12] iavf: Support for modifying UDP RSS flow hashing

[Tony Nguyen]

From: Haiyue Wang <haiyue.wang@intel.com>

Provides the ability to enable UDP RSS hashing by ethtool.

It gives users option of generating RSS hash based on the UDP source
and destination ports numbers, IPv4 or IPv6 source and destination
addresses.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 .../net/ethernet/intel/iavf/iavf_adv_rss.c    | 28 +++++++++++++++++--
 .../net/ethernet/intel/iavf/iavf_adv_rss.h    | 10 ++++++-
 .../net/ethernet/intel/iavf/iavf_ethtool.c    | 23 +++++++++++++--
 3 files changed, 56 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
index 4c5771cdc445..a8e03aaccc6b 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
@@ -56,6 +56,23 @@ iavf_fill_adv_rss_tcp_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
 		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, TCP, DST_PORT);
 }
 
+/**
+ * iavf_fill_adv_rss_udp_hdr - fill the UDP RSS protocol header
+ * @hdr: the virtchnl message protocol header data structure
+ * @hash_flds: the RSS configuration protocol hash fields
+ */
+static void
+iavf_fill_adv_rss_udp_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
+{
+	VIRTCHNL_SET_PROTO_HDR_TYPE(hdr, UDP);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, UDP, SRC_PORT);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, UDP, DST_PORT);
+}
+
 /**
  * iavf_fill_adv_rss_cfg_msg - fill the RSS configuration into virtchnl message
  * @rss_cfg: the virtchnl message to be filled with RSS configuration setting
@@ -92,6 +109,9 @@ iavf_fill_adv_rss_cfg_msg(struct virtchnl_rss_cfg *rss_cfg,
 	case IAVF_ADV_RSS_FLOW_SEG_HDR_TCP:
 		iavf_fill_adv_rss_tcp_hdr(hdr, hash_flds);
 		break;
+	case IAVF_ADV_RSS_FLOW_SEG_HDR_UDP:
+		iavf_fill_adv_rss_udp_hdr(hdr, hash_flds);
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -138,6 +158,8 @@ iavf_print_adv_rss_cfg(struct iavf_adapter *adapter, struct iavf_adv_rss *rss,
 
 	if (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_TCP)
 		proto = "TCP";
+	else if (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_UDP)
+		proto = "UDP";
 	else
 		return;
 
@@ -155,9 +177,11 @@ iavf_print_adv_rss_cfg(struct iavf_adapter *adapter, struct iavf_adv_rss *rss,
 	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_IPV4_DA |
 			 IAVF_ADV_RSS_HASH_FLD_IPV6_DA))
 		strcat(hash_opt, "IP DA,");
-	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT)
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT))
 		strcat(hash_opt, "src port,");
-	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT)
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT))
 		strcat(hash_opt, "dst port,");
 
 	if (!action)
diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
index 339ecb42938b..4681f5e8321d 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
@@ -20,6 +20,7 @@ enum iavf_adv_rss_flow_seg_hdr {
 	IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4	= 0x00000001,
 	IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6	= 0x00000002,
 	IAVF_ADV_RSS_FLOW_SEG_HDR_TCP	= 0x00000004,
+	IAVF_ADV_RSS_FLOW_SEG_HDR_UDP	= 0x00000008,
 };
 
 #define IAVF_ADV_RSS_FLOW_SEG_HDR_L3		\
@@ -27,7 +28,8 @@ enum iavf_adv_rss_flow_seg_hdr {
 	 IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6)
 
 #define IAVF_ADV_RSS_FLOW_SEG_HDR_L4		\
-	(IAVF_ADV_RSS_FLOW_SEG_HDR_TCP)
+	(IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |	\
+	 IAVF_ADV_RSS_FLOW_SEG_HDR_UDP)
 
 enum iavf_adv_rss_flow_field {
 	/* L3 */
@@ -38,6 +40,8 @@ enum iavf_adv_rss_flow_field {
 	/* L4 */
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_SRC_PORT,
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_DST_PORT,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_SRC_PORT,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_DST_PORT,
 
 	/* The total number of enums must not exceed 64 */
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_MAX
@@ -56,6 +60,10 @@ enum iavf_adv_rss_flow_field {
 	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_SRC_PORT)
 #define IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT	\
 	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_DST_PORT)
+#define IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_SRC_PORT)
+#define IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_DST_PORT)
 
 /* bookkeeping of advanced RSS configuration */
 struct iavf_adv_rss {
diff --git a/drivers/net/ethernet/intel/iavf/iavf_ethtool.c b/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
index d8dca5645c21..e6169a336694 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
@@ -1434,10 +1434,18 @@ static u32 iavf_adv_rss_parse_hdrs(struct ethtool_rxnfc *cmd)
 		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |
 			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4;
 		break;
+	case UDP_V4_FLOW:
+		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_UDP |
+			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4;
+		break;
 	case TCP_V6_FLOW:
 		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |
 			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6;
 		break;
+	case UDP_V6_FLOW:
+		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_UDP |
+			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6;
+		break;
 	default:
 		break;
 	}
@@ -1459,12 +1467,14 @@ static u64 iavf_adv_rss_parse_hash_flds(struct ethtool_rxnfc *cmd)
 	if (cmd->data & RXH_IP_SRC || cmd->data & RXH_IP_DST) {
 		switch (cmd->flow_type) {
 		case TCP_V4_FLOW:
+		case UDP_V4_FLOW:
 			if (cmd->data & RXH_IP_SRC)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV4_SA;
 			if (cmd->data & RXH_IP_DST)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV4_DA;
 			break;
 		case TCP_V6_FLOW:
+		case UDP_V6_FLOW:
 			if (cmd->data & RXH_IP_SRC)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV6_SA;
 			if (cmd->data & RXH_IP_DST)
@@ -1484,6 +1494,13 @@ static u64 iavf_adv_rss_parse_hash_flds(struct ethtool_rxnfc *cmd)
 			if (cmd->data & RXH_L4_B_2_3)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT;
 			break;
+		case UDP_V4_FLOW:
+		case UDP_V6_FLOW:
+			if (cmd->data & RXH_L4_B_0_1)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT;
+			if (cmd->data & RXH_L4_B_2_3)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT;
+			break;
 		default:
 			break;
 		}
@@ -1617,10 +1634,12 @@ iavf_get_adv_rss_hash_opt(struct iavf_adapter *adapter,
 			 IAVF_ADV_RSS_HASH_FLD_IPV6_DA))
 		cmd->data |= (u64)RXH_IP_DST;
 
-	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT)
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT))
 		cmd->data |= (u64)RXH_L4_B_0_1;
 
-	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT)
+	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT))
 		cmd->data |= (u64)RXH_L4_B_2_3;
 
 	return 0;
-- 
2.26.2

[PATCH net-next 12/12] iavf: Support for modifying SCTP RSS flow hashing

[Tony Nguyen]

From: Haiyue Wang <haiyue.wang@intel.com>

Provide the ability to enable SCTP RSS hashing by ethtool.

It gives users option of generating RSS hash based on the SCTP source
and destination ports numbers, IPv4 or IPv6 source and destination
addresses.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 .../net/ethernet/intel/iavf/iavf_adv_rss.c    | 28 +++++++++++++++++--
 .../net/ethernet/intel/iavf/iavf_adv_rss.h    | 10 ++++++-
 .../net/ethernet/intel/iavf/iavf_ethtool.c    | 23 +++++++++++++--
 3 files changed, 56 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
index a8e03aaccc6b..6edbf134b73f 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.c
@@ -73,6 +73,23 @@ iavf_fill_adv_rss_udp_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
 		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, UDP, DST_PORT);
 }
 
+/**
+ * iavf_fill_adv_rss_sctp_hdr - fill the SCTP RSS protocol header
+ * @hdr: the virtchnl message protocol header data structure
+ * @hash_flds: the RSS configuration protocol hash fields
+ */
+static void
+iavf_fill_adv_rss_sctp_hdr(struct virtchnl_proto_hdr *hdr, u64 hash_flds)
+{
+	VIRTCHNL_SET_PROTO_HDR_TYPE(hdr, SCTP);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_SCTP_SRC_PORT)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, SCTP, SRC_PORT);
+
+	if (hash_flds & IAVF_ADV_RSS_HASH_FLD_SCTP_DST_PORT)
+		VIRTCHNL_ADD_PROTO_HDR_FIELD_BIT(hdr, SCTP, DST_PORT);
+}
+
 /**
  * iavf_fill_adv_rss_cfg_msg - fill the RSS configuration into virtchnl message
  * @rss_cfg: the virtchnl message to be filled with RSS configuration setting
@@ -112,6 +129,9 @@ iavf_fill_adv_rss_cfg_msg(struct virtchnl_rss_cfg *rss_cfg,
 	case IAVF_ADV_RSS_FLOW_SEG_HDR_UDP:
 		iavf_fill_adv_rss_udp_hdr(hdr, hash_flds);
 		break;
+	case IAVF_ADV_RSS_FLOW_SEG_HDR_SCTP:
+		iavf_fill_adv_rss_sctp_hdr(hdr, hash_flds);
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -160,6 +180,8 @@ iavf_print_adv_rss_cfg(struct iavf_adapter *adapter, struct iavf_adv_rss *rss,
 		proto = "TCP";
 	else if (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_UDP)
 		proto = "UDP";
+	else if (packet_hdrs & IAVF_ADV_RSS_FLOW_SEG_HDR_SCTP)
+		proto = "SCTP";
 	else
 		return;
 
@@ -178,10 +200,12 @@ iavf_print_adv_rss_cfg(struct iavf_adapter *adapter, struct iavf_adv_rss *rss,
 			 IAVF_ADV_RSS_HASH_FLD_IPV6_DA))
 		strcat(hash_opt, "IP DA,");
 	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT |
-			 IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT))
+			 IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_SCTP_SRC_PORT))
 		strcat(hash_opt, "src port,");
 	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT |
-			 IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT))
+			 IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_SCTP_DST_PORT))
 		strcat(hash_opt, "dst port,");
 
 	if (!action)
diff --git a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
index 4681f5e8321d..4d3be11af7aa 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
+++ b/drivers/net/ethernet/intel/iavf/iavf_adv_rss.h
@@ -21,6 +21,7 @@ enum iavf_adv_rss_flow_seg_hdr {
 	IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6	= 0x00000002,
 	IAVF_ADV_RSS_FLOW_SEG_HDR_TCP	= 0x00000004,
 	IAVF_ADV_RSS_FLOW_SEG_HDR_UDP	= 0x00000008,
+	IAVF_ADV_RSS_FLOW_SEG_HDR_SCTP	= 0x00000010,
 };
 
 #define IAVF_ADV_RSS_FLOW_SEG_HDR_L3		\
@@ -29,7 +30,8 @@ enum iavf_adv_rss_flow_seg_hdr {
 
 #define IAVF_ADV_RSS_FLOW_SEG_HDR_L4		\
 	(IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |	\
-	 IAVF_ADV_RSS_FLOW_SEG_HDR_UDP)
+	 IAVF_ADV_RSS_FLOW_SEG_HDR_UDP |	\
+	 IAVF_ADV_RSS_FLOW_SEG_HDR_SCTP)
 
 enum iavf_adv_rss_flow_field {
 	/* L3 */
@@ -42,6 +44,8 @@ enum iavf_adv_rss_flow_field {
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_TCP_DST_PORT,
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_SRC_PORT,
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_DST_PORT,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_SCTP_SRC_PORT,
+	IAVF_ADV_RSS_FLOW_FIELD_IDX_SCTP_DST_PORT,
 
 	/* The total number of enums must not exceed 64 */
 	IAVF_ADV_RSS_FLOW_FIELD_IDX_MAX
@@ -64,6 +68,10 @@ enum iavf_adv_rss_flow_field {
 	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_SRC_PORT)
 #define IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT	\
 	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_UDP_DST_PORT)
+#define IAVF_ADV_RSS_HASH_FLD_SCTP_SRC_PORT	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_SCTP_SRC_PORT)
+#define IAVF_ADV_RSS_HASH_FLD_SCTP_DST_PORT	\
+	BIT_ULL(IAVF_ADV_RSS_FLOW_FIELD_IDX_SCTP_DST_PORT)
 
 /* bookkeeping of advanced RSS configuration */
 struct iavf_adv_rss {
diff --git a/drivers/net/ethernet/intel/iavf/iavf_ethtool.c b/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
index e6169a336694..3d904bc6ee76 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_ethtool.c
@@ -1438,6 +1438,10 @@ static u32 iavf_adv_rss_parse_hdrs(struct ethtool_rxnfc *cmd)
 		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_UDP |
 			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4;
 		break;
+	case SCTP_V4_FLOW:
+		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_SCTP |
+			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV4;
+		break;
 	case TCP_V6_FLOW:
 		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_TCP |
 			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6;
@@ -1446,6 +1450,10 @@ static u32 iavf_adv_rss_parse_hdrs(struct ethtool_rxnfc *cmd)
 		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_UDP |
 			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6;
 		break;
+	case SCTP_V6_FLOW:
+		hdrs |= IAVF_ADV_RSS_FLOW_SEG_HDR_SCTP |
+			IAVF_ADV_RSS_FLOW_SEG_HDR_IPV6;
+		break;
 	default:
 		break;
 	}
@@ -1468,6 +1476,7 @@ static u64 iavf_adv_rss_parse_hash_flds(struct ethtool_rxnfc *cmd)
 		switch (cmd->flow_type) {
 		case TCP_V4_FLOW:
 		case UDP_V4_FLOW:
+		case SCTP_V4_FLOW:
 			if (cmd->data & RXH_IP_SRC)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV4_SA;
 			if (cmd->data & RXH_IP_DST)
@@ -1475,6 +1484,7 @@ static u64 iavf_adv_rss_parse_hash_flds(struct ethtool_rxnfc *cmd)
 			break;
 		case TCP_V6_FLOW:
 		case UDP_V6_FLOW:
+		case SCTP_V6_FLOW:
 			if (cmd->data & RXH_IP_SRC)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_IPV6_SA;
 			if (cmd->data & RXH_IP_DST)
@@ -1501,6 +1511,13 @@ static u64 iavf_adv_rss_parse_hash_flds(struct ethtool_rxnfc *cmd)
 			if (cmd->data & RXH_L4_B_2_3)
 				hfld |= IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT;
 			break;
+		case SCTP_V4_FLOW:
+		case SCTP_V6_FLOW:
+			if (cmd->data & RXH_L4_B_0_1)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_SCTP_SRC_PORT;
+			if (cmd->data & RXH_L4_B_2_3)
+				hfld |= IAVF_ADV_RSS_HASH_FLD_SCTP_DST_PORT;
+			break;
 		default:
 			break;
 		}
@@ -1635,11 +1652,13 @@ iavf_get_adv_rss_hash_opt(struct iavf_adapter *adapter,
 		cmd->data |= (u64)RXH_IP_DST;
 
 	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_SRC_PORT |
-			 IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT))
+			 IAVF_ADV_RSS_HASH_FLD_UDP_SRC_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_SCTP_SRC_PORT))
 		cmd->data |= (u64)RXH_L4_B_0_1;
 
 	if (hash_flds & (IAVF_ADV_RSS_HASH_FLD_TCP_DST_PORT |
-			 IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT))
+			 IAVF_ADV_RSS_HASH_FLD_UDP_DST_PORT |
+			 IAVF_ADV_RSS_HASH_FLD_SCTP_DST_PORT))
 		cmd->data |= (u64)RXH_L4_B_2_3;
 
 	return 0;
-- 
2.26.2

Re: [PATCH net-next 00/12][pull request] 100GbE Intel Wired LAN Driver Updates 2021-04-22

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

On Thu, 22 Apr 2021 10:31:18 -0700 you wrote:
> This series contains updates to virtchnl header file, ice, and iavf
> drivers.
> 
> Vignesh adds support to warn about potentially malicious VFs; those that
> are overflowing the mailbox for the ice driver.
> 
> Michal adds support for an allowlist/denylist of VF commands based on
> supported capabilities for the ice driver.
> 
> [...]

Here is the summary with links:
  - [net-next,01/12] ice: warn about potentially malicious VFs
    https://git.kernel.org/netdev/net-next/c/0891c89674e8
  - [net-next,02/12] ice: Allow ignoring opcodes on specific VF
    https://git.kernel.org/netdev/net-next/c/c0dcaa55f91d
  - [net-next,03/12] ice: Advertise virtchnl UDP segmentation offload capability
    https://git.kernel.org/netdev/net-next/c/142da08c4dc0
  - [net-next,04/12] iavf: add support for UDP Segmentation Offload
    https://git.kernel.org/netdev/net-next/c/c91a4f9feb67
  - [net-next,05/12] ice: remove redundant assignment to pointer vsi
    https://git.kernel.org/netdev/net-next/c/c9b5f681fe41
  - [net-next,06/12] ice: Add helper function to get the VF's VSI
    https://git.kernel.org/netdev/net-next/c/c5afbe99b778
  - [net-next,07/12] ice: Enable RSS configure for AVF
    https://git.kernel.org/netdev/net-next/c/222a8ab01698
  - [net-next,08/12] ice: Support RSS configure removal for AVF
    https://git.kernel.org/netdev/net-next/c/ddd1f3cfed3f
  - [net-next,09/12] iavf: Add framework to enable ethtool RSS config
    https://git.kernel.org/netdev/net-next/c/0aaeb4fbc842
  - [net-next,10/12] iavf: Support for modifying TCP RSS flow hashing
    https://git.kernel.org/netdev/net-next/c/5ab91e0593a1
  - [net-next,11/12] iavf: Support for modifying UDP RSS flow hashing
    https://git.kernel.org/netdev/net-next/c/7b8f3f957b22
  - [net-next,12/12] iavf: Support for modifying SCTP RSS flow hashing
    https://git.kernel.org/netdev/net-next/c/e41985f0fe8b

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html