diff options
Diffstat (limited to 'drivers/net/ethernet/intel/idpf/idpf_txrx.c')
| -rw-r--r-- | drivers/net/ethernet/intel/idpf/idpf_txrx.c | 1874 |
1 files changed, 993 insertions, 881 deletions
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.c b/drivers/net/ethernet/intel/idpf/idpf_txrx.c index 2f8ad79ae3f0..977741c41498 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_txrx.c +++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.c @@ -1,7 +1,22 @@ // SPDX-License-Identifier: GPL-2.0-only /* Copyright (C) 2023 Intel Corporation */ +#include <net/libeth/rx.h> +#include <net/libeth/tx.h> + #include "idpf.h" +#include "idpf_virtchnl.h" + +struct idpf_tx_stash { + struct hlist_node hlist; + struct libeth_sqe buf; +}; + +#define idpf_tx_buf_compl_tag(buf) (*(u32 *)&(buf)->priv) +LIBETH_SQE_CHECK_PRIV(u32); + +static bool idpf_chk_linearize(struct sk_buff *skb, unsigned int max_bufs, + unsigned int count); /** * idpf_buf_lifo_push - push a buffer pointer onto stack @@ -55,39 +70,20 @@ void idpf_tx_timeout(struct net_device *netdev, unsigned int txqueue) } /** - * idpf_tx_buf_rel - Release a Tx buffer - * @tx_q: the queue that owns the buffer - * @tx_buf: the buffer to free - */ -static void idpf_tx_buf_rel(struct idpf_queue *tx_q, struct idpf_tx_buf *tx_buf) -{ - if (tx_buf->skb) { - if (dma_unmap_len(tx_buf, len)) - dma_unmap_single(tx_q->dev, - dma_unmap_addr(tx_buf, dma), - dma_unmap_len(tx_buf, len), - DMA_TO_DEVICE); - dev_kfree_skb_any(tx_buf->skb); - } else if (dma_unmap_len(tx_buf, len)) { - dma_unmap_page(tx_q->dev, - dma_unmap_addr(tx_buf, dma), - dma_unmap_len(tx_buf, len), - DMA_TO_DEVICE); - } - - tx_buf->next_to_watch = NULL; - tx_buf->skb = NULL; - tx_buf->compl_tag = IDPF_SPLITQ_TX_INVAL_COMPL_TAG; - dma_unmap_len_set(tx_buf, len, 0); -} - -/** * idpf_tx_buf_rel_all - Free any empty Tx buffers * @txq: queue to be cleaned */ -static void idpf_tx_buf_rel_all(struct idpf_queue *txq) +static void idpf_tx_buf_rel_all(struct idpf_tx_queue *txq) { - u16 i; + struct libeth_sq_napi_stats ss = { }; + struct idpf_buf_lifo *buf_stack; + struct idpf_tx_stash *stash; + struct libeth_cq_pp cp = { + .dev = txq->dev, + .ss = &ss, + }; + struct hlist_node *tmp; + u32 i, tag; /* Buffers already cleared, nothing to do */ if (!txq->tx_buf) @@ -95,44 +91,78 @@ static void idpf_tx_buf_rel_all(struct idpf_queue *txq) /* Free all the Tx buffer sk_buffs */ for (i = 0; i < txq->desc_count; i++) - idpf_tx_buf_rel(txq, &txq->tx_buf[i]); + libeth_tx_complete(&txq->tx_buf[i], &cp); kfree(txq->tx_buf); txq->tx_buf = NULL; - if (!txq->buf_stack.bufs) + if (!idpf_queue_has(FLOW_SCH_EN, txq)) return; - for (i = 0; i < txq->buf_stack.size; i++) - kfree(txq->buf_stack.bufs[i]); + buf_stack = &txq->stash->buf_stack; + if (!buf_stack->bufs) + return; + + /* + * If a Tx timeout occurred, there are potentially still bufs in the + * hash table, free them here. + */ + hash_for_each_safe(txq->stash->sched_buf_hash, tag, tmp, stash, + hlist) { + if (!stash) + continue; + + libeth_tx_complete(&stash->buf, &cp); + hash_del(&stash->hlist); + idpf_buf_lifo_push(buf_stack, stash); + } - kfree(txq->buf_stack.bufs); - txq->buf_stack.bufs = NULL; + for (i = 0; i < buf_stack->size; i++) + kfree(buf_stack->bufs[i]); + + kfree(buf_stack->bufs); + buf_stack->bufs = NULL; } /** * idpf_tx_desc_rel - Free Tx resources per queue * @txq: Tx descriptor ring for a specific queue - * @bufq: buffer q or completion q * * Free all transmit software resources */ -static void idpf_tx_desc_rel(struct idpf_queue *txq, bool bufq) +static void idpf_tx_desc_rel(struct idpf_tx_queue *txq) { - if (bufq) - idpf_tx_buf_rel_all(txq); + idpf_tx_buf_rel_all(txq); + netdev_tx_reset_subqueue(txq->netdev, txq->idx); if (!txq->desc_ring) return; dmam_free_coherent(txq->dev, txq->size, txq->desc_ring, txq->dma); txq->desc_ring = NULL; - txq->next_to_alloc = 0; txq->next_to_use = 0; txq->next_to_clean = 0; } /** + * idpf_compl_desc_rel - Free completion resources per queue + * @complq: completion queue + * + * Free all completion software resources. + */ +static void idpf_compl_desc_rel(struct idpf_compl_queue *complq) +{ + if (!complq->comp) + return; + + dma_free_coherent(complq->netdev->dev.parent, complq->size, + complq->comp, complq->dma); + complq->comp = NULL; + complq->next_to_use = 0; + complq->next_to_clean = 0; +} + +/** * idpf_tx_desc_rel_all - Free Tx Resources for All Queues * @vport: virtual port structure * @@ -149,10 +179,10 @@ static void idpf_tx_desc_rel_all(struct idpf_vport *vport) struct idpf_txq_group *txq_grp = &vport->txq_grps[i]; for (j = 0; j < txq_grp->num_txq; j++) - idpf_tx_desc_rel(txq_grp->txqs[j], true); + idpf_tx_desc_rel(txq_grp->txqs[j]); if (idpf_is_queue_model_split(vport->txq_model)) - idpf_tx_desc_rel(txq_grp->complq, false); + idpf_compl_desc_rel(txq_grp->complq); } } @@ -162,8 +192,9 @@ static void idpf_tx_desc_rel_all(struct idpf_vport *vport) * * Returns 0 on success, negative on failure */ -static int idpf_tx_buf_alloc_all(struct idpf_queue *tx_q) +static int idpf_tx_buf_alloc_all(struct idpf_tx_queue *tx_q) { + struct idpf_buf_lifo *buf_stack; int buf_size; int i; @@ -175,26 +206,26 @@ static int idpf_tx_buf_alloc_all(struct idpf_queue *tx_q) if (!tx_q->tx_buf) return -ENOMEM; - /* Initialize tx_bufs with invalid completion tags */ - for (i = 0; i < tx_q->desc_count; i++) - tx_q->tx_buf[i].compl_tag = IDPF_SPLITQ_TX_INVAL_COMPL_TAG; + if (!idpf_queue_has(FLOW_SCH_EN, tx_q)) + return 0; + + buf_stack = &tx_q->stash->buf_stack; /* Initialize tx buf stack for out-of-order completions if * flow scheduling offload is enabled */ - tx_q->buf_stack.bufs = - kcalloc(tx_q->desc_count, sizeof(struct idpf_tx_stash *), - GFP_KERNEL); - if (!tx_q->buf_stack.bufs) + buf_stack->bufs = kcalloc(tx_q->desc_count, sizeof(*buf_stack->bufs), + GFP_KERNEL); + if (!buf_stack->bufs) return -ENOMEM; - tx_q->buf_stack.size = tx_q->desc_count; - tx_q->buf_stack.top = tx_q->desc_count; + buf_stack->size = tx_q->desc_count; + buf_stack->top = tx_q->desc_count; for (i = 0; i < tx_q->desc_count; i++) { - tx_q->buf_stack.bufs[i] = kzalloc(sizeof(*tx_q->buf_stack.bufs[i]), - GFP_KERNEL); - if (!tx_q->buf_stack.bufs[i]) + buf_stack->bufs[i] = kzalloc(sizeof(*buf_stack->bufs[i]), + GFP_KERNEL); + if (!buf_stack->bufs[i]) return -ENOMEM; } @@ -203,28 +234,22 @@ static int idpf_tx_buf_alloc_all(struct idpf_queue *tx_q) /** * idpf_tx_desc_alloc - Allocate the Tx descriptors + * @vport: vport to allocate resources for * @tx_q: the tx ring to set up - * @bufq: buffer or completion queue * * Returns 0 on success, negative on failure */ -static int idpf_tx_desc_alloc(struct idpf_queue *tx_q, bool bufq) +static int idpf_tx_desc_alloc(const struct idpf_vport *vport, + struct idpf_tx_queue *tx_q) { struct device *dev = tx_q->dev; - u32 desc_sz; int err; - if (bufq) { - err = idpf_tx_buf_alloc_all(tx_q); - if (err) - goto err_alloc; - - desc_sz = sizeof(struct idpf_base_tx_desc); - } else { - desc_sz = sizeof(struct idpf_splitq_tx_compl_desc); - } + err = idpf_tx_buf_alloc_all(tx_q); + if (err) + goto err_alloc; - tx_q->size = tx_q->desc_count * desc_sz; + tx_q->size = tx_q->desc_count * sizeof(*tx_q->base_tx); /* Allocate descriptors also round up to nearest 4K */ tx_q->size = ALIGN(tx_q->size, 4096); @@ -237,20 +262,44 @@ static int idpf_tx_desc_alloc(struct idpf_queue *tx_q, bool bufq) goto err_alloc; } - tx_q->next_to_alloc = 0; tx_q->next_to_use = 0; tx_q->next_to_clean = 0; - set_bit(__IDPF_Q_GEN_CHK, tx_q->flags); + idpf_queue_set(GEN_CHK, tx_q); return 0; err_alloc: - idpf_tx_desc_rel(tx_q, bufq); + idpf_tx_desc_rel(tx_q); return err; } /** + * idpf_compl_desc_alloc - allocate completion descriptors + * @vport: vport to allocate resources for + * @complq: completion queue to set up + * + * Return: 0 on success, -errno on failure. + */ +static int idpf_compl_desc_alloc(const struct idpf_vport *vport, + struct idpf_compl_queue *complq) +{ + complq->size = array_size(complq->desc_count, sizeof(*complq->comp)); + + complq->comp = dma_alloc_coherent(complq->netdev->dev.parent, + complq->size, &complq->dma, + GFP_KERNEL); + if (!complq->comp) + return -ENOMEM; + + complq->next_to_use = 0; + complq->next_to_clean = 0; + idpf_queue_set(GEN_CHK, complq); + + return 0; +} + +/** * idpf_tx_desc_alloc_all - allocate all queues Tx resources * @vport: virtual port private structure * @@ -258,7 +307,6 @@ err_alloc: */ static int idpf_tx_desc_alloc_all(struct idpf_vport *vport) { - struct device *dev = &vport->adapter->pdev->dev; int err = 0; int i, j; @@ -267,13 +315,14 @@ static int idpf_tx_desc_alloc_all(struct idpf_vport *vport) */ for (i = 0; i < vport->num_txq_grp; i++) { for (j = 0; j < vport->txq_grps[i].num_txq; j++) { - struct idpf_queue *txq = vport->txq_grps[i].txqs[j]; + struct idpf_tx_queue *txq = vport->txq_grps[i].txqs[j]; u8 gen_bits = 0; u16 bufidx_mask; - err = idpf_tx_desc_alloc(txq, true); + err = idpf_tx_desc_alloc(vport, txq); if (err) { - dev_err(dev, "Allocation for Tx Queue %u failed\n", + pci_err(vport->adapter->pdev, + "Allocation for Tx Queue %u failed\n", i); goto err_out; } @@ -311,9 +360,10 @@ static int idpf_tx_desc_alloc_all(struct idpf_vport *vport) continue; /* Setup completion queues */ - err = idpf_tx_desc_alloc(vport->txq_grps[i].complq, false); + err = idpf_compl_desc_alloc(vport, vport->txq_grps[i].complq); if (err) { - dev_err(dev, "Allocation for Tx Completion Queue %u failed\n", + pci_err(vport->adapter->pdev, + "Allocation for Tx Completion Queue %u failed\n", i); goto err_out; } @@ -328,70 +378,97 @@ err_out: /** * idpf_rx_page_rel - Release an rx buffer page - * @rxq: the queue that owns the buffer * @rx_buf: the buffer to free */ -static void idpf_rx_page_rel(struct idpf_queue *rxq, struct idpf_rx_buf *rx_buf) +static void idpf_rx_page_rel(struct libeth_fqe *rx_buf) { if (unlikely(!rx_buf->page)) return; - page_pool_put_full_page(rxq->pp, rx_buf->page, false); + page_pool_put_full_page(rx_buf->page->pp, rx_buf->page, false); rx_buf->page = NULL; - rx_buf->page_offset = 0; + rx_buf->offset = 0; } /** * idpf_rx_hdr_buf_rel_all - Release header buffer memory - * @rxq: queue to use + * @bufq: queue to use */ -static void idpf_rx_hdr_buf_rel_all(struct idpf_queue *rxq) +static void idpf_rx_hdr_buf_rel_all(struct idpf_buf_queue *bufq) { - struct idpf_adapter *adapter = rxq->vport->adapter; + struct libeth_fq fq = { + .fqes = bufq->hdr_buf, + .pp = bufq->hdr_pp, + }; + + for (u32 i = 0; i < bufq->desc_count; i++) + idpf_rx_page_rel(&bufq->hdr_buf[i]); - dma_free_coherent(&adapter->pdev->dev, - rxq->desc_count * IDPF_HDR_BUF_SIZE, - rxq->rx_buf.hdr_buf_va, - rxq->rx_buf.hdr_buf_pa); - rxq->rx_buf.hdr_buf_va = NULL; + libeth_rx_fq_destroy(&fq); + bufq->hdr_buf = NULL; + bufq->hdr_pp = NULL; } /** - * idpf_rx_buf_rel_all - Free all Rx buffer resources for a queue - * @rxq: queue to be cleaned + * idpf_rx_buf_rel_bufq - Free all Rx buffer resources for a buffer queue + * @bufq: queue to be cleaned */ -static void idpf_rx_buf_rel_all(struct idpf_queue *rxq) +static void idpf_rx_buf_rel_bufq(struct idpf_buf_queue *bufq) { - u16 i; + struct libeth_fq fq = { + .fqes = bufq->buf, + .pp = bufq->pp, + }; /* queue already cleared, nothing to do */ - if (!rxq->rx_buf.buf) + if (!bufq->buf) return; /* Free all the bufs allocated and given to hw on Rx queue */ - for (i = 0; i < rxq->desc_count; i++) - idpf_rx_page_rel(rxq, &rxq->rx_buf.buf[i]); + for (u32 i = 0; i < bufq->desc_count; i++) + idpf_rx_page_rel(&bufq->buf[i]); - if (rxq->rx_hsplit_en) - idpf_rx_hdr_buf_rel_all(rxq); + if (idpf_queue_has(HSPLIT_EN, bufq)) + idpf_rx_hdr_buf_rel_all(bufq); - page_pool_destroy(rxq->pp); - rxq->pp = NULL; + libeth_rx_fq_destroy(&fq); + bufq->buf = NULL; + bufq->pp = NULL; +} - kfree(rxq->rx_buf.buf); - rxq->rx_buf.buf = NULL; +/** + * idpf_rx_buf_rel_all - Free all Rx buffer resources for a receive queue + * @rxq: queue to be cleaned + */ +static void idpf_rx_buf_rel_all(struct idpf_rx_queue *rxq) +{ + struct libeth_fq fq = { + .fqes = rxq->rx_buf, + .pp = rxq->pp, + }; + + if (!rxq->rx_buf) + return; + + for (u32 i = 0; i < rxq->desc_count; i++) + idpf_rx_page_rel(&rxq->rx_buf[i]); + + libeth_rx_fq_destroy(&fq); + rxq->rx_buf = NULL; + rxq->pp = NULL; } /** * idpf_rx_desc_rel - Free a specific Rx q resources * @rxq: queue to clean the resources from - * @bufq: buffer q or completion q - * @q_model: single or split q model + * @dev: device to free DMA memory + * @model: single or split queue model * * Free a specific rx queue resources */ -static void idpf_rx_desc_rel(struct idpf_queue *rxq, bool bufq, s32 q_model) +static void idpf_rx_desc_rel(struct idpf_rx_queue *rxq, struct device *dev, + u32 model) { if (!rxq) return; @@ -401,7 +478,7 @@ static void idpf_rx_desc_rel(struct idpf_queue *rxq, bool bufq, s32 q_model) rxq->skb = NULL; } - if (bufq || !idpf_is_queue_model_split(q_model)) + if (!idpf_is_queue_model_split(model)) idpf_rx_buf_rel_all(rxq); rxq->next_to_alloc = 0; @@ -410,11 +487,35 @@ static void idpf_rx_desc_rel(struct idpf_queue *rxq, bool bufq, s32 q_model) if (!rxq->desc_ring) return; - dmam_free_coherent(rxq->dev, rxq->size, rxq->desc_ring, rxq->dma); + dmam_free_coherent(dev, rxq->size, rxq->desc_ring, rxq->dma); rxq->desc_ring = NULL; } /** + * idpf_rx_desc_rel_bufq - free buffer queue resources + * @bufq: buffer queue to clean the resources from + * @dev: device to free DMA memory + */ +static void idpf_rx_desc_rel_bufq(struct idpf_buf_queue *bufq, + struct device *dev) +{ + if (!bufq) + return; + + idpf_rx_buf_rel_bufq(bufq); + + bufq->next_to_alloc = 0; + bufq->next_to_clean = 0; + bufq->next_to_use = 0; + + if (!bufq->split_buf) + return; + + dma_free_coherent(dev, bufq->size, bufq->split_buf, bufq->dma); + bufq->split_buf = NULL; +} + +/** * idpf_rx_desc_rel_all - Free Rx Resources for All Queues * @vport: virtual port structure * @@ -422,6 +523,7 @@ static void idpf_rx_desc_rel(struct idpf_queue *rxq, bool bufq, s32 q_model) */ static void idpf_rx_desc_rel_all(struct idpf_vport *vport) { + struct device *dev = &vport->adapter->pdev->dev; struct idpf_rxq_group *rx_qgrp; u16 num_rxq; int i, j; @@ -434,15 +536,15 @@ static void idpf_rx_desc_rel_all(struct idpf_vport *vport) if (!idpf_is_queue_model_split(vport->rxq_model)) { for (j = 0; j < rx_qgrp->singleq.num_rxq; j++) - idpf_rx_desc_rel(rx_qgrp->singleq.rxqs[j], - false, vport->rxq_model); + idpf_rx_desc_rel(rx_qgrp->singleq.rxqs[j], dev, + VIRTCHNL2_QUEUE_MODEL_SINGLE); continue; } num_rxq = rx_qgrp->splitq.num_rxq_sets; for (j = 0; j < num_rxq; j++) idpf_rx_desc_rel(&rx_qgrp->splitq.rxq_sets[j]->rxq, - false, vport->rxq_model); + dev, VIRTCHNL2_QUEUE_MODEL_SPLIT); if (!rx_qgrp->splitq.bufq_sets) continue; @@ -451,45 +553,50 @@ static void idpf_rx_desc_rel_all(struct idpf_vport *vport) struct idpf_bufq_set *bufq_set = &rx_qgrp->splitq.bufq_sets[j]; - idpf_rx_desc_rel(&bufq_set->bufq, true, - vport->rxq_model); + idpf_rx_desc_rel_bufq(&bufq_set->bufq, dev); } } } /** * idpf_rx_buf_hw_update - Store the new tail and head values - * @rxq: queue to bump + * @bufq: queue to bump * @val: new head index */ -void idpf_rx_buf_hw_update(struct idpf_queue *rxq, u32 val) +static void idpf_rx_buf_hw_update(struct idpf_buf_queue *bufq, u32 val) { - rxq->next_to_use = val; + bufq->next_to_use = val; - if (unlikely(!rxq->tail)) + if (unlikely(!bufq->tail)) return; /* writel has an implicit memory barrier */ - writel(val, rxq->tail); + writel(val, bufq->tail); } /** * idpf_rx_hdr_buf_alloc_all - Allocate memory for header buffers - * @rxq: ring to use + * @bufq: ring to use * * Returns 0 on success, negative on failure. */ -static int idpf_rx_hdr_buf_alloc_all(struct idpf_queue *rxq) +static int idpf_rx_hdr_buf_alloc_all(struct idpf_buf_queue *bufq) { - struct idpf_adapter *adapter = rxq->vport->adapter; - - rxq->rx_buf.hdr_buf_va = - dma_alloc_coherent(&adapter->pdev->dev, - IDPF_HDR_BUF_SIZE * rxq->desc_count, - &rxq->rx_buf.hdr_buf_pa, - GFP_KERNEL); - if (!rxq->rx_buf.hdr_buf_va) - return -ENOMEM; + struct libeth_fq fq = { + .count = bufq->desc_count, + .type = LIBETH_FQE_HDR, + .nid = idpf_q_vector_to_mem(bufq->q_vector), + }; + int ret; + + ret = libeth_rx_fq_create(&fq, &bufq->q_vector->napi); + if (ret) + return ret; + + bufq->hdr_pp = fq.pp; + bufq->hdr_buf = fq.fqes; + bufq->hdr_truesize = fq.truesize; + bufq->rx_hbuf_size = fq.buf_len; return 0; } @@ -501,19 +608,20 @@ static int idpf_rx_hdr_buf_alloc_all(struct idpf_queue *rxq) */ static void idpf_rx_post_buf_refill(struct idpf_sw_queue *refillq, u16 buf_id) { - u16 nta = refillq->next_to_alloc; + u32 nta = refillq->next_to_use; /* store the buffer ID and the SW maintained GEN bit to the refillq */ refillq->ring[nta] = FIELD_PREP(IDPF_RX_BI_BUFID_M, buf_id) | FIELD_PREP(IDPF_RX_BI_GEN_M, - test_bit(__IDPF_Q_GEN_CHK, refillq->flags)); + idpf_queue_has(GEN_CHK, refillq)); if (unlikely(++nta == refillq->desc_count)) { nta = 0; - change_bit(__IDPF_Q_GEN_CHK, refillq->flags); + idpf_queue_change(GEN_CHK, refillq); } - refillq->next_to_alloc = nta; + + refillq->next_to_use = nta; } /** @@ -523,24 +631,35 @@ static void idpf_rx_post_buf_refill(struct idpf_sw_queue *refillq, u16 buf_id) * * Returns false if buffer could not be allocated, true otherwise. */ -static bool idpf_rx_post_buf_desc(struct idpf_queue *bufq, u16 buf_id) +static bool idpf_rx_post_buf_desc(struct idpf_buf_queue *bufq, u16 buf_id) { struct virtchnl2_splitq_rx_buf_desc *splitq_rx_desc = NULL; + struct libeth_fq_fp fq = { + .count = bufq->desc_count, + }; u16 nta = bufq->next_to_alloc; - struct idpf_rx_buf *buf; dma_addr_t addr; - splitq_rx_desc = IDPF_SPLITQ_RX_BUF_DESC(bufq, nta); - buf = &bufq->rx_buf.buf[buf_id]; + splitq_rx_desc = &bufq->split_buf[nta]; + + if (idpf_queue_has(HSPLIT_EN, bufq)) { + fq.pp = bufq->hdr_pp; + fq.fqes = bufq->hdr_buf; + fq.truesize = bufq->hdr_truesize; - if (bufq->rx_hsplit_en) { - splitq_rx_desc->hdr_addr = - cpu_to_le64(bufq->rx_buf.hdr_buf_pa + - (u32)buf_id * IDPF_HDR_BUF_SIZE); + addr = libeth_rx_alloc(&fq, buf_id); + if (addr == DMA_MAPPING_ERROR) + return false; + + splitq_rx_desc->hdr_addr = cpu_to_le64(addr); } - addr = idpf_alloc_page(bufq->pp, buf, bufq->rx_buf_size); - if (unlikely(addr == DMA_MAPPING_ERROR)) + fq.pp = bufq->pp; + fq.fqes = bufq->buf; + fq.truesize = bufq->truesize; + + addr = libeth_rx_alloc(&fq, buf_id); + if (addr == DMA_MAPPING_ERROR) return false; splitq_rx_desc->pkt_addr = cpu_to_le64(addr); @@ -561,7 +680,8 @@ static bool idpf_rx_post_buf_desc(struct idpf_queue *bufq, u16 buf_id) * * Returns true if @working_set bufs were posted successfully, false otherwise. */ -static bool idpf_rx_post_init_bufs(struct idpf_queue *bufq, u16 working_set) +static bool idpf_rx_post_init_bufs(struct idpf_buf_queue *bufq, + u16 working_set) { int i; @@ -570,95 +690,114 @@ static bool idpf_rx_post_init_bufs(struct idpf_queue *bufq, u16 working_set) return false; } - idpf_rx_buf_hw_update(bufq, - bufq->next_to_alloc & ~(bufq->rx_buf_stride - 1)); + idpf_rx_buf_hw_update(bufq, ALIGN_DOWN(bufq->next_to_alloc, + IDPF_RX_BUF_STRIDE)); return true; } /** - * idpf_rx_create_page_pool - Create a page pool - * @rxbufq: RX queue to create page pool for + * idpf_rx_buf_alloc_singleq - Allocate memory for all buffer resources + * @rxq: queue for which the buffers are allocated * - * Returns &page_pool on success, casted -errno on failure + * Return: 0 on success, -ENOMEM on failure. */ -static struct page_pool *idpf_rx_create_page_pool(struct idpf_queue *rxbufq) +static int idpf_rx_buf_alloc_singleq(struct idpf_rx_queue *rxq) { - struct page_pool_params pp = { - .flags = PP_FLAG_DMA_MAP | PP_FLAG_DMA_SYNC_DEV, - .order = 0, - .pool_size = rxbufq->desc_count, - .nid = NUMA_NO_NODE, - .dev = rxbufq->vport->netdev->dev.parent, - .max_len = PAGE_SIZE, - .dma_dir = DMA_FROM_DEVICE, - .offset = 0, + if (idpf_rx_singleq_buf_hw_alloc_all(rxq, rxq->desc_count - 1)) + goto err; + + return 0; + +err: + idpf_rx_buf_rel_all(rxq); + + return -ENOMEM; +} + +/** + * idpf_rx_bufs_init_singleq - Initialize page pool and allocate Rx bufs + * @rxq: buffer queue to create page pool for + * + * Return: 0 on success, -errno on failure. + */ +static int idpf_rx_bufs_init_singleq(struct idpf_rx_queue *rxq) +{ + struct libeth_fq fq = { + .count = rxq->desc_count, + .type = LIBETH_FQE_MTU, + .nid = idpf_q_vector_to_mem(rxq->q_vector), }; + int ret; + + ret = libeth_rx_fq_create(&fq, &rxq->q_vector->napi); + if (ret) + return ret; + + rxq->pp = fq.pp; + rxq->rx_buf = fq.fqes; + rxq->truesize = fq.truesize; + rxq->rx_buf_size = fq.buf_len; - return page_pool_create(&pp); + return idpf_rx_buf_alloc_singleq(rxq); } /** * idpf_rx_buf_alloc_all - Allocate memory for all buffer resources - * @rxbufq: queue for which the buffers are allocated; equivalent to - * rxq when operating in singleq mode + * @rxbufq: queue for which the buffers are allocated * * Returns 0 on success, negative on failure */ -static int idpf_rx_buf_alloc_all(struct idpf_queue *rxbufq) +static int idpf_rx_buf_alloc_all(struct idpf_buf_queue *rxbufq) { int err = 0; - /* Allocate book keeping buffers */ - rxbufq->rx_buf.buf = kcalloc(rxbufq->desc_count, - sizeof(struct idpf_rx_buf), GFP_KERNEL); - if (!rxbufq->rx_buf.buf) { - err = -ENOMEM; - goto rx_buf_alloc_all_out; - } - - if (rxbufq->rx_hsplit_en) { + if (idpf_queue_has(HSPLIT_EN, rxbufq)) { err = idpf_rx_hdr_buf_alloc_all(rxbufq); if (err) goto rx_buf_alloc_all_out; } /* Allocate buffers to be given to HW. */ - if (idpf_is_queue_model_split(rxbufq->vport->rxq_model)) { - int working_set = IDPF_RX_BUFQ_WORKING_SET(rxbufq); - - if (!idpf_rx_post_init_bufs(rxbufq, working_set)) - err = -ENOMEM; - } else { - if (idpf_rx_singleq_buf_hw_alloc_all(rxbufq, - rxbufq->desc_count - 1)) - err = -ENOMEM; - } + if (!idpf_rx_post_init_bufs(rxbufq, IDPF_RX_BUFQ_WORKING_SET(rxbufq))) + err = -ENOMEM; rx_buf_alloc_all_out: if (err) - idpf_rx_buf_rel_all(rxbufq); + idpf_rx_buf_rel_bufq(rxbufq); return err; } /** * idpf_rx_bufs_init - Initialize page pool, allocate rx bufs, and post to HW - * @rxbufq: RX queue to create page pool for + * @bufq: buffer queue to create page pool for + * @type: type of Rx buffers to allocate * * Returns 0 on success, negative on failure */ -static int idpf_rx_bufs_init(struct idpf_queue *rxbufq) +static int idpf_rx_bufs_init(struct idpf_buf_queue *bufq, + enum libeth_fqe_type type) { - struct page_pool *pool; + struct libeth_fq fq = { + .truesize = bufq->truesize, + .count = bufq->desc_count, + .type = type, + .hsplit = idpf_queue_has(HSPLIT_EN, bufq), + .nid = idpf_q_vector_to_mem(bufq->q_vector), + }; + int ret; - pool = idpf_rx_create_page_pool(rxbufq); - if (IS_ERR(pool)) - return PTR_ERR(pool); + ret = libeth_rx_fq_create(&fq, &bufq->q_vector->napi); + if (ret) + return ret; - rxbufq->pp = pool; + bufq->pp = fq.pp; + bufq->buf = fq.fqes; + bufq->truesize = fq.truesize; + bufq->rx_buf_size = fq.buf_len; - return idpf_rx_buf_alloc_all(rxbufq); + return idpf_rx_buf_alloc_all(bufq); } /** @@ -669,20 +808,22 @@ static int idpf_rx_bufs_init(struct idpf_queue *rxbufq) */ int idpf_rx_bufs_init_all(struct idpf_vport *vport) { - struct idpf_rxq_group *rx_qgrp; - struct idpf_queue *q; + bool split = idpf_is_queue_model_split(vport->rxq_model); int i, j, err; for (i = 0; i < vport->num_rxq_grp; i++) { - rx_qgrp = &vport->rxq_grps[i]; + struct idpf_rxq_group *rx_qgrp = &vport->rxq_grps[i]; + u32 truesize = 0; /* Allocate bufs for the rxq itself in singleq */ - if (!idpf_is_queue_model_split(vport->rxq_model)) { + if (!split) { int num_rxq = rx_qgrp->singleq.num_rxq; for (j = 0; j < num_rxq; j++) { + struct idpf_rx_queue *q; + q = rx_qgrp->singleq.rxqs[j]; - err = idpf_rx_bufs_init(q); + err = idpf_rx_bufs_init_singleq(q); if (err) return err; } @@ -692,10 +833,19 @@ int idpf_rx_bufs_init_all(struct idpf_vport *vport) /* Otherwise, allocate bufs for the buffer queues */ for (j = 0; j < vport->num_bufqs_per_qgrp; j++) { + enum libeth_fqe_type type; + struct idpf_buf_queue *q; + q = &rx_qgrp->splitq.bufq_sets[j].bufq; - err = idpf_rx_bufs_init(q); + q->truesize = truesize; + + type = truesize ? LIBETH_FQE_SHORT : LIBETH_FQE_MTU; + + err = idpf_rx_bufs_init(q, type); if (err) return err; + + truesize = q->truesize >> 1; } } @@ -704,22 +854,17 @@ int idpf_rx_bufs_init_all(struct idpf_vport *vport) /** * idpf_rx_desc_alloc - Allocate queue Rx resources + * @vport: vport to allocate resources for * @rxq: Rx queue for which the resources are setup - * @bufq: buffer or completion queue - * @q_model: single or split queue model * * Returns 0 on success, negative on failure */ -static int idpf_rx_desc_alloc(struct idpf_queue *rxq, bool bufq, s32 q_model) +static int idpf_rx_desc_alloc(const struct idpf_vport *vport, + struct idpf_rx_queue *rxq) { - struct device *dev = rxq->dev; + struct device *dev = &vport->adapter->pdev->dev; - if (bufq) - rxq->size = rxq->desc_count * - sizeof(struct virtchnl2_splitq_rx_buf_desc); - else - rxq->size = rxq->desc_count * - sizeof(union virtchnl2_rx_desc); + rxq->size = rxq->desc_count * sizeof(union virtchnl2_rx_desc); /* Allocate descriptors and also round up to nearest 4K */ rxq->size = ALIGN(rxq->size, 4096); @@ -734,7 +879,35 @@ static int idpf_rx_desc_alloc(struct idpf_queue *rxq, bool bufq, s32 q_model) rxq->next_to_alloc = 0; rxq->next_to_clean = 0; rxq->next_to_use = 0; - set_bit(__IDPF_Q_GEN_CHK, rxq->flags); + idpf_queue_set(GEN_CHK, rxq); + + return 0; +} + +/** + * idpf_bufq_desc_alloc - Allocate buffer queue descriptor ring + * @vport: vport to allocate resources for + * @bufq: buffer queue for which the resources are set up + * + * Return: 0 on success, -ENOMEM on failure. + */ +static int idpf_bufq_desc_alloc(const struct idpf_vport *vport, + struct idpf_buf_queue *bufq) +{ + struct device *dev = &vport->adapter->pdev->dev; + + bufq->size = array_size(bufq->desc_count, sizeof(*bufq->split_buf)); + + bufq->split_buf = dma_alloc_coherent(dev, bufq->size, &bufq->dma, + GFP_KERNEL); + if (!bufq->split_buf) + return -ENOMEM; + + bufq->next_to_alloc = 0; + bufq->next_to_clean = 0; + bufq->next_to_use = 0; + + idpf_queue_set(GEN_CHK, bufq); return 0; } @@ -747,9 +920,7 @@ static int idpf_rx_desc_alloc(struct idpf_queue *rxq, bool bufq, s32 q_model) */ static int idpf_rx_desc_alloc_all(struct idpf_vport *vport) { - struct device *dev = &vport->adapter->pdev->dev; struct idpf_rxq_group *rx_qgrp; - struct idpf_queue *q; int i, j, err; u16 num_rxq; @@ -761,13 +932,17 @@ static int idpf_rx_desc_alloc_all(struct idpf_vport *vport) num_rxq = rx_qgrp->singleq.num_rxq; for (j = 0; j < num_rxq; j++) { + struct idpf_rx_queue *q; + if (idpf_is_queue_model_split(vport->rxq_model)) q = &rx_qgrp->splitq.rxq_sets[j]->rxq; else q = rx_qgrp->singleq.rxqs[j]; - err = idpf_rx_desc_alloc(q, false, vport->rxq_model); + + err = idpf_rx_desc_alloc(vport, q); if (err) { - dev_err(dev, "Memory allocation for Rx Queue %u failed\n", + pci_err(vport->adapter->pdev, + "Memory allocation for Rx Queue %u failed\n", i); goto err_out; } @@ -777,10 +952,14 @@ static int idpf_rx_desc_alloc_all(struct idpf_vport *vport) continue; for (j = 0; j < vport->num_bufqs_per_qgrp; j++) { + struct idpf_buf_queue *q; + q = &rx_qgrp->splitq.bufq_sets[j].bufq; - err = idpf_rx_desc_alloc(q, true, vport->rxq_model); + + err = idpf_bufq_desc_alloc(vport, q); if (err) { - dev_err(dev, "Memory allocation for Rx Buffer Queue %u failed\n", + pci_err(vport->adapter->pdev, + "Memory allocation for Rx Buffer Queue %u failed\n", i); goto err_out; } @@ -801,11 +980,16 @@ err_out: */ static void idpf_txq_group_rel(struct idpf_vport *vport) { + bool split, flow_sch_en; int i, j; if (!vport->txq_grps) return; + split = idpf_is_queue_model_split(vport->txq_model); + flow_sch_en = !idpf_is_cap_ena(vport->adapter, IDPF_OTHER_CAPS, + VIRTCHNL2_CAP_SPLITQ_QSCHED); + for (i = 0; i < vport->num_txq_grp; i++) { struct idpf_txq_group *txq_grp = &vport->txq_grps[i]; @@ -813,8 +997,15 @@ static void idpf_txq_group_rel(struct idpf_vport *vport) kfree(txq_grp->txqs[j]); txq_grp->txqs[j] = NULL; } + + if (!split) + continue; + kfree(txq_grp->complq); txq_grp->complq = NULL; + + if (flow_sch_en) + kfree(txq_grp->stashes); } kfree(vport->txq_grps); vport->txq_grps = NULL; @@ -918,7 +1109,7 @@ static int idpf_vport_init_fast_path_txqs(struct idpf_vport *vport) { int i, j, k = 0; - vport->txqs = kcalloc(vport->num_txq, sizeof(struct idpf_queue *), + vport->txqs = kcalloc(vport->num_txq, sizeof(*vport->txqs), GFP_KERNEL); if (!vport->txqs) @@ -966,17 +1157,11 @@ void idpf_vport_init_num_qs(struct idpf_vport *vport, /* Adjust number of buffer queues per Rx queue group. */ if (!idpf_is_queue_model_split(vport->rxq_model)) { vport->num_bufqs_per_qgrp = 0; - vport->bufq_size[0] = IDPF_RX_BUF_2048; return; } vport->num_bufqs_per_qgrp = IDPF_MAX_BUFQS_PER_RXQ_GRP; - /* Bufq[0] default buffer size is 4K - * Bufq[1] default buffer size is 2K - */ - vport->bufq_size[0] = IDPF_RX_BUF_4096; - vport->bufq_size[1] = IDPF_RX_BUF_2048; } /** @@ -1136,9 +1321,10 @@ static void idpf_vport_calc_numq_per_grp(struct idpf_vport *vport, * @q: rx queue for which descids are set * */ -static void idpf_rxq_set_descids(struct idpf_vport *vport, struct idpf_queue *q) +static void idpf_rxq_set_descids(const struct idpf_vport *vport, + struct idpf_rx_queue *q) { - if (vport->rxq_model == VIRTCHNL2_QUEUE_MODEL_SPLIT) { + if (idpf_is_queue_model_split(vport->rxq_model)) { q->rxdids = VIRTCHNL2_RXDID_2_FLEX_SPLITQ_M; } else { if (vport->base_rxd) @@ -1157,20 +1343,22 @@ static void idpf_rxq_set_descids(struct idpf_vport *vport, struct idpf_queue *q) */ static int idpf_txq_group_alloc(struct idpf_vport *vport, u16 num_txq) { - bool flow_sch_en; - int err, i; + bool split, flow_sch_en; + int i; vport->txq_grps = kcalloc(vport->num_txq_grp, sizeof(*vport->txq_grps), GFP_KERNEL); if (!vport->txq_grps) return -ENOMEM; + split = idpf_is_queue_model_split(vport->txq_model); flow_sch_en = !idpf_is_cap_ena(vport->adapter, IDPF_OTHER_CAPS, VIRTCHNL2_CAP_SPLITQ_QSCHED); for (i = 0; i < vport->num_txq_grp; i++) { struct idpf_txq_group *tx_qgrp = &vport->txq_grps[i]; struct idpf_adapter *adapter = vport->adapter; + struct idpf_txq_stash *stashes; int j; tx_qgrp->vport = vport; @@ -1179,45 +1367,62 @@ static int idpf_txq_group_alloc(struct idpf_vport *vport, u16 num_txq) for (j = 0; j < tx_qgrp->num_txq; j++) { tx_qgrp->txqs[j] = kzalloc(sizeof(*tx_qgrp->txqs[j]), GFP_KERNEL); - if (!tx_qgrp->txqs[j]) { - err = -ENOMEM; + if (!tx_qgrp->txqs[j]) goto err_alloc; - } + } + + if (split && flow_sch_en) { + stashes = kcalloc(num_txq, sizeof(*stashes), + GFP_KERNEL); + if (!stashes) + goto err_alloc; + + tx_qgrp->stashes = stashes; } for (j = 0; j < tx_qgrp->num_txq; j++) { - struct idpf_queue *q = tx_qgrp->txqs[j]; + struct idpf_tx_queue *q = tx_qgrp->txqs[j]; q->dev = &adapter->pdev->dev; q->desc_count = vport->txq_desc_count; q->tx_max_bufs = idpf_get_max_tx_bufs(adapter); q->tx_min_pkt_len = idpf_get_min_tx_pkt_len(adapter); - q->vport = vport; + q->netdev = vport->netdev; q->txq_grp = tx_qgrp; - hash_init(q->sched_buf_hash); - if (flow_sch_en) - set_bit(__IDPF_Q_FLOW_SCH_EN, q->flags); + if (!split) { + q->clean_budget = vport->compln_clean_budget; + idpf_queue_assign(CRC_EN, q, + vport->crc_enable); + } + + if (!flow_sch_en) + continue; + + if (split) { + q->stash = &stashes[j]; + hash_init(q->stash->sched_buf_hash); + } + + idpf_queue_set(FLOW_SCH_EN, q); } - if (!idpf_is_queue_model_split(vport->txq_model)) + if (!split) continue; tx_qgrp->complq = kcalloc(IDPF_COMPLQ_PER_GROUP, sizeof(*tx_qgrp->complq), GFP_KERNEL); - if (!tx_qgrp->complq) { - err = -ENOMEM; + if (!tx_qgrp->complq) goto err_alloc; - } - tx_qgrp->complq->dev = &adapter->pdev->dev; tx_qgrp->complq->desc_count = vport->complq_desc_count; - tx_qgrp->complq->vport = vport; tx_qgrp->complq->txq_grp = tx_qgrp; + tx_qgrp->complq->netdev = vport->netdev; + tx_qgrp->complq->clean_budget = vport->compln_clean_budget; if (flow_sch_en) - __set_bit(__IDPF_Q_FLOW_SCH_EN, tx_qgrp->complq->flags); + idpf_queue_set(FLOW_SCH_EN, tx_qgrp->complq); } return 0; @@ -1225,7 +1430,7 @@ static int idpf_txq_group_alloc(struct idpf_vport *vport, u16 num_txq) err_alloc: idpf_txq_group_rel(vport); - return err; + return -ENOMEM; } /** @@ -1237,8 +1442,6 @@ err_alloc: */ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq) { - struct idpf_adapter *adapter = vport->adapter; - struct idpf_queue *q; int i, k, err = 0; bool hs; @@ -1291,21 +1494,13 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq) struct idpf_bufq_set *bufq_set = &rx_qgrp->splitq.bufq_sets[j]; int swq_size = sizeof(struct idpf_sw_queue); + struct idpf_buf_queue *q; q = &rx_qgrp->splitq.bufq_sets[j].bufq; - q->dev = &adapter->pdev->dev; q->desc_count = vport->bufq_desc_count[j]; - q->vport = vport; - q->rxq_grp = rx_qgrp; - q->idx = j; - q->rx_buf_size = vport->bufq_size[j]; q->rx_buffer_low_watermark = IDPF_LOW_WATERMARK; - q->rx_buf_stride = IDPF_RX_BUF_STRIDE; - if (hs) { - q->rx_hsplit_en = true; - q->rx_hbuf_size = IDPF_HDR_BUF_SIZE; - } + idpf_queue_assign(HSPLIT_EN, q, hs); bufq_set->num_refillqs = num_rxq; bufq_set->refillqs = kcalloc(num_rxq, swq_size, @@ -1318,13 +1513,12 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq) struct idpf_sw_queue *refillq = &bufq_set->refillqs[k]; - refillq->dev = &vport->adapter->pdev->dev; refillq->desc_count = vport->bufq_desc_count[j]; - set_bit(__IDPF_Q_GEN_CHK, refillq->flags); - set_bit(__IDPF_RFLQ_GEN_CHK, refillq->flags); + idpf_queue_set(GEN_CHK, refillq); + idpf_queue_set(RFL_GEN_CHK, refillq); refillq->ring = kcalloc(refillq->desc_count, - sizeof(u16), + sizeof(*refillq->ring), GFP_KERNEL); if (!refillq->ring) { err = -ENOMEM; @@ -1335,36 +1529,30 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq) skip_splitq_rx_init: for (j = 0; j < num_rxq; j++) { + struct idpf_rx_queue *q; + if (!idpf_is_queue_model_split(vport->rxq_model)) { q = rx_qgrp->singleq.rxqs[j]; goto setup_rxq; } q = &rx_qgrp->splitq.rxq_sets[j]->rxq; - rx_qgrp->splitq.rxq_sets[j]->refillq0 = + rx_qgrp->splitq.rxq_sets[j]->refillq[0] = &rx_qgrp->splitq.bufq_sets[0].refillqs[j]; if (vport->num_bufqs_per_qgrp > IDPF_SINGLE_BUFQ_PER_RXQ_GRP) - rx_qgrp->splitq.rxq_sets[j]->refillq1 = + rx_qgrp->splitq.rxq_sets[j]->refillq[1] = &rx_qgrp->splitq.bufq_sets[1].refillqs[j]; - if (hs) { - q->rx_hsplit_en = true; - q->rx_hbuf_size = IDPF_HDR_BUF_SIZE; - } + idpf_queue_assign(HSPLIT_EN, q, hs); setup_rxq: - q->dev = &adapter->pdev->dev; q->desc_count = vport->rxq_desc_count; - q->vport = vport; - q->rxq_grp = rx_qgrp; + q->rx_ptype_lkup = vport->rx_ptype_lkup; + q->netdev = vport->netdev; + q->bufq_sets = rx_qgrp->splitq.bufq_sets; q->idx = (i * num_rxq) + j; - /* In splitq mode, RXQ buffer size should be - * set to that of the first buffer queue - * associated with this RXQ - */ - q->rx_buf_size = vport->bufq_size[0]; q->rx_buffer_low_watermark = IDPF_LOW_WATERMARK; q->rx_max_pkt_size = vport->netdev->mtu + - IDPF_PACKET_HDR_PAD; + LIBETH_RX_LL_LEN; idpf_rxq_set_descids(vport, q); } } @@ -1444,12 +1632,13 @@ err_out: * idpf_tx_handle_sw_marker - Handle queue marker packet * @tx_q: tx queue to handle software marker */ -static void idpf_tx_handle_sw_marker(struct idpf_queue *tx_q) +static void idpf_tx_handle_sw_marker(struct idpf_tx_queue *tx_q) { - struct idpf_vport *vport = tx_q->vport; + struct idpf_netdev_priv *priv = netdev_priv(tx_q->netdev); + struct idpf_vport *vport = priv->vport; int i; - clear_bit(__IDPF_Q_SW_MARKER, tx_q->flags); + idpf_queue_clear(SW_MARKER, tx_q); /* Hardware must write marker packets to all queues associated with * completion queues. So check if all queues received marker packets */ @@ -1457,7 +1646,7 @@ static void idpf_tx_handle_sw_marker(struct idpf_queue *tx_q) /* If we're still waiting on any other TXQ marker completions, * just return now since we cannot wake up the marker_wq yet. */ - if (test_bit(__IDPF_Q_SW_MARKER, vport->txqs[i]->flags)) + if (idpf_queue_has(SW_MARKER, vport->txqs[i])) return; /* Drain complete */ @@ -1466,37 +1655,6 @@ static void idpf_tx_handle_sw_marker(struct idpf_queue *tx_q) } /** - * idpf_tx_splitq_clean_hdr - Clean TX buffer resources for header portion of - * packet - * @tx_q: tx queue to clean buffer from - * @tx_buf: buffer to be cleaned - * @cleaned: pointer to stats struct to track cleaned packets/bytes - * @napi_budget: Used to determine if we are in netpoll - */ -static void idpf_tx_splitq_clean_hdr(struct idpf_queue *tx_q, - struct idpf_tx_buf *tx_buf, - struct idpf_cleaned_stats *cleaned, - int napi_budget) -{ - napi_consume_skb(tx_buf->skb, napi_budget); - - if (dma_unmap_len(tx_buf, len)) { - dma_unmap_single(tx_q->dev, - dma_unmap_addr(tx_buf, dma), - dma_unmap_len(tx_buf, len), - DMA_TO_DEVICE); - - dma_unmap_len_set(tx_buf, len, 0); - } - - /* clear tx_buf data */ - tx_buf->skb = NULL; - - cleaned->bytes += tx_buf->bytecount; - cleaned->packets += tx_buf->gso_segs; -} - -/** * idpf_tx_clean_stashed_bufs - clean bufs that were stored for * out of order completions * @txq: queue to clean @@ -1504,34 +1662,30 @@ static void idpf_tx_splitq_clean_hdr(struct idpf_queue *tx_q, * @cleaned: pointer to stats struct to track cleaned packets/bytes * @budget: Used to determine if we are in netpoll */ -static void idpf_tx_clean_stashed_bufs(struct idpf_queue *txq, u16 compl_tag, - struct idpf_cleaned_stats *cleaned, +static void idpf_tx_clean_stashed_bufs(struct idpf_tx_queue *txq, + u16 compl_tag, + struct libeth_sq_napi_stats *cleaned, int budget) { struct idpf_tx_stash *stash; struct hlist_node *tmp_buf; + struct libeth_cq_pp cp = { + .dev = txq->dev, + .ss = cleaned, + .napi = budget, + }; /* Buffer completion */ - hash_for_each_possible_safe(txq->sched_buf_hash, stash, tmp_buf, + hash_for_each_possible_safe(txq->stash->sched_buf_hash, stash, tmp_buf, hlist, compl_tag) { - if (unlikely(stash->buf.compl_tag != (int)compl_tag)) + if (unlikely(idpf_tx_buf_compl_tag(&stash->buf) != compl_tag)) continue; - if (stash->buf.skb) { - idpf_tx_splitq_clean_hdr(txq, &stash->buf, cleaned, - budget); - } else if (dma_unmap_len(&stash->buf, len)) { - dma_unmap_page(txq->dev, - dma_unmap_addr(&stash->buf, dma), - dma_unmap_len(&stash->buf, len), - DMA_TO_DEVICE); - dma_unmap_len_set(&stash->buf, len, 0); - } + hash_del(&stash->hlist); + libeth_tx_complete(&stash->buf, &cp); /* Push shadow buf back onto stack */ - idpf_buf_lifo_push(&txq->buf_stack, stash); - - hash_del(&stash->hlist); + idpf_buf_lifo_push(&txq->stash->buf_stack, stash); } } @@ -1541,49 +1695,47 @@ static void idpf_tx_clean_stashed_bufs(struct idpf_queue *txq, u16 compl_tag, * @txq: Tx queue to clean * @tx_buf: buffer to store */ -static int idpf_stash_flow_sch_buffers(struct idpf_queue *txq, +static int idpf_stash_flow_sch_buffers(struct idpf_tx_queue *txq, struct idpf_tx_buf *tx_buf) { struct idpf_tx_stash *stash; - if (unlikely(!dma_unmap_addr(tx_buf, dma) && - !dma_unmap_len(tx_buf, len))) + if (unlikely(tx_buf->type <= LIBETH_SQE_CTX)) return 0; - stash = idpf_buf_lifo_pop(&txq->buf_stack); + stash = idpf_buf_lifo_pop(&txq->stash->buf_stack); if (unlikely(!stash)) { net_err_ratelimited("%s: No out-of-order TX buffers left!\n", - txq->vport->netdev->name); + netdev_name(txq->netdev)); return -ENOMEM; } /* Store buffer params in shadow buffer */ stash->buf.skb = tx_buf->skb; - stash->buf.bytecount = tx_buf->bytecount; - stash->buf.gso_segs = tx_buf->gso_segs; + stash->buf.bytes = tx_buf->bytes; + stash->buf.packets = tx_buf->packets; + stash->buf.type = tx_buf->type; + stash->buf.nr_frags = tx_buf->nr_frags; dma_unmap_addr_set(&stash->buf, dma, dma_unmap_addr(tx_buf, dma)); dma_unmap_len_set(&stash->buf, len, dma_unmap_len(tx_buf, len)); - stash->buf.compl_tag = tx_buf->compl_tag; + idpf_tx_buf_compl_tag(&stash->buf) = idpf_tx_buf_compl_tag(tx_buf); /* Add buffer to buf_hash table to be freed later */ - hash_add(txq->sched_buf_hash, &stash->hlist, stash->buf.compl_tag); + hash_add(txq->stash->sched_buf_hash, &stash->hlist, + idpf_tx_buf_compl_tag(&stash->buf)); - memset(tx_buf, 0, sizeof(struct idpf_tx_buf)); - - /* Reinitialize buf_id portion of tag */ - tx_buf->compl_tag = IDPF_SPLITQ_TX_INVAL_COMPL_TAG; + tx_buf->type = LIBETH_SQE_EMPTY; return 0; } #define idpf_tx_splitq_clean_bump_ntc(txq, ntc, desc, buf) \ do { \ - (ntc)++; \ - if (unlikely(!(ntc))) { \ - ntc -= (txq)->desc_count; \ + if (unlikely(++(ntc) == (txq)->desc_count)) { \ + ntc = 0; \ buf = (txq)->tx_buf; \ - desc = IDPF_FLEX_TX_DESC(txq, 0); \ + desc = &(txq)->flex_tx[0]; \ } else { \ (buf)++; \ (desc)++; \ @@ -1605,69 +1757,71 @@ do { \ * Separate packet completion events will be reported on the completion queue, * and the buffers will be cleaned separately. The stats are not updated from * this function when using flow-based scheduling. + * + * Furthermore, in flow scheduling mode, check to make sure there are enough + * reserve buffers to stash the packet. If there are not, return early, which + * will leave next_to_clean pointing to the packet that failed to be stashed. + * + * Return: false in the scenario above, true otherwise. */ -static void idpf_tx_splitq_clean(struct idpf_queue *tx_q, u16 end, +static bool idpf_tx_splitq_clean(struct idpf_tx_queue *tx_q, u16 end, int napi_budget, - struct idpf_cleaned_stats *cleaned, + struct libeth_sq_napi_stats *cleaned, bool descs_only) { union idpf_tx_flex_desc *next_pending_desc = NULL; union idpf_tx_flex_desc *tx_desc; - s16 ntc = tx_q->next_to_clean; + u32 ntc = tx_q->next_to_clean; + struct libeth_cq_pp cp = { + .dev = tx_q->dev, + .ss = cleaned, + .napi = napi_budget, + }; struct idpf_tx_buf *tx_buf; + bool clean_complete = true; - tx_desc = IDPF_FLEX_TX_DESC(tx_q, ntc); - next_pending_desc = IDPF_FLEX_TX_DESC(tx_q, end); + tx_desc = &tx_q->flex_tx[ntc]; + next_pending_desc = &tx_q->flex_tx[end]; tx_buf = &tx_q->tx_buf[ntc]; - ntc -= tx_q->desc_count; while (tx_desc != next_pending_desc) { - union idpf_tx_flex_desc *eop_desc; + u32 eop_idx; /* If this entry in the ring was used as a context descriptor, - * it's corresponding entry in the buffer ring will have an - * invalid completion tag since no buffer was used. We can - * skip this descriptor since there is no buffer to clean. + * it's corresponding entry in the buffer ring is reserved. We + * can skip this descriptor since there is no buffer to clean. */ - if (unlikely(tx_buf->compl_tag == IDPF_SPLITQ_TX_INVAL_COMPL_TAG)) + if (tx_buf->type <= LIBETH_SQE_CTX) goto fetch_next_txq_desc; - eop_desc = (union idpf_tx_flex_desc *)tx_buf->next_to_watch; + if (unlikely(tx_buf->type != LIBETH_SQE_SKB)) + break; - /* clear next_to_watch to prevent false hangs */ - tx_buf->next_to_watch = NULL; + eop_idx = tx_buf->rs_idx; if (descs_only) { - if (idpf_stash_flow_sch_buffers(tx_q, tx_buf)) + if (IDPF_TX_BUF_RSV_UNUSED(tx_q) < tx_buf->nr_frags) { + clean_complete = false; goto tx_splitq_clean_out; + } + + idpf_stash_flow_sch_buffers(tx_q, tx_buf); - while (tx_desc != eop_desc) { + while (ntc != eop_idx) { idpf_tx_splitq_clean_bump_ntc(tx_q, ntc, tx_desc, tx_buf); - - if (dma_unmap_len(tx_buf, len)) { - if (idpf_stash_flow_sch_buffers(tx_q, - tx_buf)) - goto tx_splitq_clean_out; - } + idpf_stash_flow_sch_buffers(tx_q, tx_buf); } } else { - idpf_tx_splitq_clean_hdr(tx_q, tx_buf, cleaned, - napi_budget); + libeth_tx_complete(tx_buf, &cp); /* unmap remaining buffers */ - while (tx_desc != eop_desc) { + while (ntc != eop_idx) { idpf_tx_splitq_clean_bump_ntc(tx_q, ntc, tx_desc, tx_buf); /* unmap any remaining paged data */ - if (dma_unmap_len(tx_buf, len)) { - dma_unmap_page(tx_q->dev, - dma_unmap_addr(tx_buf, dma), - dma_unmap_len(tx_buf, len), - DMA_TO_DEVICE); - dma_unmap_len_set(tx_buf, len, 0); - } + libeth_tx_complete(tx_buf, &cp); } } @@ -1676,8 +1830,9 @@ fetch_next_txq_desc: } tx_splitq_clean_out: - ntc += tx_q->desc_count; tx_q->next_to_clean = ntc; + + return clean_complete; } #define idpf_tx_clean_buf_ring_bump_ntc(txq, ntc, buf) \ @@ -1702,58 +1857,69 @@ do { \ * stashed. Returns the byte/segment count for the cleaned packet associated * this completion tag. */ -static bool idpf_tx_clean_buf_ring(struct idpf_queue *txq, u16 compl_tag, - struct idpf_cleaned_stats *cleaned, +static bool idpf_tx_clean_buf_ring(struct idpf_tx_queue *txq, u16 compl_tag, + struct libeth_sq_napi_stats *cleaned, int budget) { u16 idx = compl_tag & txq->compl_tag_bufid_m; struct idpf_tx_buf *tx_buf = NULL; - u16 ntc = txq->next_to_clean; - u16 num_descs_cleaned = 0; - u16 orig_idx = idx; + struct libeth_cq_pp cp = { + .dev = txq->dev, + .ss = cleaned, + .napi = budget, + }; + u16 ntc, orig_idx = idx; tx_buf = &txq->tx_buf[idx]; - while (tx_buf->compl_tag == (int)compl_tag) { - if (tx_buf->skb) { - idpf_tx_splitq_clean_hdr(txq, tx_buf, cleaned, budget); - } else if (dma_unmap_len(tx_buf, len)) { - dma_unmap_page(txq->dev, - dma_unmap_addr(tx_buf, dma), - dma_unmap_len(tx_buf, len), - DMA_TO_DEVICE); - dma_unmap_len_set(tx_buf, len, 0); - } + if (unlikely(tx_buf->type <= LIBETH_SQE_CTX || + idpf_tx_buf_compl_tag(tx_buf) != compl_tag)) + return false; + + if (tx_buf->type == LIBETH_SQE_SKB) + libeth_tx_complete(tx_buf, &cp); - memset(tx_buf, 0, sizeof(struct idpf_tx_buf)); - tx_buf->compl_tag = IDPF_SPLITQ_TX_INVAL_COMPL_TAG; + idpf_tx_clean_buf_ring_bump_ntc(txq, idx, tx_buf); - num_descs_cleaned++; + while (idpf_tx_buf_compl_tag(tx_buf) == compl_tag) { + libeth_tx_complete(tx_buf, &cp); idpf_tx_clean_buf_ring_bump_ntc(txq, idx, tx_buf); } - /* If we didn't clean anything on the ring for this completion, there's - * nothing more to do. - */ - if (unlikely(!num_descs_cleaned)) - return false; - - /* Otherwise, if we did clean a packet on the ring directly, it's safe - * to assume that the descriptors starting from the original - * next_to_clean up until the previously cleaned packet can be reused. - * Therefore, we will go back in the ring and stash any buffers still - * in the ring into the hash table to be cleaned later. + /* + * It's possible the packet we just cleaned was an out of order + * completion, which means we can stash the buffers starting from + * the original next_to_clean and reuse the descriptors. We need + * to compare the descriptor ring next_to_clean packet's "first" buffer + * to the "first" buffer of the packet we just cleaned to determine if + * this is the case. Howevever, next_to_clean can point to either a + * reserved buffer that corresponds to a context descriptor used for the + * next_to_clean packet (TSO packet) or the "first" buffer (single + * packet). The orig_idx from the packet we just cleaned will always + * point to the "first" buffer. If next_to_clean points to a reserved + * buffer, let's bump ntc once and start the comparison from there. */ + ntc = txq->next_to_clean; tx_buf = &txq->tx_buf[ntc]; - while (tx_buf != &txq->tx_buf[orig_idx]) { - idpf_stash_flow_sch_buffers(txq, tx_buf); + + if (tx_buf->type == LIBETH_SQE_CTX) idpf_tx_clean_buf_ring_bump_ntc(txq, ntc, tx_buf); - } - /* Finally, update next_to_clean to reflect the work that was just done - * on the ring, if any. If the packet was only cleaned from the hash - * table, the ring will not be impacted, therefore we should not touch - * next_to_clean. The updated idx is used here + /* + * If ntc still points to a different "first" buffer, clean the + * descriptor ring and stash all of the buffers for later cleaning. If + * we cannot stash all of the buffers, next_to_clean will point to the + * "first" buffer of the packet that could not be stashed and cleaning + * will start there next time. + */ + if (unlikely(tx_buf != &txq->tx_buf[orig_idx] && + !idpf_tx_splitq_clean(txq, orig_idx, budget, cleaned, + true))) + return true; + + /* + * Otherwise, update next_to_clean to reflect the cleaning that was + * done above. */ txq->next_to_clean = idx; @@ -1771,17 +1937,18 @@ static bool idpf_tx_clean_buf_ring(struct idpf_queue *txq, u16 compl_tag, * * Returns bytes/packets cleaned */ -static void idpf_tx_handle_rs_completion(struct idpf_queue *txq, +static void idpf_tx_handle_rs_completion(struct idpf_tx_queue *txq, struct idpf_splitq_tx_compl_desc *desc, - struct idpf_cleaned_stats *cleaned, + struct libeth_sq_napi_stats *cleaned, int budget) { u16 compl_tag; - if (!test_bit(__IDPF_Q_FLOW_SCH_EN, txq->flags)) { + if (!idpf_queue_has(FLOW_SCH_EN, txq)) { u16 head = le16_to_cpu(desc->q_head_compl_tag.q_head); - return idpf_tx_splitq_clean(txq, head, budget, cleaned, false); + idpf_tx_splitq_clean(txq, head, budget, cleaned, false); + return; } compl_tag = le16_to_cpu(desc->q_head_compl_tag.compl_tag); @@ -1801,24 +1968,23 @@ static void idpf_tx_handle_rs_completion(struct idpf_queue *txq, * * Returns true if there's any budget left (e.g. the clean is finished) */ -static bool idpf_tx_clean_complq(struct idpf_queue *complq, int budget, +static bool idpf_tx_clean_complq(struct idpf_compl_queue *complq, int budget, int *cleaned) { struct idpf_splitq_tx_compl_desc *tx_desc; - struct idpf_vport *vport = complq->vport; s16 ntc = complq->next_to_clean; struct idpf_netdev_priv *np; unsigned int complq_budget; bool complq_ok = true; int i; - complq_budget = vport->compln_clean_budget; - tx_desc = IDPF_SPLITQ_TX_COMPLQ_DESC(complq, ntc); + complq_budget = complq->clean_budget; + tx_desc = &complq->comp[ntc]; ntc -= complq->desc_count; do { - struct idpf_cleaned_stats cleaned_stats = { }; - struct idpf_queue *tx_q; + struct libeth_sq_napi_stats cleaned_stats = { }; + struct idpf_tx_queue *tx_q; int rel_tx_qid; u16 hw_head; u8 ctype; /* completion type */ @@ -1827,7 +1993,7 @@ static bool idpf_tx_clean_complq(struct idpf_queue *complq, int budget, /* if the descriptor isn't done, no work yet to do */ gen = le16_get_bits(tx_desc->qid_comptype_gen, IDPF_TXD_COMPLQ_GEN_M); - if (test_bit(__IDPF_Q_GEN_CHK, complq->flags) != gen) + if (idpf_queue_has(GEN_CHK, complq) != gen) break; /* Find necessary info of TX queue to clean buffers */ @@ -1835,8 +2001,7 @@ static bool idpf_tx_clean_complq(struct idpf_queue *complq, int budget, IDPF_TXD_COMPLQ_QID_M); if (rel_tx_qid >= complq->txq_grp->num_txq || !complq->txq_grp->txqs[rel_tx_qid]) { - dev_err(&complq->vport->adapter->pdev->dev, - "TxQ not found\n"); + netdev_err(complq->netdev, "TxQ not found\n"); goto fetch_next_desc; } tx_q = complq->txq_grp->txqs[rel_tx_qid]; @@ -1859,15 +2024,14 @@ static bool idpf_tx_clean_complq(struct idpf_queue *complq, int budget, idpf_tx_handle_sw_marker(tx_q); break; default: - dev_err(&tx_q->vport->adapter->pdev->dev, - "Unknown TX completion type: %d\n", - ctype); + netdev_err(tx_q->netdev, + "Unknown TX completion type: %d\n", ctype); goto fetch_next_desc; } u64_stats_update_begin(&tx_q->stats_sync); - u64_stats_add(&tx_q->q_stats.tx.packets, cleaned_stats.packets); - u64_stats_add(&tx_q->q_stats.tx.bytes, cleaned_stats.bytes); + u64_stats_add(&tx_q->q_stats.packets, cleaned_stats.packets); + u64_stats_add(&tx_q->q_stats.bytes, cleaned_stats.bytes); tx_q->cleaned_pkts += cleaned_stats.packets; tx_q->cleaned_bytes += cleaned_stats.bytes; complq->num_completions++; @@ -1878,8 +2042,8 @@ fetch_next_desc: ntc++; if (unlikely(!ntc)) { ntc -= complq->desc_count; - tx_desc = IDPF_SPLITQ_TX_COMPLQ_DESC(complq, 0); - change_bit(__IDPF_Q_GEN_CHK, complq->flags); + tx_desc = &complq->comp[0]; + idpf_queue_change(GEN_CHK, complq); } prefetch(tx_desc); @@ -1895,9 +2059,9 @@ fetch_next_desc: IDPF_TX_COMPLQ_OVERFLOW_THRESH(complq))) complq_ok = false; - np = netdev_priv(complq->vport->netdev); + np = netdev_priv(complq->netdev); for (i = 0; i < complq->txq_grp->num_txq; ++i) { - struct idpf_queue *tx_q = complq->txq_grp->txqs[i]; + struct idpf_tx_queue *tx_q = complq->txq_grp->txqs[i]; struct netdev_queue *nq; bool dont_wake; @@ -1908,11 +2072,11 @@ fetch_next_desc: *cleaned += tx_q->cleaned_pkts; /* Update BQL */ - nq = netdev_get_tx_queue(tx_q->vport->netdev, tx_q->idx); + nq = netdev_get_tx_queue(tx_q->netdev, tx_q->idx); dont_wake = !complq_ok || IDPF_TX_BUF_RSV_LOW(tx_q) || np->state != __IDPF_VPORT_UP || - !netif_carrier_ok(tx_q->vport->netdev); + !netif_carrier_ok(tx_q->netdev); /* Check if the TXQ needs to and can be restarted */ __netif_txq_completed_wake(nq, tx_q->cleaned_pkts, tx_q->cleaned_bytes, IDPF_DESC_UNUSED(tx_q), IDPF_TX_WAKE_THRESH, @@ -1969,40 +2133,17 @@ void idpf_tx_splitq_build_flow_desc(union idpf_tx_flex_desc *desc, } /** - * idpf_tx_maybe_stop_common - 1st level check for common Tx stop conditions - * @tx_q: the queue to be checked - * @size: number of descriptors we want to assure is available - * - * Returns 0 if stop is not needed - */ -int idpf_tx_maybe_stop_common(struct idpf_queue *tx_q, unsigned int size) -{ - struct netdev_queue *nq; - - if (likely(IDPF_DESC_UNUSED(tx_q) >= size)) - return 0; - - u64_stats_update_begin(&tx_q->stats_sync); - u64_stats_inc(&tx_q->q_stats.tx.q_busy); - u64_stats_update_end(&tx_q->stats_sync); - - nq = netdev_get_tx_queue(tx_q->vport->netdev, tx_q->idx); - - return netif_txq_maybe_stop(nq, IDPF_DESC_UNUSED(tx_q), size, size); -} - -/** * idpf_tx_maybe_stop_splitq - 1st level check for Tx splitq stop conditions * @tx_q: the queue to be checked * @descs_needed: number of descriptors required for this packet * * Returns 0 if stop is not needed */ -static int idpf_tx_maybe_stop_splitq(struct idpf_queue *tx_q, +static int idpf_tx_maybe_stop_splitq(struct idpf_tx_queue *tx_q, unsigned int descs_needed) { if (idpf_tx_maybe_stop_common(tx_q, descs_needed)) - goto splitq_stop; + goto out; /* If there are too many outstanding completions expected on the * completion queue, stop the TX queue to give the device some time to @@ -2021,10 +2162,12 @@ static int idpf_tx_maybe_stop_splitq(struct idpf_queue *tx_q, return 0; splitq_stop: + netif_stop_subqueue(tx_q->netdev, tx_q->idx); + +out: u64_stats_update_begin(&tx_q->stats_sync); - u64_stats_inc(&tx_q->q_stats.tx.q_busy); + u64_stats_inc(&tx_q->q_stats.q_busy); u64_stats_update_end(&tx_q->stats_sync); - netif_stop_subqueue(tx_q->vport->netdev, tx_q->idx); return -EBUSY; } @@ -2039,15 +2182,19 @@ splitq_stop: * to do a register write to update our queue status. We know this can only * mean tail here as HW should be owning head for TX. */ -void idpf_tx_buf_hw_update(struct idpf_queue *tx_q, u32 val, +void idpf_tx_buf_hw_update(struct idpf_tx_queue *tx_q, u32 val, bool xmit_more) { struct netdev_queue *nq; - nq = netdev_get_tx_queue(tx_q->vport->netdev, tx_q->idx); + nq = netdev_get_tx_queue(tx_q->netdev, tx_q->idx); tx_q->next_to_use = val; - idpf_tx_maybe_stop_common(tx_q, IDPF_TX_DESC_NEEDED); + if (idpf_tx_maybe_stop_common(tx_q, IDPF_TX_DESC_NEEDED)) { + u64_stats_update_begin(&tx_q->stats_sync); + u64_stats_inc(&tx_q->q_stats.q_busy); + u64_stats_update_end(&tx_q->stats_sync); + } /* Force memory writes to complete before letting h/w * know there are new descriptors to fetch. (Only @@ -2068,7 +2215,7 @@ void idpf_tx_buf_hw_update(struct idpf_queue *tx_q, u32 val, * * Returns number of data descriptors needed for this skb. */ -unsigned int idpf_tx_desc_count_required(struct idpf_queue *txq, +unsigned int idpf_tx_desc_count_required(struct idpf_tx_queue *txq, struct sk_buff *skb) { const struct skb_shared_info *shinfo; @@ -2101,7 +2248,7 @@ unsigned int idpf_tx_desc_count_required(struct idpf_queue *txq, count = idpf_size_to_txd_count(skb->len); u64_stats_update_begin(&txq->stats_sync); - u64_stats_inc(&txq->q_stats.tx.linearize); + u64_stats_inc(&txq->q_stats.linearize); u64_stats_update_end(&txq->stats_sync); } @@ -2115,11 +2262,17 @@ unsigned int idpf_tx_desc_count_required(struct idpf_queue *txq, * @first: original first buffer info buffer for packet * @idx: starting point on ring to unwind */ -void idpf_tx_dma_map_error(struct idpf_queue *txq, struct sk_buff *skb, +void idpf_tx_dma_map_error(struct idpf_tx_queue *txq, struct sk_buff *skb, struct idpf_tx_buf *first, u16 idx) { + struct libeth_sq_napi_stats ss = { }; + struct libeth_cq_pp cp = { + .dev = txq->dev, + .ss = &ss, + }; + u64_stats_update_begin(&txq->stats_sync); - u64_stats_inc(&txq->q_stats.tx.dma_map_errs); + u64_stats_inc(&txq->q_stats.dma_map_errs); u64_stats_update_end(&txq->stats_sync); /* clear dma mappings for failed tx_buf map */ @@ -2127,7 +2280,7 @@ void idpf_tx_dma_map_error(struct idpf_queue *txq, struct sk_buff *skb, struct idpf_tx_buf *tx_buf; tx_buf = &txq->tx_buf[idx]; - idpf_tx_buf_rel(txq, tx_buf); + libeth_tx_complete(tx_buf, &cp); if (tx_buf == first) break; if (idx == 0) @@ -2142,7 +2295,7 @@ void idpf_tx_dma_map_error(struct idpf_queue *txq, struct sk_buff *skb, * used one additional descriptor for a context * descriptor. Reset that here. */ - tx_desc = IDPF_FLEX_TX_DESC(txq, idx); + tx_desc = &txq->flex_tx[idx]; memset(tx_desc, 0, sizeof(struct idpf_flex_tx_ctx_desc)); if (idx == 0) idx = txq->desc_count; @@ -2158,7 +2311,7 @@ void idpf_tx_dma_map_error(struct idpf_queue *txq, struct sk_buff *skb, * @txq: the tx ring to wrap * @ntu: ring index to bump */ -static unsigned int idpf_tx_splitq_bump_ntu(struct idpf_queue *txq, u16 ntu) +static unsigned int idpf_tx_splitq_bump_ntu(struct idpf_tx_queue *txq, u16 ntu) { ntu++; @@ -2180,7 +2333,7 @@ static unsigned int idpf_tx_splitq_bump_ntu(struct idpf_queue *txq, u16 ntu) * and gets a physical address for each memory location and programs * it and the length into the transmit flex descriptor. */ -static void idpf_tx_splitq_map(struct idpf_queue *tx_q, +static void idpf_tx_splitq_map(struct idpf_tx_queue *tx_q, struct idpf_tx_splitq_params *params, struct idpf_tx_buf *first) { @@ -2201,11 +2354,12 @@ static void idpf_tx_splitq_map(struct idpf_queue *tx_q, data_len = skb->data_len; size = skb_headlen(skb); - tx_desc = IDPF_FLEX_TX_DESC(tx_q, i); + tx_desc = &tx_q->flex_tx[i]; dma = dma_map_single(tx_q->dev, skb->data, size, DMA_TO_DEVICE); tx_buf = first; + first->nr_frags = 0; params->compl_tag = (tx_q->compl_tag_cur_gen << tx_q->compl_tag_gen_s) | i; @@ -2216,7 +2370,9 @@ static void idpf_tx_splitq_map(struct idpf_queue *tx_q, if (dma_mapping_error(tx_q->dev, dma)) return idpf_tx_dma_map_error(tx_q, skb, first, i); - tx_buf->compl_tag = params->compl_tag; + first->nr_frags++; + idpf_tx_buf_compl_tag(tx_buf) = params->compl_tag; + tx_buf->type = LIBETH_SQE_FRAG; /* record length, and DMA address */ dma_unmap_len_set(tx_buf, len, size); @@ -2270,14 +2426,15 @@ static void idpf_tx_splitq_map(struct idpf_queue *tx_q, idpf_tx_splitq_build_desc(tx_desc, params, td_cmd, max_data); - tx_desc++; - i++; - - if (i == tx_q->desc_count) { - tx_desc = IDPF_FLEX_TX_DESC(tx_q, 0); + if (unlikely(++i == tx_q->desc_count)) { + tx_buf = tx_q->tx_buf; + tx_desc = &tx_q->flex_tx[0]; i = 0; tx_q->compl_tag_cur_gen = IDPF_TX_ADJ_COMPL_TAG_GEN(tx_q); + } else { + tx_buf++; + tx_desc++; } /* Since this packet has a buffer that is going to span @@ -2290,8 +2447,8 @@ static void idpf_tx_splitq_map(struct idpf_queue *tx_q, * simply pass over these holes and finish cleaning the * rest of the packet. */ - memset(&tx_q->tx_buf[i], 0, sizeof(struct idpf_tx_buf)); - tx_q->tx_buf[i].compl_tag = params->compl_tag; + tx_buf->type = LIBETH_SQE_EMPTY; + idpf_tx_buf_compl_tag(tx_buf) = params->compl_tag; /* Adjust the DMA offset and the remaining size of the * fragment. On the first iteration of this loop, @@ -2315,13 +2472,15 @@ static void idpf_tx_splitq_map(struct idpf_queue *tx_q, break; idpf_tx_splitq_build_desc(tx_desc, params, td_cmd, size); - tx_desc++; - i++; - if (i == tx_q->desc_count) { - tx_desc = IDPF_FLEX_TX_DESC(tx_q, 0); + if (unlikely(++i == tx_q->desc_count)) { + tx_buf = tx_q->tx_buf; + tx_desc = &tx_q->flex_tx[0]; i = 0; tx_q->compl_tag_cur_gen = IDPF_TX_ADJ_COMPL_TAG_GEN(tx_q); + } else { + tx_buf++; + tx_desc++; } size = skb_frag_size(frag); @@ -2329,26 +2488,24 @@ static void idpf_tx_splitq_map(struct idpf_queue *tx_q, dma = skb_frag_dma_map(tx_q->dev, frag, 0, size, DMA_TO_DEVICE); - - tx_buf = &tx_q->tx_buf[i]; } /* record SW timestamp if HW timestamp is not available */ skb_tx_timestamp(skb); + first->type = LIBETH_SQE_SKB; + /* write last descriptor with RS and EOP bits */ + first->rs_idx = i; td_cmd |= params->eop_cmd; idpf_tx_splitq_build_desc(tx_desc, params, td_cmd, size); i = idpf_tx_splitq_bump_ntu(tx_q, i); - /* set next_to_watch value indicating a packet is present */ - first->next_to_watch = tx_desc; - tx_q->txq_grp->num_completions_pending++; /* record bytecount for BQL */ - nq = netdev_get_tx_queue(tx_q->vport->netdev, tx_q->idx); - netdev_tx_sent_queue(nq, first->bytecount); + nq = netdev_get_tx_queue(tx_q->netdev, tx_q->idx); + netdev_tx_sent_queue(nq, first->bytes); idpf_tx_buf_hw_update(tx_q, i, netdev_xmit_more()); } @@ -2524,8 +2681,8 @@ static bool __idpf_chk_linearize(struct sk_buff *skb, unsigned int max_bufs) * E.g.: a packet with 7 fragments can require 9 DMA transactions; 1 for TSO * header, 1 for segment payload, and then 7 for the fragments. */ -bool idpf_chk_linearize(struct sk_buff *skb, unsigned int max_bufs, - unsigned int count) +static bool idpf_chk_linearize(struct sk_buff *skb, unsigned int max_bufs, + unsigned int count) { if (likely(count < max_bufs)) return false; @@ -2543,16 +2700,15 @@ bool idpf_chk_linearize(struct sk_buff *skb, unsigned int max_bufs, * ring entry to reflect that this index is a context descriptor */ static struct idpf_flex_tx_ctx_desc * -idpf_tx_splitq_get_ctx_desc(struct idpf_queue *txq) +idpf_tx_splitq_get_ctx_desc(struct idpf_tx_queue *txq) { struct idpf_flex_tx_ctx_desc *desc; int i = txq->next_to_use; - memset(&txq->tx_buf[i], 0, sizeof(struct idpf_tx_buf)); - txq->tx_buf[i].compl_tag = IDPF_SPLITQ_TX_INVAL_COMPL_TAG; + txq->tx_buf[i].type = LIBETH_SQE_CTX; /* grab the next descriptor */ - desc = IDPF_FLEX_TX_CTX_DESC(txq, i); + desc = &txq->flex_ctx[i]; txq->next_to_use = idpf_tx_splitq_bump_ntu(txq, i); return desc; @@ -2563,10 +2719,10 @@ idpf_tx_splitq_get_ctx_desc(struct idpf_queue *txq) * @tx_q: queue to send buffer on * @skb: pointer to skb */ -netdev_tx_t idpf_tx_drop_skb(struct idpf_queue *tx_q, struct sk_buff *skb) +netdev_tx_t idpf_tx_drop_skb(struct idpf_tx_queue *tx_q, struct sk_buff *skb) { u64_stats_update_begin(&tx_q->stats_sync); - u64_stats_inc(&tx_q->q_stats.tx.skb_drops); + u64_stats_inc(&tx_q->q_stats.skb_drops); u64_stats_update_end(&tx_q->stats_sync); idpf_tx_buf_hw_update(tx_q, tx_q->next_to_use, false); @@ -2584,7 +2740,7 @@ netdev_tx_t idpf_tx_drop_skb(struct idpf_queue *tx_q, struct sk_buff *skb) * Returns NETDEV_TX_OK if sent, else an error code */ static netdev_tx_t idpf_tx_splitq_frame(struct sk_buff *skb, - struct idpf_queue *tx_q) + struct idpf_tx_queue *tx_q) { struct idpf_tx_splitq_params tx_params = { }; struct idpf_tx_buf *first; @@ -2624,7 +2780,7 @@ static netdev_tx_t idpf_tx_splitq_frame(struct sk_buff *skb, ctx_desc->tso.qw0.hdr_len = tx_params.offload.tso_hdr_len; u64_stats_update_begin(&tx_q->stats_sync); - u64_stats_inc(&tx_q->q_stats.tx.lso_pkts); + u64_stats_inc(&tx_q->q_stats.lso_pkts); u64_stats_update_end(&tx_q->stats_sync); } @@ -2633,15 +2789,15 @@ static netdev_tx_t idpf_tx_splitq_frame(struct sk_buff *skb, first->skb = skb; if (tso) { - first->gso_segs = tx_params.offload.tso_segs; - first->bytecount = skb->len + - ((first->gso_segs - 1) * tx_params.offload.tso_hdr_len); + first->packets = tx_params.offload.tso_segs; + first->bytes = skb->len + + ((first->packets - 1) * tx_params.offload.tso_hdr_len); } else { - first->gso_segs = 1; - first->bytecount = max_t(unsigned int, skb->len, ETH_ZLEN); + first->packets = 1; + first->bytes = max_t(unsigned int, skb->len, ETH_ZLEN); } - if (test_bit(__IDPF_Q_FLOW_SCH_EN, tx_q->flags)) { + if (idpf_queue_has(FLOW_SCH_EN, tx_q)) { tx_params.dtype = IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE; tx_params.eop_cmd = IDPF_TXD_FLEX_FLOW_CMD_EOP; /* Set the RE bit to catch any packets that may have not been @@ -2671,17 +2827,16 @@ static netdev_tx_t idpf_tx_splitq_frame(struct sk_buff *skb, } /** - * idpf_tx_splitq_start - Selects the right Tx queue to send buffer + * idpf_tx_start - Selects the right Tx queue to send buffer * @skb: send buffer * @netdev: network interface device structure * * Returns NETDEV_TX_OK if sent, else an error code */ -netdev_tx_t idpf_tx_splitq_start(struct sk_buff *skb, - struct net_device *netdev) +netdev_tx_t idpf_tx_start(struct sk_buff *skb, struct net_device *netdev) { struct idpf_vport *vport = idpf_netdev_to_vport(netdev); - struct idpf_queue *tx_q; + struct idpf_tx_queue *tx_q; if (unlikely(skb_get_queue_mapping(skb) >= vport->num_txq)) { dev_kfree_skb_any(skb); @@ -2700,31 +2855,10 @@ netdev_tx_t idpf_tx_splitq_start(struct sk_buff *skb, return NETDEV_TX_OK; } - return idpf_tx_splitq_frame(skb, tx_q); -} - -/** - * idpf_ptype_to_htype - get a hash type - * @decoded: Decoded Rx packet type related fields - * - * Returns appropriate hash type (such as PKT_HASH_TYPE_L2/L3/L4) to be used by - * skb_set_hash based on PTYPE as parsed by HW Rx pipeline and is part of - * Rx desc. - */ -enum pkt_hash_types idpf_ptype_to_htype(const struct idpf_rx_ptype_decoded *decoded) -{ - if (!decoded->known) - return PKT_HASH_TYPE_NONE; - if (decoded->payload_layer == IDPF_RX_PTYPE_PAYLOAD_LAYER_PAY2 && - decoded->inner_prot) - return PKT_HASH_TYPE_L4; - if (decoded->payload_layer == IDPF_RX_PTYPE_PAYLOAD_LAYER_PAY2 && - decoded->outer_ip) - return PKT_HASH_TYPE_L3; - if (decoded->outer_ip == IDPF_RX_PTYPE_OUTER_L2) - return PKT_HASH_TYPE_L2; - - return PKT_HASH_TYPE_NONE; + if (idpf_is_queue_model_split(vport->txq_model)) + return idpf_tx_splitq_frame(skb, tx_q); + else + return idpf_tx_singleq_frame(skb, tx_q); } /** @@ -2734,20 +2868,21 @@ enum pkt_hash_types idpf_ptype_to_htype(const struct idpf_rx_ptype_decoded *deco * @rx_desc: Receive descriptor * @decoded: Decoded Rx packet type related fields */ -static void idpf_rx_hash(struct idpf_queue *rxq, struct sk_buff *skb, - struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc, - struct idpf_rx_ptype_decoded *decoded) +static void +idpf_rx_hash(const struct idpf_rx_queue *rxq, struct sk_buff *skb, + const struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc, + struct libeth_rx_pt decoded) { u32 hash; - if (unlikely(!idpf_is_feature_ena(rxq->vport, NETIF_F_RXHASH))) + if (!libeth_rx_pt_has_hash(rxq->netdev, decoded)) return; hash = le16_to_cpu(rx_desc->hash1) | (rx_desc->ff2_mirrid_hash2.hash2 << 16) | (rx_desc->hash3 << 24); - skb_set_hash(skb, hash, idpf_ptype_to_htype(decoded)); + libeth_rx_pt_set_hash(skb, hash, decoded); } /** @@ -2759,92 +2894,83 @@ static void idpf_rx_hash(struct idpf_queue *rxq, struct sk_buff *skb, * * skb->protocol must be set before this function is called */ -static void idpf_rx_csum(struct idpf_queue *rxq, struct sk_buff *skb, - struct idpf_rx_csum_decoded *csum_bits, - struct idpf_rx_ptype_decoded *decoded) +static void idpf_rx_csum(struct idpf_rx_queue *rxq, struct sk_buff *skb, + struct idpf_rx_csum_decoded csum_bits, + struct libeth_rx_pt decoded) { bool ipv4, ipv6; /* check if Rx checksum is enabled */ - if (unlikely(!idpf_is_feature_ena(rxq->vport, NETIF_F_RXCSUM))) + if (!libeth_rx_pt_has_checksum(rxq->netdev, decoded)) return; /* check if HW has decoded the packet and checksum */ - if (!(csum_bits->l3l4p)) + if (unlikely(!csum_bits.l3l4p)) return; - ipv4 = IDPF_RX_PTYPE_TO_IPV(decoded, IDPF_RX_PTYPE_OUTER_IPV4); - ipv6 = IDPF_RX_PTYPE_TO_IPV(decoded, IDPF_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 && (csum_bits->ipe || csum_bits->eipe)) + if (unlikely(ipv4 && (csum_bits.ipe || csum_bits.eipe))) goto checksum_fail; - if (ipv6 && csum_bits->ipv6exadd) + if (unlikely(ipv6 && csum_bits.ipv6exadd)) return; /* check for L4 errors and handle packets that were not able to be * checksummed */ - if (csum_bits->l4e) + if (unlikely(csum_bits.l4e)) goto checksum_fail; - /* Only report checksum unnecessary for ICMP, TCP, UDP, or SCTP */ - switch (decoded->inner_prot) { - case IDPF_RX_PTYPE_INNER_PROT_ICMP: - case IDPF_RX_PTYPE_INNER_PROT_TCP: - case IDPF_RX_PTYPE_INNER_PROT_UDP: - if (!csum_bits->raw_csum_inv) { - u16 csum = csum_bits->raw_csum; - - skb->csum = csum_unfold((__force __sum16)~swab16(csum)); - skb->ip_summed = CHECKSUM_COMPLETE; - } else { - skb->ip_summed = CHECKSUM_UNNECESSARY; - } - break; - case IDPF_RX_PTYPE_INNER_PROT_SCTP: + if (csum_bits.raw_csum_inv || + decoded.inner_prot == LIBETH_RX_PT_INNER_SCTP) { skb->ip_summed = CHECKSUM_UNNECESSARY; - break; - default: - break; + return; } + skb->csum = csum_unfold((__force __sum16)~swab16(csum_bits.raw_csum)); + skb->ip_summed = CHECKSUM_COMPLETE; + return; checksum_fail: u64_stats_update_begin(&rxq->stats_sync); - u64_stats_inc(&rxq->q_stats.rx.hw_csum_err); + u64_stats_inc(&rxq->q_stats.hw_csum_err); u64_stats_update_end(&rxq->stats_sync); } /** * idpf_rx_splitq_extract_csum_bits - Extract checksum bits from descriptor * @rx_desc: receive descriptor - * @csum: structure to extract checksum fields * + * Return: parsed checksum status. **/ -static void idpf_rx_splitq_extract_csum_bits(struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc, - struct idpf_rx_csum_decoded *csum) +static struct idpf_rx_csum_decoded +idpf_rx_splitq_extract_csum_bits(const struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc) { + struct idpf_rx_csum_decoded csum = { }; u8 qword0, qword1; qword0 = rx_desc->status_err0_qw0; qword1 = rx_desc->status_err0_qw1; - csum->ipe = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_XSUM_IPE_M, + csum.ipe = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_XSUM_IPE_M, + qword1); + csum.eipe = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_XSUM_EIPE_M, qword1); - csum->eipe = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_XSUM_EIPE_M, + csum.l4e = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_XSUM_L4E_M, + qword1); + csum.l3l4p = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_L3L4P_M, qword1); - csum->l4e = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_XSUM_L4E_M, - qword1); - csum->l3l4p = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_L3L4P_M, - qword1); - csum->ipv6exadd = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_IPV6EXADD_M, - qword0); - csum->raw_csum_inv = + csum.ipv6exadd = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_IPV6EXADD_M, + qword0); + csum.raw_csum_inv = le16_get_bits(rx_desc->ptype_err_fflags0, VIRTCHNL2_RX_FLEX_DESC_ADV_RAW_CSUM_INV_M); - csum->raw_csum = le16_to_cpu(rx_desc->misc.raw_cs); + csum.raw_csum = le16_to_cpu(rx_desc->misc.raw_cs); + + return csum; } /** @@ -2859,36 +2985,34 @@ static void idpf_rx_splitq_extract_csum_bits(struct virtchnl2_rx_flex_desc_adv_n * Populate the skb fields with the total number of RSC segments, RSC payload * length and packet type. */ -static int idpf_rx_rsc(struct idpf_queue *rxq, struct sk_buff *skb, - struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc, - struct idpf_rx_ptype_decoded *decoded) +static int idpf_rx_rsc(struct idpf_rx_queue *rxq, struct sk_buff *skb, + const struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc, + struct libeth_rx_pt decoded) { u16 rsc_segments, rsc_seg_len; bool ipv4, ipv6; int len; - if (unlikely(!decoded->outer_ip)) + if (unlikely(libeth_rx_pt_get_ip_ver(decoded) == + LIBETH_RX_PT_OUTER_L2)) return -EINVAL; rsc_seg_len = le16_to_cpu(rx_desc->misc.rscseglen); if (unlikely(!rsc_seg_len)) return -EINVAL; - ipv4 = IDPF_RX_PTYPE_TO_IPV(decoded, IDPF_RX_PTYPE_OUTER_IPV4); - ipv6 = IDPF_RX_PTYPE_TO_IPV(decoded, IDPF_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 (unlikely(!(ipv4 ^ ipv6))) return -EINVAL; rsc_segments = DIV_ROUND_UP(skb->data_len, rsc_seg_len); - if (unlikely(rsc_segments == 1)) - return 0; NAPI_GRO_CB(skb)->count = rsc_segments; skb_shinfo(skb)->gso_size = rsc_seg_len; skb_reset_network_header(skb); - len = skb->len - skb_transport_offset(skb); if (ipv4) { struct iphdr *ipv4h = ip_hdr(skb); @@ -2897,6 +3021,7 @@ static int idpf_rx_rsc(struct idpf_queue *rxq, struct sk_buff *skb, /* Reset and set transport header offset in skb */ skb_set_transport_header(skb, sizeof(struct iphdr)); + len = skb->len - skb_transport_offset(skb); /* Compute the TCP pseudo header checksum*/ tcp_hdr(skb)->check = @@ -2906,6 +3031,7 @@ static int idpf_rx_rsc(struct idpf_queue *rxq, struct sk_buff *skb, skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6; skb_set_transport_header(skb, sizeof(struct ipv6hdr)); + len = skb->len - skb_transport_offset(skb); tcp_hdr(skb)->check = ~tcp_v6_check(len, &ipv6h->saddr, &ipv6h->daddr, 0); } @@ -2913,7 +3039,7 @@ static int idpf_rx_rsc(struct idpf_queue *rxq, struct sk_buff *skb, tcp_gro_complete(skb); u64_stats_update_begin(&rxq->stats_sync); - u64_stats_inc(&rxq->q_stats.rx.rsc_pkts); + u64_stats_inc(&rxq->q_stats.rsc_pkts); u64_stats_update_end(&rxq->stats_sync); return 0; @@ -2929,35 +3055,30 @@ static int idpf_rx_rsc(struct idpf_queue *rxq, struct sk_buff *skb, * order to populate the hash, checksum, protocol, and * other fields within the skb. */ -static int idpf_rx_process_skb_fields(struct idpf_queue *rxq, - struct sk_buff *skb, - struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc) +static int +idpf_rx_process_skb_fields(struct idpf_rx_queue *rxq, struct sk_buff *skb, + const struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc) { - struct idpf_rx_csum_decoded csum_bits = { }; - struct idpf_rx_ptype_decoded decoded; + struct idpf_rx_csum_decoded csum_bits; + struct libeth_rx_pt decoded; u16 rx_ptype; rx_ptype = le16_get_bits(rx_desc->ptype_err_fflags0, VIRTCHNL2_RX_FLEX_DESC_ADV_PTYPE_M); - - decoded = rxq->vport->rx_ptype_lkup[rx_ptype]; - /* If we don't know the ptype we can't do anything else with it. Just - * pass it up the stack as-is. - */ - if (!decoded.known) - return 0; + decoded = rxq->rx_ptype_lkup[rx_ptype]; /* process RSS/hash */ - idpf_rx_hash(rxq, skb, rx_desc, &decoded); + idpf_rx_hash(rxq, skb, rx_desc, decoded); - skb->protocol = eth_type_trans(skb, rxq->vport->netdev); + skb->protocol = eth_type_trans(skb, rxq->netdev); + skb_record_rx_queue(skb, rxq->idx); if (le16_get_bits(rx_desc->hdrlen_flags, VIRTCHNL2_RX_FLEX_DESC_ADV_RSC_M)) - return idpf_rx_rsc(rxq, skb, rx_desc, &decoded); + return idpf_rx_rsc(rxq, skb, rx_desc, decoded); - idpf_rx_splitq_extract_csum_bits(rx_desc, &csum_bits); - idpf_rx_csum(rxq, skb, &csum_bits, &decoded); + csum_bits = idpf_rx_splitq_extract_csum_bits(rx_desc); + idpf_rx_csum(rxq, skb, csum_bits, decoded); return 0; } @@ -2975,104 +3096,73 @@ static int idpf_rx_process_skb_fields(struct idpf_queue *rxq, void idpf_rx_add_frag(struct idpf_rx_buf *rx_buf, struct sk_buff *skb, unsigned int size) { - skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, rx_buf->page, - rx_buf->page_offset, size, rx_buf->truesize); + u32 hr = rx_buf->page->pp->p.offset; - rx_buf->page = NULL; + skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, rx_buf->page, + rx_buf->offset + hr, size, rx_buf->truesize); } /** - * idpf_rx_construct_skb - Allocate skb and populate it - * @rxq: Rx descriptor queue - * @rx_buf: Rx buffer to pull data from - * @size: the length of the packet + * idpf_rx_hsplit_wa - handle header buffer overflows and split errors + * @hdr: Rx buffer for the headers + * @buf: Rx buffer for the payload + * @data_len: number of bytes received to the payload buffer * - * This function allocates an skb. It then populates it with the page - * data from the current receive descriptor, taking care to set up the - * skb correctly. + * When a header buffer overflow occurs or the HW was unable do parse the + * packet type to perform header split, the whole frame gets placed to the + * payload buffer. We can't build a valid skb around a payload buffer when + * the header split is active since it doesn't reserve any head- or tailroom. + * In that case, copy either the whole frame when it's short or just the + * Ethernet header to the header buffer to be able to build an skb and adjust + * the data offset in the payload buffer, IOW emulate the header split. + * + * Return: number of bytes copied to the header buffer. */ -struct sk_buff *idpf_rx_construct_skb(struct idpf_queue *rxq, - struct idpf_rx_buf *rx_buf, - unsigned int size) +static u32 idpf_rx_hsplit_wa(const struct libeth_fqe *hdr, + struct libeth_fqe *buf, u32 data_len) { - unsigned int headlen; - struct sk_buff *skb; - void *va; - - va = page_address(rx_buf->page) + rx_buf->page_offset; - - /* prefetch first cache line of first page */ - net_prefetch(va); - /* allocate a skb to store the frags */ - skb = __napi_alloc_skb(&rxq->q_vector->napi, IDPF_RX_HDR_SIZE, - GFP_ATOMIC); - if (unlikely(!skb)) { - idpf_rx_put_page(rx_buf); - - return NULL; - } - - skb_record_rx_queue(skb, rxq->idx); - skb_mark_for_recycle(skb); + u32 copy = data_len <= L1_CACHE_BYTES ? data_len : ETH_HLEN; + const void *src; + void *dst; - /* Determine available headroom for copy */ - headlen = size; - if (headlen > IDPF_RX_HDR_SIZE) - headlen = eth_get_headlen(skb->dev, va, IDPF_RX_HDR_SIZE); - - /* align pull length to size of long to optimize memcpy performance */ - memcpy(__skb_put(skb, headlen), va, ALIGN(headlen, sizeof(long))); - - /* if we exhaust the linear part then add what is left as a frag */ - size -= headlen; - if (!size) { - idpf_rx_put_page(rx_buf); - - return skb; - } + if (!libeth_rx_sync_for_cpu(buf, copy)) + return 0; - skb_add_rx_frag(skb, 0, rx_buf->page, rx_buf->page_offset + headlen, - size, rx_buf->truesize); + dst = page_address(hdr->page) + hdr->offset + hdr->page->pp->p.offset; + src = page_address(buf->page) + buf->offset + buf->page->pp->p.offset; + memcpy(dst, src, LARGEST_ALIGN(copy)); - /* Since we're giving the page to the stack, clear our reference to it. - * We'll get a new one during buffer posting. - */ - rx_buf->page = NULL; + buf->offset += copy; - return skb; + return copy; } /** - * idpf_rx_hdr_construct_skb - Allocate skb and populate it from header buffer - * @rxq: Rx descriptor queue - * @va: Rx buffer to pull data from + * idpf_rx_build_skb - Allocate skb and populate it from header buffer + * @buf: Rx buffer to pull data from * @size: the length of the packet * * This function allocates an skb. It then populates it with the page data from * the current receive descriptor, taking care to set up the skb correctly. - * This specifically uses a header buffer to start building the skb. */ -static struct sk_buff *idpf_rx_hdr_construct_skb(struct idpf_queue *rxq, - const void *va, - unsigned int size) +struct sk_buff *idpf_rx_build_skb(const struct libeth_fqe *buf, u32 size) { + u32 hr = buf->page->pp->p.offset; struct sk_buff *skb; + void *va; - /* allocate a skb to store the frags */ - skb = __napi_alloc_skb(&rxq->q_vector->napi, size, GFP_ATOMIC); + va = page_address(buf->page) + buf->offset; + prefetch(va + hr); + + skb = napi_build_skb(va, buf->truesize); if (unlikely(!skb)) return NULL; - skb_record_rx_queue(skb, rxq->idx); - - memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long))); - - /* More than likely, a payload fragment, which will use a page from - * page_pool will be added to the SKB so mark it for recycle - * preemptively. And if not, it's inconsequential. - */ skb_mark_for_recycle(skb); + skb_reserve(skb, hr); + __skb_put(skb, size); + return skb; } @@ -3115,31 +3205,27 @@ static bool idpf_rx_splitq_is_eop(struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_de * * Returns amount of work completed */ -static int idpf_rx_splitq_clean(struct idpf_queue *rxq, int budget) +static int idpf_rx_splitq_clean(struct idpf_rx_queue *rxq, int budget) { int total_rx_bytes = 0, total_rx_pkts = 0; - struct idpf_queue *rx_bufq = NULL; + struct idpf_buf_queue *rx_bufq = NULL; struct sk_buff *skb = rxq->skb; u16 ntc = rxq->next_to_clean; /* Process Rx packets bounded by budget */ while (likely(total_rx_pkts < budget)) { struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc; + struct libeth_fqe *hdr, *rx_buf = NULL; struct idpf_sw_queue *refillq = NULL; struct idpf_rxq_set *rxq_set = NULL; - struct idpf_rx_buf *rx_buf = NULL; - union virtchnl2_rx_desc *desc; unsigned int pkt_len = 0; unsigned int hdr_len = 0; u16 gen_id, buf_id = 0; - /* Header buffer overflow only valid for header split */ - bool hbo = false; int bufq_id; u8 rxdid; /* get the Rx desc from Rx queue based on 'next_to_clean' */ - desc = IDPF_RX_DESC(rxq, ntc); - rx_desc = (struct virtchnl2_rx_flex_desc_adv_nic_3 *)desc; + rx_desc = &rxq->rx[ntc].flex_adv_nic_3_wb; /* This memory barrier is needed to keep us from reading * any other fields out of the rx_desc @@ -3150,7 +3236,7 @@ static int idpf_rx_splitq_clean(struct idpf_queue *rxq, int budget) gen_id = le16_get_bits(rx_desc->pktlen_gen_bufq_id, VIRTCHNL2_RX_FLEX_DESC_ADV_GEN_M); - if (test_bit(__IDPF_Q_GEN_CHK, rxq->flags) != gen_id) + if (idpf_queue_has(GEN_CHK, rxq) != gen_id) break; rxdid = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_RXDID_M, @@ -3158,7 +3244,7 @@ static int idpf_rx_splitq_clean(struct idpf_queue *rxq, int budget) if (rxdid != VIRTCHNL2_RXDID_2_FLEX_SPLITQ) { IDPF_RX_BUMP_NTC(rxq, ntc); u64_stats_update_begin(&rxq->stats_sync); - u64_stats_inc(&rxq->q_stats.rx.bad_descs); + u64_stats_inc(&rxq->q_stats.bad_descs); u64_stats_update_end(&rxq->stats_sync); continue; } @@ -3166,71 +3252,79 @@ static int idpf_rx_splitq_clean(struct idpf_queue *rxq, int budget) pkt_len = le16_get_bits(rx_desc->pktlen_gen_bufq_id, VIRTCHNL2_RX_FLEX_DESC_ADV_LEN_PBUF_M); - hbo = FIELD_GET(VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_HBO_M, - rx_desc->status_err0_qw1); - - if (unlikely(hbo)) { - /* If a header buffer overflow, occurs, i.e. header is - * too large to fit in the header split buffer, HW will - * put the entire packet, including headers, in the - * data/payload buffer. - */ - u64_stats_update_begin(&rxq->stats_sync); - u64_stats_inc(&rxq->q_stats.rx.hsplit_buf_ovf); - u64_stats_update_end(&rxq->stats_sync); - goto bypass_hsplit; - } - - hdr_len = le16_get_bits(rx_desc->hdrlen_flags, - VIRTCHNL2_RX_FLEX_DESC_ADV_LEN_HDR_M); - -bypass_hsplit: bufq_id = le16_get_bits(rx_desc->pktlen_gen_bufq_id, VIRTCHNL2_RX_FLEX_DESC_ADV_BUFQ_ID_M); rxq_set = container_of(rxq, struct idpf_rxq_set, rxq); - if (!bufq_id) - refillq = rxq_set->refillq0; - else - refillq = rxq_set->refillq1; + refillq = rxq_set->refillq[bufq_id]; /* retrieve buffer from the rxq */ - rx_bufq = &rxq->rxq_grp->splitq.bufq_sets[bufq_id].bufq; + rx_bufq = &rxq->bufq_sets[bufq_id].bufq; buf_id = le16_to_cpu(rx_desc->buf_id); - rx_buf = &rx_bufq->rx_buf.buf[buf_id]; + rx_buf = &rx_bufq->buf[buf_id]; + + if (!rx_bufq->hdr_pp) + goto payload; + +#define __HBO_BIT VIRTCHNL2_RX_FLEX_DESC_ADV_STATUS0_HBO_M +#define __HDR_LEN_MASK VIRTCHNL2_RX_FLEX_DESC_ADV_LEN_HDR_M + if (likely(!(rx_desc->status_err0_qw1 & __HBO_BIT))) + /* If a header buffer overflow, occurs, i.e. header is + * too large to fit in the header split buffer, HW will + * put the entire packet, including headers, in the + * data/payload buffer. + */ + hdr_len = le16_get_bits(rx_desc->hdrlen_flags, + __HDR_LEN_MASK); +#undef __HDR_LEN_MASK +#undef __HBO_BIT + + hdr = &rx_bufq->hdr_buf[buf_id]; - if (hdr_len) { - const void *va = (u8 *)rx_bufq->rx_buf.hdr_buf_va + - (u32)buf_id * IDPF_HDR_BUF_SIZE; + if (unlikely(!hdr_len && !skb)) { + hdr_len = idpf_rx_hsplit_wa(hdr, rx_buf, pkt_len); + pkt_len -= hdr_len; - skb = idpf_rx_hdr_construct_skb(rxq, va, hdr_len); u64_stats_update_begin(&rxq->stats_sync); - u64_stats_inc(&rxq->q_stats.rx.hsplit_pkts); + u64_stats_inc(&rxq->q_stats.hsplit_buf_ovf); u64_stats_update_end(&rxq->stats_sync); } - if (pkt_len) { - idpf_rx_sync_for_cpu(rx_buf, pkt_len); - if (skb) - idpf_rx_add_frag(rx_buf, skb, pkt_len); - else - skb = idpf_rx_construct_skb(rxq, rx_buf, - pkt_len); - } else { - idpf_rx_put_page(rx_buf); + if (libeth_rx_sync_for_cpu(hdr, hdr_len)) { + skb = idpf_rx_build_skb(hdr, hdr_len); + if (!skb) + break; + + u64_stats_update_begin(&rxq->stats_sync); + u64_stats_inc(&rxq->q_stats.hsplit_pkts); + u64_stats_update_end(&rxq->stats_sync); } + hdr->page = NULL; + +payload: + if (!libeth_rx_sync_for_cpu(rx_buf, pkt_len)) + goto skip_data; + + if (skb) + idpf_rx_add_frag(rx_buf, skb, pkt_len); + else + skb = idpf_rx_build_skb(rx_buf, pkt_len); + /* exit if we failed to retrieve a buffer */ if (!skb) break; - idpf_rx_post_buf_refill(refillq, buf_id); +skip_data: + rx_buf->page = NULL; + idpf_rx_post_buf_refill(refillq, buf_id); IDPF_RX_BUMP_NTC(rxq, ntc); + /* skip if it is non EOP desc */ - if (!idpf_rx_splitq_is_eop(rx_desc)) + if (!idpf_rx_splitq_is_eop(rx_desc) || unlikely(!skb)) continue; /* pad skb if needed (to make valid ethernet frame) */ @@ -3250,7 +3344,7 @@ bypass_hsplit: } /* send completed skb up the stack */ - napi_gro_receive(&rxq->q_vector->napi, skb); + napi_gro_receive(rxq->napi, skb); skb = NULL; /* update budget accounting */ @@ -3261,8 +3355,8 @@ bypass_hsplit: rxq->skb = skb; u64_stats_update_begin(&rxq->stats_sync); - u64_stats_add(&rxq->q_stats.rx.packets, total_rx_pkts); - u64_stats_add(&rxq->q_stats.rx.bytes, total_rx_bytes); + u64_stats_add(&rxq->q_stats.packets, total_rx_pkts); + u64_stats_add(&rxq->q_stats.bytes, total_rx_bytes); u64_stats_update_end(&rxq->stats_sync); /* guarantee a trip back through this routine if there was a failure */ @@ -3272,34 +3366,41 @@ bypass_hsplit: /** * idpf_rx_update_bufq_desc - Update buffer queue descriptor * @bufq: Pointer to the buffer queue - * @refill_desc: SW Refill queue descriptor containing buffer ID + * @buf_id: buffer ID * @buf_desc: Buffer queue descriptor * * Return 0 on success and negative on failure. */ -static int idpf_rx_update_bufq_desc(struct idpf_queue *bufq, u16 refill_desc, +static int idpf_rx_update_bufq_desc(struct idpf_buf_queue *bufq, u32 buf_id, struct virtchnl2_splitq_rx_buf_desc *buf_desc) { - struct idpf_rx_buf *buf; + struct libeth_fq_fp fq = { + .pp = bufq->pp, + .fqes = bufq->buf, + .truesize = bufq->truesize, + .count = bufq->desc_count, + }; dma_addr_t addr; - u16 buf_id; - - buf_id = FIELD_GET(IDPF_RX_BI_BUFID_M, refill_desc); - - buf = &bufq->rx_buf.buf[buf_id]; - addr = idpf_alloc_page(bufq->pp, buf, bufq->rx_buf_size); - if (unlikely(addr == DMA_MAPPING_ERROR)) + addr = libeth_rx_alloc(&fq, buf_id); + if (addr == DMA_MAPPING_ERROR) return -ENOMEM; buf_desc->pkt_addr = cpu_to_le64(addr); buf_desc->qword0.buf_id = cpu_to_le16(buf_id); - if (!bufq->rx_hsplit_en) + if (!idpf_queue_has(HSPLIT_EN, bufq)) return 0; - buf_desc->hdr_addr = cpu_to_le64(bufq->rx_buf.hdr_buf_pa + - (u32)buf_id * IDPF_HDR_BUF_SIZE); + fq.pp = bufq->hdr_pp; + fq.fqes = bufq->hdr_buf; + fq.truesize = bufq->hdr_truesize; + + addr = libeth_rx_alloc(&fq, buf_id); + if (addr == DMA_MAPPING_ERROR) + return -ENOMEM; + + buf_desc->hdr_addr = cpu_to_le64(addr); return 0; } @@ -3311,38 +3412,37 @@ static int idpf_rx_update_bufq_desc(struct idpf_queue *bufq, u16 refill_desc, * * This function takes care of the buffer refill management */ -static void idpf_rx_clean_refillq(struct idpf_queue *bufq, +static void idpf_rx_clean_refillq(struct idpf_buf_queue *bufq, struct idpf_sw_queue *refillq) { struct virtchnl2_splitq_rx_buf_desc *buf_desc; u16 bufq_nta = bufq->next_to_alloc; u16 ntc = refillq->next_to_clean; int cleaned = 0; - u16 gen; - buf_desc = IDPF_SPLITQ_RX_BUF_DESC(bufq, bufq_nta); + buf_desc = &bufq->split_buf[bufq_nta]; /* make sure we stop at ring wrap in the unlikely case ring is full */ while (likely(cleaned < refillq->desc_count)) { - u16 refill_desc = IDPF_SPLITQ_RX_BI_DESC(refillq, ntc); + u32 buf_id, refill_desc = refillq->ring[ntc]; bool failure; - gen = FIELD_GET(IDPF_RX_BI_GEN_M, refill_desc); - if (test_bit(__IDPF_RFLQ_GEN_CHK, refillq->flags) != gen) + if (idpf_queue_has(RFL_GEN_CHK, refillq) != + !!(refill_desc & IDPF_RX_BI_GEN_M)) break; - failure = idpf_rx_update_bufq_desc(bufq, refill_desc, - buf_desc); + buf_id = FIELD_GET(IDPF_RX_BI_BUFID_M, refill_desc); + failure = idpf_rx_update_bufq_desc(bufq, buf_id, buf_desc); if (failure) break; if (unlikely(++ntc == refillq->desc_count)) { - change_bit(__IDPF_RFLQ_GEN_CHK, refillq->flags); + idpf_queue_change(RFL_GEN_CHK, refillq); ntc = 0; } if (unlikely(++bufq_nta == bufq->desc_count)) { - buf_desc = IDPF_SPLITQ_RX_BUF_DESC(bufq, 0); + buf_desc = &bufq->split_buf[0]; bufq_nta = 0; } else { buf_desc++; @@ -3371,16 +3471,21 @@ static void idpf_rx_clean_refillq(struct idpf_queue *bufq, /** * idpf_rx_clean_refillq_all - Clean all refill queues * @bufq: buffer queue with refill queues + * @nid: ID of the closest NUMA node with memory * * Iterates through all refill queues assigned to the buffer queue assigned to * this vector. Returns true if clean is complete within budget, false * otherwise. */ -static void idpf_rx_clean_refillq_all(struct idpf_queue *bufq) +static void idpf_rx_clean_refillq_all(struct idpf_buf_queue *bufq, int nid) { struct idpf_bufq_set *bufq_set; int i; + page_pool_nid_changed(bufq->pp, nid); + if (bufq->hdr_pp) + page_pool_nid_changed(bufq->hdr_pp, nid); + bufq_set = container_of(bufq, struct idpf_bufq_set, bufq); for (i = 0; i < bufq_set->num_refillqs; i++) idpf_rx_clean_refillq(bufq, &bufq_set->refillqs[i]); @@ -3436,39 +3541,21 @@ static void idpf_vport_intr_napi_dis_all(struct idpf_vport *vport) */ void idpf_vport_intr_rel(struct idpf_vport *vport) { - int i, j, v_idx; - - for (v_idx = 0; v_idx < vport->num_q_vectors; v_idx++) { + for (u32 v_idx = 0; v_idx < vport->num_q_vectors; v_idx++) { struct idpf_q_vector *q_vector = &vport->q_vectors[v_idx]; + kfree(q_vector->complq); + q_vector->complq = NULL; kfree(q_vector->bufq); q_vector->bufq = NULL; kfree(q_vector->tx); q_vector->tx = NULL; kfree(q_vector->rx); q_vector->rx = NULL; - } - - /* Clean up the mapping of queues to vectors */ - for (i = 0; i < vport->num_rxq_grp; i++) { - struct idpf_rxq_group *rx_qgrp = &vport->rxq_grps[i]; - if (idpf_is_queue_model_split(vport->rxq_model)) - for (j = 0; j < rx_qgrp->splitq.num_rxq_sets; j++) - rx_qgrp->splitq.rxq_sets[j]->rxq.q_vector = NULL; - else - for (j = 0; j < rx_qgrp->singleq.num_rxq; j++) - rx_qgrp->singleq.rxqs[j]->q_vector = NULL; + free_cpumask_var(q_vector->affinity_mask); } - if (idpf_is_queue_model_split(vport->txq_model)) - for (i = 0; i < vport->num_txq_grp; i++) - vport->txq_grps[i].complq->q_vector = NULL; - else - for (i = 0; i < vport->num_txq_grp; i++) - for (j = 0; j < vport->txq_grps[i].num_txq; j++) - vport->txq_grps[i].txqs[j]->q_vector = NULL; - kfree(vport->q_vectors); vport->q_vectors = NULL; } @@ -3495,7 +3582,7 @@ static void idpf_vport_intr_rel_irq(struct idpf_vport *vport) /* clear the affinity_mask in the IRQ descriptor */ irq_set_affinity_hint(irq_num, NULL); - free_irq(irq_num, q_vector); + kfree(free_irq(irq_num, q_vector)); } } @@ -3515,21 +3602,31 @@ static void idpf_vport_intr_dis_irq_all(struct idpf_vport *vport) /** * idpf_vport_intr_buildreg_itr - Enable default interrupt generation settings * @q_vector: pointer to q_vector - * @type: itr index - * @itr: itr value */ -static u32 idpf_vport_intr_buildreg_itr(struct idpf_q_vector *q_vector, - const int type, u16 itr) +static u32 idpf_vport_intr_buildreg_itr(struct idpf_q_vector *q_vector) { - u32 itr_val; + u32 itr_val = q_vector->intr_reg.dyn_ctl_intena_m; + int type = IDPF_NO_ITR_UPDATE_IDX; + u16 itr = 0; + + if (q_vector->wb_on_itr) { + /* + * Trigger a software interrupt when exiting wb_on_itr, to make + * sure we catch any pending write backs that might have been + * missed due to interrupt state transition. + */ + itr_val |= q_vector->intr_reg.dyn_ctl_swint_trig_m | + q_vector->intr_reg.dyn_ctl_sw_itridx_ena_m; + type = IDPF_SW_ITR_UPDATE_IDX; + itr = IDPF_ITR_20K; + } itr &= IDPF_ITR_MASK; /* Don't clear PBA because that can cause lost interrupts that * came in while we were cleaning/polling */ - itr_val = q_vector->intr_reg.dyn_ctl_intena_m | - (type << q_vector->intr_reg.dyn_ctl_itridx_s) | - (itr << (q_vector->intr_reg.dyn_ctl_intrvl_s - 1)); + itr_val |= (type << q_vector->intr_reg.dyn_ctl_itridx_s) | + (itr << (q_vector->intr_reg.dyn_ctl_intrvl_s - 1)); return itr_val; } @@ -3579,38 +3676,38 @@ static void idpf_net_dim(struct idpf_q_vector *q_vector) goto check_rx_itr; for (i = 0, packets = 0, bytes = 0; i < q_vector->num_txq; i++) { - struct idpf_queue *txq = q_vector->tx[i]; + struct idpf_tx_queue *txq = q_vector->tx[i]; unsigned int start; do { start = u64_stats_fetch_begin(&txq->stats_sync); - packets += u64_stats_read(&txq->q_stats.tx.packets); - bytes += u64_stats_read(&txq->q_stats.tx.bytes); + packets += u64_stats_read(&txq->q_stats.packets); + bytes += u64_stats_read(&txq->q_stats.bytes); } while (u64_stats_fetch_retry(&txq->stats_sync, start)); } idpf_update_dim_sample(q_vector, &dim_sample, &q_vector->tx_dim, packets, bytes); - net_dim(&q_vector->tx_dim, dim_sample); + net_dim(&q_vector->tx_dim, &dim_sample); check_rx_itr: if (!IDPF_ITR_IS_DYNAMIC(q_vector->rx_intr_mode)) return; for (i = 0, packets = 0, bytes = 0; i < q_vector->num_rxq; i++) { - struct idpf_queue *rxq = q_vector->rx[i]; + struct idpf_rx_queue *rxq = q_vector->rx[i]; unsigned int start; do { start = u64_stats_fetch_begin(&rxq->stats_sync); - packets += u64_stats_read(&rxq->q_stats.rx.packets); - bytes += u64_stats_read(&rxq->q_stats.rx.bytes); + packets += u64_stats_read(&rxq->q_stats.packets); + bytes += u64_stats_read(&rxq->q_stats.bytes); } while (u64_stats_fetch_retry(&rxq->stats_sync, start)); } idpf_update_dim_sample(q_vector, &dim_sample, &q_vector->rx_dim, packets, bytes); - net_dim(&q_vector->rx_dim, dim_sample); + net_dim(&q_vector->rx_dim, &dim_sample); } /** @@ -3627,8 +3724,8 @@ void idpf_vport_intr_update_itr_ena_irq(struct idpf_q_vector *q_vector) /* net_dim() updates ITR out-of-band using a work item */ idpf_net_dim(q_vector); - intval = idpf_vport_intr_buildreg_itr(q_vector, - IDPF_NO_ITR_UPDATE_IDX, 0); + intval = idpf_vport_intr_buildreg_itr(q_vector); + q_vector->wb_on_itr = false; writel(intval, q_vector->intr_reg.dyn_ctl); } @@ -3636,16 +3733,19 @@ void idpf_vport_intr_update_itr_ena_irq(struct idpf_q_vector *q_vector) /** * idpf_vport_intr_req_irq - get MSI-X vectors from the OS for the vport * @vport: main vport structure - * @basename: name for the vector */ -static int idpf_vport_intr_req_irq(struct idpf_vport *vport, char *basename) +static int idpf_vport_intr_req_irq(struct idpf_vport *vport) { struct idpf_adapter *adapter = vport->adapter; + const char *drv_name, *if_name, *vec_name; int vector, err, irq_num, vidx; - const char *vec_name; + + drv_name = dev_driver_string(&adapter->pdev->dev); + if_name = netdev_name(vport->netdev); for (vector = 0; vector < vport->num_q_vectors; vector++) { struct idpf_q_vector *q_vector = &vport->q_vectors[vector]; + char *name; vidx = vport->q_vector_idxs[vector]; irq_num = adapter->msix_entries[vidx].vector; @@ -3659,18 +3759,18 @@ static int idpf_vport_intr_req_irq(struct idpf_vport *vport, char *basename) else continue; - q_vector->name = kasprintf(GFP_KERNEL, "%s-%s-%d", - basename, vec_name, vidx); + name = kasprintf(GFP_KERNEL, "%s-%s-%s-%d", drv_name, if_name, + vec_name, vidx); err = request_irq(irq_num, idpf_vport_intr_clean_queues, 0, - q_vector->name, q_vector); + name, q_vector); if (err) { netdev_err(vport->netdev, "Request_irq failed, error: %d\n", err); goto free_q_irqs; } /* assign the mask for this irq */ - irq_set_affinity_hint(irq_num, &q_vector->affinity_mask); + irq_set_affinity_hint(irq_num, q_vector->affinity_mask); } return 0; @@ -3679,7 +3779,7 @@ free_q_irqs: while (--vector >= 0) { vidx = vport->q_vector_idxs[vector]; irq_num = adapter->msix_entries[vidx].vector; - free_irq(irq_num, &vport->q_vectors[vector]); + kfree(free_irq(irq_num, &vport->q_vectors[vector])); } return err; @@ -3746,9 +3846,9 @@ static void idpf_vport_intr_ena_irq_all(struct idpf_vport *vport) */ void idpf_vport_intr_deinit(struct idpf_vport *vport) { + idpf_vport_intr_dis_irq_all(vport); idpf_vport_intr_napi_dis_all(vport); idpf_vport_intr_napi_del_all(vport); - idpf_vport_intr_dis_irq_all(vport); idpf_vport_intr_rel_irq(vport); } @@ -3846,16 +3946,17 @@ static void idpf_vport_intr_napi_ena_all(struct idpf_vport *vport) static bool idpf_tx_splitq_clean_all(struct idpf_q_vector *q_vec, int budget, int *cleaned) { - u16 num_txq = q_vec->num_txq; + u16 num_complq = q_vec->num_complq; bool clean_complete = true; int i, budget_per_q; - if (unlikely(!num_txq)) + if (unlikely(!num_complq)) return true; - budget_per_q = DIV_ROUND_UP(budget, num_txq); - for (i = 0; i < num_txq; i++) - clean_complete &= idpf_tx_clean_complq(q_vec->tx[i], + budget_per_q = DIV_ROUND_UP(budget, num_complq); + + for (i = 0; i < num_complq; i++) + clean_complete &= idpf_tx_clean_complq(q_vec->complq[i], budget_per_q, cleaned); return clean_complete; @@ -3876,13 +3977,14 @@ static bool idpf_rx_splitq_clean_all(struct idpf_q_vector *q_vec, int budget, bool clean_complete = true; int pkts_cleaned = 0; int i, budget_per_q; + int nid; /* 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_q = num_rxq ? max(budget / num_rxq, 1) : 0; for (i = 0; i < num_rxq; i++) { - struct idpf_queue *rxq = q_vec->rx[i]; + struct idpf_rx_queue *rxq = q_vec->rx[i]; int pkts_cleaned_per_q; pkts_cleaned_per_q = idpf_rx_splitq_clean(rxq, budget_per_q); @@ -3893,8 +3995,10 @@ static bool idpf_rx_splitq_clean_all(struct idpf_q_vector *q_vec, int budget, } *cleaned = pkts_cleaned; + nid = numa_mem_id(); + for (i = 0; i < q_vec->num_bufq; i++) - idpf_rx_clean_refillq_all(q_vec->bufq[i]); + idpf_rx_clean_refillq_all(q_vec->bufq[i], nid); return clean_complete; } @@ -3922,8 +4026,10 @@ static int idpf_vport_splitq_napi_poll(struct napi_struct *napi, int budget) clean_complete &= idpf_tx_splitq_clean_all(q_vector, budget, &work_done); /* If work not completed, return budget and polling will return */ - if (!clean_complete) + if (!clean_complete) { + idpf_vport_intr_set_wb_on_itr(q_vector); return budget; + } work_done = min_t(int, work_done, budget - 1); @@ -3932,13 +4038,15 @@ static int idpf_vport_splitq_napi_poll(struct napi_struct *napi, int budget) */ if (likely(napi_complete_done(napi, work_done))) idpf_vport_intr_update_itr_ena_irq(q_vector); + else + idpf_vport_intr_set_wb_on_itr(q_vector); /* Switch to poll mode in the tear-down path after sending disable * queues virtchnl message, as the interrupts will be disabled after * that */ - if (unlikely(q_vector->num_txq && test_bit(__IDPF_Q_POLL_MODE, - q_vector->tx[0]->flags))) + if (unlikely(q_vector->num_txq && idpf_queue_has(POLL_MODE, + q_vector->tx[0]))) return budget; else return work_done; @@ -3952,27 +4060,28 @@ static int idpf_vport_splitq_napi_poll(struct napi_struct *napi, int budget) */ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport) { + bool split = idpf_is_queue_model_split(vport->rxq_model); u16 num_txq_grp = vport->num_txq_grp; - int i, j, qv_idx, bufq_vidx = 0; struct idpf_rxq_group *rx_qgrp; struct idpf_txq_group *tx_qgrp; - struct idpf_queue *q, *bufq; - u16 q_index; + u32 i, qv_idx, q_index; for (i = 0, qv_idx = 0; i < vport->num_rxq_grp; i++) { u16 num_rxq; + if (qv_idx >= vport->num_q_vectors) + qv_idx = 0; + rx_qgrp = &vport->rxq_grps[i]; - if (idpf_is_queue_model_split(vport->rxq_model)) + if (split) num_rxq = rx_qgrp->splitq.num_rxq_sets; else num_rxq = rx_qgrp->singleq.num_rxq; - for (j = 0; j < num_rxq; j++) { - if (qv_idx >= vport->num_q_vectors) - qv_idx = 0; + for (u32 j = 0; j < num_rxq; j++) { + struct idpf_rx_queue *q; - if (idpf_is_queue_model_split(vport->rxq_model)) + if (split) q = &rx_qgrp->splitq.rxq_sets[j]->rxq; else q = rx_qgrp->singleq.rxqs[j]; @@ -3980,52 +4089,53 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport) q_index = q->q_vector->num_rxq; q->q_vector->rx[q_index] = q; q->q_vector->num_rxq++; - qv_idx++; + + if (split) + q->napi = &q->q_vector->napi; } - if (idpf_is_queue_model_split(vport->rxq_model)) { - for (j = 0; j < vport->num_bufqs_per_qgrp; j++) { + if (split) { + for (u32 j = 0; j < vport->num_bufqs_per_qgrp; j++) { + struct idpf_buf_queue *bufq; + bufq = &rx_qgrp->splitq.bufq_sets[j].bufq; - bufq->q_vector = &vport->q_vectors[bufq_vidx]; + bufq->q_vector = &vport->q_vectors[qv_idx]; q_index = bufq->q_vector->num_bufq; bufq->q_vector->bufq[q_index] = bufq; bufq->q_vector->num_bufq++; } - if (++bufq_vidx >= vport->num_q_vectors) - bufq_vidx = 0; } + + qv_idx++; } + split = idpf_is_queue_model_split(vport->txq_model); + for (i = 0, qv_idx = 0; i < num_txq_grp; i++) { u16 num_txq; + if (qv_idx >= vport->num_q_vectors) + qv_idx = 0; + tx_qgrp = &vport->txq_grps[i]; num_txq = tx_qgrp->num_txq; - if (idpf_is_queue_model_split(vport->txq_model)) { - if (qv_idx >= vport->num_q_vectors) - qv_idx = 0; + for (u32 j = 0; j < num_txq; j++) { + struct idpf_tx_queue *q; - q = tx_qgrp->complq; + q = tx_qgrp->txqs[j]; q->q_vector = &vport->q_vectors[qv_idx]; - q_index = q->q_vector->num_txq; - q->q_vector->tx[q_index] = q; - q->q_vector->num_txq++; - qv_idx++; - } else { - for (j = 0; j < num_txq; j++) { - if (qv_idx >= vport->num_q_vectors) - qv_idx = 0; + q->q_vector->tx[q->q_vector->num_txq++] = q; + } - q = tx_qgrp->txqs[j]; - q->q_vector = &vport->q_vectors[qv_idx]; - q_index = q->q_vector->num_txq; - q->q_vector->tx[q_index] = q; - q->q_vector->num_txq++; + if (split) { + struct idpf_compl_queue *q = tx_qgrp->complq; - qv_idx++; - } + q->q_vector = &vport->q_vectors[qv_idx]; + q->q_vector->complq[q->q_vector->num_complq++] = q; } + + qv_idx++; } } @@ -4086,7 +4196,7 @@ static void idpf_vport_intr_napi_add_all(struct idpf_vport *vport) /* only set affinity_mask if the CPU is online */ if (cpu_online(v_idx)) - cpumask_set_cpu(v_idx, &q_vector->affinity_mask); + cpumask_set_cpu(v_idx, q_vector->affinity_mask); } } @@ -4101,18 +4211,22 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport) { u16 txqs_per_vector, rxqs_per_vector, bufqs_per_vector; struct idpf_q_vector *q_vector; - int v_idx, err; + u32 complqs_per_vector, v_idx; vport->q_vectors = kcalloc(vport->num_q_vectors, sizeof(struct idpf_q_vector), GFP_KERNEL); if (!vport->q_vectors) return -ENOMEM; - txqs_per_vector = DIV_ROUND_UP(vport->num_txq, vport->num_q_vectors); - rxqs_per_vector = DIV_ROUND_UP(vport->num_rxq, vport->num_q_vectors); + txqs_per_vector = DIV_ROUND_UP(vport->num_txq_grp, + vport->num_q_vectors); + rxqs_per_vector = DIV_ROUND_UP(vport->num_rxq_grp, + vport->num_q_vectors); bufqs_per_vector = vport->num_bufqs_per_qgrp * DIV_ROUND_UP(vport->num_rxq_grp, vport->num_q_vectors); + complqs_per_vector = DIV_ROUND_UP(vport->num_txq_grp, + vport->num_q_vectors); for (v_idx = 0; v_idx < vport->num_q_vectors; v_idx++) { q_vector = &vport->q_vectors[v_idx]; @@ -4126,32 +4240,33 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport) q_vector->rx_intr_mode = IDPF_ITR_DYNAMIC; q_vector->rx_itr_idx = VIRTCHNL2_ITR_IDX_0; - q_vector->tx = kcalloc(txqs_per_vector, - sizeof(struct idpf_queue *), + if (!zalloc_cpumask_var(&q_vector->affinity_mask, GFP_KERNEL)) + goto error; + + q_vector->tx = kcalloc(txqs_per_vector, sizeof(*q_vector->tx), GFP_KERNEL); - if (!q_vector->tx) { - err = -ENOMEM; + if (!q_vector->tx) goto error; - } - q_vector->rx = kcalloc(rxqs_per_vector, - sizeof(struct idpf_queue *), + q_vector->rx = kcalloc(rxqs_per_vector, sizeof(*q_vector->rx), GFP_KERNEL); - if (!q_vector->rx) { - err = -ENOMEM; + if (!q_vector->rx) goto error; - } if (!idpf_is_queue_model_split(vport->rxq_model)) continue; q_vector->bufq = kcalloc(bufqs_per_vector, - sizeof(struct idpf_queue *), + sizeof(*q_vector->bufq), GFP_KERNEL); - if (!q_vector->bufq) { - err = -ENOMEM; + if (!q_vector->bufq) + goto error; + + q_vector->complq = kcalloc(complqs_per_vector, + sizeof(*q_vector->complq), + GFP_KERNEL); + if (!q_vector->complq) goto error; - } } return 0; @@ -4159,7 +4274,7 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport) error: idpf_vport_intr_rel(vport); - return err; + return -ENOMEM; } /** @@ -4170,7 +4285,6 @@ error: */ int idpf_vport_intr_init(struct idpf_vport *vport) { - char *int_name; int err; err = idpf_vport_intr_init_vec_idx(vport); @@ -4179,31 +4293,29 @@ int idpf_vport_intr_init(struct idpf_vport *vport) idpf_vport_intr_map_vector_to_qs(vport); idpf_vport_intr_napi_add_all(vport); - idpf_vport_intr_napi_ena_all(vport); err = vport->adapter->dev_ops.reg_ops.intr_reg_init(vport); if (err) goto unroll_vectors_alloc; - int_name = kasprintf(GFP_KERNEL, "%s-%s", - dev_driver_string(&vport->adapter->pdev->dev), - vport->netdev->name); - - err = idpf_vport_intr_req_irq(vport, int_name); + err = idpf_vport_intr_req_irq(vport); if (err) goto unroll_vectors_alloc; - idpf_vport_intr_ena_irq_all(vport); - return 0; unroll_vectors_alloc: - idpf_vport_intr_napi_dis_all(vport); idpf_vport_intr_napi_del_all(vport); return err; } +void idpf_vport_intr_ena(struct idpf_vport *vport) +{ + idpf_vport_intr_napi_ena_all(vport); + idpf_vport_intr_ena_irq_all(vport); +} + /** * idpf_config_rss - Send virtchnl messages to configure RSS * @vport: virtual port |