diff options
Diffstat (limited to 'drivers/net/ethernet/intel/i40e/i40e_txrx.c')
| -rw-r--r-- | drivers/net/ethernet/intel/i40e/i40e_txrx.c | 212 |
1 files changed, 131 insertions, 81 deletions
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c index 10a83e5385c7..b97c95f89fa0 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c @@ -3,6 +3,7 @@ #include <linux/prefetch.h> #include <linux/bpf_trace.h> +#include <net/mpls.h> #include <net/xdp.h> #include "i40e.h" #include "i40e_trace.h" @@ -371,7 +372,6 @@ static void i40e_change_filter_num(bool ipv4, bool add, u16 *ipv4_filter_num, } } -#define IP_HEADER_OFFSET 14 #define I40E_UDPIP_DUMMY_PACKET_LEN 42 #define I40E_UDPIP6_DUMMY_PACKET_LEN 62 /** @@ -830,8 +830,6 @@ void i40e_free_tx_resources(struct i40e_ring *tx_ring) i40e_clean_tx_ring(tx_ring); kfree(tx_ring->tx_bi); tx_ring->tx_bi = NULL; - kfree(tx_ring->xsk_descs); - tx_ring->xsk_descs = NULL; if (tx_ring->desc) { dma_free_coherent(tx_ring->dev, tx_ring->size, @@ -1382,8 +1380,6 @@ static void i40e_reuse_rx_page(struct i40e_ring *rx_ring, new_buff->page_offset = old_buff->page_offset; new_buff->pagecnt_bias = old_buff->pagecnt_bias; - rx_ring->rx_stats.page_reuse_count++; - /* clear contents of buffer_info */ old_buff->page = NULL; } @@ -1433,13 +1429,6 @@ int i40e_setup_tx_descriptors(struct i40e_ring *tx_ring) if (!tx_ring->tx_bi) goto err; - if (ring_is_xdp(tx_ring)) { - tx_ring->xsk_descs = kcalloc(I40E_MAX_NUM_DESCRIPTORS, sizeof(*tx_ring->xsk_descs), - GFP_KERNEL); - if (!tx_ring->xsk_descs) - goto err; - } - u64_stats_init(&tx_ring->syncp); /* round up to nearest 4K */ @@ -1463,21 +1452,11 @@ int i40e_setup_tx_descriptors(struct i40e_ring *tx_ring) return 0; err: - kfree(tx_ring->xsk_descs); - tx_ring->xsk_descs = NULL; kfree(tx_ring->tx_bi); tx_ring->tx_bi = NULL; return -ENOMEM; } -int i40e_alloc_rx_bi(struct i40e_ring *rx_ring) -{ - unsigned long sz = sizeof(*rx_ring->rx_bi) * rx_ring->count; - - rx_ring->rx_bi = kzalloc(sz, GFP_KERNEL); - return rx_ring->rx_bi ? 0 : -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); @@ -1495,10 +1474,8 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring) if (!rx_ring->rx_bi) return; - if (rx_ring->skb) { - dev_kfree_skb(rx_ring->skb); - rx_ring->skb = NULL; - } + dev_kfree_skb(rx_ring->skb); + rx_ring->skb = NULL; if (rx_ring->xsk_pool) { i40e_xsk_clean_rx_ring(rx_ring); @@ -1608,6 +1585,11 @@ int i40e_setup_rx_descriptors(struct i40e_ring *rx_ring) 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; } @@ -1675,6 +1657,8 @@ 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), @@ -1982,32 +1966,43 @@ static bool i40e_cleanup_headers(struct i40e_ring *rx_ring, struct sk_buff *skb, /** * 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 * @rx_buffer_pgcnt: buffer page refcount pre xdp_do_redirect() call * * 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 + * page freed. + * + * 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, int rx_buffer_pgcnt) { unsigned int pagecnt_bias = rx_buffer->pagecnt_bias; struct page *page = rx_buffer->page; /* Is any reuse possible? */ - if (!dev_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((rx_buffer_pgcnt - pagecnt_bias) > 1)) + if (unlikely((rx_buffer_pgcnt - 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 @@ -2204,7 +2199,7 @@ static struct sk_buff *i40e_build_skb(struct i40e_ring *rx_ring, net_prefetch(xdp->data_meta); /* build an skb around the page buffer */ - skb = build_skb(xdp->data_hard_start, truesize); + skb = napi_build_skb(xdp->data_hard_start, truesize); if (unlikely(!skb)) return NULL; @@ -2237,7 +2232,7 @@ static void i40e_put_rx_buffer(struct i40e_ring *rx_ring, struct i40e_rx_buffer *rx_buffer, int rx_buffer_pgcnt) { - if (i40e_can_reuse_rx_page(rx_buffer, rx_buffer_pgcnt)) { + if (i40e_can_reuse_rx_page(rx_buffer, &rx_ring->rx_stats, rx_buffer_pgcnt)) { /* hand second half of page back to the ring */ i40e_reuse_rx_page(rx_ring, rx_buffer); } else { @@ -2290,16 +2285,14 @@ int i40e_xmit_xdp_tx_ring(struct xdp_buff *xdp, struct i40e_ring *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 int 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 err, result = I40E_XDP_PASS; struct i40e_ring *xdp_ring; - struct bpf_prog *xdp_prog; u32 act; - xdp_prog = READ_ONCE(rx_ring->xdp_prog); - if (!xdp_prog) goto xdp_out; @@ -2322,7 +2315,7 @@ static int i40e_run_xdp(struct i40e_ring *rx_ring, struct xdp_buff *xdp) 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: @@ -2444,6 +2437,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget) unsigned int offset = rx_ring->rx_offset; struct sk_buff *skb = rx_ring->skb; unsigned int xdp_xmit = 0; + struct bpf_prog *xdp_prog; bool failure = false; struct xdp_buff xdp; int xdp_res = 0; @@ -2453,6 +2447,8 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget) #endif xdp_init_buff(&xdp, frame_sz, &rx_ring->xdp_rxq); + xdp_prog = READ_ONCE(rx_ring->xdp_prog); + while (likely(total_rx_packets < (unsigned int)budget)) { struct i40e_rx_buffer *rx_buffer; union i40e_rx_desc *rx_desc; @@ -2508,11 +2504,12 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget) hard_start = page_address(rx_buffer->page) + rx_buffer->page_offset - offset; xdp_prepare_buff(&xdp, hard_start, offset, size, true); + xdp_buff_clear_frags_flag(&xdp); #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 - xdp_res = i40e_run_xdp(rx_ring, &xdp); + xdp_res = i40e_run_xdp(rx_ring, &xdp, xdp_prog); } if (xdp_res) { @@ -3015,6 +3012,7 @@ 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; @@ -3026,7 +3024,7 @@ 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) @@ -3039,15 +3037,23 @@ 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 | @@ -3100,10 +3106,9 @@ static int i40e_tso(struct i40e_tx_buffer *first, u8 *hdr_len, /* 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 */ @@ -3187,13 +3192,29 @@ 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; @@ -3373,6 +3394,8 @@ 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; @@ -3662,7 +3685,8 @@ u16 i40e_lan_select_queue(struct net_device *netdev, u8 prio; /* is DCB enabled at all? */ - if (vsi->tc_config.numtc == 1) + if (vsi->tc_config.numtc == 1 || + i40e_is_tc_mqprio_enabled(vsi->back)) return netdev_pick_tx(netdev, skb, sb_dev); prio = skb->priority; @@ -3688,35 +3712,55 @@ u16 i40e_lan_select_queue(struct net_device *netdev, static int i40e_xmit_xdp_ring(struct xdp_frame *xdpf, struct i40e_ring *xdp_ring) { - u16 i = xdp_ring->next_to_use; - struct i40e_tx_buffer *tx_bi; - struct i40e_tx_desc *tx_desc; + 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; - dma_addr_t dma; - if (!unlikely(I40E_DESC_UNUSED(xdp_ring))) { + 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, data, size, DMA_TO_DEVICE); - if (dma_mapping_error(xdp_ring->dev, dma)) - return I40E_XDP_CONSUMED; - tx_bi = &xdp_ring->tx_bi[i]; - tx_bi->bytecount = size; - tx_bi->gso_segs = 1; - tx_bi->xdpf = xdpf; + tx_head->bytecount = xdp_get_frame_len(xdpf); + tx_head->gso_segs = 1; + tx_head->xdpf = xdpf; - /* record length, and DMA address */ - dma_unmap_len_set(tx_bi, len, size); - dma_unmap_addr_set(tx_bi, dma, dma); + for (;;) { + dma_addr_t dma; - 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); + dma = dma_map_single(xdp_ring->dev, data, size, DMA_TO_DEVICE); + if (dma_mapping_error(xdp_ring->dev, dma)) + goto unmap; + + /* 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. @@ -3724,14 +3768,30 @@ static int i40e_xmit_xdp_ring(struct xdp_frame *xdpf, smp_wmb(); xdp_ring->xdp_tx_active++; - i++; - if (i == xdp_ring->count) - i = 0; - 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; } /** @@ -3749,7 +3809,6 @@ 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; @@ -3791,15 +3850,6 @@ 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) |