hinic_rx.c

/*
 * Huawei HiNIC PCI Express Linux driver
 * Copyright(c) 2017 Huawei Technologies Co., Ltd
 *
 * 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.
 *
 */

#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/pci.h>
#include <linux/device.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/u64_stats_sync.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/skbuff.h>
#include <linux/dma-mapping.h>
#include <linux/prefetch.h>
#include <linux/cpumask.h>
#include <asm/barrier.h>

#include "hinic_common.h"
#include "hinic_hw_if.h"
#include "hinic_hw_wqe.h"
#include "hinic_hw_wq.h"
#include "hinic_hw_qp.h"
#include "hinic_hw_dev.h"
#include "hinic_rx.h"
#include "hinic_dev.h"

#define RX_IRQ_NO_PENDING               0
#define RX_IRQ_NO_COALESC               0
#define RX_IRQ_NO_LLI_TIMER             0
#define RX_IRQ_NO_CREDIT                0
#define RX_IRQ_NO_RESEND_TIMER          0

/**
 * hinic_rxq_clean_stats - Clean the statistics of specific queue
 * @rxq: Logical Rx Queue
 **/
void hinic_rxq_clean_stats(struct hinic_rxq *rxq)
{
	struct hinic_rxq_stats *rxq_stats = &rxq->rxq_stats;

	u64_stats_update_begin(&rxq_stats->syncp);
	rxq_stats->pkts  = 0;
	rxq_stats->bytes = 0;
	u64_stats_update_end(&rxq_stats->syncp);
}

/**
 * hinic_rxq_get_stats - get statistics of Rx Queue
 * @rxq: Logical Rx Queue
 * @stats: return updated stats here
 **/
void hinic_rxq_get_stats(struct hinic_rxq *rxq, struct hinic_rxq_stats *stats)
{
	struct hinic_rxq_stats *rxq_stats = &rxq->rxq_stats;
	unsigned int start;

	u64_stats_update_begin(&stats->syncp);
	do {
		start = u64_stats_fetch_begin(&rxq_stats->syncp);
		stats->pkts = rxq_stats->pkts;
		stats->bytes = rxq_stats->bytes;
	} while (u64_stats_fetch_retry(&rxq_stats->syncp, start));
	u64_stats_update_end(&stats->syncp);
}

/**
 * rxq_stats_init - Initialize the statistics of specific queue
 * @rxq: Logical Rx Queue
 **/
static void rxq_stats_init(struct hinic_rxq *rxq)
{
	struct hinic_rxq_stats *rxq_stats = &rxq->rxq_stats;

	u64_stats_init(&rxq_stats->syncp);
	hinic_rxq_clean_stats(rxq);
}

/**
 * rx_alloc_skb - allocate skb and map it to dma address
 * @rxq: rx queue
 * @dma_addr: returned dma address for the skb
 *
 * Return skb
 **/
static struct sk_buff *rx_alloc_skb(struct hinic_rxq *rxq,
				    dma_addr_t *dma_addr)
{
	struct hinic_dev *nic_dev = netdev_priv(rxq->netdev);
	struct hinic_hwdev *hwdev = nic_dev->hwdev;
	struct hinic_hwif *hwif = hwdev->hwif;
	struct pci_dev *pdev = hwif->pdev;
	struct sk_buff *skb;
	dma_addr_t addr;
	int err;

	skb = netdev_alloc_skb_ip_align(rxq->netdev, rxq->rq->buf_sz);
	if (!skb) {
		netdev_err(rxq->netdev, "Failed to allocate Rx SKB\n");
		return NULL;
	}

	addr = dma_map_single(&pdev->dev, skb->data, rxq->rq->buf_sz,
			      DMA_FROM_DEVICE);
	err = dma_mapping_error(&pdev->dev, addr);
	if (err) {
		dev_err(&pdev->dev, "Failed to map Rx DMA, err = %d\n", err);
		goto err_rx_map;
	}

	*dma_addr = addr;
	return skb;

err_rx_map:
	dev_kfree_skb_any(skb);
	return NULL;
}

/**
 * rx_unmap_skb - unmap the dma address of the skb
 * @rxq: rx queue
 * @dma_addr: dma address of the skb
 **/
static void rx_unmap_skb(struct hinic_rxq *rxq, dma_addr_t dma_addr)
{
	struct hinic_dev *nic_dev = netdev_priv(rxq->netdev);
	struct hinic_hwdev *hwdev = nic_dev->hwdev;
	struct hinic_hwif *hwif = hwdev->hwif;
	struct pci_dev *pdev = hwif->pdev;

	dma_unmap_single(&pdev->dev, dma_addr, rxq->rq->buf_sz,
			 DMA_FROM_DEVICE);
}

/**
 * rx_free_skb - unmap and free skb
 * @rxq: rx queue
 * @skb: skb to free
 * @dma_addr: dma address of the skb
 **/
static void rx_free_skb(struct hinic_rxq *rxq, struct sk_buff *skb,
			dma_addr_t dma_addr)
{
	rx_unmap_skb(rxq, dma_addr);
	dev_kfree_skb_any(skb);
}

/**
 * rx_alloc_pkts - allocate pkts in rx queue
 * @rxq: rx queue
 *
 * Return number of skbs allocated
 **/
static int rx_alloc_pkts(struct hinic_rxq *rxq)
{
	struct hinic_dev *nic_dev = netdev_priv(rxq->netdev);
	struct hinic_rq_wqe *rq_wqe;
	unsigned int free_wqebbs;
	struct hinic_sge sge;
	dma_addr_t dma_addr;
	struct sk_buff *skb;
	u16 prod_idx;
	int i;

	free_wqebbs = hinic_get_rq_free_wqebbs(rxq->rq);

	/* Limit the allocation chunks */
	if (free_wqebbs > nic_dev->rx_weight)
		free_wqebbs = nic_dev->rx_weight;

	for (i = 0; i < free_wqebbs; i++) {
		skb = rx_alloc_skb(rxq, &dma_addr);
		if (!skb) {
			netdev_err(rxq->netdev, "Failed to alloc Rx skb\n");
			goto skb_out;
		}

		hinic_set_sge(&sge, dma_addr, skb->len);

		rq_wqe = hinic_rq_get_wqe(rxq->rq, HINIC_RQ_WQE_SIZE,
					  &prod_idx);
		if (!rq_wqe) {
			rx_free_skb(rxq, skb, dma_addr);
			goto skb_out;
		}

		hinic_rq_prepare_wqe(rxq->rq, prod_idx, rq_wqe, &sge);

		hinic_rq_write_wqe(rxq->rq, prod_idx, rq_wqe, skb);
	}

skb_out:
	if (i) {
		wmb();  /* write all the wqes before update PI */

		hinic_rq_update(rxq->rq, prod_idx);
	}

	tasklet_schedule(&rxq->rx_task);
	return i;
}

/**
 * free_all_rx_skbs - free all skbs in rx queue
 * @rxq: rx queue
 **/
static void free_all_rx_skbs(struct hinic_rxq *rxq)
{
	struct hinic_rq *rq = rxq->rq;
	struct hinic_hw_wqe *hw_wqe;
	struct hinic_sge sge;
	u16 ci;

	while ((hw_wqe = hinic_read_wqe(rq->wq, HINIC_RQ_WQE_SIZE, &ci))) {
		if (IS_ERR(hw_wqe))
			break;

		hinic_rq_get_sge(rq, &hw_wqe->rq_wqe, ci, &sge);

		hinic_put_wqe(rq->wq, HINIC_RQ_WQE_SIZE);

		rx_free_skb(rxq, rq->saved_skb[ci], hinic_sge_to_dma(&sge));
	}
}

/**
 * rx_alloc_task - tasklet for queue allocation
 * @data: rx queue
 **/
static void rx_alloc_task(unsigned long data)
{
	struct hinic_rxq *rxq = (struct hinic_rxq *)data;

	(void)rx_alloc_pkts(rxq);
}

/**
 * rx_recv_jumbo_pkt - Rx handler for jumbo pkt
 * @rxq: rx queue
 * @head_skb: the first skb in the list
 * @left_pkt_len: left size of the pkt exclude head skb
 * @ci: consumer index
 *
 * Return number of wqes that used for the left of the pkt
 **/
static int rx_recv_jumbo_pkt(struct hinic_rxq *rxq, struct sk_buff *head_skb,
			     unsigned int left_pkt_len, u16 ci)
{
	struct sk_buff *skb, *curr_skb = head_skb;
	struct hinic_rq_wqe *rq_wqe;
	unsigned int curr_len;
	struct hinic_sge sge;
	int num_wqes = 0;

	while (left_pkt_len > 0) {
		rq_wqe = hinic_rq_read_next_wqe(rxq->rq, HINIC_RQ_WQE_SIZE,
						&skb, &ci);

		num_wqes++;

		hinic_rq_get_sge(rxq->rq, rq_wqe, ci, &sge);

		rx_unmap_skb(rxq, hinic_sge_to_dma(&sge));

		prefetch(skb->data);

		curr_len = (left_pkt_len > HINIC_RX_BUF_SZ) ? HINIC_RX_BUF_SZ :
			    left_pkt_len;

		left_pkt_len -= curr_len;

		__skb_put(skb, curr_len);

		if (curr_skb == head_skb)
			skb_shinfo(head_skb)->frag_list = skb;
		else
			curr_skb->next = skb;

		head_skb->len += skb->len;
		head_skb->data_len += skb->len;
		head_skb->truesize += skb->truesize;

		curr_skb = skb;
	}

	return num_wqes;
}

/**
 * rxq_recv - Rx handler
 * @rxq: rx queue
 * @budget: maximum pkts to process
 *
 * Return number of pkts received
 **/
static int rxq_recv(struct hinic_rxq *rxq, int budget)
{
	struct hinic_qp *qp = container_of(rxq->rq, struct hinic_qp, rq);
	u64 pkt_len = 0, rx_bytes = 0;
	struct hinic_rq_wqe *rq_wqe;
	int num_wqes, pkts = 0;
	struct hinic_sge sge;
	struct sk_buff *skb;
	u16 ci;

	while (pkts < budget) {
		num_wqes = 0;

		rq_wqe = hinic_rq_read_wqe(rxq->rq, HINIC_RQ_WQE_SIZE, &skb,
					   &ci);
		if (!rq_wqe)
			break;

		hinic_rq_get_sge(rxq->rq, rq_wqe, ci, &sge);

		rx_unmap_skb(rxq, hinic_sge_to_dma(&sge));

		prefetch(skb->data);

		pkt_len = sge.len;

		if (pkt_len <= HINIC_RX_BUF_SZ) {
			__skb_put(skb, pkt_len);
		} else {
			__skb_put(skb, HINIC_RX_BUF_SZ);
			num_wqes = rx_recv_jumbo_pkt(rxq, skb, pkt_len -
						     HINIC_RX_BUF_SZ, ci);
		}

		hinic_rq_put_wqe(rxq->rq, ci,
				 (num_wqes + 1) * HINIC_RQ_WQE_SIZE);

		skb_record_rx_queue(skb, qp->q_id);
		skb->protocol = eth_type_trans(skb, rxq->netdev);

		napi_gro_receive(&rxq->napi, skb);

		pkts++;
		rx_bytes += pkt_len;
	}

	if (pkts)
		tasklet_schedule(&rxq->rx_task); /* rx_alloc_pkts */

	u64_stats_update_begin(&rxq->rxq_stats.syncp);
	rxq->rxq_stats.pkts += pkts;
	rxq->rxq_stats.bytes += rx_bytes;
	u64_stats_update_end(&rxq->rxq_stats.syncp);

	return pkts;
}

static int rx_poll(struct napi_struct *napi, int budget)
{
	struct hinic_rxq *rxq = container_of(napi, struct hinic_rxq, napi);
	struct hinic_rq *rq = rxq->rq;
	int pkts;

	pkts = rxq_recv(rxq, budget);
	if (pkts >= budget)
		return budget;

	napi_complete(napi);
	enable_irq(rq->irq);
	return pkts;
}

static void rx_add_napi(struct hinic_rxq *rxq)
{
	struct hinic_dev *nic_dev = netdev_priv(rxq->netdev);

	netif_napi_add(rxq->netdev, &rxq->napi, rx_poll, nic_dev->rx_weight);
	napi_enable(&rxq->napi);
}

static void rx_del_napi(struct hinic_rxq *rxq)
{
	napi_disable(&rxq->napi);
	netif_napi_del(&rxq->napi);
}

static irqreturn_t rx_irq(int irq, void *data)
{
	struct hinic_rxq *rxq = (struct hinic_rxq *)data;
	struct hinic_rq *rq = rxq->rq;
	struct hinic_dev *nic_dev;

	/* Disable the interrupt until napi will be completed */
	disable_irq_nosync(rq->irq);

	nic_dev = netdev_priv(rxq->netdev);
	hinic_hwdev_msix_cnt_set(nic_dev->hwdev, rq->msix_entry);

	napi_schedule(&rxq->napi);
	return IRQ_HANDLED;
}

static int rx_request_irq(struct hinic_rxq *rxq)
{
	struct hinic_dev *nic_dev = netdev_priv(rxq->netdev);
	struct hinic_hwdev *hwdev = nic_dev->hwdev;
	struct hinic_rq *rq = rxq->rq;
	struct hinic_qp *qp;
	struct cpumask mask;
	int err;

	rx_add_napi(rxq);

	hinic_hwdev_msix_set(hwdev, rq->msix_entry,
			     RX_IRQ_NO_PENDING, RX_IRQ_NO_COALESC,
			     RX_IRQ_NO_LLI_TIMER, RX_IRQ_NO_CREDIT,
			     RX_IRQ_NO_RESEND_TIMER);

	err = request_irq(rq->irq, rx_irq, 0, rxq->irq_name, rxq);
	if (err) {
		rx_del_napi(rxq);
		return err;
	}

	qp = container_of(rq, struct hinic_qp, rq);
	cpumask_set_cpu(qp->q_id % num_online_cpus(), &mask);
	return irq_set_affinity_hint(rq->irq, &mask);
}

static void rx_free_irq(struct hinic_rxq *rxq)
{
	struct hinic_rq *rq = rxq->rq;

	free_irq(rq->irq, rxq);
	rx_del_napi(rxq);
}

/**
 * hinic_init_rxq - Initialize the Rx Queue
 * @rxq: Logical Rx Queue
 * @rq: Hardware Rx Queue to connect the Logical queue with
 * @netdev: network device to connect the Logical queue with
 *
 * Return 0 - Success, negative - Failure
 **/
int hinic_init_rxq(struct hinic_rxq *rxq, struct hinic_rq *rq,
		   struct net_device *netdev)
{
	struct hinic_qp *qp = container_of(rq, struct hinic_qp, rq);
	int err, pkts, irqname_len;

	rxq->netdev = netdev;
	rxq->rq = rq;

	rxq_stats_init(rxq);

	irqname_len = snprintf(NULL, 0, "hinic_rxq%d", qp->q_id) + 1;
	rxq->irq_name = devm_kzalloc(&netdev->dev, irqname_len, GFP_KERNEL);
	if (!rxq->irq_name)
		return -ENOMEM;

	sprintf(rxq->irq_name, "hinic_rxq%d", qp->q_id);

	tasklet_init(&rxq->rx_task, rx_alloc_task, (unsigned long)rxq);

	pkts = rx_alloc_pkts(rxq);
	if (!pkts) {
		err = -ENOMEM;
		goto err_rx_pkts;
	}

	err = rx_request_irq(rxq);
	if (err) {
		netdev_err(netdev, "Failed to request Rx irq\n");
		goto err_req_rx_irq;
	}

	return 0;

err_req_rx_irq:
err_rx_pkts:
	tasklet_kill(&rxq->rx_task);
	free_all_rx_skbs(rxq);
	devm_kfree(&netdev->dev, rxq->irq_name);
	return err;
}

/**
 * hinic_clean_rxq - Clean the Rx Queue
 * @rxq: Logical Rx Queue
 **/
void hinic_clean_rxq(struct hinic_rxq *rxq)
{
	struct net_device *netdev = rxq->netdev;

	rx_free_irq(rxq);

	tasklet_kill(&rxq->rx_task);
	free_all_rx_skbs(rxq);
	devm_kfree(&netdev->dev, rxq->irq_name);
}