diff options
Diffstat (limited to 'drivers/net/ethernet/intel/i40e/i40e_txrx.c')
| -rw-r--r-- | drivers/net/ethernet/intel/i40e/i40e_txrx.c | 2718 |
1 files changed, 1660 insertions, 1058 deletions
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c index b936febc315a..cc0b9efc2637 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c @@ -1,45 +1,16 @@ -/******************************************************************************* - * - * Intel Ethernet Controller XL710 Family Linux Driver - * Copyright(c) 2013 - 2016 Intel Corporation. - * - * 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/>. - * - * The full GNU General Public License is included in this distribution in - * the file called "COPYING". - * - * Contact Information: - * e1000-devel Mailing List <e1000-devel@lists.sourceforge.net> - * Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497 - * - ******************************************************************************/ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright(c) 2013 - 2018 Intel Corporation. */ -#include <linux/prefetch.h> -#include <net/busy_poll.h> #include <linux/bpf_trace.h> -#include "i40e.h" +#include <linux/net/intel/libie/pctype.h> +#include <linux/net/intel/libie/rx.h> +#include <linux/prefetch.h> +#include <linux/sctp.h> +#include <net/mpls.h> +#include <net/xdp.h> +#include "i40e_txrx_common.h" #include "i40e_trace.h" -#include "i40e_prototype.h" - -static inline __le64 build_ctob(u32 td_cmd, u32 td_offset, unsigned int size, - u32 td_tag) -{ - return cpu_to_le64(I40E_TX_DESC_DTYPE_DATA | - ((u64)td_cmd << I40E_TXD_QW1_CMD_SHIFT) | - ((u64)td_offset << I40E_TXD_QW1_OFFSET_SHIFT) | - ((u64)size << I40E_TXD_QW1_TX_BUF_SZ_SHIFT) | - ((u64)td_tag << I40E_TXD_QW1_L2TAG1_SHIFT)); -} +#include "i40e_xsk.h" #define I40E_TXD_CMD (I40E_TX_DESC_CMD_EOP | I40E_TX_DESC_CMD_RS) /** @@ -54,7 +25,7 @@ static void i40e_fdir(struct i40e_ring *tx_ring, { struct i40e_filter_program_desc *fdir_desc; struct i40e_pf *pf = tx_ring->vsi->back; - u32 flex_ptype, dtype_cmd; + u32 flex_ptype, dtype_cmd, vsi_id; u16 i; /* grab the next descriptor */ @@ -64,22 +35,16 @@ static void i40e_fdir(struct i40e_ring *tx_ring, i++; tx_ring->next_to_use = (i < tx_ring->count) ? i : 0; - flex_ptype = I40E_TXD_FLTR_QW0_QINDEX_MASK & - (fdata->q_index << I40E_TXD_FLTR_QW0_QINDEX_SHIFT); + flex_ptype = FIELD_PREP(I40E_TXD_FLTR_QW0_QINDEX_MASK, fdata->q_index); - flex_ptype |= I40E_TXD_FLTR_QW0_FLEXOFF_MASK & - (fdata->flex_off << I40E_TXD_FLTR_QW0_FLEXOFF_SHIFT); + flex_ptype |= FIELD_PREP(I40E_TXD_FLTR_QW0_FLEXOFF_MASK, + fdata->flex_off); - flex_ptype |= I40E_TXD_FLTR_QW0_PCTYPE_MASK & - (fdata->pctype << I40E_TXD_FLTR_QW0_PCTYPE_SHIFT); - - flex_ptype |= I40E_TXD_FLTR_QW0_PCTYPE_MASK & - (fdata->flex_offset << I40E_TXD_FLTR_QW0_FLEXOFF_SHIFT); + flex_ptype |= FIELD_PREP(I40E_TXD_FLTR_QW0_PCTYPE_MASK, fdata->pctype); /* Use LAN VSI Id if not programmed by user */ - flex_ptype |= I40E_TXD_FLTR_QW0_DEST_VSI_MASK & - ((u32)(fdata->dest_vsi ? : pf->vsi[pf->lan_vsi]->id) << - I40E_TXD_FLTR_QW0_DEST_VSI_SHIFT); + vsi_id = fdata->dest_vsi ? : i40e_pf_get_main_vsi(pf)->id; + flex_ptype |= FIELD_PREP(I40E_TXD_FLTR_QW0_DEST_VSI_MASK, vsi_id); dtype_cmd = I40E_TX_DESC_DTYPE_FILTER_PROG; @@ -89,17 +54,15 @@ static void i40e_fdir(struct i40e_ring *tx_ring, I40E_FILTER_PROGRAM_DESC_PCMD_REMOVE << I40E_TXD_FLTR_QW1_PCMD_SHIFT; - dtype_cmd |= I40E_TXD_FLTR_QW1_DEST_MASK & - (fdata->dest_ctl << I40E_TXD_FLTR_QW1_DEST_SHIFT); + dtype_cmd |= FIELD_PREP(I40E_TXD_FLTR_QW1_DEST_MASK, fdata->dest_ctl); - dtype_cmd |= I40E_TXD_FLTR_QW1_FD_STATUS_MASK & - (fdata->fd_status << I40E_TXD_FLTR_QW1_FD_STATUS_SHIFT); + dtype_cmd |= FIELD_PREP(I40E_TXD_FLTR_QW1_FD_STATUS_MASK, + fdata->fd_status); if (fdata->cnt_index) { dtype_cmd |= I40E_TXD_FLTR_QW1_CNT_ENA_MASK; - dtype_cmd |= I40E_TXD_FLTR_QW1_CNTINDEX_MASK & - ((u32)fdata->cnt_index << - I40E_TXD_FLTR_QW1_CNTINDEX_SHIFT); + dtype_cmd |= FIELD_PREP(I40E_TXD_FLTR_QW1_CNTINDEX_MASK, + fdata->cnt_index); } fdir_desc->qindex_flex_ptype_vsi = cpu_to_le32(flex_ptype); @@ -191,59 +154,180 @@ dma_fail: return -1; } -#define IP_HEADER_OFFSET 14 -#define I40E_UDPIP_DUMMY_PACKET_LEN 42 /** - * i40e_add_del_fdir_udpv4 - Add/Remove UDPv4 filters - * @vsi: pointer to the targeted VSI - * @fd_data: the flow director data required for the FDir descriptor - * @add: true adds a filter, false removes it + * i40e_create_dummy_packet - Constructs dummy packet for HW + * @dummy_packet: preallocated space for dummy packet + * @ipv4: is layer 3 packet of version 4 or 6 + * @l4proto: next level protocol used in data portion of l3 + * @data: filter data * - * Returns 0 if the filters were successfully added or removed + * Returns address of layer 4 protocol dummy packet. **/ -static int i40e_add_del_fdir_udpv4(struct i40e_vsi *vsi, - struct i40e_fdir_filter *fd_data, - bool add) +static char *i40e_create_dummy_packet(u8 *dummy_packet, bool ipv4, u8 l4proto, + struct i40e_fdir_filter *data) +{ + bool is_vlan = !!data->vlan_tag; + struct vlan_hdr vlan = {}; + struct ipv6hdr ipv6 = {}; + struct ethhdr eth = {}; + struct iphdr ip = {}; + u8 *tmp; + + if (ipv4) { + eth.h_proto = cpu_to_be16(ETH_P_IP); + ip.protocol = l4proto; + ip.version = 0x4; + ip.ihl = 0x5; + + ip.daddr = data->dst_ip; + ip.saddr = data->src_ip; + } else { + eth.h_proto = cpu_to_be16(ETH_P_IPV6); + ipv6.nexthdr = l4proto; + ipv6.version = 0x6; + + memcpy(&ipv6.saddr.in6_u.u6_addr32, data->src_ip6, + sizeof(__be32) * 4); + memcpy(&ipv6.daddr.in6_u.u6_addr32, data->dst_ip6, + sizeof(__be32) * 4); + } + + if (is_vlan) { + vlan.h_vlan_TCI = data->vlan_tag; + vlan.h_vlan_encapsulated_proto = eth.h_proto; + eth.h_proto = data->vlan_etype; + } + + tmp = dummy_packet; + memcpy(tmp, ð, sizeof(eth)); + tmp += sizeof(eth); + + if (is_vlan) { + memcpy(tmp, &vlan, sizeof(vlan)); + tmp += sizeof(vlan); + } + + if (ipv4) { + memcpy(tmp, &ip, sizeof(ip)); + tmp += sizeof(ip); + } else { + memcpy(tmp, &ipv6, sizeof(ipv6)); + tmp += sizeof(ipv6); + } + + return tmp; +} + +/** + * i40e_create_dummy_udp_packet - helper function to create UDP packet + * @raw_packet: preallocated space for dummy packet + * @ipv4: is layer 3 packet of version 4 or 6 + * @l4proto: next level protocol used in data portion of l3 + * @data: filter data + * + * Helper function to populate udp fields. + **/ +static void i40e_create_dummy_udp_packet(u8 *raw_packet, bool ipv4, u8 l4proto, + struct i40e_fdir_filter *data) { - struct i40e_pf *pf = vsi->back; struct udphdr *udp; - struct iphdr *ip; - u8 *raw_packet; - int ret; - static char packet[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0x08, 0, - 0x45, 0, 0, 0x1c, 0, 0, 0x40, 0, 0x40, 0x11, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; + u8 *tmp; - raw_packet = kzalloc(I40E_FDIR_MAX_RAW_PACKET_SIZE, GFP_KERNEL); - if (!raw_packet) - return -ENOMEM; - memcpy(raw_packet, packet, I40E_UDPIP_DUMMY_PACKET_LEN); + tmp = i40e_create_dummy_packet(raw_packet, ipv4, IPPROTO_UDP, data); + udp = (struct udphdr *)(tmp); + udp->dest = data->dst_port; + udp->source = data->src_port; +} + +/** + * i40e_create_dummy_tcp_packet - helper function to create TCP packet + * @raw_packet: preallocated space for dummy packet + * @ipv4: is layer 3 packet of version 4 or 6 + * @l4proto: next level protocol used in data portion of l3 + * @data: filter data + * + * Helper function to populate tcp fields. + **/ +static void i40e_create_dummy_tcp_packet(u8 *raw_packet, bool ipv4, u8 l4proto, + struct i40e_fdir_filter *data) +{ + struct tcphdr *tcp; + u8 *tmp; + /* Dummy tcp packet */ + static const char tcp_packet[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0x50, 0x11, 0x0, 0x72, 0, 0, 0, 0}; - ip = (struct iphdr *)(raw_packet + IP_HEADER_OFFSET); - udp = (struct udphdr *)(raw_packet + IP_HEADER_OFFSET - + sizeof(struct iphdr)); + tmp = i40e_create_dummy_packet(raw_packet, ipv4, IPPROTO_TCP, data); - ip->daddr = fd_data->dst_ip; - udp->dest = fd_data->dst_port; - ip->saddr = fd_data->src_ip; - udp->source = fd_data->src_port; + tcp = (struct tcphdr *)tmp; + memcpy(tcp, tcp_packet, sizeof(tcp_packet)); + tcp->dest = data->dst_port; + tcp->source = data->src_port; +} + +/** + * i40e_create_dummy_sctp_packet - helper function to create SCTP packet + * @raw_packet: preallocated space for dummy packet + * @ipv4: is layer 3 packet of version 4 or 6 + * @l4proto: next level protocol used in data portion of l3 + * @data: filter data + * + * Helper function to populate sctp fields. + **/ +static void i40e_create_dummy_sctp_packet(u8 *raw_packet, bool ipv4, + u8 l4proto, + struct i40e_fdir_filter *data) +{ + struct sctphdr *sctp; + u8 *tmp; + + tmp = i40e_create_dummy_packet(raw_packet, ipv4, IPPROTO_SCTP, data); + + sctp = (struct sctphdr *)tmp; + sctp->dest = data->dst_port; + sctp->source = data->src_port; +} + +/** + * i40e_prepare_fdir_filter - Prepare and program fdir filter + * @pf: physical function to attach filter to + * @fd_data: filter data + * @add: add or delete filter + * @packet_addr: address of dummy packet, used in filtering + * @payload_offset: offset from dummy packet address to user defined data + * @pctype: Packet type for which filter is used + * + * Helper function to offset data of dummy packet, program it and + * handle errors. + **/ +static int i40e_prepare_fdir_filter(struct i40e_pf *pf, + struct i40e_fdir_filter *fd_data, + bool add, char *packet_addr, + int payload_offset, u8 pctype) +{ + int ret; if (fd_data->flex_filter) { - u8 *payload = raw_packet + I40E_UDPIP_DUMMY_PACKET_LEN; + u8 *payload; __be16 pattern = fd_data->flex_word; u16 off = fd_data->flex_offset; + payload = packet_addr + payload_offset; + + /* If user provided vlan, offset payload by vlan header length */ + if (!!fd_data->vlan_tag) + payload += VLAN_HLEN; + *((__force __be16 *)(payload + off)) = pattern; } - fd_data->pctype = I40E_FILTER_PCTYPE_NONF_IPV4_UDP; - ret = i40e_program_fdir_filter(fd_data, raw_packet, pf, add); + fd_data->pctype = pctype; + ret = i40e_program_fdir_filter(fd_data, packet_addr, pf, add); if (ret) { dev_info(&pf->pdev->dev, "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n", fd_data->pctype, fd_data->fd_id, ret); /* Free the packet buffer since it wasn't added to the ring */ - kfree(raw_packet); return -EOPNOTSUPP; } else if (I40E_DEBUG_FD & pf->hw.debug_mask) { if (add) @@ -256,281 +340,316 @@ static int i40e_add_del_fdir_udpv4(struct i40e_vsi *vsi, fd_data->pctype, fd_data->fd_id); } - if (add) - pf->fd_udp4_filter_cnt++; - else - pf->fd_udp4_filter_cnt--; + return ret; +} - return 0; +/** + * i40e_change_filter_num - Prepare and program fdir filter + * @ipv4: is layer 3 packet of version 4 or 6 + * @add: add or delete filter + * @ipv4_filter_num: field to update + * @ipv6_filter_num: field to update + * + * Update filter number field for pf. + **/ +static void i40e_change_filter_num(bool ipv4, bool add, u16 *ipv4_filter_num, + u16 *ipv6_filter_num) +{ + if (add) { + if (ipv4) + (*ipv4_filter_num)++; + else + (*ipv6_filter_num)++; + } else { + if (ipv4) + (*ipv4_filter_num)--; + else + (*ipv6_filter_num)--; + } } -#define I40E_TCPIP_DUMMY_PACKET_LEN 54 +#define I40E_UDPIP_DUMMY_PACKET_LEN 42 +#define I40E_UDPIP6_DUMMY_PACKET_LEN 62 /** - * i40e_add_del_fdir_tcpv4 - Add/Remove TCPv4 filters + * i40e_add_del_fdir_udp - Add/Remove UDP filters * @vsi: pointer to the targeted VSI * @fd_data: the flow director data required for the FDir descriptor * @add: true adds a filter, false removes it + * @ipv4: true is v4, false is v6 * * Returns 0 if the filters were successfully added or removed **/ -static int i40e_add_del_fdir_tcpv4(struct i40e_vsi *vsi, - struct i40e_fdir_filter *fd_data, - bool add) +static int i40e_add_del_fdir_udp(struct i40e_vsi *vsi, + struct i40e_fdir_filter *fd_data, + bool add, + bool ipv4) { struct i40e_pf *pf = vsi->back; - struct tcphdr *tcp; - struct iphdr *ip; u8 *raw_packet; int ret; - /* Dummy packet */ - static char packet[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0x08, 0, - 0x45, 0, 0, 0x28, 0, 0, 0x40, 0, 0x40, 0x6, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0x80, 0x11, - 0x0, 0x72, 0, 0, 0, 0}; raw_packet = kzalloc(I40E_FDIR_MAX_RAW_PACKET_SIZE, GFP_KERNEL); if (!raw_packet) return -ENOMEM; - memcpy(raw_packet, packet, I40E_TCPIP_DUMMY_PACKET_LEN); - - ip = (struct iphdr *)(raw_packet + IP_HEADER_OFFSET); - tcp = (struct tcphdr *)(raw_packet + IP_HEADER_OFFSET - + sizeof(struct iphdr)); - ip->daddr = fd_data->dst_ip; - tcp->dest = fd_data->dst_port; - ip->saddr = fd_data->src_ip; - tcp->source = fd_data->src_port; + i40e_create_dummy_udp_packet(raw_packet, ipv4, IPPROTO_UDP, fd_data); - if (fd_data->flex_filter) { - u8 *payload = raw_packet + I40E_TCPIP_DUMMY_PACKET_LEN; - __be16 pattern = fd_data->flex_word; - u16 off = fd_data->flex_offset; + if (ipv4) + ret = i40e_prepare_fdir_filter + (pf, fd_data, add, raw_packet, + I40E_UDPIP_DUMMY_PACKET_LEN, + LIBIE_FILTER_PCTYPE_NONF_IPV4_UDP); + else + ret = i40e_prepare_fdir_filter + (pf, fd_data, add, raw_packet, + I40E_UDPIP6_DUMMY_PACKET_LEN, + LIBIE_FILTER_PCTYPE_NONF_IPV6_UDP); - *((__force __be16 *)(payload + off)) = pattern; + if (ret) { + kfree(raw_packet); + return ret; } - fd_data->pctype = I40E_FILTER_PCTYPE_NONF_IPV4_TCP; - ret = i40e_program_fdir_filter(fd_data, raw_packet, pf, add); + i40e_change_filter_num(ipv4, add, &pf->fd_udp4_filter_cnt, + &pf->fd_udp6_filter_cnt); + + return 0; +} + +#define I40E_TCPIP_DUMMY_PACKET_LEN 54 +#define I40E_TCPIP6_DUMMY_PACKET_LEN 74 +/** + * i40e_add_del_fdir_tcp - Add/Remove TCPv4 filters + * @vsi: pointer to the targeted VSI + * @fd_data: the flow director data required for the FDir descriptor + * @add: true adds a filter, false removes it + * @ipv4: true is v4, false is v6 + * + * Returns 0 if the filters were successfully added or removed + **/ +static int i40e_add_del_fdir_tcp(struct i40e_vsi *vsi, + struct i40e_fdir_filter *fd_data, + bool add, + bool ipv4) +{ + struct i40e_pf *pf = vsi->back; + u8 *raw_packet; + int ret; + + raw_packet = kzalloc(I40E_FDIR_MAX_RAW_PACKET_SIZE, GFP_KERNEL); + if (!raw_packet) + return -ENOMEM; + + i40e_create_dummy_tcp_packet(raw_packet, ipv4, IPPROTO_TCP, fd_data); + if (ipv4) + ret = i40e_prepare_fdir_filter + (pf, fd_data, add, raw_packet, + I40E_TCPIP_DUMMY_PACKET_LEN, + LIBIE_FILTER_PCTYPE_NONF_IPV4_TCP); + else + ret = i40e_prepare_fdir_filter + (pf, fd_data, add, raw_packet, + I40E_TCPIP6_DUMMY_PACKET_LEN, + LIBIE_FILTER_PCTYPE_NONF_IPV6_TCP); + if (ret) { - dev_info(&pf->pdev->dev, - "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n", - fd_data->pctype, fd_data->fd_id, ret); - /* Free the packet buffer since it wasn't added to the ring */ kfree(raw_packet); - return -EOPNOTSUPP; - } else if (I40E_DEBUG_FD & pf->hw.debug_mask) { - if (add) - dev_info(&pf->pdev->dev, "Filter OK for PCTYPE %d loc = %d)\n", - fd_data->pctype, fd_data->fd_id); - else - dev_info(&pf->pdev->dev, - "Filter deleted for PCTYPE %d loc = %d\n", - fd_data->pctype, fd_data->fd_id); + return ret; } + i40e_change_filter_num(ipv4, add, &pf->fd_tcp4_filter_cnt, + &pf->fd_tcp6_filter_cnt); + if (add) { - pf->fd_tcp4_filter_cnt++; - if ((pf->flags & I40E_FLAG_FD_ATR_ENABLED) && + if (test_bit(I40E_FLAG_FD_ATR_ENA, pf->flags) && I40E_DEBUG_FD & pf->hw.debug_mask) dev_info(&pf->pdev->dev, "Forcing ATR off, sideband rules for TCP/IPv4 flow being applied\n"); - pf->flags |= I40E_FLAG_FD_ATR_AUTO_DISABLED; - } else { - pf->fd_tcp4_filter_cnt--; + set_bit(__I40E_FD_ATR_AUTO_DISABLED, pf->state); } - return 0; } -#define I40E_SCTPIP_DUMMY_PACKET_LEN 46 +#define I40E_SCTPIP_DUMMY_PACKET_LEN 46 +#define I40E_SCTPIP6_DUMMY_PACKET_LEN 66 /** - * i40e_add_del_fdir_sctpv4 - Add/Remove SCTPv4 Flow Director filters for + * i40e_add_del_fdir_sctp - Add/Remove SCTPv4 Flow Director filters for * a specific flow spec * @vsi: pointer to the targeted VSI * @fd_data: the flow director data required for the FDir descriptor * @add: true adds a filter, false removes it + * @ipv4: true is v4, false is v6 * * Returns 0 if the filters were successfully added or removed **/ -static int i40e_add_del_fdir_sctpv4(struct i40e_vsi *vsi, - struct i40e_fdir_filter *fd_data, - bool add) +static int i40e_add_del_fdir_sctp(struct i40e_vsi *vsi, + struct i40e_fdir_filter *fd_data, + bool add, + bool ipv4) { struct i40e_pf *pf = vsi->back; - struct sctphdr *sctp; - struct iphdr *ip; u8 *raw_packet; int ret; - /* Dummy packet */ - static char packet[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0x08, 0, - 0x45, 0, 0, 0x20, 0, 0, 0x40, 0, 0x40, 0x84, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; raw_packet = kzalloc(I40E_FDIR_MAX_RAW_PACKET_SIZE, GFP_KERNEL); if (!raw_packet) return -ENOMEM; - memcpy(raw_packet, packet, I40E_SCTPIP_DUMMY_PACKET_LEN); - ip = (struct iphdr *)(raw_packet + IP_HEADER_OFFSET); - sctp = (struct sctphdr *)(raw_packet + IP_HEADER_OFFSET - + sizeof(struct iphdr)); + i40e_create_dummy_sctp_packet(raw_packet, ipv4, IPPROTO_SCTP, fd_data); - ip->daddr = fd_data->dst_ip; - sctp->dest = fd_data->dst_port; - ip->saddr = fd_data->src_ip; - sctp->source = fd_data->src_port; - - if (fd_data->flex_filter) { - u8 *payload = raw_packet + I40E_SCTPIP_DUMMY_PACKET_LEN; - __be16 pattern = fd_data->flex_word; - u16 off = fd_data->flex_offset; - - *((__force __be16 *)(payload + off)) = pattern; - } + if (ipv4) + ret = i40e_prepare_fdir_filter + (pf, fd_data, add, raw_packet, + I40E_SCTPIP_DUMMY_PACKET_LEN, + LIBIE_FILTER_PCTYPE_NONF_IPV4_SCTP); + else + ret = i40e_prepare_fdir_filter + (pf, fd_data, add, raw_packet, + I40E_SCTPIP6_DUMMY_PACKET_LEN, + LIBIE_FILTER_PCTYPE_NONF_IPV6_SCTP); - fd_data->pctype = I40E_FILTER_PCTYPE_NONF_IPV4_SCTP; - ret = i40e_program_fdir_filter(fd_data, raw_packet, pf, add); if (ret) { - dev_info(&pf->pdev->dev, - "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n", - fd_data->pctype, fd_data->fd_id, ret); - /* Free the packet buffer since it wasn't added to the ring */ kfree(raw_packet); - return -EOPNOTSUPP; - } else if (I40E_DEBUG_FD & pf->hw.debug_mask) { - if (add) - dev_info(&pf->pdev->dev, - "Filter OK for PCTYPE %d loc = %d\n", - fd_data->pctype, fd_data->fd_id); - else - dev_info(&pf->pdev->dev, - "Filter deleted for PCTYPE %d loc = %d\n", - fd_data->pctype, fd_data->fd_id); + return ret; } - if (add) - pf->fd_sctp4_filter_cnt++; - else - pf->fd_sctp4_filter_cnt--; + i40e_change_filter_num(ipv4, add, &pf->fd_sctp4_filter_cnt, + &pf->fd_sctp6_filter_cnt); return 0; } -#define I40E_IP_DUMMY_PACKET_LEN 34 +#define I40E_IP_DUMMY_PACKET_LEN 34 +#define I40E_IP6_DUMMY_PACKET_LEN 54 /** - * i40e_add_del_fdir_ipv4 - Add/Remove IPv4 Flow Director filters for + * i40e_add_del_fdir_ip - Add/Remove IPv4 Flow Director filters for * a specific flow spec * @vsi: pointer to the targeted VSI * @fd_data: the flow director data required for the FDir descriptor * @add: true adds a filter, false removes it + * @ipv4: true is v4, false is v6 * * Returns 0 if the filters were successfully added or removed **/ -static int i40e_add_del_fdir_ipv4(struct i40e_vsi *vsi, - struct i40e_fdir_filter *fd_data, - bool add) +static int i40e_add_del_fdir_ip(struct i40e_vsi *vsi, + struct i40e_fdir_filter *fd_data, + bool add, + bool ipv4) { struct i40e_pf *pf = vsi->back; - struct iphdr *ip; + int payload_offset; u8 *raw_packet; + int iter_start; + int iter_end; int ret; int i; - static char packet[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0x08, 0, - 0x45, 0, 0, 0x14, 0, 0, 0x40, 0, 0x40, 0x10, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0}; - for (i = I40E_FILTER_PCTYPE_NONF_IPV4_OTHER; - i <= I40E_FILTER_PCTYPE_FRAG_IPV4; i++) { + if (ipv4) { + iter_start = LIBIE_FILTER_PCTYPE_NONF_IPV4_OTHER; + iter_end = LIBIE_FILTER_PCTYPE_FRAG_IPV4; + } else { + iter_start = LIBIE_FILTER_PCTYPE_NONF_IPV6_OTHER; + iter_end = LIBIE_FILTER_PCTYPE_FRAG_IPV6; + } + + for (i = iter_start; i <= iter_end; i++) { raw_packet = kzalloc(I40E_FDIR_MAX_RAW_PACKET_SIZE, GFP_KERNEL); if (!raw_packet) return -ENOMEM; - memcpy(raw_packet, packet, I40E_IP_DUMMY_PACKET_LEN); - ip = (struct iphdr *)(raw_packet + IP_HEADER_OFFSET); - - ip->saddr = fd_data->src_ip; - ip->daddr = fd_data->dst_ip; - ip->protocol = 0; - if (fd_data->flex_filter) { - u8 *payload = raw_packet + I40E_IP_DUMMY_PACKET_LEN; - __be16 pattern = fd_data->flex_word; - u16 off = fd_data->flex_offset; + /* IPv6 no header option differs from IPv4 */ + (void)i40e_create_dummy_packet + (raw_packet, ipv4, (ipv4) ? IPPROTO_IP : IPPROTO_NONE, + fd_data); - *((__force __be16 *)(payload + off)) = pattern; - } - - fd_data->pctype = i; - ret = i40e_program_fdir_filter(fd_data, raw_packet, pf, add); - if (ret) { - dev_info(&pf->pdev->dev, - "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n", - fd_data->pctype, fd_data->fd_id, ret); - /* The packet buffer wasn't added to the ring so we - * need to free it now. - */ - kfree(raw_packet); - return -EOPNOTSUPP; - } else if (I40E_DEBUG_FD & pf->hw.debug_mask) { - if (add) - dev_info(&pf->pdev->dev, - "Filter OK for PCTYPE %d loc = %d\n", - fd_data->pctype, fd_data->fd_id); - else - dev_info(&pf->pdev->dev, - "Filter deleted for PCTYPE %d loc = %d\n", - fd_data->pctype, fd_data->fd_id); - } + payload_offset = (ipv4) ? I40E_IP_DUMMY_PACKET_LEN : + I40E_IP6_DUMMY_PACKET_LEN; + ret = i40e_prepare_fdir_filter(pf, fd_data, add, raw_packet, + payload_offset, i); + if (ret) + goto err; } - if (add) - pf->fd_ip4_filter_cnt++; - else - pf->fd_ip4_filter_cnt--; + i40e_change_filter_num(ipv4, add, &pf->fd_ip4_filter_cnt, + &pf->fd_ip6_filter_cnt); return 0; +err: + kfree(raw_packet); + return ret; } /** * i40e_add_del_fdir - Build raw packets to add/del fdir filter * @vsi: pointer to the targeted VSI - * @cmd: command to get or set RX flow classification rules + * @input: filter to add or delete * @add: true adds a filter, false removes it * **/ int i40e_add_del_fdir(struct i40e_vsi *vsi, struct i40e_fdir_filter *input, bool add) { + enum ip_ver { ipv6 = 0, ipv4 = 1 }; struct i40e_pf *pf = vsi->back; int ret; switch (input->flow_type & ~FLOW_EXT) { case TCP_V4_FLOW: - ret = i40e_add_del_fdir_tcpv4(vsi, input, add); + ret = i40e_add_del_fdir_tcp(vsi, input, add, ipv4); break; case UDP_V4_FLOW: - ret = i40e_add_del_fdir_udpv4(vsi, input, add); + ret = i40e_add_del_fdir_udp(vsi, input, add, ipv4); break; case SCTP_V4_FLOW: - ret = i40e_add_del_fdir_sctpv4(vsi, input, add); + ret = i40e_add_del_fdir_sctp(vsi, input, add, ipv4); + break; + case TCP_V6_FLOW: + ret = i40e_add_del_fdir_tcp(vsi, input, add, ipv6); + break; + case UDP_V6_FLOW: + ret = i40e_add_del_fdir_udp(vsi, input, add, ipv6); + break; + case SCTP_V6_FLOW: + ret = i40e_add_del_fdir_sctp(vsi, input, add, ipv6); break; case IP_USER_FLOW: - switch (input->ip4_proto) { + switch (input->ipl4_proto) { case IPPROTO_TCP: - ret = i40e_add_del_fdir_tcpv4(vsi, input, add); + ret = i40e_add_del_fdir_tcp(vsi, input, add, ipv4); break; case IPPROTO_UDP: - ret = i40e_add_del_fdir_udpv4(vsi, input, add); + ret = i40e_add_del_fdir_udp(vsi, input, add, ipv4); break; case IPPROTO_SCTP: - ret = i40e_add_del_fdir_sctpv4(vsi, input, add); + ret = i40e_add_del_fdir_sctp(vsi, input, add, ipv4); break; case IPPROTO_IP: - ret = i40e_add_del_fdir_ipv4(vsi, input, add); + ret = i40e_add_del_fdir_ip(vsi, input, add, ipv4); break; default: /* We cannot support masking based on protocol */ dev_info(&pf->pdev->dev, "Unsupported IPv4 protocol 0x%02x\n", - input->ip4_proto); + input->ipl4_proto); + return -EINVAL; + } + break; + case IPV6_USER_FLOW: + switch (input->ipl4_proto) { + case IPPROTO_TCP: + ret = i40e_add_del_fdir_tcp(vsi, input, add, ipv6); + break; + case IPPROTO_UDP: + ret = i40e_add_del_fdir_udp(vsi, input, add, ipv6); + break; + case IPPROTO_SCTP: + ret = i40e_add_del_fdir_sctp(vsi, input, add, ipv6); + break; + case IPPROTO_IP: + ret = i40e_add_del_fdir_ip(vsi, input, add, ipv6); + break; + default: + /* We cannot support masking based on protocol */ + dev_info(&pf->pdev->dev, "Unsupported IPv6 protocol 0x%02x\n", + input->ipl4_proto); return -EINVAL; } break; @@ -552,28 +671,28 @@ int i40e_add_del_fdir(struct i40e_vsi *vsi, /** * i40e_fd_handle_status - check the Programming Status for FD * @rx_ring: the Rx ring for this descriptor - * @rx_desc: the Rx descriptor for programming Status, not a packet descriptor. + * @qword0_raw: qword0 + * @qword1: qword1 after le_to_cpu * @prog_id: the id originally used for programming * * This is used to verify if the FD programming or invalidation * requested by SW to the HW is successful or not and take actions accordingly. **/ -static void i40e_fd_handle_status(struct i40e_ring *rx_ring, - union i40e_rx_desc *rx_desc, u8 prog_id) +static void i40e_fd_handle_status(struct i40e_ring *rx_ring, u64 qword0_raw, + u64 qword1, u8 prog_id) { struct i40e_pf *pf = rx_ring->vsi->back; struct pci_dev *pdev = pf->pdev; + struct i40e_16b_rx_wb_qw0 *qw0; u32 fcnt_prog, fcnt_avail; u32 error; - u64 qw; - qw = le64_to_cpu(rx_desc->wb.qword1.status_error_len); - error = (qw & I40E_RX_PROG_STATUS_DESC_QW1_ERROR_MASK) >> - I40E_RX_PROG_STATUS_DESC_QW1_ERROR_SHIFT; + qw0 = (struct i40e_16b_rx_wb_qw0 *)&qword0_raw; + error = FIELD_GET(I40E_RX_PROG_STATUS_DESC_QW1_ERROR_MASK, qword1); if (error == BIT(I40E_RX_PROG_STATUS_DESC_FD_TBL_FULL_SHIFT)) { - pf->fd_inv = le32_to_cpu(rx_desc->wb.qword0.hi_dword.fd_id); - if ((rx_desc->wb.qword0.hi_dword.fd_id != 0) || + pf->fd_inv = le32_to_cpu(qw0->hi_dword.fd_id); + if (qw0->hi_dword.fd_id != 0 || (I40E_DEBUG_FD & pf->hw.debug_mask)) dev_warn(&pdev->dev, "ntuple filter loc = %d, could not be added\n", pf->fd_inv); @@ -591,9 +710,15 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring, /* store the current atr filter count */ pf->fd_atr_cnt = i40e_get_current_atr_cnt(pf); - if ((rx_desc->wb.qword0.hi_dword.fd_id == 0) && - pf->flags & I40E_FLAG_FD_SB_AUTO_DISABLED) { - pf->flags |= I40E_FLAG_FD_ATR_AUTO_DISABLED; + if (qw0->hi_dword.fd_id == 0 && + test_bit(__I40E_FD_SB_AUTO_DISABLED, pf->state)) { + /* These set_bit() calls aren't atomic with the + * test_bit() here, but worse case we potentially + * disable ATR and queue a flush right after SB + * support is re-enabled. That shouldn't cause an + * issue in practice + */ + set_bit(__I40E_FD_ATR_AUTO_DISABLED, pf->state); set_bit(__I40E_FD_FLUSH_REQUESTED, pf->state); } @@ -605,17 +730,16 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring, * FD ATR/SB and then re-enable it when there is room. */ if (fcnt_prog >= (fcnt_avail - I40E_FDIR_BUFFER_FULL_MARGIN)) { - if ((pf->flags & I40E_FLAG_FD_SB_ENABLED) && - !(pf->flags & I40E_FLAG_FD_SB_AUTO_DISABLED)) { - pf->flags |= I40E_FLAG_FD_SB_AUTO_DISABLED; + if (test_bit(I40E_FLAG_FD_SB_ENA, pf->flags) && + !test_and_set_bit(__I40E_FD_SB_AUTO_DISABLED, + pf->state)) if (I40E_DEBUG_FD & pf->hw.debug_mask) dev_warn(&pdev->dev, "FD filter space full, new ntuple rules will not be added\n"); - } } } else if (error == BIT(I40E_RX_PROG_STATUS_DESC_NO_FD_ENTRY_SHIFT)) { if (I40E_DEBUG_FD & pf->hw.debug_mask) dev_info(&pdev->dev, "ntuple filter fd_id = %d, could not be removed\n", - rx_desc->wb.qword0.hi_dword.fd_id); + qw0->hi_dword.fd_id); } } @@ -631,7 +755,7 @@ static void i40e_unmap_and_free_tx_resource(struct i40e_ring *ring, if (tx_buffer->tx_flags & I40E_TX_FLAGS_FD_SB) kfree(tx_buffer->raw_buf); else if (ring_is_xdp(ring)) - page_frag_free(tx_buffer->raw_buf); + xdp_return_frame(tx_buffer->xdpf); else dev_kfree_skb_any(tx_buffer->skb); if (dma_unmap_len(tx_buffer, len)) @@ -661,13 +785,18 @@ void i40e_clean_tx_ring(struct i40e_ring *tx_ring) unsigned long bi_size; u16 i; - /* ring already cleared, nothing to do */ - if (!tx_ring->tx_bi) - return; + if (ring_is_xdp(tx_ring) && tx_ring->xsk_pool) { + i40e_xsk_clean_tx_ring(tx_ring); + } else { + /* ring already cleared, nothing to do */ + if (!tx_ring->tx_bi) + return; - /* Free all the Tx ring sk_buffs */ - for (i = 0; i < tx_ring->count; i++) - i40e_unmap_and_free_tx_resource(tx_ring, &tx_ring->tx_bi[i]); + /* Free all the Tx ring sk_buffs */ + for (i = 0; i < tx_ring->count; i++) + i40e_unmap_and_free_tx_resource(tx_ring, + &tx_ring->tx_bi[i]); + } bi_size = sizeof(struct i40e_tx_buffer) * tx_ring->count; memset(tx_ring->tx_bi, 0, bi_size); @@ -706,17 +835,23 @@ void i40e_free_tx_resources(struct i40e_ring *tx_ring) /** * i40e_get_tx_pending - how many tx descriptors not processed - * @tx_ring: the ring of descriptors + * @ring: the ring of descriptors + * @in_sw: use SW variables * * Since there is no access to the ring head register * in XL710, we need to use our local copies **/ -u32 i40e_get_tx_pending(struct i40e_ring *ring) +u32 i40e_get_tx_pending(struct i40e_ring *ring, bool in_sw) { u32 head, tail; - head = i40e_get_head(ring); - tail = readl(ring->tail); + if (!in_sw) { + head = i40e_get_head(ring); + tail = readl(ring->tail); + } else { + head = ring->next_to_clean; + tail = ring->next_to_use; + } if (head != tail) return (head < tail) ? @@ -725,20 +860,75 @@ u32 i40e_get_tx_pending(struct i40e_ring *ring) return 0; } -#define WB_STRIDE 4 +/** + * i40e_detect_recover_hung - Function to detect and recover hung_queues + * @pf: pointer to PF struct + * + * LAN VSI has netdev and netdev has TX queues. This function is to check + * each of those TX queues if they are hung, trigger recovery by issuing + * SW interrupt. + **/ +void i40e_detect_recover_hung(struct i40e_pf *pf) +{ + struct i40e_vsi *vsi = i40e_pf_get_main_vsi(pf); + struct i40e_ring *tx_ring = NULL; + struct net_device *netdev; + unsigned int i; + int packets; + + if (!vsi) + return; + + if (test_bit(__I40E_VSI_DOWN, vsi->state)) + return; + + netdev = vsi->netdev; + if (!netdev) + return; + + if (!netif_carrier_ok(netdev)) + return; + + for (i = 0; i < vsi->num_queue_pairs; i++) { + tx_ring = vsi->tx_rings[i]; + if (tx_ring && tx_ring->desc) { + /* If packet counter has not changed the queue is + * likely stalled, so force an interrupt for this + * queue. + * + * prev_pkt_ctr would be negative if there was no + * pending work. + */ + packets = tx_ring->stats.packets & INT_MAX; + if (tx_ring->tx_stats.prev_pkt_ctr == packets) { + i40e_force_wb(vsi, tx_ring->q_vector); + continue; + } + + /* Memory barrier between read of packet count and call + * to i40e_get_tx_pending() + */ + smp_rmb(); + tx_ring->tx_stats.prev_pkt_ctr = + i40e_get_tx_pending(tx_ring, true) ? packets : -1; + } + } +} /** * i40e_clean_tx_irq - Reclaim resources after transmit completes * @vsi: the VSI we care about * @tx_ring: Tx ring to clean * @napi_budget: Used to determine if we are in netpoll + * @tx_cleaned: Out parameter set to the number of TXes cleaned * * Returns true if there's any budget left (e.g. the clean is finished) **/ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi, - struct i40e_ring *tx_ring, int napi_budget) + struct i40e_ring *tx_ring, int napi_budget, + unsigned int *tx_cleaned) { - u16 i = tx_ring->next_to_clean; + int i = tx_ring->next_to_clean; struct i40e_tx_buffer *tx_buf; struct i40e_tx_desc *tx_head; struct i40e_tx_desc *tx_desc; @@ -758,9 +948,6 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi, if (!eop_desc) break; - /* prevent any other reads prior to eop_desc */ - read_barrier_depends(); - i40e_trace(clean_tx_irq, tx_ring, tx_desc, tx_buf); /* we have caught up to head, no work left to do */ if (tx_head == tx_desc) @@ -775,7 +962,7 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi, /* free the skb/XDP data */ if (ring_is_xdp(tx_ring)) - page_frag_free(tx_buf->raw_buf); + xdp_return_frame(tx_buf->xdpf); else napi_consume_skb(tx_buf->skb, napi_budget); @@ -831,27 +1018,8 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi, i += tx_ring->count; tx_ring->next_to_clean = i; - u64_stats_update_begin(&tx_ring->syncp); - tx_ring->stats.bytes += total_bytes; - tx_ring->stats.packets += total_packets; - u64_stats_update_end(&tx_ring->syncp); - tx_ring->q_vector->tx.total_bytes += total_bytes; - tx_ring->q_vector->tx.total_packets += total_packets; - - if (tx_ring->flags & I40E_TXR_FLAGS_WB_ON_ITR) { - /* check to see if there are < 4 descriptors - * waiting to be written back, then kick the hardware to force - * them to be written back in case we stay in NAPI. - * In this mode on X722 we do not enable Interrupt. - */ - unsigned int j = i40e_get_tx_pending(tx_ring); - - if (budget && - ((j / WB_STRIDE) == 0) && (j > 0) && - !test_bit(__I40E_VSI_DOWN, vsi->state) && - (I40E_DESC_UNUSED(tx_ring) != tx_ring->count)) - tx_ring->arm_wb = true; - } + i40e_update_tx_stats(tx_ring, total_packets, total_bytes); + i40e_arm_wb(tx_ring, vsi, budget); if (ring_is_xdp(tx_ring)) return !!budget; @@ -860,7 +1028,7 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi, netdev_tx_completed_queue(txring_txq(tx_ring), total_packets, total_bytes); -#define TX_WAKE_THRESHOLD (DESC_NEEDED * 2) +#define TX_WAKE_THRESHOLD ((s16)(DESC_NEEDED * 2)) if (unlikely(total_packets && netif_carrier_ok(tx_ring->netdev) && (I40E_DESC_UNUSED(tx_ring) >= TX_WAKE_THRESHOLD))) { /* Make sure that anybody stopping the queue after this @@ -876,6 +1044,7 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi, } } + *tx_cleaned = total_packets; return !!budget; } @@ -897,12 +1066,12 @@ static void i40e_enable_wb_on_itr(struct i40e_vsi *vsi, if (q_vector->arm_wb_state) return; - if (vsi->back->flags & I40E_FLAG_MSIX_ENABLED) { + if (test_bit(I40E_FLAG_MSIX_ENA, vsi->back->flags)) { val = I40E_PFINT_DYN_CTLN_WB_ON_ITR_MASK | I40E_PFINT_DYN_CTLN_ITR_INDX_MASK; /* set noitr */ wr32(&vsi->back->hw, - I40E_PFINT_DYN_CTLN(q_vector->v_idx + vsi->base_vector - 1), + I40E_PFINT_DYN_CTLN(q_vector->reg_idx), val); } else { val = I40E_PFINT_DYN_CTL0_WB_ON_ITR_MASK | @@ -921,7 +1090,7 @@ static void i40e_enable_wb_on_itr(struct i40e_vsi *vsi, **/ void i40e_force_wb(struct i40e_vsi *vsi, struct i40e_q_vector *q_vector) { - if (vsi->back->flags & I40E_FLAG_MSIX_ENABLED) { + if (test_bit(I40E_FLAG_MSIX_ENA, vsi->back->flags)) { u32 val = I40E_PFINT_DYN_CTLN_INTENA_MASK | I40E_PFINT_DYN_CTLN_ITR_INDX_MASK | /* set noitr */ I40E_PFINT_DYN_CTLN_SWINT_TRIG_MASK | @@ -929,8 +1098,7 @@ void i40e_force_wb(struct i40e_vsi *vsi, struct i40e_q_vector *q_vector) /* allow 00 to be written to the index */ wr32(&vsi->back->hw, - I40E_PFINT_DYN_CTLN(q_vector->v_idx + - vsi->base_vector - 1), val); + I40E_PFINT_DYN_CTLN(q_vector->reg_idx), val); } else { u32 val = I40E_PFINT_DYN_CTL0_INTENA_MASK | I40E_PFINT_DYN_CTL0_ITR_INDX_MASK | /* set noitr */ @@ -942,154 +1110,298 @@ void i40e_force_wb(struct i40e_vsi *vsi, struct i40e_q_vector *q_vector) } } +static inline bool i40e_container_is_rx(struct i40e_q_vector *q_vector, + struct i40e_ring_container *rc) +{ + return &q_vector->rx == rc; +} + +static inline unsigned int i40e_itr_divisor(struct i40e_q_vector *q_vector) +{ + unsigned int divisor; + + switch (q_vector->vsi->back->hw.phy.link_info.link_speed) { + case I40E_LINK_SPEED_40GB: + divisor = I40E_ITR_ADAPTIVE_MIN_INC * 1024; + break; + case I40E_LINK_SPEED_25GB: + case I40E_LINK_SPEED_20GB: + divisor = I40E_ITR_ADAPTIVE_MIN_INC * 512; + break; + default: + case I40E_LINK_SPEED_10GB: + divisor = I40E_ITR_ADAPTIVE_MIN_INC * 256; + break; + case I40E_LINK_SPEED_1GB: + case I40E_LINK_SPEED_100MB: + divisor = I40E_ITR_ADAPTIVE_MIN_INC * 32; + break; + } + + return divisor; +} + /** - * i40e_set_new_dynamic_itr - Find new ITR level + * i40e_update_itr - update the dynamic ITR value based on statistics + * @q_vector: structure containing interrupt and ring information * @rc: structure containing ring performance data * - * Returns true if ITR changed, false if not - * - * Stores a new ITR value based on packets and byte counts during - * the last interrupt. The advantage of per interrupt computation - * is faster updates and more accurate ITR for the current traffic - * pattern. Constants in this function were computed based on - * theoretical maximum wire speed and thresholds were set based on - * testing data as well as attempting to minimize response time + * Stores a new ITR value based on packets and byte + * counts during the last interrupt. The advantage of per interrupt + * computation is faster updates and more accurate ITR for the current + * traffic pattern. Constants in this function were computed + * based on theoretical maximum wire speed and thresholds were set based + * on testing data as well as attempting to minimize response time * while increasing bulk throughput. **/ -static bool i40e_set_new_dynamic_itr(struct i40e_ring_container *rc) +static void i40e_update_itr(struct i40e_q_vector *q_vector, + struct i40e_ring_container *rc) { - enum i40e_latency_range new_latency_range = rc->latency_range; - struct i40e_q_vector *qv = rc->ring->q_vector; - u32 new_itr = rc->itr; - int bytes_per_int; - int usecs; + unsigned int avg_wire_size, packets, bytes, itr; + unsigned long next_update = jiffies; - if (rc->total_packets == 0 || !rc->itr) - return false; + /* If we don't have any rings just leave ourselves set for maximum + * possible latency so we take ourselves out of the equation. + */ + if (!rc->ring || !ITR_IS_DYNAMIC(rc->ring->itr_setting)) + return; - /* simple throttlerate management - * 0-10MB/s lowest (50000 ints/s) - * 10-20MB/s low (20000 ints/s) - * 20-1249MB/s bulk (18000 ints/s) - * > 40000 Rx packets per second (8000 ints/s) + /* For Rx we want to push the delay up and default to low latency. + * for Tx we want to pull the delay down and default to high latency. + */ + itr = i40e_container_is_rx(q_vector, rc) ? + I40E_ITR_ADAPTIVE_MIN_USECS | I40E_ITR_ADAPTIVE_LATENCY : + I40E_ITR_ADAPTIVE_MAX_USECS | I40E_ITR_ADAPTIVE_LATENCY; + + /* If we didn't update within up to 1 - 2 jiffies we can assume + * that either packets are coming in so slow there hasn't been + * any work, or that there is so much work that NAPI is dealing + * with interrupt moderation and we don't need to do anything. + */ + if (time_after(next_update, rc->next_update)) + goto clear_counts; + + /* If itr_countdown is set it means we programmed an ITR within + * the last 4 interrupt cycles. This has a side effect of us + * potentially firing an early interrupt. In order to work around + * this we need to throw out any data received for a few + * interrupts following the update. + */ + if (q_vector->itr_countdown) { + itr = rc->target_itr; + goto clear_counts; + } + + packets = rc->total_packets; + bytes = rc->total_bytes; + + if (i40e_container_is_rx(q_vector, rc)) { + /* If Rx there are 1 to 4 packets and bytes are less than + * 9000 assume insufficient data to use bulk rate limiting + * approach unless Tx is already in bulk rate limiting. We + * are likely latency driven. + */ + if (packets && packets < 4 && bytes < 9000 && + (q_vector->tx.target_itr & I40E_ITR_ADAPTIVE_LATENCY)) { + itr = I40E_ITR_ADAPTIVE_LATENCY; + goto adjust_by_size; + } + } else if (packets < 4) { + /* If we have Tx and Rx ITR maxed and Tx ITR is running in + * bulk mode and we are receiving 4 or fewer packets just + * reset the ITR_ADAPTIVE_LATENCY bit for latency mode so + * that the Rx can relax. + */ + if (rc->target_itr == I40E_ITR_ADAPTIVE_MAX_USECS && + (q_vector->rx.target_itr & I40E_ITR_MASK) == + I40E_ITR_ADAPTIVE_MAX_USECS) + goto clear_counts; + } else if (packets > 32) { + /* If we have processed over 32 packets in a single interrupt + * for Tx assume we need to switch over to "bulk" mode. + */ + rc->target_itr &= ~I40E_ITR_ADAPTIVE_LATENCY; + } + + /* We have no packets to actually measure against. This means + * either one of the other queues on this vector is active or + * we are a Tx queue doing TSO with too high of an interrupt rate. * - * The math works out because the divisor is in 10^(-6) which - * turns the bytes/us input value into MB/s values, but - * make sure to use usecs, as the register values written - * are in 2 usec increments in the ITR registers, and make sure - * to use the smoothed values that the countdown timer gives us. + * Between 4 and 56 we can assume that our current interrupt delay + * is only slightly too low. As such we should increase it by a small + * fixed amount. */ - usecs = (rc->itr << 1) * ITR_COUNTDOWN_START; - bytes_per_int = rc->total_bytes / usecs; + if (packets < 56) { + itr = rc->target_itr + I40E_ITR_ADAPTIVE_MIN_INC; + if ((itr & I40E_ITR_MASK) > I40E_ITR_ADAPTIVE_MAX_USECS) { + itr &= I40E_ITR_ADAPTIVE_LATENCY; + itr += I40E_ITR_ADAPTIVE_MAX_USECS; + } + goto clear_counts; + } - switch (new_latency_range) { - case I40E_LOWEST_LATENCY: - if (bytes_per_int > 10) - new_latency_range = I40E_LOW_LATENCY; - break; - case I40E_LOW_LATENCY: - if (bytes_per_int > 20) - new_latency_range = I40E_BULK_LATENCY; - else if (bytes_per_int <= 10) - new_latency_range = I40E_LOWEST_LATENCY; - break; - case I40E_BULK_LATENCY: - case I40E_ULTRA_LATENCY: - default: - if (bytes_per_int <= 20) - new_latency_range = I40E_LOW_LATENCY; - break; + if (packets <= 256) { + itr = min(q_vector->tx.current_itr, q_vector->rx.current_itr); + itr &= I40E_ITR_MASK; + + /* Between 56 and 112 is our "goldilocks" zone where we are + * working out "just right". Just report that our current + * ITR is good for us. + */ + if (packets <= 112) + goto clear_counts; + + /* If packet count is 128 or greater we are likely looking + * at a slight overrun of the delay we want. Try halving + * our delay to see if that will cut the number of packets + * in half per interrupt. + */ + itr /= 2; + itr &= I40E_ITR_MASK; + if (itr < I40E_ITR_ADAPTIVE_MIN_USECS) + itr = I40E_ITR_ADAPTIVE_MIN_USECS; + + goto clear_counts; } - /* this is to adjust RX more aggressively when streaming small - * packets. The value of 40000 was picked as it is just beyond - * what the hardware can receive per second if in low latency - * mode. + /* The paths below assume we are dealing with a bulk ITR since + * number of packets is greater than 256. We are just going to have + * to compute a value and try to bring the count under control, + * though for smaller packet sizes there isn't much we can do as + * NAPI polling will likely be kicking in sooner rather than later. */ -#define RX_ULTRA_PACKET_RATE 40000 + itr = I40E_ITR_ADAPTIVE_BULK; - if ((((rc->total_packets * 1000000) / usecs) > RX_ULTRA_PACKET_RATE) && - (&qv->rx == rc)) - new_latency_range = I40E_ULTRA_LATENCY; +adjust_by_size: + /* If packet counts are 256 or greater we can assume we have a gross + * overestimation of what the rate should be. Instead of trying to fine + * tune it just use the formula below to try and dial in an exact value + * give the current packet size of the frame. + */ + avg_wire_size = bytes / packets; - rc->latency_range = new_latency_range; + /* The following is a crude approximation of: + * wmem_default / (size + overhead) = desired_pkts_per_int + * rate / bits_per_byte / (size + ethernet overhead) = pkt_rate + * (desired_pkt_rate / pkt_rate) * usecs_per_sec = ITR value + * + * Assuming wmem_default is 212992 and overhead is 640 bytes per + * packet, (256 skb, 64 headroom, 320 shared info), we can reduce the + * formula down to + * + * (170 * (size + 24)) / (size + 640) = ITR + * + * We first do some math on the packet size and then finally bitshift + * by 8 after rounding up. We also have to account for PCIe link speed + * difference as ITR scales based on this. + */ + if (avg_wire_size <= 60) { + /* Start at 250k ints/sec */ + avg_wire_size = 4096; + } else if (avg_wire_size <= 380) { + /* 250K ints/sec to 60K ints/sec */ + avg_wire_size *= 40; + avg_wire_size += 1696; + } else if (avg_wire_size <= 1084) { + /* 60K ints/sec to 36K ints/sec */ + avg_wire_size *= 15; + avg_wire_size += 11452; + } else if (avg_wire_size <= 1980) { + /* 36K ints/sec to 30K ints/sec */ + avg_wire_size *= 5; + avg_wire_size += 22420; + } else { + /* plateau at a limit of 30K ints/sec */ + avg_wire_size = 32256; + } - switch (new_latency_range) { - case I40E_LOWEST_LATENCY: - new_itr = I40E_ITR_50K; - break; - case I40E_LOW_LATENCY: - new_itr = I40E_ITR_20K; - break; - case I40E_BULK_LATENCY: - new_itr = I40E_ITR_18K; - break; - case I40E_ULTRA_LATENCY: - new_itr = I40E_ITR_8K; - break; - default: - break; + /* If we are in low latency mode halve our delay which doubles the + * rate to somewhere between 100K to 16K ints/sec + */ + if (itr & I40E_ITR_ADAPTIVE_LATENCY) + avg_wire_size /= 2; + + /* Resultant value is 256 times larger than it needs to be. This + * gives us room to adjust the value as needed to either increase + * or decrease the value based on link speeds of 10G, 2.5G, 1G, etc. + * + * Use addition as we have already recorded the new latency flag + * for the ITR value. + */ + itr += DIV_ROUND_UP(avg_wire_size, i40e_itr_divisor(q_vector)) * + I40E_ITR_ADAPTIVE_MIN_INC; + + if ((itr & I40E_ITR_MASK) > I40E_ITR_ADAPTIVE_MAX_USECS) { + itr &= I40E_ITR_ADAPTIVE_LATENCY; + itr += I40E_ITR_ADAPTIVE_MAX_USECS; } +clear_counts: + /* write back value */ + rc->target_itr = itr; + + /* next update should occur within next jiffy */ + rc->next_update = next_update + 1; + rc->total_bytes = 0; rc->total_packets = 0; +} - if (new_itr != rc->itr) { - rc->itr = new_itr; - return true; - } - - return false; +static struct i40e_rx_buffer *i40e_rx_bi(struct i40e_ring *rx_ring, u32 idx) +{ + return &rx_ring->rx_bi[idx]; } /** - * i40e_rx_is_programming_status - check for programming status descriptor - * @qw: qword representing status_error_len in CPU ordering + * i40e_reuse_rx_page - page flip buffer and store it back on the ring + * @rx_ring: rx descriptor ring to store buffers on + * @old_buff: donor buffer to have page reused * - * The value of in the descriptor length field indicate if this - * is a programming status descriptor for flow director or FCoE - * by the value of I40E_RX_PROG_STATUS_DESC_LENGTH, otherwise - * it is a packet descriptor. + * Synchronizes page for reuse by the adapter **/ -static inline bool i40e_rx_is_programming_status(u64 qw) +static void i40e_reuse_rx_page(struct i40e_ring *rx_ring, + struct i40e_rx_buffer *old_buff) { - /* The Rx filter programming status and SPH bit occupy the same - * spot in the descriptor. Since we don't support packet split we - * can just reuse the bit as an indication that this is a - * programming status descriptor. - */ - return qw & I40E_RXD_QW1_LENGTH_SPH_MASK; + struct i40e_rx_buffer *new_buff; + u16 nta = rx_ring->next_to_alloc; + + new_buff = i40e_rx_bi(rx_ring, nta); + + /* update, and store next to alloc */ + nta++; + rx_ring->next_to_alloc = (nta < rx_ring->count) ? nta : 0; + + /* transfer page from old buffer to new buffer */ + new_buff->dma = old_buff->dma; + new_buff->page = old_buff->page; + new_buff->page_offset = old_buff->page_offset; + new_buff->pagecnt_bias = old_buff->pagecnt_bias; + + /* clear contents of buffer_info */ + old_buff->page = NULL; } /** * i40e_clean_programming_status - clean the programming status descriptor * @rx_ring: the rx ring that has this descriptor - * @rx_desc: the rx descriptor written back by HW - * @qw: qword representing status_error_len in CPU ordering + * @qword0_raw: qword0 + * @qword1: qword1 representing status_error_len in CPU ordering * * Flow director should handle FD_FILTER_STATUS to check its filter programming * status being successful or not and take actions accordingly. FCoE should * handle its context/filter programming/invalidation status and take actions. * + * Returns an i40e_rx_buffer to reuse if the cleanup occurred, otherwise NULL. **/ -static void i40e_clean_programming_status(struct i40e_ring *rx_ring, - union i40e_rx_desc *rx_desc, - u64 qw) +void i40e_clean_programming_status(struct i40e_ring *rx_ring, u64 qword0_raw, + u64 qword1) { - u32 ntc = rx_ring->next_to_clean + 1; u8 id; - /* fetch, update, and store next to clean */ - ntc = (ntc < rx_ring->count) ? ntc : 0; - rx_ring->next_to_clean = ntc; - - prefetch(I40E_RX_DESC(rx_ring, ntc)); - - id = (qw & I40E_RX_PROG_STATUS_DESC_QW1_PROGID_MASK) >> - I40E_RX_PROG_STATUS_DESC_QW1_PROGID_SHIFT; + id = FIELD_GET(I40E_RX_PROG_STATUS_DESC_QW1_PROGID_MASK, qword1); if (id == I40E_RX_PROG_STATUS_DESC_FD_FILTER_STATUS) - i40e_fd_handle_status(rx_ring, rx_desc, id); + i40e_fd_handle_status(rx_ring, qword0_raw, qword1, id); } /** @@ -1113,6 +1425,8 @@ int i40e_setup_tx_descriptors(struct i40e_ring *tx_ring) if (!tx_ring->tx_bi) goto err; + u64_stats_init(&tx_ring->syncp); + /* round up to nearest 4K */ tx_ring->size = tx_ring->count * sizeof(struct i40e_tx_desc); /* add u32 for head writeback, align after this takes care of @@ -1130,6 +1444,7 @@ int i40e_setup_tx_descriptors(struct i40e_ring *tx_ring) tx_ring->next_to_use = 0; tx_ring->next_to_clean = 0; + tx_ring->tx_stats.prev_pkt_ctr = -1; return 0; err: @@ -1138,27 +1453,31 @@ err: return -ENOMEM; } +static void i40e_clear_rx_bi(struct i40e_ring *rx_ring) +{ + memset(rx_ring->rx_bi, 0, sizeof(*rx_ring->rx_bi) * rx_ring->count); +} + /** * i40e_clean_rx_ring - Free Rx buffers * @rx_ring: ring to be cleaned **/ void i40e_clean_rx_ring(struct i40e_ring *rx_ring) { - unsigned long bi_size; u16 i; /* ring already cleared, nothing to do */ if (!rx_ring->rx_bi) return; - if (rx_ring->skb) { - dev_kfree_skb(rx_ring->skb); - rx_ring->skb = NULL; + if (rx_ring->xsk_pool) { + i40e_xsk_clean_rx_ring(rx_ring); + goto skip_free; } /* Free all the Rx ring sk_buffs */ for (i = 0; i < rx_ring->count; i++) { - struct i40e_rx_buffer *rx_bi = &rx_ring->rx_bi[i]; + struct i40e_rx_buffer *rx_bi = i40e_rx_bi(rx_ring, i); if (!rx_bi->page) continue; @@ -1184,14 +1503,18 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring) rx_bi->page_offset = 0; } - bi_size = sizeof(struct i40e_rx_buffer) * rx_ring->count; - memset(rx_ring->rx_bi, 0, bi_size); +skip_free: + if (rx_ring->xsk_pool) + i40e_clear_rx_bi_zc(rx_ring); + else + i40e_clear_rx_bi(rx_ring); /* Zero out the descriptor ring */ memset(rx_ring->desc, 0, rx_ring->size); rx_ring->next_to_alloc = 0; rx_ring->next_to_clean = 0; + rx_ring->next_to_process = 0; rx_ring->next_to_use = 0; } @@ -1204,6 +1527,8 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring) void i40e_free_rx_resources(struct i40e_ring *rx_ring) { i40e_clean_rx_ring(rx_ring); + if (rx_ring->vsi->type == I40E_VSI_MAIN) + xdp_rxq_info_unreg(&rx_ring->xdp_rxq); rx_ring->xdp_prog = NULL; kfree(rx_ring->rx_bi); rx_ring->rx_bi = NULL; @@ -1224,19 +1549,11 @@ void i40e_free_rx_resources(struct i40e_ring *rx_ring) int i40e_setup_rx_descriptors(struct i40e_ring *rx_ring) { struct device *dev = rx_ring->dev; - int bi_size; - - /* warn if we are about to overwrite the pointer */ - WARN_ON(rx_ring->rx_bi); - bi_size = sizeof(struct i40e_rx_buffer) * rx_ring->count; - rx_ring->rx_bi = kzalloc(bi_size, GFP_KERNEL); - if (!rx_ring->rx_bi) - goto err; u64_stats_init(&rx_ring->syncp); /* Round up to nearest 4K */ - rx_ring->size = rx_ring->count * sizeof(union i40e_32byte_rx_desc); + rx_ring->size = rx_ring->count * sizeof(union i40e_rx_desc); rx_ring->size = ALIGN(rx_ring->size, 4096); rx_ring->desc = dma_alloc_coherent(dev, rx_ring->size, &rx_ring->dma, GFP_KERNEL); @@ -1244,20 +1561,22 @@ int i40e_setup_rx_descriptors(struct i40e_ring *rx_ring) if (!rx_ring->desc) { dev_info(dev, "Unable to allocate memory for the Rx descriptor ring, size=%d\n", rx_ring->size); - goto err; + return -ENOMEM; } rx_ring->next_to_alloc = 0; rx_ring->next_to_clean = 0; + rx_ring->next_to_process = 0; rx_ring->next_to_use = 0; rx_ring->xdp_prog = rx_ring->vsi->xdp_prog; + rx_ring->rx_bi = + kcalloc(rx_ring->count, sizeof(*rx_ring->rx_bi), GFP_KERNEL); + if (!rx_ring->rx_bi) + return -ENOMEM; + return 0; -err: - kfree(rx_ring->rx_bi); - rx_ring->rx_bi = NULL; - return -ENOMEM; } /** @@ -1265,7 +1584,7 @@ err: * @rx_ring: ring to bump * @val: new head index **/ -static inline void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val) +void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val) { rx_ring->next_to_use = val; @@ -1281,16 +1600,19 @@ static inline void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val) writel(val, rx_ring->tail); } -/** - * i40e_rx_offset - Return expected offset into page to access data - * @rx_ring: Ring we are requesting offset of - * - * Returns the offset value for ring into the data buffer. - */ -static inline unsigned int i40e_rx_offset(struct i40e_ring *rx_ring) +#if (PAGE_SIZE >= 8192) +static unsigned int i40e_rx_frame_truesize(struct i40e_ring *rx_ring, + unsigned int size) { - return ring_uses_build_skb(rx_ring) ? I40E_SKB_PAD : 0; + unsigned int truesize; + + truesize = rx_ring->rx_offset ? + SKB_DATA_ALIGN(size + rx_ring->rx_offset) + + SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) : + SKB_DATA_ALIGN(size); + return truesize; } +#endif /** * i40e_alloc_mapped_page - recycle or make a new page @@ -1319,6 +1641,8 @@ static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring, return false; } + rx_ring->rx_stats.page_alloc_count++; + /* map page for use */ dma = dma_map_page_attrs(rx_ring->dev, page, 0, i40e_rx_pg_size(rx_ring), @@ -1336,33 +1660,14 @@ static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring, bi->dma = dma; bi->page = page; - bi->page_offset = i40e_rx_offset(rx_ring); - - /* initialize pagecnt_bias to 1 representing we fully own page */ - bi->pagecnt_bias = 1; + bi->page_offset = rx_ring->rx_offset; + page_ref_add(page, USHRT_MAX - 1); + bi->pagecnt_bias = USHRT_MAX; return true; } /** - * i40e_receive_skb - Send a completed packet up the stack - * @rx_ring: rx ring in play - * @skb: packet to send up - * @vlan_tag: vlan tag for packet - **/ -static void i40e_receive_skb(struct i40e_ring *rx_ring, - struct sk_buff *skb, u16 vlan_tag) -{ - struct i40e_q_vector *q_vector = rx_ring->q_vector; - - if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) && - (vlan_tag & VLAN_VID_MASK)) - __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag); - - napi_gro_receive(&q_vector->napi, skb); -} - -/** * i40e_alloc_rx_buffers - Replace used receive buffers * @rx_ring: ring to place buffers on * @cleaned_count: number of buffers to replace @@ -1380,7 +1685,7 @@ bool i40e_alloc_rx_buffers(struct i40e_ring *rx_ring, u16 cleaned_count) return false; rx_desc = I40E_RX_DESC(rx_ring, ntu); - bi = &rx_ring->rx_bi[ntu]; + bi = i40e_rx_bi(rx_ring, ntu); do { if (!i40e_alloc_mapped_page(rx_ring, bi)) @@ -1402,7 +1707,7 @@ bool i40e_alloc_rx_buffers(struct i40e_ring *rx_ring, u16 cleaned_count) ntu++; if (unlikely(ntu == rx_ring->count)) { rx_desc = I40E_RX_DESC(rx_ring, 0); - bi = rx_ring->rx_bi; + bi = i40e_rx_bi(rx_ring, 0); ntu = 0; } @@ -1437,40 +1742,30 @@ static inline void i40e_rx_checksum(struct i40e_vsi *vsi, struct sk_buff *skb, union i40e_rx_desc *rx_desc) { - struct i40e_rx_ptype_decoded decoded; + struct libeth_rx_pt decoded; u32 rx_error, rx_status; bool ipv4, ipv6; u8 ptype; u64 qword; - qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len); - ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >> I40E_RXD_QW1_PTYPE_SHIFT; - rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >> - I40E_RXD_QW1_ERROR_SHIFT; - rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >> - I40E_RXD_QW1_STATUS_SHIFT; - decoded = decode_rx_desc_ptype(ptype); - skb->ip_summed = CHECKSUM_NONE; - skb_checksum_none_assert(skb); + qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len); + ptype = FIELD_GET(I40E_RXD_QW1_PTYPE_MASK, qword); - /* Rx csum enabled and ip headers found? */ - if (!(vsi->netdev->features & NETIF_F_RXCSUM)) + decoded = libie_rx_pt_parse(ptype); + if (!libeth_rx_pt_has_checksum(vsi->netdev, decoded)) return; + rx_error = FIELD_GET(I40E_RXD_QW1_ERROR_MASK, qword); + rx_status = FIELD_GET(I40E_RXD_QW1_STATUS_MASK, qword); + /* did the hardware decode the packet and checksum? */ if (!(rx_status & BIT(I40E_RX_DESC_STATUS_L3L4P_SHIFT))) return; - /* both known and outer_ip must be set for the below code to work */ - if (!(decoded.known && decoded.outer_ip)) - return; - - ipv4 = (decoded.outer_ip == I40E_RX_PTYPE_OUTER_IP) && - (decoded.outer_ip_ver == I40E_RX_PTYPE_OUTER_IPV4); - ipv6 = (decoded.outer_ip == I40E_RX_PTYPE_OUTER_IP) && - (decoded.outer_ip_ver == I40E_RX_PTYPE_OUTER_IPV6); + ipv4 = libeth_rx_pt_get_ip_ver(decoded) == LIBETH_RX_PT_OUTER_IPV4; + ipv6 = libeth_rx_pt_get_ip_ver(decoded) == LIBETH_RX_PT_OUTER_IPV6; if (ipv4 && (rx_error & (BIT(I40E_RX_DESC_ERROR_IPE_SHIFT) | @@ -1498,20 +1793,10 @@ static inline void i40e_rx_checksum(struct i40e_vsi *vsi, * we need to bump the checksum level by 1 to reflect the fact that * we are indicating we validated the inner checksum. */ - if (decoded.tunnel_type >= I40E_RX_PTYPE_TUNNEL_IP_GRENAT) + if (decoded.tunnel_type >= LIBETH_RX_PT_TUNNEL_IP_GRENAT) skb->csum_level = 1; - /* Only report checksum unnecessary for TCP, UDP, or SCTP */ - switch (decoded.inner_prot) { - case I40E_RX_PTYPE_INNER_PROT_TCP: - case I40E_RX_PTYPE_INNER_PROT_UDP: - case I40E_RX_PTYPE_INNER_PROT_SCTP: - skb->ip_summed = CHECKSUM_UNNECESSARY; - /* fall though */ - default: - break; - } - + skb->ip_summed = CHECKSUM_UNNECESSARY; return; checksum_fail: @@ -1519,49 +1804,30 @@ checksum_fail: } /** - * i40e_ptype_to_htype - get a hash type - * @ptype: the ptype value from the descriptor - * - * Returns a hash type to be used by skb_set_hash - **/ -static inline int i40e_ptype_to_htype(u8 ptype) -{ - struct i40e_rx_ptype_decoded decoded = decode_rx_desc_ptype(ptype); - - if (!decoded.known) - return PKT_HASH_TYPE_NONE; - - if (decoded.outer_ip == I40E_RX_PTYPE_OUTER_IP && - decoded.payload_layer == I40E_RX_PTYPE_PAYLOAD_LAYER_PAY4) - return PKT_HASH_TYPE_L4; - else if (decoded.outer_ip == I40E_RX_PTYPE_OUTER_IP && - decoded.payload_layer == I40E_RX_PTYPE_PAYLOAD_LAYER_PAY3) - return PKT_HASH_TYPE_L3; - else - return PKT_HASH_TYPE_L2; -} - -/** * i40e_rx_hash - set the hash value in the skb * @ring: descriptor ring * @rx_desc: specific descriptor + * @skb: skb currently being received and modified + * @rx_ptype: Rx packet type **/ static inline void i40e_rx_hash(struct i40e_ring *ring, union i40e_rx_desc *rx_desc, struct sk_buff *skb, u8 rx_ptype) { + struct libeth_rx_pt decoded; u32 hash; const __le64 rss_mask = cpu_to_le64((u64)I40E_RX_DESC_FLTSTAT_RSS_HASH << I40E_RX_DESC_STATUS_FLTSTAT_SHIFT); - if (!(ring->netdev->features & NETIF_F_RXHASH)) + decoded = libie_rx_pt_parse(rx_ptype); + if (!libeth_rx_pt_has_hash(ring->netdev, decoded)) return; if ((rx_desc->wb.qword1.status_error_len & rss_mask) == rss_mask) { hash = le32_to_cpu(rx_desc->wb.qword0.hi_dword.rss); - skb_set_hash(skb, hash, i40e_ptype_to_htype(rx_ptype)); + libeth_rx_pt_set_hash(skb, hash, decoded); } } @@ -1570,23 +1836,19 @@ static inline void i40e_rx_hash(struct i40e_ring *ring, * @rx_ring: rx descriptor ring packet is being transacted on * @rx_desc: pointer to the EOP Rx descriptor * @skb: pointer to current skb being populated - * @rx_ptype: the packet type decoded by hardware * * This function checks the ring, descriptor, and packet information in * order to populate the hash, checksum, VLAN, protocol, and * other fields within the skb. **/ -static inline void i40e_process_skb_fields(struct i40e_ring *rx_ring, - union i40e_rx_desc *rx_desc, struct sk_buff *skb, - u8 rx_ptype) + union i40e_rx_desc *rx_desc, struct sk_buff *skb) { u64 qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len); - u32 rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >> - I40E_RXD_QW1_STATUS_SHIFT; + u32 rx_status = FIELD_GET(I40E_RXD_QW1_STATUS_MASK, qword); u32 tsynvalid = rx_status & I40E_RXD_QW1_STATUS_TSYNVALID_MASK; - u32 tsyn = (rx_status & I40E_RXD_QW1_STATUS_TSYNINDX_MASK) >> - I40E_RXD_QW1_STATUS_TSYNINDX_SHIFT; + u32 tsyn = FIELD_GET(I40E_RXD_QW1_STATUS_TSYNINDX_MASK, rx_status); + u8 rx_ptype = FIELD_GET(I40E_RXD_QW1_PTYPE_MASK, qword); if (unlikely(tsynvalid)) i40e_ptp_rx_hwtstamp(rx_ring->vsi->back, skb, tsyn); @@ -1597,6 +1859,13 @@ void i40e_process_skb_fields(struct i40e_ring *rx_ring, skb_record_rx_queue(skb, rx_ring->queue_index); + if (qword & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)) { + __le16 vlan_tag = rx_desc->wb.qword0.lo_dword.l2tag1; + + __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), + le16_to_cpu(vlan_tag)); + } + /* modifies the skb - consumes the enet header */ skb->protocol = eth_type_trans(skb, rx_ring->netdev); } @@ -1607,9 +1876,6 @@ void i40e_process_skb_fields(struct i40e_ring *rx_ring, * @skb: pointer to current skb being fixed * @rx_desc: pointer to the EOP Rx descriptor * - * Also address the case where we are pulling data in on pages only - * and as such no data is present in the skb header. - * * In addition if skb is not at least 60 bytes we need to pad it so that * it is large enough to qualify as a valid Ethernet frame. * @@ -1619,10 +1885,6 @@ static bool i40e_cleanup_headers(struct i40e_ring *rx_ring, struct sk_buff *skb, union i40e_rx_desc *rx_desc) { - /* XDP packets use error pointer so abort at this point */ - if (IS_ERR(skb)) - return true; - /* ERR_MASK will only have valid bits if EOP set, and * what we are doing here is actually checking * I40E_RX_DESC_ERROR_RXE_SHIFT, since it is the zeroth bit in @@ -1642,97 +1904,51 @@ static bool i40e_cleanup_headers(struct i40e_ring *rx_ring, struct sk_buff *skb, } /** - * i40e_reuse_rx_page - page flip buffer and store it back on the ring - * @rx_ring: rx descriptor ring to store buffers on - * @old_buff: donor buffer to have page reused - * - * Synchronizes page for reuse by the adapter - **/ -static void i40e_reuse_rx_page(struct i40e_ring *rx_ring, - struct i40e_rx_buffer *old_buff) -{ - struct i40e_rx_buffer *new_buff; - u16 nta = rx_ring->next_to_alloc; - - new_buff = &rx_ring->rx_bi[nta]; - - /* update, and store next to alloc */ - nta++; - rx_ring->next_to_alloc = (nta < rx_ring->count) ? nta : 0; - - /* transfer page from old buffer to new buffer */ - new_buff->dma = old_buff->dma; - new_buff->page = old_buff->page; - new_buff->page_offset = old_buff->page_offset; - new_buff->pagecnt_bias = old_buff->pagecnt_bias; -} - -/** - * i40e_page_is_reusable - check if any reuse is possible - * @page: page struct to check - * - * A page is not reusable if it was allocated under low memory - * conditions, or it's not in the same NUMA node as this CPU. - */ -static inline bool i40e_page_is_reusable(struct page *page) -{ - return (page_to_nid(page) == numa_mem_id()) && - !page_is_pfmemalloc(page); -} - -/** - * i40e_can_reuse_rx_page - Determine if this page can be reused by - * the adapter for another receive - * + * i40e_can_reuse_rx_page - Determine if page can be reused for another Rx * @rx_buffer: buffer containing the page + * @rx_stats: rx stats structure for the rx ring * - * If page is reusable, rx_buffer->page_offset is adjusted to point to - * an unused region in the page. - * - * For small pages, @truesize will be a constant value, half the size - * of the memory at page. We'll attempt to alternate between high and - * low halves of the page, with one half ready for use by the hardware - * and the other half being consumed by the stack. We use the page - * ref count to determine whether the stack has finished consuming the - * portion of this page that was passed up with a previous packet. If - * the page ref count is >1, we'll assume the "other" half page is - * still busy, and this page cannot be reused. + * If page is reusable, we have a green light for calling i40e_reuse_rx_page, + * which will assign the current buffer to the buffer that next_to_alloc is + * pointing to; otherwise, the DMA mapping needs to be destroyed and + * page freed. * - * For larger pages, @truesize will be the actual space used by the - * received packet (adjusted upward to an even multiple of the cache - * line size). This will advance through the page by the amount - * actually consumed by the received packets while there is still - * space for a buffer. Each region of larger pages will be used at - * most once, after which the page will not be reused. - * - * In either case, if the page is reusable its refcount is increased. - **/ -static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer) + * rx_stats will be updated to indicate whether the page was waived + * or busy if it could not be reused. + */ +static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer, + struct i40e_rx_queue_stats *rx_stats) { unsigned int pagecnt_bias = rx_buffer->pagecnt_bias; struct page *page = rx_buffer->page; /* Is any reuse possible? */ - if (unlikely(!i40e_page_is_reusable(page))) + if (!dev_page_is_reusable(page)) { + rx_stats->page_waive_count++; return false; + } #if (PAGE_SIZE < 8192) /* if we are only owner of page we can reuse it */ - if (unlikely((page_count(page) - pagecnt_bias) > 1)) + if (unlikely((rx_buffer->page_count - pagecnt_bias) > 1)) { + rx_stats->page_busy_count++; return false; + } #else #define I40E_LAST_OFFSET \ (SKB_WITH_OVERHEAD(PAGE_SIZE) - I40E_RXBUFFER_2048) - if (rx_buffer->page_offset > I40E_LAST_OFFSET) + if (rx_buffer->page_offset > I40E_LAST_OFFSET) { + rx_stats->page_busy_count++; return false; + } #endif /* If we have drained the page fragment pool we need to update * the pagecnt_bias and page count so that we fully restock the * number of references the driver holds. */ - if (unlikely(!pagecnt_bias)) { - page_ref_add(page, USHRT_MAX); + if (unlikely(pagecnt_bias == 1)) { + page_ref_add(page, USHRT_MAX - 1); rx_buffer->pagecnt_bias = USHRT_MAX; } @@ -1740,33 +1956,14 @@ static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer) } /** - * i40e_add_rx_frag - Add contents of Rx buffer to sk_buff - * @rx_ring: rx descriptor ring to transact packets on - * @rx_buffer: buffer containing page to add - * @skb: sk_buff to place the data into - * @size: packet length from rx_desc - * - * This function will add the data contained in rx_buffer->page to the skb. - * It will just attach the page as a frag to the skb. - * - * The function will then update the page offset. + * i40e_rx_buffer_flip - adjusted rx_buffer to point to an unused region + * @rx_buffer: Rx buffer to adjust + * @truesize: Size of adjustment **/ -static void i40e_add_rx_frag(struct i40e_ring *rx_ring, - struct i40e_rx_buffer *rx_buffer, - struct sk_buff *skb, - unsigned int size) +static void i40e_rx_buffer_flip(struct i40e_rx_buffer *rx_buffer, + unsigned int truesize) { #if (PAGE_SIZE < 8192) - unsigned int truesize = i40e_rx_pg_size(rx_ring) / 2; -#else - unsigned int truesize = SKB_DATA_ALIGN(size + i40e_rx_offset(rx_ring)); -#endif - - skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, rx_buffer->page, - rx_buffer->page_offset, size, truesize); - - /* page is being used so we must update the page offset */ -#if (PAGE_SIZE < 8192) rx_buffer->page_offset ^= truesize; #else rx_buffer->page_offset += truesize; @@ -1786,8 +1983,14 @@ static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring, { struct i40e_rx_buffer *rx_buffer; - rx_buffer = &rx_ring->rx_bi[rx_ring->next_to_clean]; - prefetchw(rx_buffer->page); + rx_buffer = i40e_rx_bi(rx_ring, rx_ring->next_to_process); + rx_buffer->page_count = +#if (PAGE_SIZE < 8192) + page_count(rx_buffer->page); +#else + 0; +#endif + prefetch_page_address(rx_buffer->page); /* we are reusing so sync this buffer for CPU use */ dma_sync_single_range_for_cpu(rx_ring->dev, @@ -1803,9 +2006,70 @@ static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring, } /** - * i40e_construct_skb - Allocate skb and populate it + * i40e_put_rx_buffer - Clean up used buffer and either recycle or free * @rx_ring: rx descriptor ring to transact packets on * @rx_buffer: rx buffer to pull data from + * + * This function will clean up the contents of the rx_buffer. It will + * either recycle the buffer or unmap it and free the associated resources. + */ +static void i40e_put_rx_buffer(struct i40e_ring *rx_ring, + struct i40e_rx_buffer *rx_buffer) +{ + if (i40e_can_reuse_rx_page(rx_buffer, &rx_ring->rx_stats)) { + /* hand second half of page back to the ring */ + i40e_reuse_rx_page(rx_ring, rx_buffer); + } else { + /* we are not reusing the buffer so unmap it */ + dma_unmap_page_attrs(rx_ring->dev, rx_buffer->dma, + i40e_rx_pg_size(rx_ring), + DMA_FROM_DEVICE, I40E_RX_DMA_ATTR); + __page_frag_cache_drain(rx_buffer->page, + rx_buffer->pagecnt_bias); + /* clear contents of buffer_info */ + rx_buffer->page = NULL; + } +} + +/** + * i40e_process_rx_buffs- Processing of buffers post XDP prog or on error + * @rx_ring: Rx descriptor ring to transact packets on + * @xdp_res: Result of the XDP program + * @xdp: xdp_buff pointing to the data + **/ +static void i40e_process_rx_buffs(struct i40e_ring *rx_ring, int xdp_res, + struct xdp_buff *xdp) +{ + u32 nr_frags = xdp_get_shared_info_from_buff(xdp)->nr_frags; + u32 next = rx_ring->next_to_clean, i = 0; + struct i40e_rx_buffer *rx_buffer; + + xdp->flags = 0; + + while (1) { + rx_buffer = i40e_rx_bi(rx_ring, next); + if (++next == rx_ring->count) + next = 0; + + if (!rx_buffer->page) + continue; + + if (xdp_res != I40E_XDP_CONSUMED) + i40e_rx_buffer_flip(rx_buffer, xdp->frame_sz); + else if (i++ <= nr_frags) + rx_buffer->pagecnt_bias++; + + /* EOP buffer will be put in i40e_clean_rx_irq() */ + if (next == rx_ring->next_to_process) + return; + + i40e_put_rx_buffer(rx_ring, rx_buffer); + } +} + +/** + * i40e_construct_skb - Allocate skb and populate it + * @rx_ring: rx descriptor ring to transact packets on * @xdp: xdp_buff pointing to the data * * This function allocates an skb. It then populates it with the page @@ -1813,158 +2077,159 @@ static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring, * skb correctly. */ static struct sk_buff *i40e_construct_skb(struct i40e_ring *rx_ring, - struct i40e_rx_buffer *rx_buffer, struct xdp_buff *xdp) { unsigned int size = xdp->data_end - xdp->data; -#if (PAGE_SIZE < 8192) - unsigned int truesize = i40e_rx_pg_size(rx_ring) / 2; -#else - unsigned int truesize = SKB_DATA_ALIGN(size); -#endif + struct i40e_rx_buffer *rx_buffer; + struct skb_shared_info *sinfo; unsigned int headlen; struct sk_buff *skb; + u32 nr_frags = 0; /* prefetch first cache line of first page */ - prefetch(xdp->data); -#if L1_CACHE_BYTES < 128 - prefetch(xdp->data + L1_CACHE_BYTES); -#endif + net_prefetch(xdp->data); + + /* Note, we get here by enabling legacy-rx via: + * + * ethtool --set-priv-flags <dev> legacy-rx on + * + * In this mode, we currently get 0 extra XDP headroom as + * opposed to having legacy-rx off, where we process XDP + * packets going to stack via i40e_build_skb(). The latter + * provides us currently with 192 bytes of headroom. + * + * For i40e_construct_skb() mode it means that the + * xdp->data_meta will always point to xdp->data, since + * the helper cannot expand the head. Should this ever + * change in future for legacy-rx mode on, then lets also + * add xdp->data_meta handling here. + */ /* allocate a skb to store the frags */ - skb = __napi_alloc_skb(&rx_ring->q_vector->napi, - I40E_RX_HDR_SIZE, - GFP_ATOMIC | __GFP_NOWARN); + skb = napi_alloc_skb(&rx_ring->q_vector->napi, I40E_RX_HDR_SIZE); if (unlikely(!skb)) return NULL; /* Determine available headroom for copy */ headlen = size; if (headlen > I40E_RX_HDR_SIZE) - headlen = eth_get_headlen(xdp->data, I40E_RX_HDR_SIZE); + headlen = eth_get_headlen(skb->dev, xdp->data, + I40E_RX_HDR_SIZE); /* align pull length to size of long to optimize memcpy performance */ memcpy(__skb_put(skb, headlen), xdp->data, ALIGN(headlen, sizeof(long))); + if (unlikely(xdp_buff_has_frags(xdp))) { + sinfo = xdp_get_shared_info_from_buff(xdp); + nr_frags = sinfo->nr_frags; + } + rx_buffer = i40e_rx_bi(rx_ring, rx_ring->next_to_clean); /* update all of the pointers */ size -= headlen; if (size) { + if (unlikely(nr_frags >= MAX_SKB_FRAGS)) { + dev_kfree_skb(skb); + return NULL; + } skb_add_rx_frag(skb, 0, rx_buffer->page, rx_buffer->page_offset + headlen, - size, truesize); - + size, xdp->frame_sz); /* buffer is used by skb, update page_offset */ -#if (PAGE_SIZE < 8192) - rx_buffer->page_offset ^= truesize; -#else - rx_buffer->page_offset += truesize; -#endif + i40e_rx_buffer_flip(rx_buffer, xdp->frame_sz); } else { /* buffer is unused, reset bias back to rx_buffer */ rx_buffer->pagecnt_bias++; } + if (unlikely(xdp_buff_has_frags(xdp))) { + struct skb_shared_info *skinfo = skb_shinfo(skb); + + memcpy(&skinfo->frags[skinfo->nr_frags], &sinfo->frags[0], + sizeof(skb_frag_t) * nr_frags); + + xdp_update_skb_frags_info(skb, skinfo->nr_frags + nr_frags, + sinfo->xdp_frags_size, + nr_frags * xdp->frame_sz, + xdp_buff_get_skb_flags(xdp)); + + /* First buffer has already been processed, so bump ntc */ + if (++rx_ring->next_to_clean == rx_ring->count) + rx_ring->next_to_clean = 0; + + i40e_process_rx_buffs(rx_ring, I40E_XDP_PASS, xdp); + } + return skb; } /** * i40e_build_skb - Build skb around an existing buffer * @rx_ring: Rx descriptor ring to transact packets on - * @rx_buffer: Rx buffer to pull data from * @xdp: xdp_buff pointing to the data * * This function builds an skb around an existing Rx buffer, taking care * to set up the skb correctly and avoid any memcpy overhead. */ static struct sk_buff *i40e_build_skb(struct i40e_ring *rx_ring, - struct i40e_rx_buffer *rx_buffer, struct xdp_buff *xdp) { - unsigned int size = xdp->data_end - xdp->data; -#if (PAGE_SIZE < 8192) - unsigned int truesize = i40e_rx_pg_size(rx_ring) / 2; -#else - unsigned int truesize = SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) + - SKB_DATA_ALIGN(I40E_SKB_PAD + size); -#endif + unsigned int metasize = xdp->data - xdp->data_meta; + struct skb_shared_info *sinfo; struct sk_buff *skb; + u32 nr_frags; + + /* Prefetch first cache line of first page. If xdp->data_meta + * is unused, this points exactly as xdp->data, otherwise we + * likely have a consumer accessing first few bytes of meta + * data, and then actual data. + */ + net_prefetch(xdp->data_meta); + + if (unlikely(xdp_buff_has_frags(xdp))) { + sinfo = xdp_get_shared_info_from_buff(xdp); + nr_frags = sinfo->nr_frags; + } - /* prefetch first cache line of first page */ - prefetch(xdp->data); -#if L1_CACHE_BYTES < 128 - prefetch(xdp->data + L1_CACHE_BYTES); -#endif /* build an skb around the page buffer */ - skb = build_skb(xdp->data_hard_start, truesize); + skb = napi_build_skb(xdp->data_hard_start, xdp->frame_sz); if (unlikely(!skb)) return NULL; /* update pointers within the skb to store the data */ - skb_reserve(skb, I40E_SKB_PAD); - __skb_put(skb, size); - - /* buffer is used by skb, update page_offset */ -#if (PAGE_SIZE < 8192) - rx_buffer->page_offset ^= truesize; -#else - rx_buffer->page_offset += truesize; -#endif + skb_reserve(skb, xdp->data - xdp->data_hard_start); + __skb_put(skb, xdp->data_end - xdp->data); + if (metasize) + skb_metadata_set(skb, metasize); - return skb; -} + if (unlikely(xdp_buff_has_frags(xdp))) { + xdp_update_skb_frags_info(skb, nr_frags, sinfo->xdp_frags_size, + nr_frags * xdp->frame_sz, + xdp_buff_get_skb_flags(xdp)); -/** - * i40e_put_rx_buffer - Clean up used buffer and either recycle or free - * @rx_ring: rx descriptor ring to transact packets on - * @rx_buffer: rx buffer to pull data from - * - * This function will clean up the contents of the rx_buffer. It will - * either recycle the bufer or unmap it and free the associated resources. - */ -static void i40e_put_rx_buffer(struct i40e_ring *rx_ring, - struct i40e_rx_buffer *rx_buffer) -{ - if (i40e_can_reuse_rx_page(rx_buffer)) { - /* hand second half of page back to the ring */ - i40e_reuse_rx_page(rx_ring, rx_buffer); - rx_ring->rx_stats.page_reuse_count++; + i40e_process_rx_buffs(rx_ring, I40E_XDP_PASS, xdp); } else { - /* we are not reusing the buffer so unmap it */ - dma_unmap_page_attrs(rx_ring->dev, rx_buffer->dma, - i40e_rx_pg_size(rx_ring), - DMA_FROM_DEVICE, I40E_RX_DMA_ATTR); - __page_frag_cache_drain(rx_buffer->page, - rx_buffer->pagecnt_bias); + struct i40e_rx_buffer *rx_buffer; + + rx_buffer = i40e_rx_bi(rx_ring, rx_ring->next_to_clean); + /* buffer is used by skb, update page_offset */ + i40e_rx_buffer_flip(rx_buffer, xdp->frame_sz); } - /* clear contents of buffer_info */ - rx_buffer->page = NULL; + return skb; } /** * i40e_is_non_eop - process handling of non-EOP buffers * @rx_ring: Rx ring being processed * @rx_desc: Rx descriptor for current buffer - * @skb: Current socket buffer containing buffer in progress * - * This function updates next to clean. If the buffer is an EOP buffer - * this function exits returning false, otherwise it will place the - * sk_buff in the next buffer to be chained and return true indicating - * that this is in fact a non-EOP buffer. - **/ -static bool i40e_is_non_eop(struct i40e_ring *rx_ring, - union i40e_rx_desc *rx_desc, - struct sk_buff *skb) + * If the buffer is an EOP buffer, this function exits returning false, + * otherwise return true indicating that this is in fact a non-EOP buffer. + */ +bool i40e_is_non_eop(struct i40e_ring *rx_ring, + union i40e_rx_desc *rx_desc) { - u32 ntc = rx_ring->next_to_clean + 1; - - /* fetch, update, and store next to clean */ - ntc = (ntc < rx_ring->count) ? ntc : 0; - rx_ring->next_to_clean = ntc; - - prefetch(I40E_RX_DESC(rx_ring, ntc)); - /* if we are the last buffer then there is nothing else to do */ #define I40E_RXD_EOF BIT(I40E_RX_DESC_STATUS_EOF_SHIFT) if (likely(i40e_test_staterr(rx_desc, I40E_RXD_EOF))) @@ -1975,79 +2240,191 @@ static bool i40e_is_non_eop(struct i40e_ring *rx_ring, return true; } -#define I40E_XDP_PASS 0 -#define I40E_XDP_CONSUMED 1 -#define I40E_XDP_TX 2 - -static int i40e_xmit_xdp_ring(struct xdp_buff *xdp, +static int i40e_xmit_xdp_ring(struct xdp_frame *xdpf, struct i40e_ring *xdp_ring); +int i40e_xmit_xdp_tx_ring(struct xdp_buff *xdp, struct i40e_ring *xdp_ring) +{ + struct xdp_frame *xdpf = xdp_convert_buff_to_frame(xdp); + + if (unlikely(!xdpf)) + return I40E_XDP_CONSUMED; + + return i40e_xmit_xdp_ring(xdpf, xdp_ring); +} + /** * i40e_run_xdp - run an XDP program * @rx_ring: Rx ring being processed * @xdp: XDP buffer containing the frame + * @xdp_prog: XDP program to run **/ -static struct sk_buff *i40e_run_xdp(struct i40e_ring *rx_ring, - struct xdp_buff *xdp) +static int i40e_run_xdp(struct i40e_ring *rx_ring, struct xdp_buff *xdp, struct bpf_prog *xdp_prog) { - int result = I40E_XDP_PASS; + int err, result = I40E_XDP_PASS; struct i40e_ring *xdp_ring; - struct bpf_prog *xdp_prog; u32 act; - rcu_read_lock(); - xdp_prog = READ_ONCE(rx_ring->xdp_prog); - if (!xdp_prog) goto xdp_out; + prefetchw(xdp->data_hard_start); /* xdp_frame write */ + act = bpf_prog_run_xdp(xdp_prog, xdp); switch (act) { case XDP_PASS: break; case XDP_TX: xdp_ring = rx_ring->vsi->xdp_rings[rx_ring->queue_index]; - result = i40e_xmit_xdp_ring(xdp, xdp_ring); + result = i40e_xmit_xdp_tx_ring(xdp, xdp_ring); + if (result == I40E_XDP_CONSUMED) + goto out_failure; + break; + case XDP_REDIRECT: + err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog); + if (err) + goto out_failure; + result = I40E_XDP_REDIR; break; default: - bpf_warn_invalid_xdp_action(act); + bpf_warn_invalid_xdp_action(rx_ring->netdev, xdp_prog, act); + fallthrough; case XDP_ABORTED: +out_failure: trace_xdp_exception(rx_ring->netdev, xdp_prog, act); - /* fallthrough -- handle aborts by dropping packet */ + fallthrough; /* handle aborts by dropping packet */ case XDP_DROP: result = I40E_XDP_CONSUMED; break; } xdp_out: - rcu_read_unlock(); - return ERR_PTR(-result); + return result; } /** - * i40e_rx_buffer_flip - adjusted rx_buffer to point to an unused region + * i40e_xdp_ring_update_tail - Updates the XDP Tx ring tail register + * @xdp_ring: XDP Tx ring + * + * This function updates the XDP Tx ring tail register. + **/ +void i40e_xdp_ring_update_tail(struct i40e_ring *xdp_ring) +{ + /* Force memory writes to complete before letting h/w + * know there are new descriptors to fetch. + */ + wmb(); + writel_relaxed(xdp_ring->next_to_use, xdp_ring->tail); +} + +/** + * i40e_update_rx_stats - Update Rx ring statistics + * @rx_ring: rx descriptor ring + * @total_rx_bytes: number of bytes received + * @total_rx_packets: number of packets received + * + * This function updates the Rx ring statistics. + **/ +void i40e_update_rx_stats(struct i40e_ring *rx_ring, + unsigned int total_rx_bytes, + unsigned int total_rx_packets) +{ + u64_stats_update_begin(&rx_ring->syncp); + rx_ring->stats.packets += total_rx_packets; + rx_ring->stats.bytes += total_rx_bytes; + u64_stats_update_end(&rx_ring->syncp); + rx_ring->q_vector->rx.total_packets += total_rx_packets; + rx_ring->q_vector->rx.total_bytes += total_rx_bytes; +} + +/** + * i40e_finalize_xdp_rx - Bump XDP Tx tail and/or flush redirect map * @rx_ring: Rx ring - * @rx_buffer: Rx buffer to adjust - * @size: Size of adjustment + * @xdp_res: Result of the receive batch + * + * This function bumps XDP Tx tail and/or flush redirect map, and + * should be called when a batch of packets has been processed in the + * napi loop. **/ -static void i40e_rx_buffer_flip(struct i40e_ring *rx_ring, - struct i40e_rx_buffer *rx_buffer, - unsigned int size) +void i40e_finalize_xdp_rx(struct i40e_ring *rx_ring, unsigned int xdp_res) { -#if (PAGE_SIZE < 8192) - unsigned int truesize = i40e_rx_pg_size(rx_ring) / 2; + if (xdp_res & I40E_XDP_REDIR) + xdp_do_flush(); - rx_buffer->page_offset ^= truesize; -#else - unsigned int truesize = SKB_DATA_ALIGN(i40e_rx_offset(rx_ring) + size); + if (xdp_res & I40E_XDP_TX) { + struct i40e_ring *xdp_ring = + rx_ring->vsi->xdp_rings[rx_ring->queue_index]; - rx_buffer->page_offset += truesize; -#endif + i40e_xdp_ring_update_tail(xdp_ring); + } +} + +/** + * i40e_inc_ntp: Advance the next_to_process index + * @rx_ring: Rx ring + **/ +static void i40e_inc_ntp(struct i40e_ring *rx_ring) +{ + u32 ntp = rx_ring->next_to_process + 1; + + ntp = (ntp < rx_ring->count) ? ntp : 0; + rx_ring->next_to_process = ntp; + prefetch(I40E_RX_DESC(rx_ring, ntp)); +} + +/** + * i40e_add_xdp_frag: Add a frag to xdp_buff + * @xdp: xdp_buff pointing to the data + * @nr_frags: return number of buffers for the packet + * @rx_buffer: rx_buffer holding data of the current frag + * @size: size of data of current frag + */ +static int i40e_add_xdp_frag(struct xdp_buff *xdp, u32 *nr_frags, + struct i40e_rx_buffer *rx_buffer, u32 size) +{ + struct skb_shared_info *sinfo = xdp_get_shared_info_from_buff(xdp); + + if (!xdp_buff_has_frags(xdp)) { + sinfo->nr_frags = 0; + sinfo->xdp_frags_size = 0; + xdp_buff_set_frags_flag(xdp); + } else if (unlikely(sinfo->nr_frags >= MAX_SKB_FRAGS)) { + /* Overflowing packet: All frags need to be dropped */ + return -ENOMEM; + } + + __skb_fill_page_desc_noacc(sinfo, sinfo->nr_frags++, rx_buffer->page, + rx_buffer->page_offset, size); + + sinfo->xdp_frags_size += size; + + if (page_is_pfmemalloc(rx_buffer->page)) + xdp_buff_set_frag_pfmemalloc(xdp); + *nr_frags = sinfo->nr_frags; + + return 0; +} + +/** + * i40e_consume_xdp_buff - Consume all the buffers of the packet and update ntc + * @rx_ring: rx descriptor ring to transact packets on + * @xdp: xdp_buff pointing to the data + * @rx_buffer: rx_buffer of eop desc + */ +static void i40e_consume_xdp_buff(struct i40e_ring *rx_ring, + struct xdp_buff *xdp, + struct i40e_rx_buffer *rx_buffer) +{ + i40e_process_rx_buffs(rx_ring, I40E_XDP_CONSUMED, xdp); + i40e_put_rx_buffer(rx_ring, rx_buffer); + rx_ring->next_to_clean = rx_ring->next_to_process; + xdp->data = NULL; } /** * i40e_clean_rx_irq - Clean completed descriptors from Rx ring - bounce buf * @rx_ring: rx descriptor ring to transact packets on * @budget: Total limit on number of packets to process + * @rx_cleaned: Out parameter of the number of packets processed * * This function provides a "bounce buffer" approach to Rx interrupt * processing. The advantage to this is that on systems that have @@ -2056,30 +2433,39 @@ static void i40e_rx_buffer_flip(struct i40e_ring *rx_ring, * * Returns amount of work completed **/ -static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget) +static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget, + unsigned int *rx_cleaned) { unsigned int total_rx_bytes = 0, total_rx_packets = 0; - struct sk_buff *skb = rx_ring->skb; u16 cleaned_count = I40E_DESC_UNUSED(rx_ring); - bool failure = false, xdp_xmit = false; + u16 clean_threshold = rx_ring->count / 2; + unsigned int offset = rx_ring->rx_offset; + struct xdp_buff *xdp = &rx_ring->xdp; + unsigned int xdp_xmit = 0; + struct bpf_prog *xdp_prog; + bool failure = false; + int xdp_res = 0; - while (likely(total_rx_packets < budget)) { + xdp_prog = READ_ONCE(rx_ring->xdp_prog); + + while (likely(total_rx_packets < (unsigned int)budget)) { + u16 ntp = rx_ring->next_to_process; struct i40e_rx_buffer *rx_buffer; union i40e_rx_desc *rx_desc; - struct xdp_buff xdp; + struct sk_buff *skb; unsigned int size; - u16 vlan_tag; - u8 rx_ptype; + u32 nfrags = 0; + bool neop; u64 qword; /* return some buffers to hardware, one at a time is too slow */ - if (cleaned_count >= I40E_RX_BUFFER_WRITE) { + if (cleaned_count >= clean_threshold) { failure = failure || i40e_alloc_rx_buffers(rx_ring, cleaned_count); cleaned_count = 0; } - rx_desc = I40E_RX_DESC(rx_ring, rx_ring->next_to_clean); + rx_desc = I40E_RX_DESC(rx_ring, ntp); /* status_error_len will always be zero for unused descriptors * because it's cleared in cleanup, and overlaps with hdr_addr @@ -2094,136 +2480,179 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget) */ dma_rmb(); - if (unlikely(i40e_rx_is_programming_status(qword))) { - i40e_clean_programming_status(rx_ring, rx_desc, qword); + if (i40e_rx_is_programming_status(qword)) { + i40e_clean_programming_status(rx_ring, + rx_desc->raw.qword[0], + qword); + rx_buffer = i40e_rx_bi(rx_ring, ntp); + i40e_inc_ntp(rx_ring); + i40e_reuse_rx_page(rx_ring, rx_buffer); + /* Update ntc and bump cleaned count if not in the + * middle of mb packet. + */ + if (rx_ring->next_to_clean == ntp) { + rx_ring->next_to_clean = + rx_ring->next_to_process; + cleaned_count++; + } continue; } - size = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >> - I40E_RXD_QW1_LENGTH_PBUF_SHIFT; + + size = FIELD_GET(I40E_RXD_QW1_LENGTH_PBUF_MASK, qword); if (!size) break; - i40e_trace(clean_rx_irq, rx_ring, rx_desc, skb); + i40e_trace(clean_rx_irq, rx_ring, rx_desc, xdp); + /* retrieve a buffer from the ring */ rx_buffer = i40e_get_rx_buffer(rx_ring, size); - /* retrieve a buffer from the ring */ - if (!skb) { - xdp.data = page_address(rx_buffer->page) + - rx_buffer->page_offset; - xdp.data_hard_start = xdp.data - - i40e_rx_offset(rx_ring); - xdp.data_end = xdp.data + size; - - skb = i40e_run_xdp(rx_ring, &xdp); - } + neop = i40e_is_non_eop(rx_ring, rx_desc); + i40e_inc_ntp(rx_ring); - if (IS_ERR(skb)) { - if (PTR_ERR(skb) == -I40E_XDP_TX) { - xdp_xmit = true; - i40e_rx_buffer_flip(rx_ring, rx_buffer, size); - } else { - rx_buffer->pagecnt_bias++; - } - total_rx_bytes += size; - total_rx_packets++; - } else if (skb) { - i40e_add_rx_frag(rx_ring, rx_buffer, skb, size); - } else if (ring_uses_build_skb(rx_ring)) { - skb = i40e_build_skb(rx_ring, rx_buffer, &xdp); - } else { - skb = i40e_construct_skb(rx_ring, rx_buffer, &xdp); - } + if (!xdp->data) { + unsigned char *hard_start; - /* exit if we failed to retrieve a buffer */ - if (!skb) { - rx_ring->rx_stats.alloc_buff_failed++; - rx_buffer->pagecnt_bias++; + hard_start = page_address(rx_buffer->page) + + rx_buffer->page_offset - offset; + xdp_prepare_buff(xdp, hard_start, offset, size, true); +#if (PAGE_SIZE > 4096) + /* At larger PAGE_SIZE, frame_sz depend on len size */ + xdp->frame_sz = i40e_rx_frame_truesize(rx_ring, size); +#endif + } else if (i40e_add_xdp_frag(xdp, &nfrags, rx_buffer, size) && + !neop) { + /* Overflowing packet: Drop all frags on EOP */ + i40e_consume_xdp_buff(rx_ring, xdp, rx_buffer); break; } - i40e_put_rx_buffer(rx_ring, rx_buffer); - cleaned_count++; - - if (i40e_is_non_eop(rx_ring, rx_desc, skb)) + if (neop) continue; - if (i40e_cleanup_headers(rx_ring, skb, rx_desc)) { - skb = NULL; - continue; - } + xdp_res = i40e_run_xdp(rx_ring, xdp, xdp_prog); - /* probably a little skewed due to removing CRC */ - total_rx_bytes += skb->len; + if (xdp_res) { + xdp_xmit |= xdp_res & (I40E_XDP_TX | I40E_XDP_REDIR); - qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len); - rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >> - I40E_RXD_QW1_PTYPE_SHIFT; - - /* populate checksum, VLAN, and protocol */ - i40e_process_skb_fields(rx_ring, rx_desc, skb, rx_ptype); + if (unlikely(xdp_buff_has_frags(xdp))) { + i40e_process_rx_buffs(rx_ring, xdp_res, xdp); + size = xdp_get_buff_len(xdp); + } else if (xdp_res & (I40E_XDP_TX | I40E_XDP_REDIR)) { + i40e_rx_buffer_flip(rx_buffer, xdp->frame_sz); + } else { + rx_buffer->pagecnt_bias++; + } + total_rx_bytes += size; + } else { + if (ring_uses_build_skb(rx_ring)) + skb = i40e_build_skb(rx_ring, xdp); + else + skb = i40e_construct_skb(rx_ring, xdp); - vlan_tag = (qword & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)) ? - le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1) : 0; + /* drop if we failed to retrieve a buffer */ + if (!skb) { + rx_ring->rx_stats.alloc_buff_failed++; + i40e_consume_xdp_buff(rx_ring, xdp, rx_buffer); + break; + } - i40e_trace(clean_rx_irq_rx, rx_ring, rx_desc, skb); - i40e_receive_skb(rx_ring, skb, vlan_tag); - skb = NULL; + if (i40e_cleanup_headers(rx_ring, skb, rx_desc)) + goto process_next; - /* update budget accounting */ - total_rx_packets++; - } + /* probably a little skewed due to removing CRC */ + total_rx_bytes += skb->len; - if (xdp_xmit) { - struct i40e_ring *xdp_ring; + /* populate checksum, VLAN, and protocol */ + i40e_process_skb_fields(rx_ring, rx_desc, skb); - xdp_ring = rx_ring->vsi->xdp_rings[rx_ring->queue_index]; + i40e_trace(clean_rx_irq_rx, rx_ring, rx_desc, xdp); + napi_gro_receive(&rx_ring->q_vector->napi, skb); + } - /* Force memory writes to complete before letting h/w - * know there are new descriptors to fetch. - */ - wmb(); + /* update budget accounting */ + total_rx_packets++; +process_next: + cleaned_count += nfrags + 1; + i40e_put_rx_buffer(rx_ring, rx_buffer); + rx_ring->next_to_clean = rx_ring->next_to_process; - writel(xdp_ring->next_to_use, xdp_ring->tail); + xdp->data = NULL; } - rx_ring->skb = skb; + i40e_finalize_xdp_rx(rx_ring, xdp_xmit); - u64_stats_update_begin(&rx_ring->syncp); - rx_ring->stats.packets += total_rx_packets; - rx_ring->stats.bytes += total_rx_bytes; - u64_stats_update_end(&rx_ring->syncp); - rx_ring->q_vector->rx.total_packets += total_rx_packets; - rx_ring->q_vector->rx.total_bytes += total_rx_bytes; + i40e_update_rx_stats(rx_ring, total_rx_bytes, total_rx_packets); + + *rx_cleaned = total_rx_packets; /* guarantee a trip back through this routine if there was a failure */ - return failure ? budget : total_rx_packets; + return failure ? budget : (int)total_rx_packets; } -static u32 i40e_buildreg_itr(const int type, const u16 itr) +/** + * i40e_buildreg_itr - build a value for writing to I40E_PFINT_DYN_CTLN register + * @itr_idx: interrupt throttling index + * @interval: interrupt throttling interval value in usecs + * @force_swint: force software interrupt + * + * The function builds a value for I40E_PFINT_DYN_CTLN register that + * is used to update interrupt throttling interval for specified ITR index + * and optionally enforces a software interrupt. If the @itr_idx is equal + * to I40E_ITR_NONE then no interval change is applied and only @force_swint + * parameter is taken into account. If the interval change and enforced + * software interrupt are not requested then the built value just enables + * appropriate vector interrupt. + **/ +static u32 i40e_buildreg_itr(enum i40e_dyn_idx itr_idx, u16 interval, + bool force_swint) { u32 val; + /* We don't bother with setting the CLEARPBA bit as the data sheet + * points out doing so is "meaningless since it was already + * auto-cleared". The auto-clearing happens when the interrupt is + * asserted. + * + * Hardware errata 28 for also indicates that writing to a + * xxINT_DYN_CTLx CSR with INTENA_MSK (bit 31) set to 0 will clear + * an event in the PBA anyway so we need to rely on the automask + * to hold pending events for us until the interrupt is re-enabled + * + * We have to shift the given value as it is reported in microseconds + * and the register value is recorded in 2 microsecond units. + */ + interval >>= 1; + + /* 1. Enable vector interrupt + * 2. Update the interval for the specified ITR index + * (I40E_ITR_NONE in the register is used to indicate that + * no interval update is requested) + */ val = I40E_PFINT_DYN_CTLN_INTENA_MASK | - /* Don't clear PBA because that can cause lost interrupts that - * came in while we were cleaning/polling - */ - (type << I40E_PFINT_DYN_CTLN_ITR_INDX_SHIFT) | - (itr << I40E_PFINT_DYN_CTLN_INTERVAL_SHIFT); + FIELD_PREP(I40E_PFINT_DYN_CTLN_ITR_INDX_MASK, itr_idx) | + FIELD_PREP(I40E_PFINT_DYN_CTLN_INTERVAL_MASK, interval); - return val; -} + /* 3. Enforce software interrupt trigger if requested + * (These software interrupts rate is limited by ITR2 that is + * set to 20K interrupts per second) + */ + if (force_swint) + val |= I40E_PFINT_DYN_CTLN_SWINT_TRIG_MASK | + I40E_PFINT_DYN_CTLN_SW_ITR_INDX_ENA_MASK | + FIELD_PREP(I40E_PFINT_DYN_CTLN_SW_ITR_INDX_MASK, + I40E_SW_ITR); -/* a small macro to shorten up some long lines */ -#define INTREG I40E_PFINT_DYN_CTLN -static inline int get_rx_itr(struct i40e_vsi *vsi, int idx) -{ - return vsi->rx_rings[idx]->rx_itr_setting; + return val; } -static inline int get_tx_itr(struct i40e_vsi *vsi, int idx) -{ - return vsi->tx_rings[idx]->tx_itr_setting; -} +/* The act of updating the ITR will cause it to immediately trigger. In order + * to prevent this from throwing off adaptive update statistics we defer the + * update so that it can only happen so often. So after either Tx or Rx are + * updated we make the adaptive scheme wait until either the ITR completely + * expires via the next_update expiration or we have been through at least + * 3 interrupts. + */ +#define ITR_COUNTDOWN_START 3 /** * i40e_update_enable_itr - Update itr and re-enable MSIX interrupt @@ -2234,74 +2663,71 @@ static inline int get_tx_itr(struct i40e_vsi *vsi, int idx) static inline void i40e_update_enable_itr(struct i40e_vsi *vsi, struct i40e_q_vector *q_vector) { + enum i40e_dyn_idx itr_idx = I40E_ITR_NONE; struct i40e_hw *hw = &vsi->back->hw; - bool rx = false, tx = false; - u32 rxval, txval; - int vector; - int idx = q_vector->v_idx; - int rx_itr_setting, tx_itr_setting; - - vector = (q_vector->v_idx + vsi->base_vector); - - /* avoid dynamic calculation if in countdown mode OR if - * all dynamic is disabled - */ - rxval = txval = i40e_buildreg_itr(I40E_ITR_NONE, 0); - - rx_itr_setting = get_rx_itr(vsi, idx); - tx_itr_setting = get_tx_itr(vsi, idx); + u16 interval = 0; + u32 itr_val; - if (q_vector->itr_countdown > 0 || - (!ITR_IS_DYNAMIC(rx_itr_setting) && - !ITR_IS_DYNAMIC(tx_itr_setting))) { - goto enable_int; - } - - if (ITR_IS_DYNAMIC(tx_itr_setting)) { - rx = i40e_set_new_dynamic_itr(&q_vector->rx); - rxval = i40e_buildreg_itr(I40E_RX_ITR, q_vector->rx.itr); - } - - if (ITR_IS_DYNAMIC(tx_itr_setting)) { - tx = i40e_set_new_dynamic_itr(&q_vector->tx); - txval = i40e_buildreg_itr(I40E_TX_ITR, q_vector->tx.itr); + /* If we don't have MSIX, then we only need to re-enable icr0 */ + if (!test_bit(I40E_FLAG_MSIX_ENA, vsi->back->flags)) { + i40e_irq_dynamic_enable_icr0(vsi->back); + return; } - if (rx || tx) { - /* get the higher of the two ITR adjustments and - * use the same value for both ITR registers - * when in adaptive mode (Rx and/or Tx) - */ - u16 itr = max(q_vector->tx.itr, q_vector->rx.itr); - - q_vector->tx.itr = q_vector->rx.itr = itr; - txval = i40e_buildreg_itr(I40E_TX_ITR, itr); - tx = true; - rxval = i40e_buildreg_itr(I40E_RX_ITR, itr); - rx = true; - } + /* These will do nothing if dynamic updates are not enabled */ + i40e_update_itr(q_vector, &q_vector->tx); + i40e_update_itr(q_vector, &q_vector->rx); - /* only need to enable the interrupt once, but need - * to possibly update both ITR values + /* This block of logic allows us to get away with only updating + * one ITR value with each interrupt. The idea is to perform a + * pseudo-lazy update with the following criteria. + * + * 1. Rx is given higher priority than Tx if both are in same state + * 2. If we must reduce an ITR that is given highest priority. + * 3. We then give priority to increasing ITR based on amount. */ - if (rx) { - /* set the INTENA_MSK_MASK so that this first write - * won't actually enable the interrupt, instead just - * updating the ITR (it's bit 31 PF and VF) + if (q_vector->rx.target_itr < q_vector->rx.current_itr) { + /* Rx ITR needs to be reduced, this is highest priority */ + itr_idx = I40E_RX_ITR; + interval = q_vector->rx.target_itr; + q_vector->rx.current_itr = q_vector->rx.target_itr; + q_vector->itr_countdown = ITR_COUNTDOWN_START; + } else if ((q_vector->tx.target_itr < q_vector->tx.current_itr) || + ((q_vector->rx.target_itr - q_vector->rx.current_itr) < + (q_vector->tx.target_itr - q_vector->tx.current_itr))) { + /* Tx ITR needs to be reduced, this is second priority + * Tx ITR needs to be increased more than Rx, fourth priority */ - rxval |= BIT(31); - /* don't check _DOWN because interrupt isn't being enabled */ - wr32(hw, INTREG(vector - 1), rxval); + itr_idx = I40E_TX_ITR; + interval = q_vector->tx.target_itr; + q_vector->tx.current_itr = q_vector->tx.target_itr; + q_vector->itr_countdown = ITR_COUNTDOWN_START; + } else if (q_vector->rx.current_itr != q_vector->rx.target_itr) { + /* Rx ITR needs to be increased, third priority */ + itr_idx = I40E_RX_ITR; + interval = q_vector->rx.target_itr; + q_vector->rx.current_itr = q_vector->rx.target_itr; + q_vector->itr_countdown = ITR_COUNTDOWN_START; + } else { + /* No ITR update, lowest priority */ + if (q_vector->itr_countdown) + q_vector->itr_countdown--; } -enable_int: - if (!test_bit(__I40E_VSI_DOWN, vsi->state)) - wr32(hw, INTREG(vector - 1), txval); + /* Do not update interrupt control register if VSI is down */ + if (test_bit(__I40E_VSI_DOWN, vsi->state)) + return; - if (q_vector->itr_countdown) - q_vector->itr_countdown--; - else - q_vector->itr_countdown = ITR_COUNTDOWN_START; + /* Update ITR interval if necessary and enforce software interrupt + * if we are exiting busy poll. + */ + if (q_vector->in_busy_poll) { + itr_val = i40e_buildreg_itr(itr_idx, interval, true); + q_vector->in_busy_poll = false; + } else { + itr_val = i40e_buildreg_itr(itr_idx, interval, false); + } + wr32(hw, I40E_PFINT_DYN_CTLN(q_vector->reg_idx), itr_val); } /** @@ -2319,6 +2745,10 @@ int i40e_napi_poll(struct napi_struct *napi, int budget) container_of(napi, struct i40e_q_vector, napi); struct i40e_vsi *vsi = q_vector->vsi; struct i40e_ring *ring; + bool tx_clean_complete = true; + bool rx_clean_complete = true; + unsigned int tx_cleaned = 0; + unsigned int rx_cleaned = 0; bool clean_complete = true; bool arm_wb = false; int budget_per_ring; @@ -2333,8 +2763,12 @@ int i40e_napi_poll(struct napi_struct *napi, int budget) * budget and be more aggressive about cleaning up the Tx descriptors. */ i40e_for_each_ring(ring, q_vector->tx) { - if (!i40e_clean_tx_irq(vsi, ring, budget)) { - clean_complete = false; + bool wd = ring->xsk_pool ? + i40e_clean_xdp_tx_irq(vsi, ring) : + i40e_clean_tx_irq(vsi, ring, budget, &tx_cleaned); + + if (!wd) { + clean_complete = tx_clean_complete = false; continue; } arm_wb |= ring->arm_wb; @@ -2345,23 +2779,34 @@ int i40e_napi_poll(struct napi_struct *napi, int budget) if (budget <= 0) goto tx_only; - /* We attempt to distribute budget to each Rx queue fairly, but don't - * allow the budget to go below 1 because that would exit polling early. - */ - budget_per_ring = max(budget/q_vector->num_ringpairs, 1); + /* normally we have 1 Rx ring per q_vector */ + if (unlikely(q_vector->num_ringpairs > 1)) + /* We attempt to distribute budget to each Rx queue fairly, but + * don't allow the budget to go below 1 because that would exit + * polling early. + */ + budget_per_ring = max_t(int, budget / q_vector->num_ringpairs, 1); + else + /* Max of 1 Rx ring in this q_vector so give it the budget */ + budget_per_ring = budget; i40e_for_each_ring(ring, q_vector->rx) { - int cleaned = i40e_clean_rx_irq(ring, budget_per_ring); + int cleaned = ring->xsk_pool ? + i40e_clean_rx_irq_zc(ring, budget_per_ring) : + i40e_clean_rx_irq(ring, budget_per_ring, &rx_cleaned); work_done += cleaned; /* if we clean as many as budgeted, we must not be done */ if (cleaned >= budget_per_ring) - clean_complete = false; + clean_complete = rx_clean_complete = false; } + if (!i40e_enabled_xdp_vsi(vsi)) + trace_i40e_napi_poll(napi, q_vector, budget, budget_per_ring, rx_cleaned, + tx_cleaned, rx_clean_complete, tx_clean_complete); + /* If work not completed, return budget and polling will return */ if (!clean_complete) { - const cpumask_t *aff_mask = &q_vector->affinity_mask; int cpu_id = smp_processor_id(); /* It is possible that the interrupt affinity has changed but, @@ -2371,33 +2816,34 @@ int i40e_napi_poll(struct napi_struct *napi, int budget) * continue to poll, otherwise we must stop polling so the * interrupt can move to the correct cpu. */ - if (likely(cpumask_test_cpu(cpu_id, aff_mask) || - !(vsi->back->flags & I40E_FLAG_MSIX_ENABLED))) { + if (!cpumask_test_cpu(cpu_id, &q_vector->affinity_mask)) { + /* Tell napi that we are done polling */ + napi_complete_done(napi, work_done); + + /* Force an interrupt */ + i40e_force_wb(vsi, q_vector); + + /* Return budget-1 so that polling stops */ + return budget - 1; + } tx_only: - if (arm_wb) { - q_vector->tx.ring[0].tx_stats.tx_force_wb++; - i40e_enable_wb_on_itr(vsi, q_vector); - } - return budget; + if (arm_wb) { + q_vector->tx.ring[0].tx_stats.tx_force_wb++; + i40e_enable_wb_on_itr(vsi, q_vector); } + return budget; } - if (vsi->back->flags & I40E_TXR_FLAGS_WB_ON_ITR) + if (q_vector->tx.ring[0].flags & I40E_TXR_FLAGS_WB_ON_ITR) q_vector->arm_wb_state = false; - /* Work is done so exit the polling mode and re-enable the interrupt */ - napi_complete_done(napi, work_done); - - /* If we're prematurely stopping polling to fix the interrupt - * affinity we want to make sure polling starts back up so we - * issue a call to i40e_force_wb which triggers a SW interrupt. + /* Exit the polling mode, but don't re-enable interrupts if stack might + * poll us due to busy-polling */ - if (!clean_complete) - i40e_force_wb(vsi, q_vector); - else if (!(vsi->back->flags & I40E_FLAG_MSIX_ENABLED)) - i40e_irq_dynamic_enable_icr0(vsi->back, false); - else + if (likely(napi_complete_done(napi, work_done))) i40e_update_enable_itr(vsi, q_vector); + else + q_vector->in_busy_poll = true; return min(work_done, budget - 1); } @@ -2425,10 +2871,10 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb, u16 i; /* make sure ATR is enabled */ - if (!(pf->flags & I40E_FLAG_FD_ATR_ENABLED)) + if (!test_bit(I40E_FLAG_FD_ATR_ENA, pf->flags)) return; - if (pf->flags & I40E_FLAG_FD_ATR_AUTO_DISABLED) + if (test_bit(__I40E_FD_ATR_AUTO_DISABLED, pf->state)) return; /* if sampling is disabled do nothing */ @@ -2451,9 +2897,15 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb, hlen = (hdr.network[0] & 0x0F) << 2; l4_proto = hdr.ipv4->protocol; } else { - hlen = hdr.network - skb->data; - l4_proto = ipv6_find_hdr(skb, &hlen, IPPROTO_TCP, NULL, NULL); - hlen -= hdr.network - skb->data; + /* find the start of the innermost ipv6 header */ + unsigned int inner_hlen = hdr.network - skb->data; + unsigned int h_offset = inner_hlen; + + /* this function updates h_offset to the end of the header */ + l4_proto = + ipv6_find_hdr(skb, &h_offset, IPPROTO_TCP, NULL, NULL); + /* hlen will contain our best estimate of the tcp header */ + hlen = h_offset - inner_hlen; } if (l4_proto != IPPROTO_TCP) @@ -2462,9 +2914,9 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb, th = (struct tcphdr *)(hdr.network + hlen); /* Due to lack of space, no more new filters can be programmed */ - if (th->syn && (pf->flags & I40E_FLAG_FD_ATR_AUTO_DISABLED)) + if (th->syn && test_bit(__I40E_FD_ATR_AUTO_DISABLED, pf->state)) return; - if (pf->flags & I40E_FLAG_HW_ATR_EVICT_ENABLED) { + if (test_bit(I40E_FLAG_HW_ATR_EVICT_ENA, pf->flags)) { /* HW ATR eviction will take care of removing filters on FIN * and RST packets. */ @@ -2490,12 +2942,12 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb, i++; tx_ring->next_to_use = (i < tx_ring->count) ? i : 0; - flex_ptype = (tx_ring->queue_index << I40E_TXD_FLTR_QW0_QINDEX_SHIFT) & - I40E_TXD_FLTR_QW0_QINDEX_MASK; + flex_ptype = FIELD_PREP(I40E_TXD_FLTR_QW0_QINDEX_MASK, + tx_ring->queue_index); flex_ptype |= (tx_flags & I40E_TX_FLAGS_IPV4) ? - (I40E_FILTER_PCTYPE_NONF_IPV4_TCP << + (LIBIE_FILTER_PCTYPE_NONF_IPV4_TCP << I40E_TXD_FLTR_QW0_PCTYPE_SHIFT) : - (I40E_FILTER_PCTYPE_NONF_IPV6_TCP << + (LIBIE_FILTER_PCTYPE_NONF_IPV6_TCP << I40E_TXD_FLTR_QW0_PCTYPE_SHIFT); flex_ptype |= tx_ring->vsi->id << I40E_TXD_FLTR_QW0_DEST_VSI_SHIFT; @@ -2517,16 +2969,14 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb, dtype_cmd |= I40E_TXD_FLTR_QW1_CNT_ENA_MASK; if (!(tx_flags & I40E_TX_FLAGS_UDP_TUNNEL)) dtype_cmd |= - ((u32)I40E_FD_ATR_STAT_IDX(pf->hw.pf_id) << - I40E_TXD_FLTR_QW1_CNTINDEX_SHIFT) & - I40E_TXD_FLTR_QW1_CNTINDEX_MASK; + FIELD_PREP(I40E_TXD_FLTR_QW1_CNTINDEX_MASK, + I40E_FD_ATR_STAT_IDX(pf->hw.pf_id)); else dtype_cmd |= - ((u32)I40E_FD_ATR_TUNNEL_STAT_IDX(pf->hw.pf_id) << - I40E_TXD_FLTR_QW1_CNTINDEX_SHIFT) & - I40E_TXD_FLTR_QW1_CNTINDEX_MASK; + FIELD_PREP(I40E_TXD_FLTR_QW1_CNTINDEX_MASK, + I40E_FD_ATR_TUNNEL_STAT_IDX(pf->hw.pf_id)); - if (pf->flags & I40E_FLAG_HW_ATR_EVICT_ENABLED) + if (test_bit(I40E_FLAG_HW_ATR_EVICT_ENA, pf->flags)) dtype_cmd |= I40E_TXD_FLTR_QW1_ATR_MASK; fdir_desc->qindex_flex_ptype_vsi = cpu_to_le32(flex_ptype); @@ -2584,7 +3034,7 @@ static inline int i40e_tx_prepare_vlan_flags(struct sk_buff *skb, tx_flags |= I40E_TX_FLAGS_SW_VLAN; } - if (!(tx_ring->vsi->back->flags & I40E_FLAG_DCB_ENABLED)) + if (!test_bit(I40E_FLAG_DCB_ENA, tx_ring->vsi->back->flags)) goto out; /* Insert 802.1p priority into VLAN header */ @@ -2600,7 +3050,7 @@ static inline int i40e_tx_prepare_vlan_flags(struct sk_buff *skb, rc = skb_cow_head(skb, 0); if (rc < 0) return rc; - vhdr = (struct vlan_ethhdr *)skb->data; + vhdr = skb_vlan_eth_hdr(skb); vhdr->h_vlan_TCI = htons(tx_flags >> I40E_TX_FLAGS_VLAN_SHIFT); } else { @@ -2626,6 +3076,7 @@ static int i40e_tso(struct i40e_tx_buffer *first, u8 *hdr_len, { struct sk_buff *skb = first->skb; u64 cd_cmd, cd_tso_len, cd_mss; + __be16 protocol; union { struct iphdr *v4; struct ipv6hdr *v6; @@ -2637,7 +3088,7 @@ static int i40e_tso(struct i40e_tx_buffer *first, u8 *hdr_len, unsigned char *hdr; } l4; u32 paylen, l4_offset; - u16 gso_segs, gso_size; + u16 gso_size; int err; if (skb->ip_summed != CHECKSUM_PARTIAL) @@ -2650,15 +3101,23 @@ static int i40e_tso(struct i40e_tx_buffer *first, u8 *hdr_len, if (err < 0) return err; - ip.hdr = skb_network_header(skb); - l4.hdr = skb_transport_header(skb); + protocol = vlan_get_protocol(skb); + + if (eth_p_mpls(protocol)) + ip.hdr = skb_inner_network_header(skb); + else + ip.hdr = skb_network_header(skb); + l4.hdr = skb_checksum_start(skb); /* initialize outer IP header fields */ if (ip.v4->version == 4) { ip.v4->tot_len = 0; ip.v4->check = 0; + + first->tx_flags |= I40E_TX_FLAGS_TSO; } else { ip.v6->payload_len = 0; + first->tx_flags |= I40E_TX_FLAGS_TSO; } if (skb_shinfo(skb)->gso_type & (SKB_GSO_GRE | @@ -2698,17 +3157,22 @@ static int i40e_tso(struct i40e_tx_buffer *first, u8 *hdr_len, /* 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 segmentation header */ + *hdr_len = sizeof(*l4.udp) + l4_offset; + } else { + csum_replace_by_diff(&l4.tcp->check, (__force __wsum)htonl(paylen)); + /* compute length of segmentation header */ + *hdr_len = (l4.tcp->doff * 4) + l4_offset; + } /* pull values out of skb_shinfo */ gso_size = skb_shinfo(skb)->gso_size; - gso_segs = skb_shinfo(skb)->gso_segs; /* update GSO size and bytecount with header size */ - first->gso_segs = gso_segs; + first->gso_segs = skb_shinfo(skb)->gso_segs; first->bytecount += (first->gso_segs - 1) * *hdr_len; /* find the field values */ @@ -2746,7 +3210,7 @@ static int i40e_tsyn(struct i40e_ring *tx_ring, struct sk_buff *skb, * we are not already transmitting a packet to be timestamped */ pf = i40e_netdev_to_pf(tx_ring->netdev); - if (!(pf->flags & I40E_FLAG_PTP)) + if (!test_bit(I40E_FLAG_PTP_ENA, pf->flags)) return 0; if (pf->ptp_tx && @@ -2792,13 +3256,29 @@ static int i40e_tx_enable_csum(struct sk_buff *skb, u32 *tx_flags, unsigned char *exthdr; u32 offset, cmd = 0; __be16 frag_off; + __be16 protocol; u8 l4_proto = 0; if (skb->ip_summed != CHECKSUM_PARTIAL) return 0; - ip.hdr = skb_network_header(skb); - l4.hdr = skb_transport_header(skb); + protocol = vlan_get_protocol(skb); + + if (eth_p_mpls(protocol)) { + ip.hdr = skb_inner_network_header(skb); + l4.hdr = skb_checksum_start(skb); + } else { + ip.hdr = skb_network_header(skb); + l4.hdr = skb_transport_header(skb); + } + + /* set the tx_flags to indicate the IP protocol type. this is + * required so that checksum header computation below is accurate. + */ + if (ip.v4->version == 4) + *tx_flags |= I40E_TX_FLAGS_IPV4; + else + *tx_flags |= I40E_TX_FLAGS_IPV6; /* compute outer L2 header size */ offset = ((ip.hdr - skb->data) / 2) << I40E_TX_DESC_LENGTH_MACLEN_SHIFT; @@ -2813,13 +3293,16 @@ static int i40e_tx_enable_csum(struct sk_buff *skb, u32 *tx_flags, l4_proto = ip.v4->protocol; } else if (*tx_flags & I40E_TX_FLAGS_IPV6) { + int ret; + tunnel |= I40E_TX_CTX_EXT_IP_IPV6; exthdr = ip.hdr + sizeof(*ip.v6); l4_proto = ip.v6->nexthdr; - if (l4.hdr != exthdr) - ipv6_skip_exthdr(skb, exthdr - skb->data, - &l4_proto, &frag_off); + ret = ipv6_skip_exthdr(skb, exthdr - skb->data, + &l4_proto, &frag_off); + if (ret < 0) + return -1; } /* define outer transport */ @@ -2932,7 +3415,7 @@ static int i40e_tx_enable_csum(struct sk_buff *skb, u32 *tx_flags, } /** - * i40e_create_tx_ctx Build the Tx context descriptor + * i40e_create_tx_ctx - Build the Tx context descriptor * @tx_ring: ring to create the descriptor on * @cd_type_cmd_tso_mss: Quad Word 1 * @cd_tunneling: Quad Word 0 - bits 0-31 @@ -2975,6 +3458,8 @@ int __i40e_maybe_stop_tx(struct i40e_ring *tx_ring, int size) /* Memory barrier before checking head and tail */ smp_mb(); + ++tx_ring->tx_stats.tx_stopped; + /* Check again in a case another CPU has just made room available. */ if (likely(I40E_DESC_UNUSED(tx_ring) < size)) return -EBUSY; @@ -3000,7 +3485,7 @@ int __i40e_maybe_stop_tx(struct i40e_ring *tx_ring, int size) **/ bool __i40e_chk_linearize(struct sk_buff *skb) { - const struct skb_frag_struct *frag, *stale; + const skb_frag_t *frag, *stale; int nr_frags, sum; /* no need to check if number of frags is less than 7 */ @@ -3032,10 +3517,30 @@ bool __i40e_chk_linearize(struct sk_buff *skb) /* Walk through fragments adding latest fragment, testing it, and * then removing stale fragments from the sum. */ - stale = &skb_shinfo(skb)->frags[0]; - for (;;) { + for (stale = &skb_shinfo(skb)->frags[0];; stale++) { + int stale_size = skb_frag_size(stale); + sum += skb_frag_size(frag++); + /* The stale fragment may present us with a smaller + * descriptor than the actual fragment size. To account + * for that we need to remove all the data on the front and + * figure out what the remainder would be in the last + * descriptor associated with the fragment. + */ + if (stale_size > I40E_MAX_DATA_PER_TXD) { + int align_pad = -(skb_frag_off(stale)) & + (I40E_MAX_READ_REQ_SIZE - 1); + + sum -= align_pad; + stale_size -= align_pad; + + do { + sum -= I40E_MAX_DATA_PER_TXD_ALIGNED; + stale_size -= I40E_MAX_DATA_PER_TXD_ALIGNED; + } while (stale_size > I40E_MAX_DATA_PER_TXD); + } + /* if sum is negative we failed to make sufficient progress */ if (sum < 0) return true; @@ -3043,7 +3548,7 @@ bool __i40e_chk_linearize(struct sk_buff *skb) if (!nr_frags--) break; - sum -= skb_frag_size(stale++); + sum -= stale_size; } return false; @@ -3067,7 +3572,7 @@ static inline int i40e_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb, { unsigned int data_len = skb->data_len; unsigned int size = skb_headlen(skb); - struct skb_frag_struct *frag; + skb_frag_t *frag; struct i40e_tx_buffer *tx_bi; struct i40e_tx_desc *tx_desc; u16 i = tx_ring->next_to_use; @@ -3077,8 +3582,7 @@ static inline int i40e_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb, if (tx_flags & I40E_TX_FLAGS_HW_VLAN) { td_cmd |= I40E_TX_DESC_CMD_IL2TAG1; - td_tag = (tx_flags & I40E_TX_FLAGS_VLAN_MASK) >> - I40E_TX_FLAGS_VLAN_SHIFT; + td_tag = FIELD_GET(I40E_TX_FLAGS_VLAN_MASK, tx_flags); } first->tx_flags = tx_flags; @@ -3160,38 +3664,12 @@ static inline int i40e_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb, /* write last descriptor with EOP bit */ td_cmd |= I40E_TX_DESC_CMD_EOP; - /* We can OR these values together as they both are checked against - * 4 below and at this point desc_count will be used as a boolean value - * after this if/else block. + /* We OR these values together to check both against 4 (WB_STRIDE) + * below. This is safe since we don't re-use desc_count afterwards. */ desc_count |= ++tx_ring->packet_stride; - /* Algorithm to optimize tail and RS bit setting: - * if queue is stopped - * mark RS bit - * reset packet counter - * else if xmit_more is supported and is true - * advance packet counter to 4 - * reset desc_count to 0 - * - * if desc_count >= 4 - * mark RS bit - * reset packet counter - * if desc_count > 0 - * update tail - * - * Note: If there are less than 4 descriptors - * pending and interrupts were disabled the service task will - * trigger a force WB. - */ - if (netif_xmit_stopped(txring_txq(tx_ring))) { - goto do_rs; - } else if (skb->xmit_more) { - /* set stride to arm on next packet and reset desc_count */ - tx_ring->packet_stride = WB_STRIDE; - desc_count = 0; - } else if (desc_count >= WB_STRIDE) { -do_rs: + if (desc_count >= WB_STRIDE) { /* write last descriptor with RS bit set */ td_cmd |= I40E_TX_DESC_CMD_RS; tx_ring->packet_stride = 0; @@ -3200,6 +3678,8 @@ do_rs: tx_desc->cmd_type_offset_bsz = build_ctob(td_cmd, td_offset, size, td_tag); + skb_tx_timestamp(skb); + /* Force memory writes to complete before letting h/w know there * are new descriptors to fetch. * @@ -3212,13 +3692,8 @@ do_rs: first->next_to_watch = tx_desc; /* notify HW of packet */ - if (desc_count) { + if (netif_xmit_stopped(txring_txq(tx_ring)) || !netdev_xmit_more()) { writel(i, tx_ring->tail); - - /* we need this if more than one processor can write to our tail - * at a time, it synchronizes IO on IA64/Altix systems - */ - mmiowb(); } return 0; @@ -3242,57 +3717,144 @@ dma_error: return -1; } +static u16 i40e_swdcb_skb_tx_hash(struct net_device *dev, + const struct sk_buff *skb, + u16 num_tx_queues) +{ + u32 jhash_initval_salt = 0xd631614b; + u32 hash; + + if (skb->sk && skb->sk->sk_hash) + hash = skb->sk->sk_hash; + else + hash = (__force u16)skb->protocol ^ skb->hash; + + hash = jhash_1word(hash, jhash_initval_salt); + + return (u16)(((u64)hash * num_tx_queues) >> 32); +} + +u16 i40e_lan_select_queue(struct net_device *netdev, + struct sk_buff *skb, + struct net_device __always_unused *sb_dev) +{ + struct i40e_netdev_priv *np = netdev_priv(netdev); + struct i40e_vsi *vsi = np->vsi; + struct i40e_hw *hw; + u16 qoffset; + u16 qcount; + u8 tclass; + u16 hash; + u8 prio; + + /* is DCB enabled at all? */ + if (vsi->tc_config.numtc == 1 || + i40e_is_tc_mqprio_enabled(vsi->back)) + return netdev_pick_tx(netdev, skb, sb_dev); + + prio = skb->priority; + hw = &vsi->back->hw; + tclass = hw->local_dcbx_config.etscfg.prioritytable[prio]; + /* sanity check */ + if (unlikely(!(vsi->tc_config.enabled_tc & BIT(tclass)))) + tclass = 0; + + /* select a queue assigned for the given TC */ + qcount = vsi->tc_config.tc_info[tclass].qcount; + hash = i40e_swdcb_skb_tx_hash(netdev, skb, qcount); + + qoffset = vsi->tc_config.tc_info[tclass].qoffset; + return qoffset + hash; +} + /** * i40e_xmit_xdp_ring - transmits an XDP buffer to an XDP Tx ring - * @xdp: data to transmit + * @xdpf: data to transmit * @xdp_ring: XDP Tx ring **/ -static int i40e_xmit_xdp_ring(struct xdp_buff *xdp, +static int i40e_xmit_xdp_ring(struct xdp_frame *xdpf, struct i40e_ring *xdp_ring) { - u32 size = xdp->data_end - xdp->data; - u16 i = xdp_ring->next_to_use; - struct i40e_tx_buffer *tx_bi; - struct i40e_tx_desc *tx_desc; - dma_addr_t dma; - - if (!unlikely(I40E_DESC_UNUSED(xdp_ring))) { + struct skb_shared_info *sinfo = xdp_get_shared_info_from_frame(xdpf); + u8 nr_frags = unlikely(xdp_frame_has_frags(xdpf)) ? sinfo->nr_frags : 0; + u16 i = 0, index = xdp_ring->next_to_use; + struct i40e_tx_buffer *tx_head = &xdp_ring->tx_bi[index]; + struct i40e_tx_buffer *tx_bi = tx_head; + struct i40e_tx_desc *tx_desc = I40E_TX_DESC(xdp_ring, index); + void *data = xdpf->data; + u32 size = xdpf->len; + + if (unlikely(I40E_DESC_UNUSED(xdp_ring) < 1 + nr_frags)) { xdp_ring->tx_stats.tx_busy++; return I40E_XDP_CONSUMED; } - dma = dma_map_single(xdp_ring->dev, xdp->data, size, DMA_TO_DEVICE); - if (dma_mapping_error(xdp_ring->dev, dma)) - return I40E_XDP_CONSUMED; + tx_head->bytecount = xdp_get_frame_len(xdpf); + tx_head->gso_segs = 1; + tx_head->xdpf = xdpf; - tx_bi = &xdp_ring->tx_bi[i]; - tx_bi->bytecount = size; - tx_bi->gso_segs = 1; - tx_bi->raw_buf = xdp->data; + for (;;) { + dma_addr_t dma; - /* record length, and DMA address */ - dma_unmap_len_set(tx_bi, len, size); - dma_unmap_addr_set(tx_bi, dma, dma); + dma = dma_map_single(xdp_ring->dev, data, size, DMA_TO_DEVICE); + if (dma_mapping_error(xdp_ring->dev, dma)) + goto unmap; - tx_desc = I40E_TX_DESC(xdp_ring, i); - tx_desc->buffer_addr = cpu_to_le64(dma); - tx_desc->cmd_type_offset_bsz = build_ctob(I40E_TX_DESC_CMD_ICRC - | I40E_TXD_CMD, - 0, size, 0); + /* record length, and DMA address */ + dma_unmap_len_set(tx_bi, len, size); + dma_unmap_addr_set(tx_bi, dma, dma); + + tx_desc->buffer_addr = cpu_to_le64(dma); + tx_desc->cmd_type_offset_bsz = + build_ctob(I40E_TX_DESC_CMD_ICRC, 0, size, 0); + + if (++index == xdp_ring->count) + index = 0; + + if (i == nr_frags) + break; + + tx_bi = &xdp_ring->tx_bi[index]; + tx_desc = I40E_TX_DESC(xdp_ring, index); + + data = skb_frag_address(&sinfo->frags[i]); + size = skb_frag_size(&sinfo->frags[i]); + i++; + } + + tx_desc->cmd_type_offset_bsz |= + cpu_to_le64(I40E_TXD_CMD << I40E_TXD_QW1_CMD_SHIFT); /* Make certain all of the status bits have been updated * before next_to_watch is written. */ smp_wmb(); - i++; - if (i == xdp_ring->count) - i = 0; + xdp_ring->xdp_tx_active++; - tx_bi->next_to_watch = tx_desc; - xdp_ring->next_to_use = i; + tx_head->next_to_watch = tx_desc; + xdp_ring->next_to_use = index; return I40E_XDP_TX; + +unmap: + for (;;) { + tx_bi = &xdp_ring->tx_bi[index]; + if (dma_unmap_len(tx_bi, len)) + dma_unmap_page(xdp_ring->dev, + dma_unmap_addr(tx_bi, dma), + dma_unmap_len(tx_bi, len), + DMA_TO_DEVICE); + dma_unmap_len_set(tx_bi, len, 0); + if (tx_bi == tx_head) + break; + + if (!index) + index += xdp_ring->count; + index--; + } + + return I40E_XDP_CONSUMED; } /** @@ -3310,7 +3872,6 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb, struct i40e_tx_buffer *first; u32 td_offset = 0; u32 tx_flags = 0; - __be16 protocol; u32 td_cmd = 0; u8 hdr_len = 0; int tso, count; @@ -3352,15 +3913,6 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb, if (i40e_tx_prepare_vlan_flags(skb, tx_ring, &tx_flags)) goto out_drop; - /* obtain protocol of skb */ - protocol = vlan_get_protocol(skb); - - /* setup IPv4/IPv6 offloads */ - if (protocol == htons(ETH_P_IP)) - tx_flags |= I40E_TX_FLAGS_IPV4; - else if (protocol == htons(ETH_P_IPV6)) - tx_flags |= I40E_TX_FLAGS_IPV6; - tso = i40e_tso(first, &hdr_len, &cd_type_cmd_tso_mss); if (tso < 0) @@ -3379,8 +3931,6 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb, if (tsyn) tx_flags |= I40E_TX_FLAGS_TSYN; - skb_tx_timestamp(skb); - /* always enable CRC insertion offload */ td_cmd |= I40E_TX_DESC_CMD_ICRC; @@ -3436,3 +3986,55 @@ netdev_tx_t i40e_lan_xmit_frame(struct sk_buff *skb, struct net_device *netdev) return i40e_xmit_frame_ring(skb, tx_ring); } + +/** + * i40e_xdp_xmit - Implements ndo_xdp_xmit + * @dev: netdev + * @n: number of frames + * @frames: array of XDP buffer pointers + * @flags: XDP extra info + * + * Returns number of frames successfully sent. Failed frames + * will be free'ed by XDP core. + * + * For error cases, a negative errno code is returned and no-frames + * are transmitted (caller must handle freeing frames). + **/ +int i40e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, + u32 flags) +{ + struct i40e_netdev_priv *np = netdev_priv(dev); + unsigned int queue_index = smp_processor_id(); + struct i40e_vsi *vsi = np->vsi; + struct i40e_pf *pf = vsi->back; + struct i40e_ring *xdp_ring; + int nxmit = 0; + int i; + + if (test_bit(__I40E_VSI_DOWN, vsi->state)) + return -ENETDOWN; + + if (!i40e_enabled_xdp_vsi(vsi) || queue_index >= vsi->num_queue_pairs || + test_bit(__I40E_CONFIG_BUSY, pf->state)) + return -ENXIO; + + if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK)) + return -EINVAL; + + xdp_ring = vsi->xdp_rings[queue_index]; + + for (i = 0; i < n; i++) { + struct xdp_frame *xdpf = frames[i]; + int err; + + err = i40e_xmit_xdp_ring(xdpf, xdp_ring); + if (err != I40E_XDP_TX) + break; + nxmit++; + } + + if (unlikely(flags & XDP_XMIT_FLUSH)) + i40e_xdp_ring_update_tail(xdp_ring); + + return nxmit; +} |