nvethernet: tctool support for EST and FPE

Add support for ndo_setup_tc in network
device operations for kernel 5.10 or above.

Bug 200622871

Change-Id: Id5f96001be3b00925d26efaca1b39e26b959e952
Signed-off-by: rakesh goyal <rgoyal@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvidia/+/2550396
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>

drivers/net/ethernet/nvidia/nvethernet/Makefile

@@ -27,6 +27,7 @@ ccflags-y += -DMACSEC_SUPPORT -DNET30 -DMACSEC_DEBUG -DOSI_DMA_DEBUG
 nvethernet-objs:= ether_linux.o \
 		  osd.o \
 		  ethtool.o \
+		  ether_tc.o \
 		  sysfs.o \
 		  ioctl.o \
 		  ptp.o \

drivers/net/ethernet/nvidia/nvethernet/ether_linux.c

@@ -3670,6 +3670,43 @@ static int ether_vlan_rx_kill_vid(struct net_device *ndev, __be16 vlan_proto,
 
 	return ret;
 }
+
+#if (KERNEL_VERSION(5, 10, 0) <= LINUX_VERSION_CODE)
+/**
+ * @brief ether_setup_tc - TC HW offload support
+ *
+ * Algorithm:
+ * 1) Check the TC setup type
+ * 2) Call appropriate function based on type.
+ *
+ * @param[in] ndev: Network device structure
+ * @param[in] type: qdisc type
+ * @param[in] type_data: void pointer having user passed configuration
+ *
+ * @note Ethernet interface should be up
+ *
+ * @retval 0 on success
+ * @retval "negative value" on failure.
+ */
+static int ether_setup_tc(struct net_device *ndev, enum tc_setup_type type,
+			  void *type_data)
+{
+	struct ether_priv_data *pdata = netdev_priv(ndev);
+
+	if (!netif_running(ndev)) {
+		dev_err(pdata->dev, "Not Allowed. Ether interface is not up\n");
+		return -EOPNOTSUPP;
+	}
+
+	switch (type) {
+	case TC_SETUP_QDISC_TAPRIO:
+		return ether_tc_setup_taprio(pdata, type_data);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+#endif
+
 /**
  * @brief Ethernet network device operations
  */
@@ -3685,6 +3722,9 @@ static const struct net_device_ops ether_netdev_ops = {
 	.ndo_set_rx_mode = ether_set_rx_mode,
 	.ndo_vlan_rx_add_vid = ether_vlan_rx_add_vid,
 	.ndo_vlan_rx_kill_vid = ether_vlan_rx_kill_vid,
+#if (KERNEL_VERSION(5, 10, 0) <= LINUX_VERSION_CODE)
+	.ndo_setup_tc = ether_setup_tc,
+#endif
 };
 
 /**

drivers/net/ethernet/nvidia/nvethernet/ether_linux.h

@@ -44,6 +44,7 @@
 #include <linux/hrtimer.h>
 #include <linux/version.h>
 #include <linux/list.h>
+#include <net/pkt_sched.h>
 #include <linux/tegra-ivc.h>
 #if (KERNEL_VERSION(5, 4, 0) > LINUX_VERSION_CODE)
 #include <soc/tegra/chip-id.h>
@@ -584,8 +585,8 @@ void ether_ptp_remove(struct ether_priv_data *pdata);
  *
  * Algorithm: This function is used to handle the hardware PTP settings.
  *
- * @param[in] pdata Pointer to private data structure.
- * @param[in] ifr Interface request structure used for socket ioctl
+ * @param[in] pdata: Pointer to private data structure.
+ * @param[in] ifr: Interface request structure used for socket ioctl
  *
  * @note PTP clock driver need to be successfully registered during
  * 	 initialization and HW need to support PTP functionality.
@@ -659,5 +660,25 @@ void ether_assign_osd_ops(struct osi_core_priv_data *osi_core,
  */
 int osd_ivc_send_cmd(void *priv, ivc_msg_common_t *ivc_buf,
 		     unsigned int len);
+
 void ether_set_rx_mode(struct net_device *dev);
+
+#if (KERNEL_VERSION(5, 10, 0) <= LINUX_VERSION_CODE)
+/**
+ * @brief Function to configure traffic class
+ *
+ * Algorithm: This function is used to handle the hardware TC
+ * settings.
+ *
+ * @param[in] pdata: Pointer to private data structure.
+ * @param[in] qopt:  Pointer to qdisc taprio offload data.
+ *
+ * @note MAC interface should be up.
+ *
+ * @retval 0 on success
+ * @retval "negative value" on Failure
+ */
+int ether_tc_setup_taprio(struct ether_priv_data *pdata,
+			  struct tc_taprio_qopt_offload *qopt);
+#endif
 #endif /* ETHER_LINUX_H */

drivers/net/ethernet/nvidia/nvethernet/ether_tc.c

@@ -0,0 +1,173 @@
+/*
+ * Copyright (c) 2021, NVIDIA CORPORATION.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "ether_linux.h"
+
+#if (KERNEL_VERSION(5, 10, 0) <= LINUX_VERSION_CODE)
+int ether_tc_setup_taprio(struct ether_priv_data *pdata,
+			  struct tc_taprio_qopt_offload *qopt)
+{
+	struct osi_core_priv_data *osi_core = pdata->osi_core;
+	unsigned int fpe_required = OSI_DISABLE;
+	struct osi_ioctl fpe_ioctl_data = {};
+	struct osi_ioctl est_ioctl_data = {};
+	unsigned long cycle_time = 0x0U;
+	/* Hardcode width base on current HW config, input parameter validation
+	 * will be done by OSI code any way
+	 */
+	/* total GCL enatry will have 32 valid bits */
+	unsigned int wid_val = OSI_MAX_32BITS;
+	unsigned int gates = 0x0U;
+	/* time is 24 bit long */
+	unsigned int wid = 24U;
+	struct timespec64 time;
+	unsigned long ctr;
+	int i, ret = 0;
+
+	if (qopt == NULL) {
+		netdev_err(pdata->ndev, "invalid input argument\n");
+		return -EINVAL;
+	}
+
+	if ((osi_core->hw_feature != OSI_NULL) &&
+	    (pdata->hw_feat.est_sel == OSI_DISABLE)) {
+		netdev_err(pdata->ndev, "EST not supported in HW\n");
+		ret = -EOPNOTSUPP;
+		goto done;
+	}
+
+	if (qopt->num_entries >= OSI_GCL_SIZE_256) {
+		netdev_err(pdata->ndev, "invalid number of GCL entries\n");
+		ret = -ERANGE;
+		goto done;
+	}
+
+	if (!qopt->base_time) {
+		netdev_err(pdata->ndev, "invalid base time\n");
+		ret = -ERANGE;
+		goto done;
+	}
+
+	if (!qopt->cycle_time) {
+		netdev_err(pdata->ndev, "invalid Cycle time\n");
+		ret = -ERANGE;
+		goto done;
+	}
+
+	memset(&est_ioctl_data.est, 0x0, sizeof(struct osi_est_config));
+	memset(&est_ioctl_data.fpe, 0x0, sizeof(struct osi_fpe_config));
+
+	/* This code is to disable TSN, User space is asking to disable
+	 */
+	if (!qopt->enable) {
+		goto disable;
+	}
+
+	est_ioctl_data.est.llr = qopt->num_entries;
+	est_ioctl_data.est.en_dis = qopt->enable;
+
+	for (i = 0U; i < est_ioctl_data.est.llr; i++) {
+		cycle_time = qopt->entries[i].interval;
+		gates = qopt->entries[i].gate_mask;
+
+		switch (qopt->entries[i].command) {
+		case TC_TAPRIO_CMD_SET_GATES:
+			if (fpe_required == OSI_ENABLE) {
+				netdev_err(pdata->ndev,
+					   "with set-and-hold/release, only set command is not expected\n");
+				ret = -EINVAL;
+				goto done;
+			}
+			break;
+		case TC_TAPRIO_CMD_SET_AND_HOLD:
+			gates |= OSI_BIT(0);
+			fpe_required = OSI_ENABLE;
+			break;
+		case TC_TAPRIO_CMD_SET_AND_RELEASE:
+			gates &= ~OSI_BIT(0);
+			fpe_required = OSI_ENABLE;
+			break;
+		default:
+			netdev_err(pdata->ndev, "invalid command\n");
+			ret = -EOPNOTSUPP;
+			goto done;
+		}
+
+		est_ioctl_data.est.gcl[i] = cycle_time | (gates << wid);
+		if (est_ioctl_data.est.gcl[i] > wid_val) {
+			netdev_err(pdata->ndev, "invalid GCL creation\n");
+			ret = -EINVAL;
+			goto done;
+		}
+	}
+
+	/* Adjust for real system time TODO better to have
+	 * some offset to avoid BTRE
+	 */
+	time = ktime_to_timespec64(qopt->base_time);
+	est_ioctl_data.est.btr[0] = (unsigned int)time.tv_nsec;
+	est_ioctl_data.est.btr[1] = (unsigned int)time.tv_sec;
+	est_ioctl_data.est.btr_offset[0] = 0;
+	est_ioctl_data.est.btr_offset[1] = 0;
+
+	ctr = qopt->cycle_time;
+	est_ioctl_data.est.ctr[0] = do_div(ctr, NSEC_PER_SEC);
+	est_ioctl_data.est.ctr[1] = (unsigned int)ctr;
+
+	if ((!pdata->hw_feat.fpe_sel) && (fpe_required == OSI_ENABLE)) {
+		netdev_err(pdata->ndev, "FPE not supported in HW\n");
+		ret = -EOPNOTSUPP;
+		goto done;
+	}
+
+	if (fpe_required == OSI_ENABLE) {
+		fpe_ioctl_data.fpe.rq = osi_core->residual_queue;
+		fpe_ioctl_data.fpe.tx_queue_preemption_enable = 0x1;
+		fpe_ioctl_data.cmd = OSI_CMD_CONFIG_FPE;
+		ret = osi_handle_ioctl(osi_core, &fpe_ioctl_data);
+		if (ret < 0) {
+			netdev_err(pdata->ndev,
+				   "failed to enable Frame Preemption\n");
+			goto done;
+		} else {
+			netdev_info(pdata->ndev, "configured FPE\n");
+		}
+	}
+
+	est_ioctl_data.cmd = OSI_CMD_CONFIG_EST;
+	ret = osi_handle_ioctl(osi_core, &est_ioctl_data);
+	if (ret < 0) {
+		netdev_err(pdata->ndev, "failed to configure EST\n");
+		goto disable;
+	}
+
+	netdev_info(pdata->ndev, "configured EST\n");
+	return 0;
+
+disable:
+	est_ioctl_data.est.en_dis = false;
+	est_ioctl_data.cmd = OSI_CMD_CONFIG_EST;
+	ret = osi_handle_ioctl(osi_core, &est_ioctl_data);
+	if ((ret >= 0) && (fpe_required == OSI_ENABLE)) {
+		fpe_ioctl_data.fpe.tx_queue_preemption_enable = 0x0;
+		fpe_ioctl_data.cmd = OSI_CMD_CONFIG_FPE;
+		ret = osi_handle_ioctl(osi_core, &fpe_ioctl_data);
+	}
+
+done:
+	return ret;
+}
+#endif