diff options
Diffstat (limited to 'drivers/net/ethernet/qlogic/qede')
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede.h | 97 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_dcbnl.c | 5 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_ethtool.c | 85 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_filter.c | 536 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_fp.c | 93 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_main.c | 356 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_ptp.c | 188 | ||||
| -rw-r--r-- | drivers/net/ethernet/qlogic/qede/qede_ptp.h | 6 |
8 files changed, 1014 insertions, 352 deletions
diff --git a/drivers/net/ethernet/qlogic/qede/qede.h b/drivers/net/ethernet/qlogic/qede/qede.h index f2aaef2cfb86..9b4f08b6f9b9 100644 --- a/drivers/net/ethernet/qlogic/qede/qede.h +++ b/drivers/net/ethernet/qlogic/qede/qede.h @@ -41,6 +41,9 @@ #include <linux/mutex.h> #include <linux/bpf.h> #include <linux/io.h> +#ifdef CONFIG_RFS_ACCEL +#include <linux/cpu_rmap.h> +#endif #include <linux/qed/common_hsi.h> #include <linux/qed/eth_common.h> #include <linux/qed/qed_if.h> @@ -50,7 +53,7 @@ #define QEDE_MAJOR_VERSION 8 #define QEDE_MINOR_VERSION 10 #define QEDE_REVISION_VERSION 10 -#define QEDE_ENGINEERING_VERSION 20 +#define QEDE_ENGINEERING_VERSION 21 #define DRV_MODULE_VERSION __stringify(QEDE_MAJOR_VERSION) "." \ __stringify(QEDE_MINOR_VERSION) "." \ __stringify(QEDE_REVISION_VERSION) "." \ @@ -58,7 +61,7 @@ #define DRV_MODULE_SYM qede -struct qede_stats { +struct qede_stats_common { u64 no_buff_discards; u64 packet_too_big_discard; u64 ttl0_discard; @@ -90,11 +93,6 @@ struct qede_stats { u64 rx_256_to_511_byte_packets; u64 rx_512_to_1023_byte_packets; u64 rx_1024_to_1518_byte_packets; - u64 rx_1519_to_1522_byte_packets; - u64 rx_1519_to_2047_byte_packets; - u64 rx_2048_to_4095_byte_packets; - u64 rx_4096_to_9216_byte_packets; - u64 rx_9217_to_16383_byte_packets; u64 rx_crc_errors; u64 rx_mac_crtl_frames; u64 rx_pause_frames; @@ -111,17 +109,39 @@ struct qede_stats { u64 tx_256_to_511_byte_packets; u64 tx_512_to_1023_byte_packets; u64 tx_1024_to_1518_byte_packets; + u64 tx_pause_frames; + u64 tx_pfc_frames; + u64 brb_truncates; + u64 brb_discards; + u64 tx_mac_ctrl_frames; +}; + +struct qede_stats_bb { + u64 rx_1519_to_1522_byte_packets; + u64 rx_1519_to_2047_byte_packets; + u64 rx_2048_to_4095_byte_packets; + u64 rx_4096_to_9216_byte_packets; + u64 rx_9217_to_16383_byte_packets; u64 tx_1519_to_2047_byte_packets; u64 tx_2048_to_4095_byte_packets; u64 tx_4096_to_9216_byte_packets; u64 tx_9217_to_16383_byte_packets; - u64 tx_pause_frames; - u64 tx_pfc_frames; u64 tx_lpi_entry_count; u64 tx_total_collisions; - u64 brb_truncates; - u64 brb_discards; - u64 tx_mac_ctrl_frames; +}; + +struct qede_stats_ah { + u64 rx_1519_to_max_byte_packets; + u64 tx_1519_to_max_byte_packets; +}; + +struct qede_stats { + struct qede_stats_common common; + + union { + struct qede_stats_bb bb; + struct qede_stats_ah ah; + }; }; struct qede_vlan { @@ -147,10 +167,11 @@ struct qede_dev { u32 dp_module; u8 dp_level; - u32 flags; -#define QEDE_FLAG_IS_VF BIT(0) + unsigned long flags; +#define QEDE_FLAG_IS_VF BIT(0) #define IS_VF(edev) (!!((edev)->flags & QEDE_FLAG_IS_VF)) #define QEDE_TX_TIMESTAMPING_EN BIT(1) +#define QEDE_FLAGS_PTP_TX_IN_PRORGESS BIT(2) const struct qed_eth_ops *ops; struct qede_ptp *ptp; @@ -158,6 +179,10 @@ struct qede_dev { struct qed_dev_eth_info dev_info; #define QEDE_MAX_RSS_CNT(edev) ((edev)->dev_info.num_queues) #define QEDE_MAX_TSS_CNT(edev) ((edev)->dev_info.num_queues) +#define QEDE_IS_BB(edev) \ + ((edev)->dev_info.common.dev_type == QED_DEV_TYPE_BB) +#define QEDE_IS_AH(edev) \ + ((edev)->dev_info.common.dev_type == QED_DEV_TYPE_AH) struct qede_fastpath *fp_array; u8 req_num_tx; @@ -216,7 +241,10 @@ struct qede_dev { u16 vxlan_dst_port; u16 geneve_dst_port; - bool wol_enabled; +#ifdef CONFIG_RFS_ACCEL + struct qede_arfs *arfs; +#endif + bool wol_enabled; struct qede_rdma_dev rdma_info; @@ -292,21 +320,24 @@ struct qede_rx_queue { u8 data_direction; u8 rxq_id; + /* Used once per each NAPI run */ + u16 num_rx_buffers; + + u16 rx_headroom; + u32 rx_buf_size; u32 rx_buf_seg_size; - u64 rcv_pkts; - struct sw_rx_data *sw_rx_ring; struct qed_chain rx_bd_ring; struct qed_chain rx_comp_ring ____cacheline_aligned; - /* Used once per each NAPI run */ - u16 num_rx_buffers; - /* GRO */ struct qede_agg_info tpa_info[ETH_TPA_MAX_AGGS_NUM]; + /* Used once per each NAPI run */ + u64 rcv_pkts; + u64 rx_hw_errors; u64 rx_alloc_errors; u64 rx_ip_frags; @@ -328,6 +359,11 @@ struct sw_tx_bd { #define QEDE_TSO_SPLIT_BD BIT(0) }; +struct sw_tx_xdp { + struct page *page; + dma_addr_t mapping; +}; + struct qede_tx_queue { u8 is_xdp; bool is_legacy; @@ -351,11 +387,11 @@ struct qede_tx_queue { #define QEDE_TXQ_IDX_TO_XDP(edev, idx) ((idx) + QEDE_MAX_TSS_CNT(edev)) /* Regular Tx requires skb + metadata for release purpose, - * while XDP requires only the pages themselves. + * while XDP requires the pages and the mapped address. */ union { struct sw_tx_bd *skbs; - struct page **pages; + struct sw_tx_xdp *xdp; } sw_tx_ring; struct qed_chain tx_pbl; @@ -407,8 +443,20 @@ struct qede_fastpath { #define QEDE_TUNN_CSUM_UNNECESSARY BIT(2) #define QEDE_SP_RX_MODE 1 -#define QEDE_SP_VXLAN_PORT_CONFIG 2 -#define QEDE_SP_GENEVE_PORT_CONFIG 3 + +#ifdef CONFIG_RFS_ACCEL +int qede_rx_flow_steer(struct net_device *dev, const struct sk_buff *skb, + u16 rxq_index, u32 flow_id); +void qede_process_arfs_filters(struct qede_dev *edev, bool free_fltr); +void qede_poll_for_freeing_arfs_filters(struct qede_dev *edev); +void qede_arfs_filter_op(void *dev, void *filter, u8 fw_rc); +void qede_free_arfs(struct qede_dev *edev); +int qede_alloc_arfs(struct qede_dev *edev); + +#define QEDE_SP_ARFS_CONFIG 4 +#define QEDE_SP_TASK_POLL_DELAY (5 * HZ) +#define QEDE_RFS_MAX_FLTR 256 +#endif struct qede_reload_args { void (*func)(struct qede_dev *edev, struct qede_reload_args *args); @@ -433,6 +481,7 @@ irqreturn_t qede_msix_fp_int(int irq, void *fp_cookie); /* Filtering function definitions */ void qede_force_mac(void *dev, u8 *mac, bool forced); +void qede_udp_ports_update(void *dev, u16 vxlan_port, u16 geneve_port); int qede_set_mac_addr(struct net_device *ndev, void *p); int qede_vlan_rx_add_vid(struct net_device *dev, __be16 proto, u16 vid); diff --git a/drivers/net/ethernet/qlogic/qede/qede_dcbnl.c b/drivers/net/ethernet/qlogic/qede/qede_dcbnl.c index 03e8c0212433..a9e7379313db 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_dcbnl.c +++ b/drivers/net/ethernet/qlogic/qede/qede_dcbnl.c @@ -281,6 +281,11 @@ static int qede_dcbnl_ieee_setapp(struct net_device *netdev, struct dcb_app *app) { struct qede_dev *edev = netdev_priv(netdev); + int err; + + err = dcb_ieee_setapp(netdev, app); + if (err) + return err; return edev->ops->dcb->ieee_setapp(edev->cdev, app); } diff --git a/drivers/net/ethernet/qlogic/qede/qede_ethtool.c b/drivers/net/ethernet/qlogic/qede/qede_ethtool.c index 897953133245..4dcfe9614731 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_ethtool.c +++ b/drivers/net/ethernet/qlogic/qede/qede_ethtool.c @@ -75,16 +75,33 @@ static const struct { QEDE_TQSTAT(stopped_cnt), }; -#define QEDE_STAT_OFFSET(stat_name) (offsetof(struct qede_stats, stat_name)) -#define QEDE_STAT_STRING(stat_name) (#stat_name) -#define _QEDE_STAT(stat_name, pf_only) \ - {QEDE_STAT_OFFSET(stat_name), QEDE_STAT_STRING(stat_name), pf_only} -#define QEDE_PF_STAT(stat_name) _QEDE_STAT(stat_name, true) -#define QEDE_STAT(stat_name) _QEDE_STAT(stat_name, false) +#define QEDE_STAT_OFFSET(stat_name, type, base) \ + (offsetof(type, stat_name) + (base)) +#define QEDE_STAT_STRING(stat_name) (#stat_name) +#define _QEDE_STAT(stat_name, type, base, attr) \ + {QEDE_STAT_OFFSET(stat_name, type, base), \ + QEDE_STAT_STRING(stat_name), \ + attr} +#define QEDE_STAT(stat_name) \ + _QEDE_STAT(stat_name, struct qede_stats_common, 0, 0x0) +#define QEDE_PF_STAT(stat_name) \ + _QEDE_STAT(stat_name, struct qede_stats_common, 0, \ + BIT(QEDE_STAT_PF_ONLY)) +#define QEDE_PF_BB_STAT(stat_name) \ + _QEDE_STAT(stat_name, struct qede_stats_bb, \ + offsetof(struct qede_stats, bb), \ + BIT(QEDE_STAT_PF_ONLY) | BIT(QEDE_STAT_BB_ONLY)) +#define QEDE_PF_AH_STAT(stat_name) \ + _QEDE_STAT(stat_name, struct qede_stats_ah, \ + offsetof(struct qede_stats, ah), \ + BIT(QEDE_STAT_PF_ONLY) | BIT(QEDE_STAT_AH_ONLY)) static const struct { u64 offset; char string[ETH_GSTRING_LEN]; - bool pf_only; + unsigned long attr; +#define QEDE_STAT_PF_ONLY 0 +#define QEDE_STAT_BB_ONLY 1 +#define QEDE_STAT_AH_ONLY 2 } qede_stats_arr[] = { QEDE_STAT(rx_ucast_bytes), QEDE_STAT(rx_mcast_bytes), @@ -106,22 +123,23 @@ static const struct { QEDE_PF_STAT(rx_256_to_511_byte_packets), QEDE_PF_STAT(rx_512_to_1023_byte_packets), QEDE_PF_STAT(rx_1024_to_1518_byte_packets), - QEDE_PF_STAT(rx_1519_to_1522_byte_packets), - QEDE_PF_STAT(rx_1519_to_2047_byte_packets), - QEDE_PF_STAT(rx_2048_to_4095_byte_packets), - QEDE_PF_STAT(rx_4096_to_9216_byte_packets), - QEDE_PF_STAT(rx_9217_to_16383_byte_packets), + QEDE_PF_BB_STAT(rx_1519_to_1522_byte_packets), + QEDE_PF_BB_STAT(rx_1519_to_2047_byte_packets), + QEDE_PF_BB_STAT(rx_2048_to_4095_byte_packets), + QEDE_PF_BB_STAT(rx_4096_to_9216_byte_packets), + QEDE_PF_BB_STAT(rx_9217_to_16383_byte_packets), + QEDE_PF_AH_STAT(rx_1519_to_max_byte_packets), QEDE_PF_STAT(tx_64_byte_packets), QEDE_PF_STAT(tx_65_to_127_byte_packets), QEDE_PF_STAT(tx_128_to_255_byte_packets), QEDE_PF_STAT(tx_256_to_511_byte_packets), QEDE_PF_STAT(tx_512_to_1023_byte_packets), QEDE_PF_STAT(tx_1024_to_1518_byte_packets), - QEDE_PF_STAT(tx_1519_to_2047_byte_packets), - QEDE_PF_STAT(tx_2048_to_4095_byte_packets), - QEDE_PF_STAT(tx_4096_to_9216_byte_packets), - QEDE_PF_STAT(tx_9217_to_16383_byte_packets), - + QEDE_PF_BB_STAT(tx_1519_to_2047_byte_packets), + QEDE_PF_BB_STAT(tx_2048_to_4095_byte_packets), + QEDE_PF_BB_STAT(tx_4096_to_9216_byte_packets), + QEDE_PF_BB_STAT(tx_9217_to_16383_byte_packets), + QEDE_PF_AH_STAT(tx_1519_to_max_byte_packets), QEDE_PF_STAT(rx_mac_crtl_frames), QEDE_PF_STAT(tx_mac_ctrl_frames), QEDE_PF_STAT(rx_pause_frames), @@ -136,8 +154,8 @@ static const struct { QEDE_PF_STAT(rx_jabbers), QEDE_PF_STAT(rx_undersize_packets), QEDE_PF_STAT(rx_fragments), - QEDE_PF_STAT(tx_lpi_entry_count), - QEDE_PF_STAT(tx_total_collisions), + QEDE_PF_BB_STAT(tx_lpi_entry_count), + QEDE_PF_BB_STAT(tx_total_collisions), QEDE_PF_STAT(brb_truncates), QEDE_PF_STAT(brb_discards), QEDE_STAT(no_buff_discards), @@ -155,6 +173,12 @@ static const struct { }; #define QEDE_NUM_STATS ARRAY_SIZE(qede_stats_arr) +#define QEDE_STAT_IS_PF_ONLY(i) \ + test_bit(QEDE_STAT_PF_ONLY, &qede_stats_arr[i].attr) +#define QEDE_STAT_IS_BB_ONLY(i) \ + test_bit(QEDE_STAT_BB_ONLY, &qede_stats_arr[i].attr) +#define QEDE_STAT_IS_AH_ONLY(i) \ + test_bit(QEDE_STAT_AH_ONLY, &qede_stats_arr[i].attr) enum { QEDE_PRI_FLAG_CMT, @@ -213,6 +237,13 @@ static void qede_get_strings_stats_rxq(struct qede_dev *edev, } } +static bool qede_is_irrelevant_stat(struct qede_dev *edev, int stat_index) +{ + return (IS_VF(edev) && QEDE_STAT_IS_PF_ONLY(stat_index)) || + (QEDE_IS_BB(edev) && QEDE_STAT_IS_AH_ONLY(stat_index)) || + (QEDE_IS_AH(edev) && QEDE_STAT_IS_BB_ONLY(stat_index)); +} + static void qede_get_strings_stats(struct qede_dev *edev, u8 *buf) { struct qede_fastpath *fp; @@ -234,7 +265,7 @@ static void qede_get_strings_stats(struct qede_dev *edev, u8 *buf) /* Account for non-queue statistics */ for (i = 0; i < QEDE_NUM_STATS; i++) { - if (IS_VF(edev) && qede_stats_arr[i].pf_only) + if (qede_is_irrelevant_stat(edev, i)) continue; strcpy(buf, qede_stats_arr[i].string); buf += ETH_GSTRING_LEN; @@ -309,7 +340,7 @@ static void qede_get_ethtool_stats(struct net_device *dev, } for (i = 0; i < QEDE_NUM_STATS; i++) { - if (IS_VF(edev) && qede_stats_arr[i].pf_only) + if (qede_is_irrelevant_stat(edev, i)) continue; *buf = *((u64 *)(((void *)&edev->stats) + qede_stats_arr[i].offset)); @@ -323,17 +354,13 @@ static void qede_get_ethtool_stats(struct net_device *dev, static int qede_get_sset_count(struct net_device *dev, int stringset) { struct qede_dev *edev = netdev_priv(dev); - int num_stats = QEDE_NUM_STATS; + int num_stats = QEDE_NUM_STATS, i; switch (stringset) { case ETH_SS_STATS: - if (IS_VF(edev)) { - int i; - - for (i = 0; i < QEDE_NUM_STATS; i++) - if (qede_stats_arr[i].pf_only) - num_stats--; - } + for (i = 0; i < QEDE_NUM_STATS; i++) + if (qede_is_irrelevant_stat(edev, i)) + num_stats--; /* Account for the Regular Tx statistics */ num_stats += QEDE_TSS_COUNT(edev) * QEDE_NUM_TQSTATS; diff --git a/drivers/net/ethernet/qlogic/qede/qede_filter.c b/drivers/net/ethernet/qlogic/qede/qede_filter.c index 107c3fda4792..eb5652073ca8 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_filter.c +++ b/drivers/net/ethernet/qlogic/qede/qede_filter.c @@ -38,6 +38,459 @@ #include <linux/qed/qed_if.h> #include "qede.h" +#ifdef CONFIG_RFS_ACCEL +struct qede_arfs_tuple { + union { + __be32 src_ipv4; + struct in6_addr src_ipv6; + }; + union { + __be32 dst_ipv4; + struct in6_addr dst_ipv6; + }; + __be16 src_port; + __be16 dst_port; + __be16 eth_proto; + u8 ip_proto; +}; + +struct qede_arfs_fltr_node { +#define QEDE_FLTR_VALID 0 + unsigned long state; + + /* pointer to aRFS packet buffer */ + void *data; + + /* dma map address of aRFS packet buffer */ + dma_addr_t mapping; + + /* length of aRFS packet buffer */ + int buf_len; + + /* tuples to hold from aRFS packet buffer */ + struct qede_arfs_tuple tuple; + + u32 flow_id; + u16 sw_id; + u16 rxq_id; + u16 next_rxq_id; + bool filter_op; + bool used; + struct hlist_node node; +}; + +struct qede_arfs { +#define QEDE_ARFS_POLL_COUNT 100 +#define QEDE_RFS_FLW_BITSHIFT (4) +#define QEDE_RFS_FLW_MASK ((1 << QEDE_RFS_FLW_BITSHIFT) - 1) + struct hlist_head arfs_hl_head[1 << QEDE_RFS_FLW_BITSHIFT]; + + /* lock for filter list access */ + spinlock_t arfs_list_lock; + unsigned long *arfs_fltr_bmap; + int filter_count; + bool enable; +}; + +static void qede_configure_arfs_fltr(struct qede_dev *edev, + struct qede_arfs_fltr_node *n, + u16 rxq_id, bool add_fltr) +{ + const struct qed_eth_ops *op = edev->ops; + + if (n->used) + return; + + DP_VERBOSE(edev, NETIF_MSG_RX_STATUS, + "%s arfs filter flow_id=%d, sw_id=%d, src_port=%d, dst_port=%d, rxq=%d\n", + add_fltr ? "Adding" : "Deleting", + n->flow_id, n->sw_id, ntohs(n->tuple.src_port), + ntohs(n->tuple.dst_port), rxq_id); + + n->used = true; + n->filter_op = add_fltr; + op->ntuple_filter_config(edev->cdev, n, n->mapping, n->buf_len, 0, + rxq_id, add_fltr); +} + +static void +qede_free_arfs_filter(struct qede_dev *edev, struct qede_arfs_fltr_node *fltr) +{ + kfree(fltr->data); + clear_bit(fltr->sw_id, edev->arfs->arfs_fltr_bmap); + kfree(fltr); +} + +void qede_arfs_filter_op(void *dev, void *filter, u8 fw_rc) +{ + struct qede_arfs_fltr_node *fltr = filter; + struct qede_dev *edev = dev; + + if (fw_rc) { + DP_NOTICE(edev, + "Failed arfs filter configuration fw_rc=%d, flow_id=%d, sw_id=%d, src_port=%d, dst_port=%d, rxq=%d\n", + fw_rc, fltr->flow_id, fltr->sw_id, + ntohs(fltr->tuple.src_port), + ntohs(fltr->tuple.dst_port), fltr->rxq_id); + + spin_lock_bh(&edev->arfs->arfs_list_lock); + + fltr->used = false; + clear_bit(QEDE_FLTR_VALID, &fltr->state); + + spin_unlock_bh(&edev->arfs->arfs_list_lock); + return; + } + + spin_lock_bh(&edev->arfs->arfs_list_lock); + + fltr->used = false; + + if (fltr->filter_op) { + set_bit(QEDE_FLTR_VALID, &fltr->state); + if (fltr->rxq_id != fltr->next_rxq_id) + qede_configure_arfs_fltr(edev, fltr, fltr->rxq_id, + false); + } else { + clear_bit(QEDE_FLTR_VALID, &fltr->state); + if (fltr->rxq_id != fltr->next_rxq_id) { + fltr->rxq_id = fltr->next_rxq_id; + qede_configure_arfs_fltr(edev, fltr, + fltr->rxq_id, true); + } + } + + spin_unlock_bh(&edev->arfs->arfs_list_lock); +} + +/* Should be called while qede_lock is held */ +void qede_process_arfs_filters(struct qede_dev *edev, bool free_fltr) +{ + int i; + + for (i = 0; i <= QEDE_RFS_FLW_MASK; i++) { + struct hlist_node *temp; + struct hlist_head *head; + struct qede_arfs_fltr_node *fltr; + + head = &edev->arfs->arfs_hl_head[i]; + + hlist_for_each_entry_safe(fltr, temp, head, node) { + bool del = false; + + if (edev->state != QEDE_STATE_OPEN) + del = true; + + spin_lock_bh(&edev->arfs->arfs_list_lock); + + if ((!test_bit(QEDE_FLTR_VALID, &fltr->state) && + !fltr->used) || free_fltr) { + hlist_del(&fltr->node); + dma_unmap_single(&edev->pdev->dev, + fltr->mapping, + fltr->buf_len, DMA_TO_DEVICE); + qede_free_arfs_filter(edev, fltr); + edev->arfs->filter_count--; + } else { + if ((rps_may_expire_flow(edev->ndev, + fltr->rxq_id, + fltr->flow_id, + fltr->sw_id) || del) && + !free_fltr) + qede_configure_arfs_fltr(edev, fltr, + fltr->rxq_id, + false); + } + + spin_unlock_bh(&edev->arfs->arfs_list_lock); + } + } + + spin_lock_bh(&edev->arfs->arfs_list_lock); + + if (!edev->arfs->filter_count) { + if (edev->arfs->enable) { + edev->arfs->enable = false; + edev->ops->configure_arfs_searcher(edev->cdev, false); + } + } else { + set_bit(QEDE_SP_ARFS_CONFIG, &edev->sp_flags); + schedule_delayed_work(&edev->sp_task, + QEDE_SP_TASK_POLL_DELAY); + } + + spin_unlock_bh(&edev->arfs->arfs_list_lock); +} + +/* This function waits until all aRFS filters get deleted and freed. + * On timeout it frees all filters forcefully. + */ +void qede_poll_for_freeing_arfs_filters(struct qede_dev *edev) +{ + int count = QEDE_ARFS_POLL_COUNT; + + while (count) { + qede_process_arfs_filters(edev, false); + + if (!edev->arfs->filter_count) + break; + + msleep(100); + count--; + } + + if (!count) { + DP_NOTICE(edev, "Timeout in polling for arfs filter free\n"); + + /* Something is terribly wrong, free forcefully */ + qede_process_arfs_filters(edev, true); + } +} + +int qede_alloc_arfs(struct qede_dev *edev) +{ + int i; + + edev->arfs = vzalloc(sizeof(*edev->arfs)); + if (!edev->arfs) + return -ENOMEM; + + spin_lock_init(&edev->arfs->arfs_list_lock); + + for (i = 0; i <= QEDE_RFS_FLW_MASK; i++) + INIT_HLIST_HEAD(&edev->arfs->arfs_hl_head[i]); + + edev->ndev->rx_cpu_rmap = alloc_irq_cpu_rmap(QEDE_RSS_COUNT(edev)); + if (!edev->ndev->rx_cpu_rmap) { + vfree(edev->arfs); + edev->arfs = NULL; + return -ENOMEM; + } + + edev->arfs->arfs_fltr_bmap = vzalloc(BITS_TO_LONGS(QEDE_RFS_MAX_FLTR) * + sizeof(long)); + if (!edev->arfs->arfs_fltr_bmap) { + free_irq_cpu_rmap(edev->ndev->rx_cpu_rmap); + edev->ndev->rx_cpu_rmap = NULL; + vfree(edev->arfs); + edev->arfs = NULL; + return -ENOMEM; + } + + return 0; +} + +void qede_free_arfs(struct qede_dev *edev) +{ + if (!edev->arfs) + return; + + if (edev->ndev->rx_cpu_rmap) + free_irq_cpu_rmap(edev->ndev->rx_cpu_rmap); + + edev->ndev->rx_cpu_rmap = NULL; + vfree(edev->arfs->arfs_fltr_bmap); + edev->arfs->arfs_fltr_bmap = NULL; + vfree(edev->arfs); + edev->arfs = NULL; +} + +static bool qede_compare_ip_addr(struct qede_arfs_fltr_node *tpos, + const struct sk_buff *skb) +{ + if (skb->protocol == htons(ETH_P_IP)) { + if (tpos->tuple.src_ipv4 == ip_hdr(skb)->saddr && + tpos->tuple.dst_ipv4 == ip_hdr(skb)->daddr) + return true; + else + return false; + } else { + struct in6_addr *src = &tpos->tuple.src_ipv6; + u8 size = sizeof(struct in6_addr); + + if (!memcmp(src, &ipv6_hdr(skb)->saddr, size) && + !memcmp(&tpos->tuple.dst_ipv6, &ipv6_hdr(skb)->daddr, size)) + return true; + else + return false; + } +} + +static struct qede_arfs_fltr_node * +qede_arfs_htbl_key_search(struct hlist_head *h, const struct sk_buff *skb, + __be16 src_port, __be16 dst_port, u8 ip_proto) +{ + struct qede_arfs_fltr_node *tpos; + + hlist_for_each_entry(tpos, h, node) + if (tpos->tuple.ip_proto == ip_proto && + tpos->tuple.eth_proto == skb->protocol && + qede_compare_ip_addr(tpos, skb) && + tpos->tuple.src_port == src_port && + tpos->tuple.dst_port == dst_port) + return tpos; + + return NULL; +} + +static struct qede_arfs_fltr_node * +qede_alloc_filter(struct qede_dev *edev, int min_hlen) +{ + struct qede_arfs_fltr_node *n; + int bit_id; + + bit_id = find_first_zero_bit(edev->arfs->arfs_fltr_bmap, + QEDE_RFS_MAX_FLTR); + + if (bit_id >= QEDE_RFS_MAX_FLTR) + return NULL; + + n = kzalloc(sizeof(*n), GFP_ATOMIC); + if (!n) + return NULL; + + n->data = kzalloc(min_hlen, GFP_ATOMIC); + if (!n->data) { + kfree(n); + return NULL; + } + + n->sw_id = (u16)bit_id; + set_bit(bit_id, edev->arfs->arfs_fltr_bmap); + return n; +} + +int qede_rx_flow_steer(struct net_device *dev, const struct sk_buff *skb, + u16 rxq_index, u32 flow_id) +{ + struct qede_dev *edev = netdev_priv(dev); + struct qede_arfs_fltr_node *n; + int min_hlen, rc, tp_offset; + struct ethhdr *eth; + __be16 *ports; + u16 tbl_idx; + u8 ip_proto; + + if (skb->encapsulation) + return -EPROTONOSUPPORT; + + if (skb->protocol != htons(ETH_P_IP) && + skb->protocol != htons(ETH_P_IPV6)) + return -EPROTONOSUPPORT; + + if (skb->protocol == htons(ETH_P_IP)) { + ip_proto = ip_hdr(skb)->protocol; + tp_offset = sizeof(struct iphdr); + } else { + ip_proto = ipv6_hdr(skb)->nexthdr; + tp_offset = sizeof(struct ipv6hdr); + } + + if (ip_proto != IPPROTO_TCP && ip_proto != IPPROTO_UDP) + return -EPROTONOSUPPORT; + + ports = (__be16 *)(skb->data + tp_offset); + tbl_idx = skb_get_hash_raw(skb) & QEDE_RFS_FLW_MASK; + + spin_lock_bh(&edev->arfs->arfs_list_lock); + + n = qede_arfs_htbl_key_search(&edev->arfs->arfs_hl_head[tbl_idx], + skb, ports[0], ports[1], ip_proto); + + if (n) { + /* Filter match */ + n->next_rxq_id = rxq_index; + + if (test_bit(QEDE_FLTR_VALID, &n->state)) { + if (n->rxq_id != rxq_index) + qede_configure_arfs_fltr(edev, n, n->rxq_id, + false); + } else { + if (!n->used) { + n->rxq_id = rxq_index; + qede_configure_arfs_fltr(edev, n, n->rxq_id, + true); + } + } + + rc = n->sw_id; + goto ret_unlock; + } + + min_hlen = ETH_HLEN + skb_headlen(skb); + + n = qede_alloc_filter(edev, min_hlen); + if (!n) { + rc = -ENOMEM; + goto ret_unlock; + } + + n->buf_len = min_hlen; + n->rxq_id = rxq_index; + n->next_rxq_id = rxq_index; + n->tuple.src_port = ports[0]; + n->tuple.dst_port = ports[1]; + n->flow_id = flow_id; + + if (skb->protocol == htons(ETH_P_IP)) { + n->tuple.src_ipv4 = ip_hdr(skb)->saddr; + n->tuple.dst_ipv4 = ip_hdr(skb)->daddr; + } else { + memcpy(&n->tuple.src_ipv6, &ipv6_hdr(skb)->saddr, + sizeof(struct in6_addr)); + memcpy(&n->tuple.dst_ipv6, &ipv6_hdr(skb)->daddr, + sizeof(struct in6_addr)); + } + + eth = (struct ethhdr *)n->data; + eth->h_proto = skb->protocol; + n->tuple.eth_proto = skb->protocol; + n->tuple.ip_proto = ip_proto; + memcpy(n->data + ETH_HLEN, skb->data, skb_headlen(skb)); + + n->mapping = dma_map_single(&edev->pdev->dev, n->data, + n->buf_len, DMA_TO_DEVICE); + if (dma_mapping_error(&edev->pdev->dev, n->mapping)) { + DP_NOTICE(edev, "Failed to map DMA memory for arfs\n"); + qede_free_arfs_filter(edev, n); + rc = -ENOMEM; + goto ret_unlock; + } + + INIT_HLIST_NODE(&n->node); + hlist_add_head(&n->node, &edev->arfs->arfs_hl_head[tbl_idx]); + edev->arfs->filter_count++; + + if (edev->arfs->filter_count == 1 && !edev->arfs->enable) { + edev->ops->configure_arfs_searcher(edev->cdev, true); + edev->arfs->enable = true; + } + + qede_configure_arfs_fltr(edev, n, n->rxq_id, true); + + spin_unlock_bh(&edev->arfs->arfs_list_lock); + + set_bit(QEDE_SP_ARFS_CONFIG, &edev->sp_flags); + schedule_delayed_work(&edev->sp_task, 0); + return n->sw_id; + +ret_unlock: + spin_unlock_bh(&edev->arfs->arfs_list_lock); + return rc; +} +#endif + +void qede_udp_ports_update(void *dev, u16 vxlan_port, u16 geneve_port) +{ + struct qede_dev *edev = dev; + + if (edev->vxlan_dst_port != vxlan_port) + edev->vxlan_dst_port = 0; + + if (edev->geneve_dst_port != geneve_port) + edev->geneve_dst_port = 0; +} + void qede_force_mac(void *dev, u8 *mac, bool forced) { struct qede_dev *edev = dev; @@ -441,69 +894,112 @@ int qede_set_features(struct net_device *dev, netdev_features_t features) void qede_udp_tunnel_add(struct net_device *dev, struct udp_tunnel_info *ti) { struct qede_dev *edev = netdev_priv(dev); + struct qed_tunn_params tunn_params; u16 t_port = ntohs(ti->port); + int rc; + + memset(&tunn_params, 0, sizeof(tunn_params)); switch (ti->type) { case UDP_TUNNEL_TYPE_VXLAN: + if (!edev->dev_info.common.vxlan_enable) + return; + if (edev->vxlan_dst_port) return; - edev->vxlan_dst_port = t_port; + tunn_params.update_vxlan_port = 1; + tunn_params.vxlan_port = t_port; - DP_VERBOSE(edev, QED_MSG_DEBUG, "Added vxlan port=%d\n", - t_port); + __qede_lock(edev); + rc = edev->ops->tunn_config(edev->cdev, &tunn_params); + __qede_unlock(edev); + + if (!rc) { + edev->vxlan_dst_port = t_port; + DP_VERBOSE(edev, QED_MSG_DEBUG, "Added vxlan port=%d\n", + t_port); + } else { + DP_NOTICE(edev, "Failed to add vxlan UDP port=%d\n", + t_port); + } - set_bit(QEDE_SP_VXLAN_PORT_CONFIG, &edev->sp_flags); break; case UDP_TUNNEL_TYPE_GENEVE: + if (!edev->dev_info.common.geneve_enable) + return; + if (edev->geneve_dst_port) return; - edev->geneve_dst_port = t_port; + tunn_params.update_geneve_port = 1; + tunn_params.geneve_port = t_port; + + __qede_lock(edev); + rc = edev->ops->tunn_config(edev->cdev, &tunn_params); + __qede_unlock(edev); + + if (!rc) { + edev->geneve_dst_port = t_port; + DP_VERBOSE(edev, QED_MSG_DEBUG, + "Added geneve port=%d\n", t_port); + } else { + DP_NOTICE(edev, "Failed to add geneve UDP port=%d\n", + t_port); + } - DP_VERBOSE(edev, QED_MSG_DEBUG, "Added geneve port=%d\n", - t_port); - set_bit(QEDE_SP_GENEVE_PORT_CONFIG, &edev->sp_flags); break; default: return; } - - schedule_delayed_work(&edev->sp_task, 0); } -void qede_udp_tunnel_del(struct net_device *dev, struct udp_tunnel_info *ti) +void qede_udp_tunnel_del(struct net_device *dev, + struct udp_tunnel_info *ti) { struct qede_dev *edev = netdev_priv(dev); + struct qed_tunn_params tunn_params; u16 t_port = ntohs(ti->port); + memset(&tunn_params, 0, sizeof(tunn_params)); + switch (ti->type) { case UDP_TUNNEL_TYPE_VXLAN: if (t_port != edev->vxlan_dst_port) return; + tunn_params.update_vxlan_port = 1; + tunn_params.vxlan_port = 0; + + __qede_lock(edev); + edev->ops->tunn_config(edev->cdev, &tunn_params); + __qede_unlock(edev); + edev->vxlan_dst_port = 0; DP_VERBOSE(edev, QED_MSG_DEBUG, "Deleted vxlan port=%d\n", t_port); - set_bit(QEDE_SP_VXLAN_PORT_CONFIG, &edev->sp_flags); break; case UDP_TUNNEL_TYPE_GENEVE: if (t_port != edev->geneve_dst_port) return; + tunn_params.update_geneve_port = 1; + tunn_params.geneve_port = 0; + + __qede_lock(edev); + edev->ops->tunn_config(edev->cdev, &tunn_params); + __qede_unlock(edev); + edev->geneve_dst_port = 0; DP_VERBOSE(edev, QED_MSG_DEBUG, "Deleted geneve port=%d\n", t_port); - set_bit(QEDE_SP_GENEVE_PORT_CONFIG, &edev->sp_flags); break; default: return; } - - schedule_delayed_work(&edev->sp_task, 0); } static void qede_xdp_reload_func(struct qede_dev *edev, @@ -520,11 +1016,6 @@ static int qede_xdp_set(struct qede_dev *edev, struct bpf_prog *prog) { struct qede_reload_args args; - if (prog && prog->xdp_adjust_head) { - DP_ERR(edev, "Does not support bpf_xdp_adjust_head()\n"); - return -EOPNOTSUPP; - } - /* If we're called, there was already a bpf reference increment */ args.func = &qede_xdp_reload_func; args.u.new_prog = prog; @@ -537,6 +1028,11 @@ int qede_xdp(struct net_device *dev, struct netdev_xdp *xdp) { struct qede_dev *edev = netdev_priv(dev); + if (IS_VF(edev)) { + DP_NOTICE(edev, "VFs don't support XDP\n"); + return -EOPNOTSUPP; + } + switch (xdp->command) { case XDP_SETUP_PROG: return qede_xdp_set(edev, xdp->prog); diff --git a/drivers/net/ethernet/qlogic/qede/qede_fp.c b/drivers/net/ethernet/qlogic/qede/qede_fp.c index 1e65038c8fc0..7b6f41d06245 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_fp.c +++ b/drivers/net/ethernet/qlogic/qede/qede_fp.c @@ -87,7 +87,8 @@ int qede_alloc_rx_buffer(struct qede_rx_queue *rxq, bool allow_lazy) rx_bd = (struct eth_rx_bd *)qed_chain_produce(&rxq->rx_bd_ring); WARN_ON(!rx_bd); rx_bd->addr.hi = cpu_to_le32(upper_32_bits(mapping)); - rx_bd->addr.lo = cpu_to_le32(lower_32_bits(mapping)); + rx_bd->addr.lo = cpu_to_le32(lower_32_bits(mapping) + + rxq->rx_headroom); rxq->sw_rx_prod++; rxq->filled_buffers++; @@ -360,7 +361,8 @@ static int qede_xdp_xmit(struct qede_dev *edev, struct qede_fastpath *fp, metadata->mapping + padding, length, PCI_DMA_TODEVICE); - txq->sw_tx_ring.pages[idx] = metadata->data; + txq->sw_tx_ring.xdp[idx].page = metadata->data; + txq->sw_tx_ring.xdp[idx].mapping = metadata->mapping; txq->sw_tx_prod++; /* Mark the fastpath for future XDP doorbell */ @@ -384,19 +386,19 @@ int qede_txq_has_work(struct qede_tx_queue *txq) static void qede_xdp_tx_int(struct qede_dev *edev, struct qede_tx_queue *txq) { - struct eth_tx_1st_bd *bd; - u16 hw_bd_cons; + u16 hw_bd_cons, idx; hw_bd_cons = le16_to_cpu(*txq->hw_cons_ptr); barrier(); while (hw_bd_cons != qed_chain_get_cons_idx(&txq->tx_pbl)) { - bd = (struct eth_tx_1st_bd *)qed_chain_consume(&txq->tx_pbl); + qed_chain_consume(&txq->tx_pbl); + idx = txq->sw_tx_cons & NUM_TX_BDS_MAX; - dma_unmap_single(&edev->pdev->dev, BD_UNMAP_ADDR(bd), - PAGE_SIZE, DMA_BIDIRECTIONAL); - __free_page(txq->sw_tx_ring.pages[txq->sw_tx_cons & - NUM_TX_BDS_MAX]); + dma_unmap_page(&edev->pdev->dev, + txq->sw_tx_ring.xdp[idx].mapping, + PAGE_SIZE, DMA_BIDIRECTIONAL); + __free_page(txq->sw_tx_ring.xdp[idx].page); txq->sw_tx_cons++; txq->xmit_pkts++; @@ -508,7 +510,8 @@ static inline void qede_reuse_page(struct qede_rx_queue *rxq, new_mapping = curr_prod->mapping + curr_prod->page_offset; rx_bd_prod->addr.hi = cpu_to_le32(upper_32_bits(new_mapping)); - rx_bd_prod->addr.lo = cpu_to_le32(lower_32_bits(new_mapping)); + rx_bd_prod->addr.lo = cpu_to_le32(lower_32_bits(new_mapping) + + rxq->rx_headroom); rxq->sw_rx_prod++; curr_cons->data = NULL; @@ -624,7 +627,6 @@ static inline void qede_skb_receive(struct qede_dev *edev, __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag); napi_gro_receive(&fp->napi, skb); - rxq->rcv_pkts++; } static void qede_set_gro_params(struct qede_dev *edev, @@ -884,9 +886,9 @@ static inline void qede_tpa_cont(struct qede_dev *edev, "Strange - TPA cont with more than a single len_list entry\n"); } -static void qede_tpa_end(struct qede_dev *edev, - struct qede_fastpath *fp, - struct eth_fast_path_rx_tpa_end_cqe *cqe) +static int qede_tpa_end(struct qede_dev *edev, + struct qede_fastpath *fp, + struct eth_fast_path_rx_tpa_end_cqe *cqe) { struct qede_rx_queue *rxq = fp->rxq; struct qede_agg_info *tpa_info; @@ -934,11 +936,12 @@ static void qede_tpa_end(struct qede_dev *edev, tpa_info->state = QEDE_AGG_STATE_NONE; - return; + return 1; err: tpa_info->state = QEDE_AGG_STATE_NONE; dev_kfree_skb_any(tpa_info->skb); tpa_info->skb = NULL; + return 0; } static u8 qede_check_notunn_csum(u16 flag) @@ -990,14 +993,15 @@ static bool qede_rx_xdp(struct qede_dev *edev, struct qede_rx_queue *rxq, struct bpf_prog *prog, struct sw_rx_data *bd, - struct eth_fast_path_rx_reg_cqe *cqe) + struct eth_fast_path_rx_reg_cqe *cqe, + u16 *data_offset, u16 *len) { - u16 len = le16_to_cpu(cqe->len_on_first_bd); struct xdp_buff xdp; enum xdp_action act; - xdp.data = page_address(bd->data) + cqe->placement_offset; - xdp.data_end = xdp.data + len; + xdp.data_hard_start = page_address(bd->data); + xdp.data = xdp.data_hard_start + *data_offset; + xdp.data_end = xdp.data + *len; /* Queues always have a full reset currently, so for the time * being until there's atomic program replace just mark read @@ -1007,6 +1011,10 @@ static bool qede_rx_xdp(struct qede_dev *edev, act = bpf_prog_run_xdp(prog, &xdp); rcu_read_unlock(); + /* Recalculate, as XDP might have changed the headers */ + *data_offset = xdp.data - xdp.data_hard_start; + *len = xdp.data_end - xdp.data; + if (act == XDP_PASS) return true; @@ -1025,7 +1033,7 @@ static bool qede_rx_xdp(struct qede_dev *edev, /* Now if there's a transmission problem, we'd still have to * throw current buffer, as replacement was already allocated. */ - if (qede_xdp_xmit(edev, fp, bd, cqe->placement_offset, len)) { + if (qede_xdp_xmit(edev, fp, bd, *data_offset, *len)) { dma_unmap_page(rxq->dev, bd->mapping, PAGE_SIZE, DMA_BIDIRECTIONAL); __free_page(bd->data); @@ -1052,7 +1060,7 @@ static struct sk_buff *qede_rx_allocate_skb(struct qede_dev *edev, struct sw_rx_data *bd, u16 len, u16 pad) { - unsigned int offset = bd->page_offset; + unsigned int offset = bd->page_offset + pad; struct skb_frag_struct *frag; struct page *page = bd->data; unsigned int pull_len; @@ -1069,7 +1077,7 @@ static struct sk_buff *qede_rx_allocate_skb(struct qede_dev *edev, */ if (len + pad <= edev->rx_copybreak) { memcpy(skb_put(skb, len), - page_address(page) + pad + offset, len); + page_address(page) + offset, len); qede_reuse_page(rxq, bd); goto out; } @@ -1077,7 +1085,7 @@ static struct sk_buff *qede_rx_allocate_skb(struct qede_dev *edev, frag = &skb_shinfo(skb)->frags[0]; skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, - page, pad + offset, len, rxq->rx_buf_seg_size); + page, offset, len, rxq->rx_buf_seg_size); va = skb_frag_address(frag); pull_len = eth_get_headlen(va, QEDE_RX_HDR_SIZE); @@ -1178,8 +1186,7 @@ static int qede_rx_process_tpa_cqe(struct qede_dev *edev, qede_tpa_cont(edev, rxq, &cqe->fast_path_tpa_cont); return 0; case ETH_RX_CQE_TYPE_TPA_END: - qede_tpa_end(edev, fp, &cqe->fast_path_tpa_end); - return 1; + return qede_tpa_end(edev, fp, &cqe->fast_path_tpa_end); default: return 0; } @@ -1224,12 +1231,13 @@ static int qede_rx_process_cqe(struct qede_dev *edev, fp_cqe = &cqe->fast_path_regular; len = le16_to_cpu(fp_cqe->len_on_first_bd); - pad = fp_cqe->placement_offset; + pad = fp_cqe->placement_offset + rxq->rx_headroom; /* Run eBPF program if one is attached */ if (xdp_prog) - if (!qede_rx_xdp(edev, fp, rxq, xdp_prog, bd, fp_cqe)) - return 1; + if (!qede_rx_xdp(edev, fp, rxq, xdp_prog, bd, fp_cqe, + &pad, &len)) + return 0; /* If this is an error packet then drop it */ flags = cqe->fast_path_regular.pars_flags.flags; @@ -1290,8 +1298,8 @@ static int qede_rx_int(struct qede_fastpath *fp, int budget) { struct qede_rx_queue *rxq = fp->rxq; struct qede_dev *edev = fp->edev; + int work_done = 0, rcv_pkts = 0; u16 hw_comp_cons, sw_comp_cons; - int work_done = 0; hw_comp_cons = le16_to_cpu(*rxq->hw_cons_ptr); sw_comp_cons = qed_chain_get_cons_idx(&rxq->rx_comp_ring); @@ -1305,12 +1313,14 @@ static int qede_rx_int(struct qede_fastpath *fp, int budget) /* Loop to complete all indicated BDs */ while ((sw_comp_cons != hw_comp_cons) && (work_done < budget)) { - qede_rx_process_cqe(edev, fp, rxq); + rcv_pkts += qede_rx_process_cqe(edev, fp, rxq); qed_chain_recycle_consumed(&rxq->rx_comp_ring); sw_comp_cons = qed_chain_get_cons_idx(&rxq->rx_comp_ring); work_done++; } + rxq->rcv_pkts += rcv_pkts; + /* Allocate replacement buffers */ while (rxq->num_rx_buffers - rxq->filled_buffers) if (qede_alloc_rx_buffer(rxq, false)) @@ -1687,13 +1697,24 @@ netdev_features_t qede_features_check(struct sk_buff *skb, } /* Disable offloads for geneve tunnels, as HW can't parse - * the geneve header which has option length greater than 32B. + * the geneve header which has option length greater than 32b + * and disable offloads for the ports which are not offloaded. */ - if ((l4_proto == IPPROTO_UDP) && - ((skb_inner_mac_header(skb) - - skb_transport_header(skb)) > QEDE_MAX_TUN_HDR_LEN)) - return features & ~(NETIF_F_CSUM_MASK | - NETIF_F_GSO_MASK); + if (l4_proto == IPPROTO_UDP) { + struct qede_dev *edev = netdev_priv(dev); + u16 hdrlen, vxln_port, gnv_port; + + hdrlen = QEDE_MAX_TUN_HDR_LEN; + vxln_port = edev->vxlan_dst_port; + gnv_port = edev->geneve_dst_port; + + if ((skb_inner_mac_header(skb) - + skb_transport_header(skb)) > hdrlen || + (ntohs(udp_hdr(skb)->dest) != vxln_port && + ntohs(udp_hdr(skb)->dest) != gnv_port)) + return features & ~(NETIF_F_CSUM_MASK | + NETIF_F_GSO_MASK); + } } return features; diff --git a/drivers/net/ethernet/qlogic/qede/qede_main.c b/drivers/net/ethernet/qlogic/qede/qede_main.c index 3a78c3f25157..b9ba23d71c61 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_main.c +++ b/drivers/net/ethernet/qlogic/qede/qede_main.c @@ -84,6 +84,8 @@ static const struct qed_eth_ops *qed_ops; #define CHIP_NUM_57980S_50 0x1654 #define CHIP_NUM_57980S_25 0x1656 #define CHIP_NUM_57980S_IOV 0x1664 +#define CHIP_NUM_AH 0x8070 +#define CHIP_NUM_AH_IOV 0x8090 #ifndef PCI_DEVICE_ID_NX2_57980E #define PCI_DEVICE_ID_57980S_40 CHIP_NUM_57980S_40 @@ -93,6 +95,9 @@ static const struct qed_eth_ops *qed_ops; #define PCI_DEVICE_ID_57980S_50 CHIP_NUM_57980S_50 #define PCI_DEVICE_ID_57980S_25 CHIP_NUM_57980S_25 #define PCI_DEVICE_ID_57980S_IOV CHIP_NUM_57980S_IOV +#define PCI_DEVICE_ID_AH CHIP_NUM_AH +#define PCI_DEVICE_ID_AH_IOV CHIP_NUM_AH_IOV + #endif enum qede_pci_private { @@ -110,6 +115,10 @@ static const struct pci_device_id qede_pci_tbl[] = { #ifdef CONFIG_QED_SRIOV {PCI_VDEVICE(QLOGIC, PCI_DEVICE_ID_57980S_IOV), QEDE_PRIVATE_VF}, #endif + {PCI_VDEVICE(QLOGIC, PCI_DEVICE_ID_AH), QEDE_PRIVATE_PF}, +#ifdef CONFIG_QED_SRIOV + {PCI_VDEVICE(QLOGIC, PCI_DEVICE_ID_AH_IOV), QEDE_PRIVATE_VF}, +#endif { 0 } }; @@ -216,9 +225,13 @@ static struct pci_driver qede_pci_driver = { static struct qed_eth_cb_ops qede_ll_ops = { { +#ifdef CONFIG_RFS_ACCEL + .arfs_filter_op = qede_arfs_filter_op, +#endif .link_update = qede_link_update, }, .force_mac = qede_force_mac, + .ports_update = qede_udp_ports_update, }; static int qede_netdev_event(struct notifier_block *this, unsigned long event, @@ -314,122 +327,135 @@ static int qede_close(struct net_device *ndev); void qede_fill_by_demand_stats(struct qede_dev *edev) { + struct qede_stats_common *p_common = &edev->stats.common; struct qed_eth_stats stats; edev->ops->get_vport_stats(edev->cdev, &stats); - edev->stats.no_buff_discards = stats.no_buff_discards; - edev->stats.packet_too_big_discard = stats.packet_too_big_discard; - edev->stats.ttl0_discard = stats.ttl0_discard; - edev->stats.rx_ucast_bytes = stats.rx_ucast_bytes; - edev->stats.rx_mcast_bytes = stats.rx_mcast_bytes; - edev->stats.rx_bcast_bytes = stats.rx_bcast_bytes; - edev->stats.rx_ucast_pkts = stats.rx_ucast_pkts; - edev->stats.rx_mcast_pkts = stats.rx_mcast_pkts; - edev->stats.rx_bcast_pkts = stats.rx_bcast_pkts; - edev->stats.mftag_filter_discards = stats.mftag_filter_discards; - edev->stats.mac_filter_discards = stats.mac_filter_discards; - - edev->stats.tx_ucast_bytes = stats.tx_ucast_bytes; - edev->stats.tx_mcast_bytes = stats.tx_mcast_bytes; - edev->stats.tx_bcast_bytes = stats.tx_bcast_bytes; - edev->stats.tx_ucast_pkts = stats.tx_ucast_pkts; - edev->stats.tx_mcast_pkts = stats.tx_mcast_pkts; - edev->stats.tx_bcast_pkts = stats.tx_bcast_pkts; - edev->stats.tx_err_drop_pkts = stats.tx_err_drop_pkts; - edev->stats.coalesced_pkts = stats.tpa_coalesced_pkts; - edev->stats.coalesced_events = stats.tpa_coalesced_events; - edev->stats.coalesced_aborts_num = stats.tpa_aborts_num; - edev->stats.non_coalesced_pkts = stats.tpa_not_coalesced_pkts; - edev->stats.coalesced_bytes = stats.tpa_coalesced_bytes; - - edev->stats.rx_64_byte_packets = stats.rx_64_byte_packets; - edev->stats.rx_65_to_127_byte_packets = stats.rx_65_to_127_byte_packets; - edev->stats.rx_128_to_255_byte_packets = - stats.rx_128_to_255_byte_packets; - edev->stats.rx_256_to_511_byte_packets = - stats.rx_256_to_511_byte_packets; - edev->stats.rx_512_to_1023_byte_packets = - stats.rx_512_to_1023_byte_packets; - edev->stats.rx_1024_to_1518_byte_packets = - stats.rx_1024_to_1518_byte_packets; - edev->stats.rx_1519_to_1522_byte_packets = - stats.rx_1519_to_1522_byte_packets; - edev->stats.rx_1519_to_2047_byte_packets = - stats.rx_1519_to_2047_byte_packets; - edev->stats.rx_2048_to_4095_byte_packets = - stats.rx_2048_to_4095_byte_packets; - edev->stats.rx_4096_to_9216_byte_packets = - stats.rx_4096_to_9216_byte_packets; - edev->stats.rx_9217_to_16383_byte_packets = - stats.rx_9217_to_16383_byte_packets; - edev->stats.rx_crc_errors = stats.rx_crc_errors; - edev->stats.rx_mac_crtl_frames = stats.rx_mac_crtl_frames; - edev->stats.rx_pause_frames = stats.rx_pause_frames; - edev->stats.rx_pfc_frames = stats.rx_pfc_frames; - edev->stats.rx_align_errors = stats.rx_align_errors; - edev->stats.rx_carrier_errors = stats.rx_carrier_errors; - edev->stats.rx_oversize_packets = stats.rx_oversize_packets; - edev->stats.rx_jabbers = stats.rx_jabbers; - edev->stats.rx_undersize_packets = stats.rx_undersize_packets; - edev->stats.rx_fragments = stats.rx_fragments; - edev->stats.tx_64_byte_packets = stats.tx_64_byte_packets; - edev->stats.tx_65_to_127_byte_packets = stats.tx_65_to_127_byte_packets; - edev->stats.tx_128_to_255_byte_packets = - stats.tx_128_to_255_byte_packets; - edev->stats.tx_256_to_511_byte_packets = - stats.tx_256_to_511_byte_packets; - edev->stats.tx_512_to_1023_byte_packets = - stats.tx_512_to_1023_byte_packets; - edev->stats.tx_1024_to_1518_byte_packets = - stats.tx_1024_to_1518_byte_packets; - edev->stats.tx_1519_to_2047_byte_packets = - stats.tx_1519_to_2047_byte_packets; - edev->stats.tx_2048_to_4095_byte_packets = - stats.tx_2048_to_4095_byte_packets; - edev->stats.tx_4096_to_9216_byte_packets = - stats.tx_4096_to_9216_byte_packets; - edev->stats.tx_9217_to_16383_byte_packets = - stats.tx_9217_to_16383_byte_packets; - edev->stats.tx_pause_frames = stats.tx_pause_frames; - edev->stats.tx_pfc_frames = stats.tx_pfc_frames; - edev->stats.tx_lpi_entry_count = stats.tx_lpi_entry_count; - edev->stats.tx_total_collisions = stats.tx_total_collisions; - edev->stats.brb_truncates = stats.brb_truncates; - edev->stats.brb_discards = stats.brb_discards; - edev->stats.tx_mac_ctrl_frames = stats.tx_mac_ctrl_frames; + + p_common->no_buff_discards = stats.common.no_buff_discards; + p_common->packet_too_big_discard = stats.common.packet_too_big_discard; + p_common->ttl0_discard = stats.common.ttl0_discard; + p_common->rx_ucast_bytes = stats.common.rx_ucast_bytes; + p_common->rx_mcast_bytes = stats.common.rx_mcast_bytes; + p_common->rx_bcast_bytes = stats.common.rx_bcast_bytes; + p_common->rx_ucast_pkts = stats.common.rx_ucast_pkts; + p_common->rx_mcast_pkts = stats.common.rx_mcast_pkts; + p_common->rx_bcast_pkts = stats.common.rx_bcast_pkts; + p_common->mftag_filter_discards = stats.common.mftag_filter_discards; + p_common->mac_filter_discards = stats.common.mac_filter_discards; + + p_common->tx_ucast_bytes = stats.common.tx_ucast_bytes; + p_common->tx_mcast_bytes = stats.common.tx_mcast_bytes; + p_common->tx_bcast_bytes = stats.common.tx_bcast_bytes; + p_common->tx_ucast_pkts = stats.common.tx_ucast_pkts; + p_common->tx_mcast_pkts = stats.common.tx_mcast_pkts; + p_common->tx_bcast_pkts = stats.common.tx_bcast_pkts; + p_common->tx_err_drop_pkts = stats.common.tx_err_drop_pkts; + p_common->coalesced_pkts = stats.common.tpa_coalesced_pkts; + p_common->coalesced_events = stats.common.tpa_coalesced_events; + p_common->coalesced_aborts_num = stats.common.tpa_aborts_num; + p_common->non_coalesced_pkts = stats.common.tpa_not_coalesced_pkts; + p_common->coalesced_bytes = stats.common.tpa_coalesced_bytes; + + p_common->rx_64_byte_packets = stats.common.rx_64_byte_packets; + p_common->rx_65_to_127_byte_packets = + stats.common.rx_65_to_127_byte_packets; + p_common->rx_128_to_255_byte_packets = + stats.common.rx_128_to_255_byte_packets; + p_common->rx_256_to_511_byte_packets = + stats.common.rx_256_to_511_byte_packets; + p_common->rx_512_to_1023_byte_packets = + stats.common.rx_512_to_1023_byte_packets; + p_common->rx_1024_to_1518_byte_packets = + stats.common.rx_1024_to_1518_byte_packets; + p_common->rx_crc_errors = stats.common.rx_crc_errors; + p_common->rx_mac_crtl_frames = stats.common.rx_mac_crtl_frames; + p_common->rx_pause_frames = stats.common.rx_pause_frames; + p_common->rx_pfc_frames = stats.common.rx_pfc_frames; + p_common->rx_align_errors = stats.common.rx_align_errors; + p_common->rx_carrier_errors = stats.common.rx_carrier_errors; + p_common->rx_oversize_packets = stats.common.rx_oversize_packets; + p_common->rx_jabbers = stats.common.rx_jabbers; + p_common->rx_undersize_packets = stats.common.rx_undersize_packets; + p_common->rx_fragments = stats.common.rx_fragments; + p_common->tx_64_byte_packets = stats.common.tx_64_byte_packets; + p_common->tx_65_to_127_byte_packets = + stats.common.tx_65_to_127_byte_packets; + p_common->tx_128_to_255_byte_packets = + stats.common.tx_128_to_255_byte_packets; + p_common->tx_256_to_511_byte_packets = + stats.common.tx_256_to_511_byte_packets; + p_common->tx_512_to_1023_byte_packets = + stats.common.tx_512_to_1023_byte_packets; + p_common->tx_1024_to_1518_byte_packets = + stats.common.tx_1024_to_1518_byte_packets; + p_common->tx_pause_frames = stats.common.tx_pause_frames; + p_common->tx_pfc_frames = stats.common.tx_pfc_frames; + p_common->brb_truncates = stats.common.brb_truncates; + p_common->brb_discards = stats.common.brb_discards; + p_common->tx_mac_ctrl_frames = stats.common.tx_mac_ctrl_frames; + + if (QEDE_IS_BB(edev)) { + struct qede_stats_bb *p_bb = &edev->stats.bb; + + p_bb->rx_1519_to_1522_byte_packets = + stats.bb.rx_1519_to_1522_byte_packets; + p_bb->rx_1519_to_2047_byte_packets = + stats.bb.rx_1519_to_2047_byte_packets; + p_bb->rx_2048_to_4095_byte_packets = + stats.bb.rx_2048_to_4095_byte_packets; + p_bb->rx_4096_to_9216_byte_packets = + stats.bb.rx_4096_to_9216_byte_packets; + p_bb->rx_9217_to_16383_byte_packets = + stats.bb.rx_9217_to_16383_byte_packets; + p_bb->tx_1519_to_2047_byte_packets = + stats.bb.tx_1519_to_2047_byte_packets; + p_bb->tx_2048_to_4095_byte_packets = + stats.bb.tx_2048_to_4095_byte_packets; + p_bb->tx_4096_to_9216_byte_packets = + stats.bb.tx_4096_to_9216_byte_packets; + p_bb->tx_9217_to_16383_byte_packets = + stats.bb.tx_9217_to_16383_byte_packets; + p_bb->tx_lpi_entry_count = stats.bb.tx_lpi_entry_count; + p_bb->tx_total_collisions = stats.bb.tx_total_collisions; + } else { + struct qede_stats_ah *p_ah = &edev->stats.ah; + + p_ah->rx_1519_to_max_byte_packets = + stats.ah.rx_1519_to_max_byte_packets; + p_ah->tx_1519_to_max_byte_packets = + stats.ah.tx_1519_to_max_byte_packets; + } } static void qede_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats) { struct qede_dev *edev = netdev_priv(dev); + struct qede_stats_common *p_common; qede_fill_by_demand_stats(edev); + p_common = &edev->stats.common; - stats->rx_packets = edev->stats.rx_ucast_pkts + - edev->stats.rx_mcast_pkts + - edev->stats.rx_bcast_pkts; - stats->tx_packets = edev->stats.tx_ucast_pkts + - edev->stats.tx_mcast_pkts + - edev->stats.tx_bcast_pkts; - - stats->rx_bytes = edev->stats.rx_ucast_bytes + - edev->stats.rx_mcast_bytes + - edev->stats.rx_bcast_bytes; + stats->rx_packets = p_common->rx_ucast_pkts + p_common->rx_mcast_pkts + + p_common->rx_bcast_pkts; + stats->tx_packets = p_common->tx_ucast_pkts + p_common->tx_mcast_pkts + + p_common->tx_bcast_pkts; - stats->tx_bytes = edev->stats.tx_ucast_bytes + - edev->stats.tx_mcast_bytes + - edev->stats.tx_bcast_bytes; + stats->rx_bytes = p_common->rx_ucast_bytes + p_common->rx_mcast_bytes + + p_common->rx_bcast_bytes; + stats->tx_bytes = p_common->tx_ucast_bytes + p_common->tx_mcast_bytes + + p_common->tx_bcast_bytes; - stats->tx_errors = edev->stats.tx_err_drop_pkts; - stats->multicast = edev->stats.rx_mcast_pkts + - edev->stats.rx_bcast_pkts; + stats->tx_errors = p_common->tx_err_drop_pkts; + stats->multicast = p_common->rx_mcast_pkts + p_common->rx_bcast_pkts; - stats->rx_fifo_errors = edev->stats.no_buff_discards; + stats->rx_fifo_errors = p_common->no_buff_discards; - stats->collisions = edev->stats.tx_total_collisions; - stats->rx_crc_errors = edev->stats.rx_crc_errors; - stats->rx_frame_errors = edev->stats.rx_align_errors; + if (QEDE_IS_BB(edev)) + stats->collisions = edev->stats.bb.tx_total_collisions; + stats->rx_crc_errors = p_common->rx_crc_errors; + stats->rx_frame_errors = p_common->rx_align_errors; } #ifdef CONFIG_QED_SRIOV @@ -532,6 +558,9 @@ static const struct net_device_ops qede_netdev_ops = { .ndo_udp_tunnel_del = qede_udp_tunnel_del, .ndo_features_check = qede_features_check, .ndo_xdp = qede_xdp, +#ifdef CONFIG_RFS_ACCEL + .ndo_rx_flow_steer = qede_rx_flow_steer, +#endif }; /* ------------------------------------------------------------------------- @@ -581,7 +610,8 @@ static void qede_init_ndev(struct qede_dev *edev) { struct net_device *ndev = edev->ndev; struct pci_dev *pdev = edev->pdev; - u32 hw_features; + bool udp_tunnel_enable = false; + netdev_features_t hw_features; pci_set_drvdata(pdev, ndev); @@ -603,16 +633,33 @@ static void qede_init_ndev(struct qede_dev *edev) NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM | NETIF_F_TSO | NETIF_F_TSO6; - /* Encap features*/ - hw_features |= NETIF_F_GSO_GRE | NETIF_F_GSO_UDP_TUNNEL | - NETIF_F_TSO_ECN | NETIF_F_GSO_UDP_TUNNEL_CSUM | - NETIF_F_GSO_GRE_CSUM; - ndev->hw_enc_features = NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM | - NETIF_F_SG | NETIF_F_TSO | NETIF_F_TSO_ECN | - NETIF_F_TSO6 | NETIF_F_GSO_GRE | - NETIF_F_GSO_UDP_TUNNEL | NETIF_F_RXCSUM | - NETIF_F_GSO_UDP_TUNNEL_CSUM | - NETIF_F_GSO_GRE_CSUM; + if (!IS_VF(edev) && edev->dev_info.common.num_hwfns == 1) + hw_features |= NETIF_F_NTUPLE; + + if (edev->dev_info.common.vxlan_enable || + edev->dev_info.common.geneve_enable) + udp_tunnel_enable = true; + + if (udp_tunnel_enable || edev->dev_info.common.gre_enable) { + hw_features |= NETIF_F_TSO_ECN; + ndev->hw_enc_features = NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM | + NETIF_F_SG | NETIF_F_TSO | + NETIF_F_TSO_ECN | NETIF_F_TSO6 | + NETIF_F_RXCSUM; + } + + if (udp_tunnel_enable) { + hw_features |= (NETIF_F_GSO_UDP_TUNNEL | + NETIF_F_GSO_UDP_TUNNEL_CSUM); + ndev->hw_enc_features |= (NETIF_F_GSO_UDP_TUNNEL | + NETIF_F_GSO_UDP_TUNNEL_CSUM); + } + + if (edev->dev_info.common.gre_enable) { + hw_features |= (NETIF_F_GSO_GRE | NETIF_F_GSO_GRE_CSUM); + ndev->hw_enc_features |= (NETIF_F_GSO_GRE | + NETIF_F_GSO_GRE_CSUM); + } ndev->vlan_features = hw_features | NETIF_F_RXHASH | NETIF_F_RXCSUM | NETIF_F_HIGHDMA; @@ -750,7 +797,6 @@ static void qede_sp_task(struct work_struct *work) { struct qede_dev *edev = container_of(work, struct qede_dev, sp_task.work); - struct qed_dev *cdev = edev->cdev; __qede_lock(edev); @@ -758,24 +804,12 @@ static void qede_sp_task(struct work_struct *work) if (edev->state == QEDE_STATE_OPEN) qede_config_rx_mode(edev->ndev); - if (test_and_clear_bit(QEDE_SP_VXLAN_PORT_CONFIG, &edev->sp_flags)) { - struct qed_tunn_params tunn_params; - - memset(&tunn_params, 0, sizeof(tunn_params)); - tunn_params.update_vxlan_port = 1; - tunn_params.vxlan_port = edev->vxlan_dst_port; - qed_ops->tunn_config(cdev, &tunn_params); - } - - if (test_and_clear_bit(QEDE_SP_GENEVE_PORT_CONFIG, &edev->sp_flags)) { - struct qed_tunn_params tunn_params; - - memset(&tunn_params, 0, sizeof(tunn_params)); - tunn_params.update_geneve_port = 1; - tunn_params.geneve_port = edev->geneve_dst_port; - qed_ops->tunn_config(cdev, &tunn_params); +#ifdef CONFIG_RFS_ACCEL + if (test_and_clear_bit(QEDE_SP_ARFS_CONFIG, &edev->sp_flags)) { + if (edev->state == QEDE_STATE_OPEN) + qede_process_arfs_filters(edev, false); } - +#endif __qede_unlock(edev); } @@ -786,6 +820,9 @@ static void qede_update_pf_params(struct qed_dev *cdev) /* 64 rx + 64 tx + 64 XDP */ memset(&pf_params, 0, sizeof(struct qed_pf_params)); pf_params.eth_pf_params.num_cons = (MAX_SB_PER_PF_MIMD - 1) * 3; +#ifdef CONFIG_RFS_ACCEL + pf_params.eth_pf_params.num_arfs_filters = QEDE_RFS_MAX_FLTR; +#endif qed_ops->common->update_pf_params(cdev, &pf_params); } @@ -870,13 +907,8 @@ static int __qede_probe(struct pci_dev *pdev, u32 dp_module, u8 dp_level, edev->ops->common->set_id(cdev, edev->ndev->name, DRV_MODULE_VERSION); /* PTP not supported on VFs */ - if (!is_vf) { - rc = qede_ptp_register_phc(edev); - if (rc) { - DP_NOTICE(edev, "Cannot register PHC\n"); - goto err5; - } - } + if (!is_vf) + qede_ptp_enable(edev, true); edev->ops->register_ops(cdev, &qede_ll_ops, edev); @@ -891,8 +923,6 @@ static int __qede_probe(struct pci_dev *pdev, u32 dp_module, u8 dp_level, return 0; -err5: - unregister_netdev(edev->ndev); err4: qede_roce_dev_remove(edev); err3: @@ -940,11 +970,10 @@ static void __qede_remove(struct pci_dev *pdev, enum qede_remove_mode mode) DP_INFO(edev, "Starting qede_remove\n"); - cancel_delayed_work_sync(&edev->sp_task); - unregister_netdev(ndev); + cancel_delayed_work_sync(&edev->sp_task); - qede_ptp_remove(edev); + qede_ptp_disable(edev); qede_roce_dev_remove(edev); @@ -1165,9 +1194,11 @@ static int qede_alloc_mem_rxq(struct qede_dev *edev, struct qede_rx_queue *rxq) rxq->num_rx_buffers = edev->q_num_rx_buffers; rxq->rx_buf_size = NET_IP_ALIGN + ETH_OVERHEAD + edev->ndev->mtu; + rxq->rx_headroom = edev->xdp_prog ? XDP_PACKET_HEADROOM : 0; - if (rxq->rx_buf_size > PAGE_SIZE) - rxq->rx_buf_size = PAGE_SIZE; + /* Make sure that the headroom and payload fit in a single page */ + if (rxq->rx_buf_size + rxq->rx_headroom > PAGE_SIZE) + rxq->rx_buf_size = PAGE_SIZE - rxq->rx_headroom; /* Segment size to spilt a page in multiple equal parts, * unless XDP is used in which case we'd use the entire page. @@ -1229,7 +1260,7 @@ static void qede_free_mem_txq(struct qede_dev *edev, struct qede_tx_queue *txq) { /* Free the parallel SW ring */ if (txq->is_xdp) - kfree(txq->sw_tx_ring.pages); + kfree(txq->sw_tx_ring.xdp); else kfree(txq->sw_tx_ring.skbs); @@ -1247,9 +1278,9 @@ static int qede_alloc_mem_txq(struct qede_dev *edev, struct qede_tx_queue *txq) /* Allocate the parallel driver ring for Tx buffers */ if (txq->is_xdp) { - size = sizeof(*txq->sw_tx_ring.pages) * TX_RING_SIZE; - txq->sw_tx_ring.pages = kzalloc(size, GFP_KERNEL); - if (!txq->sw_tx_ring.pages) + size = sizeof(*txq->sw_tx_ring.xdp) * TX_RING_SIZE; + txq->sw_tx_ring.xdp = kzalloc(size, GFP_KERNEL); + if (!txq->sw_tx_ring.xdp) goto err; } else { size = sizeof(*txq->sw_tx_ring.skbs) * TX_RING_SIZE; @@ -1466,6 +1497,18 @@ static int qede_req_msix_irqs(struct qede_dev *edev) } for (i = 0; i < QEDE_QUEUE_CNT(edev); i++) { +#ifdef CONFIG_RFS_ACCEL + struct qede_fastpath *fp = &edev->fp_array[i]; + + if (edev->ndev->rx_cpu_rmap && (fp->type & QEDE_FASTPATH_RX)) { + rc = irq_cpu_rmap_add(edev->ndev->rx_cpu_rmap, + edev->int_info.msix[i].vector); + if (rc) { + DP_ERR(edev, "Failed to add CPU rmap\n"); + qede_free_arfs(edev); + } + } +#endif rc = request_irq(edev->int_info.msix[i].vector, qede_msix_fp_int, 0, edev->fp_array[i].name, &edev->fp_array[i]); @@ -1827,8 +1870,6 @@ static void qede_unload(struct qede_dev *edev, enum qede_unload_mode mode, qede_roce_dev_event_close(edev); edev->state = QEDE_STATE_CLOSED; - qede_ptp_stop(edev); - /* Close OS Tx */ netif_tx_disable(edev->ndev); netif_carrier_off(edev->ndev); @@ -1847,7 +1888,12 @@ static void qede_unload(struct qede_dev *edev, enum qede_unload_mode mode, qede_vlan_mark_nonconfigured(edev); edev->ops->fastpath_stop(edev->cdev); - +#ifdef CONFIG_RFS_ACCEL + if (!IS_VF(edev) && edev->dev_info.common.num_hwfns == 1) { + qede_poll_for_freeing_arfs_filters(edev); + qede_free_arfs(edev); + } +#endif /* Release the interrupts */ qede_sync_free_irqs(edev); edev->ops->common->set_fp_int(edev->cdev, 0); @@ -1899,6 +1945,13 @@ static int qede_load(struct qede_dev *edev, enum qede_load_mode mode, if (rc) goto err2; +#ifdef CONFIG_RFS_ACCEL + if (!IS_VF(edev) && edev->dev_info.common.num_hwfns == 1) { + rc = qede_alloc_arfs(edev); + if (rc) + DP_NOTICE(edev, "aRFS memory allocation failed\n"); + } +#endif qede_napi_add_enable(edev); DP_INFO(edev, "Napi added and enabled\n"); @@ -1925,13 +1978,10 @@ static int qede_load(struct qede_dev *edev, enum qede_load_mode mode, qede_roce_dev_event_open(edev); - qede_ptp_start(edev, (mode == QEDE_LOAD_NORMAL)); - edev->state = QEDE_STATE_OPEN; DP_INFO(edev, "Ending successfully qede load\n"); - goto out; err4: qede_sync_free_irqs(edev); diff --git a/drivers/net/ethernet/qlogic/qede/qede_ptp.c b/drivers/net/ethernet/qlogic/qede/qede_ptp.c index 2e62dec09bd7..24f06e2ef43e 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_ptp.c +++ b/drivers/net/ethernet/qlogic/qede/qede_ptp.c @@ -181,6 +181,7 @@ static void qede_ptp_task(struct work_struct *work) skb_tstamp_tx(ptp->tx_skb, &shhwtstamps); dev_kfree_skb_any(ptp->tx_skb); ptp->tx_skb = NULL; + clear_bit_unlock(QEDE_FLAGS_PTP_TX_IN_PRORGESS, &edev->flags); DP_VERBOSE(edev, QED_MSG_DEBUG, "Tx timestamp, timestamp cycles = %llu, ns = %llu\n", @@ -206,23 +207,10 @@ static u64 qede_ptp_read_cc(const struct cyclecounter *cc) return phc_cycles; } -static void qede_ptp_init_cc(struct qede_dev *edev) -{ - struct qede_ptp *ptp; - - ptp = edev->ptp; - if (!ptp) - return; - - memset(&ptp->cc, 0, sizeof(ptp->cc)); - ptp->cc.read = qede_ptp_read_cc; - ptp->cc.mask = CYCLECOUNTER_MASK(64); - ptp->cc.shift = 0; - ptp->cc.mult = 1; -} - static int qede_ptp_cfg_filters(struct qede_dev *edev) { + enum qed_ptp_hwtstamp_tx_type tx_type = QED_PTP_HWTSTAMP_TX_ON; + enum qed_ptp_filter_type rx_filter = QED_PTP_FILTER_NONE; struct qede_ptp *ptp = edev->ptp; if (!ptp) @@ -236,7 +224,12 @@ static int qede_ptp_cfg_filters(struct qede_dev *edev) switch (ptp->tx_type) { case HWTSTAMP_TX_ON: edev->flags |= QEDE_TX_TIMESTAMPING_EN; - ptp->ops->hwtstamp_tx_on(edev->cdev); + tx_type = QED_PTP_HWTSTAMP_TX_ON; + break; + + case HWTSTAMP_TX_OFF: + edev->flags &= ~QEDE_TX_TIMESTAMPING_EN; + tx_type = QED_PTP_HWTSTAMP_TX_OFF; break; case HWTSTAMP_TX_ONESTEP_SYNC: @@ -247,42 +240,57 @@ static int qede_ptp_cfg_filters(struct qede_dev *edev) spin_lock_bh(&ptp->lock); switch (ptp->rx_filter) { case HWTSTAMP_FILTER_NONE: + rx_filter = QED_PTP_FILTER_NONE; break; case HWTSTAMP_FILTER_ALL: case HWTSTAMP_FILTER_SOME: ptp->rx_filter = HWTSTAMP_FILTER_NONE; + rx_filter = QED_PTP_FILTER_ALL; break; case HWTSTAMP_FILTER_PTP_V1_L4_EVENT: + ptp->rx_filter = HWTSTAMP_FILTER_PTP_V1_L4_EVENT; + rx_filter = QED_PTP_FILTER_V1_L4_EVENT; + break; case HWTSTAMP_FILTER_PTP_V1_L4_SYNC: case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ: ptp->rx_filter = HWTSTAMP_FILTER_PTP_V1_L4_EVENT; /* Initialize PTP detection for UDP/IPv4 events */ - ptp->ops->cfg_rx_filters(edev->cdev, QED_PTP_FILTER_IPV4); + rx_filter = QED_PTP_FILTER_V1_L4_GEN; break; case HWTSTAMP_FILTER_PTP_V2_L4_EVENT: + ptp->rx_filter = HWTSTAMP_FILTER_PTP_V2_L4_EVENT; + rx_filter = QED_PTP_FILTER_V2_L4_EVENT; + break; case HWTSTAMP_FILTER_PTP_V2_L4_SYNC: case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ: ptp->rx_filter = HWTSTAMP_FILTER_PTP_V2_L4_EVENT; /* Initialize PTP detection for UDP/IPv4 or UDP/IPv6 events */ - ptp->ops->cfg_rx_filters(edev->cdev, QED_PTP_FILTER_IPV4_IPV6); + rx_filter = QED_PTP_FILTER_V2_L4_GEN; break; case HWTSTAMP_FILTER_PTP_V2_L2_EVENT: + ptp->rx_filter = HWTSTAMP_FILTER_PTP_V2_L2_EVENT; + rx_filter = QED_PTP_FILTER_V2_L2_EVENT; + break; case HWTSTAMP_FILTER_PTP_V2_L2_SYNC: case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ: ptp->rx_filter = HWTSTAMP_FILTER_PTP_V2_L2_EVENT; /* Initialize PTP detection L2 events */ - ptp->ops->cfg_rx_filters(edev->cdev, QED_PTP_FILTER_L2); + rx_filter = QED_PTP_FILTER_V2_L2_GEN; break; case HWTSTAMP_FILTER_PTP_V2_EVENT: + ptp->rx_filter = HWTSTAMP_FILTER_PTP_V2_EVENT; + rx_filter = QED_PTP_FILTER_V2_EVENT; + break; case HWTSTAMP_FILTER_PTP_V2_SYNC: case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ: ptp->rx_filter = HWTSTAMP_FILTER_PTP_V2_EVENT; /* Initialize PTP detection L2, UDP/IPv4 or UDP/IPv6 events */ - ptp->ops->cfg_rx_filters(edev->cdev, - QED_PTP_FILTER_L2_IPV4_IPV6); + rx_filter = QED_PTP_FILTER_V2_GEN; break; } + ptp->ops->cfg_filters(edev->cdev, rx_filter, tx_type); + spin_unlock_bh(&ptp->lock); return 0; @@ -324,61 +332,6 @@ int qede_ptp_hw_ts(struct qede_dev *edev, struct ifreq *ifr) sizeof(config)) ? -EFAULT : 0; } -/* Called during load, to initialize PTP-related stuff */ -static void qede_ptp_init(struct qede_dev *edev, bool init_tc) -{ - struct qede_ptp *ptp; - int rc; - - ptp = edev->ptp; - if (!ptp) - return; - - spin_lock_init(&ptp->lock); - - /* Configure PTP in HW */ - rc = ptp->ops->enable(edev->cdev); - if (rc) { - DP_ERR(edev, "Stopping PTP initialization\n"); - return; - } - - /* Init work queue for Tx timestamping */ - INIT_WORK(&ptp->work, qede_ptp_task); - - /* Init cyclecounter and timecounter. This is done only in the first - * load. If done in every load, PTP application will fail when doing - * unload / load (e.g. MTU change) while it is running. - */ - if (init_tc) { - qede_ptp_init_cc(edev); - timecounter_init(&ptp->tc, &ptp->cc, - ktime_to_ns(ktime_get_real())); - } - - DP_VERBOSE(edev, QED_MSG_DEBUG, "PTP initialization is successful\n"); -} - -void qede_ptp_start(struct qede_dev *edev, bool init_tc) -{ - qede_ptp_init(edev, init_tc); - qede_ptp_cfg_filters(edev); -} - -void qede_ptp_remove(struct qede_dev *edev) -{ - struct qede_ptp *ptp; - - ptp = edev->ptp; - if (ptp && ptp->clock) { - ptp_clock_unregister(ptp->clock); - ptp->clock = NULL; - } - - kfree(ptp); - edev->ptp = NULL; -} - int qede_ptp_get_ts_info(struct qede_dev *edev, struct ethtool_ts_info *info) { struct qede_ptp *ptp = edev->ptp; @@ -417,8 +370,7 @@ int qede_ptp_get_ts_info(struct qede_dev *edev, struct ethtool_ts_info *info) return 0; } -/* Called during unload, to stop PTP-related stuff */ -void qede_ptp_stop(struct qede_dev *edev) +void qede_ptp_disable(struct qede_dev *edev) { struct qede_ptp *ptp; @@ -426,6 +378,11 @@ void qede_ptp_stop(struct qede_dev *edev) if (!ptp) return; + if (ptp->clock) { + ptp_clock_unregister(ptp->clock); + ptp->clock = NULL; + } + /* Cancel PTP work queue. Should be done after the Tx queues are * drained to prevent additional scheduling. */ @@ -439,11 +396,54 @@ void qede_ptp_stop(struct qede_dev *edev) spin_lock_bh(&ptp->lock); ptp->ops->disable(edev->cdev); spin_unlock_bh(&ptp->lock); + + kfree(ptp); + edev->ptp = NULL; +} + +static int qede_ptp_init(struct qede_dev *edev, bool init_tc) +{ + struct qede_ptp *ptp; + int rc; + + ptp = edev->ptp; + if (!ptp) + return -EINVAL; + + spin_lock_init(&ptp->lock); + + /* Configure PTP in HW */ + rc = ptp->ops->enable(edev->cdev); + if (rc) { + DP_INFO(edev, "PTP HW enable failed\n"); + return rc; + } + + /* Init work queue for Tx timestamping */ + INIT_WORK(&ptp->work, qede_ptp_task); + + /* Init cyclecounter and timecounter. This is done only in the first + * load. If done in every load, PTP application will fail when doing + * unload / load (e.g. MTU change) while it is running. + */ + if (init_tc) { + memset(&ptp->cc, 0, sizeof(ptp->cc)); + ptp->cc.read = qede_ptp_read_cc; + ptp->cc.mask = CYCLECOUNTER_MASK(64); + ptp->cc.shift = 0; + ptp->cc.mult = 1; + + timecounter_init(&ptp->tc, &ptp->cc, + ktime_to_ns(ktime_get_real())); + } + + return rc; } -int qede_ptp_register_phc(struct qede_dev *edev) +int qede_ptp_enable(struct qede_dev *edev, bool init_tc) { struct qede_ptp *ptp; + int rc; ptp = kzalloc(sizeof(*ptp), GFP_KERNEL); if (!ptp) { @@ -454,14 +454,19 @@ int qede_ptp_register_phc(struct qede_dev *edev) ptp->edev = edev; ptp->ops = edev->ops->ptp; if (!ptp->ops) { - kfree(ptp); - edev->ptp = NULL; - DP_ERR(edev, "PTP clock registeration failed\n"); - return -EIO; + DP_INFO(edev, "PTP enable failed\n"); + rc = -EIO; + goto err1; } edev->ptp = ptp; + rc = qede_ptp_init(edev, init_tc); + if (rc) + goto err1; + + qede_ptp_cfg_filters(edev); + /* Fill the ptp_clock_info struct and register PTP clock */ ptp->clock_info.owner = THIS_MODULE; snprintf(ptp->clock_info.name, 16, "%s", edev->ndev->name); @@ -478,13 +483,21 @@ int qede_ptp_register_phc(struct qede_dev *edev) ptp->clock = ptp_clock_register(&ptp->clock_info, &edev->pdev->dev); if (IS_ERR(ptp->clock)) { - ptp->clock = NULL; - kfree(ptp); - edev->ptp = NULL; + rc = -EINVAL; DP_ERR(edev, "PTP clock registeration failed\n"); + goto err2; } return 0; + +err2: + qede_ptp_disable(edev); + ptp->clock = NULL; +err1: + kfree(ptp); + edev->ptp = NULL; + + return rc; } void qede_ptp_tx_ts(struct qede_dev *edev, struct sk_buff *skb) @@ -495,6 +508,9 @@ void qede_ptp_tx_ts(struct qede_dev *edev, struct sk_buff *skb) if (!ptp) return; + if (test_and_set_bit_lock(QEDE_FLAGS_PTP_TX_IN_PRORGESS, &edev->flags)) + return; + if (unlikely(!(edev->flags & QEDE_TX_TIMESTAMPING_EN))) { DP_NOTICE(edev, "Tx timestamping was not enabled, this packet will not be timestamped\n"); diff --git a/drivers/net/ethernet/qlogic/qede/qede_ptp.h b/drivers/net/ethernet/qlogic/qede/qede_ptp.h index f328f9bba53a..691a14c4b2c5 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_ptp.h +++ b/drivers/net/ethernet/qlogic/qede/qede_ptp.h @@ -40,10 +40,8 @@ void qede_ptp_rx_ts(struct qede_dev *edev, struct sk_buff *skb); void qede_ptp_tx_ts(struct qede_dev *edev, struct sk_buff *skb); int qede_ptp_hw_ts(struct qede_dev *edev, struct ifreq *req); -void qede_ptp_start(struct qede_dev *edev, bool init_tc); -void qede_ptp_stop(struct qede_dev *edev); -void qede_ptp_remove(struct qede_dev *edev); -int qede_ptp_register_phc(struct qede_dev *edev); +void qede_ptp_disable(struct qede_dev *edev); +int qede_ptp_enable(struct qede_dev *edev, bool init_tc); int qede_ptp_get_ts_info(struct qede_dev *edev, struct ethtool_ts_info *ts); static inline void qede_ptp_record_rx_ts(struct qede_dev *edev, |