From b32bfa17246d836125958e39996a674653e899a5 Mon Sep 17 00:00:00 2001 From: Jesse Brandeburg Date: Mon, 18 Apr 2016 11:33:42 -0700 Subject: [PATCH] i40e: Drop packet split receive routine As part of preparation for the rx-refactor, remove the packet split receive routine and ancillary code. Signed-off-by: Jesse Brandeburg Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/i40e/i40e.h | 3 - drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 4 +- drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 19 -- drivers/net/ethernet/intel/i40e/i40e_main.c | 49 +---- drivers/net/ethernet/intel/i40e/i40e_txrx.c | 245 +------------------------ drivers/net/ethernet/intel/i40e/i40e_txrx.h | 7 - 6 files changed, 10 insertions(+), 317 deletions(-) diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h index 00c4738..ea6a69a 100644 --- a/drivers/net/ethernet/intel/i40e/i40e.h +++ b/drivers/net/ethernet/intel/i40e/i40e.h @@ -101,7 +101,6 @@ #define I40E_PRIV_FLAGS_LINKPOLL_FLAG BIT(1) #define I40E_PRIV_FLAGS_FD_ATR BIT(2) #define I40E_PRIV_FLAGS_VEB_STATS BIT(3) -#define I40E_PRIV_FLAGS_PS BIT(4) #define I40E_PRIV_FLAGS_HW_ATR_EVICT BIT(5) #define I40E_NVM_VERSION_LO_SHIFT 0 @@ -320,8 +319,6 @@ struct i40e_pf { #define I40E_FLAG_RX_CSUM_ENABLED BIT_ULL(1) #define I40E_FLAG_MSI_ENABLED BIT_ULL(2) #define I40E_FLAG_MSIX_ENABLED BIT_ULL(3) -#define I40E_FLAG_RX_1BUF_ENABLED BIT_ULL(4) -#define I40E_FLAG_RX_PS_ENABLED BIT_ULL(5) #define I40E_FLAG_RSS_ENABLED BIT_ULL(6) #define I40E_FLAG_VMDQ_ENABLED BIT_ULL(7) #define I40E_FLAG_FDIR_REQUIRES_REINIT BIT_ULL(8) diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c index 83dccf1..f119a74 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c +++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c @@ -273,8 +273,8 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid) rx_ring->rx_buf_len, rx_ring->dtype); dev_info(&pf->pdev->dev, - " rx_rings[%i]: hsplit = %d, next_to_use = %d, next_to_clean = %d, ring_active = %i\n", - i, ring_is_ps_enabled(rx_ring), + " rx_rings[%i]: next_to_use = %d, next_to_clean = %d, ring_active = %i\n", + i, rx_ring->next_to_use, rx_ring->next_to_clean, rx_ring->ring_active); diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c index 8e56c43..858e169 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c +++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c @@ -2829,8 +2829,6 @@ static u32 i40e_get_priv_flags(struct net_device *dev) I40E_PRIV_FLAGS_FD_ATR : 0; ret_flags |= pf->flags & I40E_FLAG_VEB_STATS_ENABLED ? I40E_PRIV_FLAGS_VEB_STATS : 0; - ret_flags |= pf->flags & I40E_FLAG_RX_PS_ENABLED ? - I40E_PRIV_FLAGS_PS : 0; ret_flags |= pf->auto_disable_flags & I40E_FLAG_HW_ATR_EVICT_CAPABLE ? 0 : I40E_PRIV_FLAGS_HW_ATR_EVICT; @@ -2851,23 +2849,6 @@ static int i40e_set_priv_flags(struct net_device *dev, u32 flags) /* NOTE: MFP is not settable */ - /* allow the user to control the method of receive - * buffer DMA, whether the packet is split at header - * boundaries into two separate buffers. In some cases - * one routine or the other will perform better. - */ - if ((flags & I40E_PRIV_FLAGS_PS) && - !(pf->flags & I40E_FLAG_RX_PS_ENABLED)) { - pf->flags |= I40E_FLAG_RX_PS_ENABLED; - pf->flags &= ~I40E_FLAG_RX_1BUF_ENABLED; - reset_required = true; - } else if (!(flags & I40E_PRIV_FLAGS_PS) && - (pf->flags & I40E_FLAG_RX_PS_ENABLED)) { - pf->flags &= ~I40E_FLAG_RX_PS_ENABLED; - pf->flags |= I40E_FLAG_RX_1BUF_ENABLED; - reset_required = true; - } - if (flags & I40E_PRIV_FLAGS_LINKPOLL_FLAG) pf->flags |= I40E_FLAG_LINK_POLLING_ENABLED; else diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index f6da6b7..84e8d4e 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -2871,18 +2871,9 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring) } rx_ctx.dtype = vsi->dtype; - if (vsi->dtype) { - set_ring_ps_enabled(ring); - rx_ctx.hsplit_0 = I40E_RX_SPLIT_L2 | - I40E_RX_SPLIT_IP | - I40E_RX_SPLIT_TCP_UDP | - I40E_RX_SPLIT_SCTP; - } else { - rx_ctx.hsplit_0 = 0; - } + rx_ctx.hsplit_0 = 0; - rx_ctx.rxmax = min_t(u16, vsi->max_frame, - (chain_len * ring->rx_buf_len)); + rx_ctx.rxmax = min_t(u16, vsi->max_frame, chain_len * ring->rx_buf_len); if (hw->revision_id == 0) rx_ctx.lrxqthresh = 0; else @@ -2919,12 +2910,7 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring) ring->tail = hw->hw_addr + I40E_QRX_TAIL(pf_q); writel(0, ring->tail); - if (ring_is_ps_enabled(ring)) { - i40e_alloc_rx_headers(ring); - i40e_alloc_rx_buffers_ps(ring, I40E_DESC_UNUSED(ring)); - } else { - i40e_alloc_rx_buffers_1buf(ring, I40E_DESC_UNUSED(ring)); - } + i40e_alloc_rx_buffers_1buf(ring, I40E_DESC_UNUSED(ring)); return 0; } @@ -2963,25 +2949,9 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi) else vsi->max_frame = I40E_RXBUFFER_2048; - /* figure out correct receive buffer length */ - switch (vsi->back->flags & (I40E_FLAG_RX_1BUF_ENABLED | - I40E_FLAG_RX_PS_ENABLED)) { - case I40E_FLAG_RX_1BUF_ENABLED: - vsi->rx_hdr_len = 0; - vsi->rx_buf_len = vsi->max_frame; - vsi->dtype = I40E_RX_DTYPE_NO_SPLIT; - break; - case I40E_FLAG_RX_PS_ENABLED: - vsi->rx_hdr_len = I40E_RX_HDR_SIZE; - vsi->rx_buf_len = I40E_RXBUFFER_2048; - vsi->dtype = I40E_RX_DTYPE_HEADER_SPLIT; - break; - default: - vsi->rx_hdr_len = I40E_RX_HDR_SIZE; - vsi->rx_buf_len = I40E_RXBUFFER_2048; - vsi->dtype = I40E_RX_DTYPE_SPLIT_ALWAYS; - break; - } + vsi->rx_hdr_len = 0; + vsi->rx_buf_len = vsi->max_frame; + vsi->dtype = I40E_RX_DTYPE_NO_SPLIT; #ifdef I40E_FCOE /* setup rx buffer for FCoE */ @@ -8460,11 +8430,6 @@ static int i40e_sw_init(struct i40e_pf *pf) I40E_FLAG_MSI_ENABLED | I40E_FLAG_MSIX_ENABLED; - if (iommu_present(&pci_bus_type)) - pf->flags |= I40E_FLAG_RX_PS_ENABLED; - else - pf->flags |= I40E_FLAG_RX_1BUF_ENABLED; - /* Set default ITR */ pf->rx_itr_default = I40E_ITR_DYNAMIC | I40E_ITR_RX_DEF; pf->tx_itr_default = I40E_ITR_DYNAMIC | I40E_ITR_TX_DEF; @@ -10699,7 +10664,7 @@ static void i40e_print_features(struct i40e_pf *pf) i += snprintf(&buf[i], REMAIN(i), " VSIs: %d QP: %d RX: %s", pf->hw.func_caps.num_vsis, pf->vsi[pf->lan_vsi]->num_queue_pairs, - pf->flags & I40E_FLAG_RX_PS_ENABLED ? "PS" : "1BUF"); + "1BUF"); if (pf->flags & I40E_FLAG_RSS_ENABLED) i += snprintf(&buf[i], REMAIN(i), " RSS"); diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c index dab733c..450ecdd 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c @@ -1032,22 +1032,6 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring) if (!rx_ring->rx_bi) return; - if (ring_is_ps_enabled(rx_ring)) { - int bufsz = ALIGN(rx_ring->rx_hdr_len, 256) * rx_ring->count; - - rx_bi = &rx_ring->rx_bi[0]; - if (rx_bi->hdr_buf) { - dma_free_coherent(dev, - bufsz, - rx_bi->hdr_buf, - rx_bi->dma); - for (i = 0; i < rx_ring->count; i++) { - rx_bi = &rx_ring->rx_bi[i]; - rx_bi->dma = 0; - rx_bi->hdr_buf = NULL; - } - } - } /* Free all the Rx ring sk_buffs */ for (i = 0; i < rx_ring->count; i++) { rx_bi = &rx_ring->rx_bi[i]; @@ -1503,230 +1487,6 @@ static inline void i40e_rx_hash(struct i40e_ring *ring, } /** - * i40e_clean_rx_irq_ps - Reclaim resources after receive; packet split - * @rx_ring: rx ring to clean - * @budget: how many cleans we're allowed - * - * Returns true if there's any budget left (e.g. the clean is finished) - **/ -static int i40e_clean_rx_irq_ps(struct i40e_ring *rx_ring, const int budget) -{ - unsigned int total_rx_bytes = 0, total_rx_packets = 0; - u16 rx_packet_len, rx_header_len, rx_sph, rx_hbo; - u16 cleaned_count = I40E_DESC_UNUSED(rx_ring); - struct i40e_vsi *vsi = rx_ring->vsi; - u16 i = rx_ring->next_to_clean; - union i40e_rx_desc *rx_desc; - u32 rx_error, rx_status; - bool failure = false; - u8 rx_ptype; - u64 qword; - u32 copysize; - - if (budget <= 0) - return 0; - - do { - struct i40e_rx_buffer *rx_bi; - struct sk_buff *skb; - u16 vlan_tag; - /* return some buffers to hardware, one at a time is too slow */ - if (cleaned_count >= I40E_RX_BUFFER_WRITE) { - failure = failure || - i40e_alloc_rx_buffers_ps(rx_ring, - cleaned_count); - cleaned_count = 0; - } - - i = rx_ring->next_to_clean; - rx_desc = I40E_RX_DESC(rx_ring, i); - qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len); - rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >> - I40E_RXD_QW1_STATUS_SHIFT; - - if (!(rx_status & BIT(I40E_RX_DESC_STATUS_DD_SHIFT))) - break; - - /* This memory barrier is needed to keep us from reading - * any other fields out of the rx_desc until we know the - * DD bit is set. - */ - dma_rmb(); - /* sync header buffer for reading */ - dma_sync_single_range_for_cpu(rx_ring->dev, - rx_ring->rx_bi[0].dma, - i * rx_ring->rx_hdr_len, - rx_ring->rx_hdr_len, - DMA_FROM_DEVICE); - if (i40e_rx_is_programming_status(qword)) { - i40e_clean_programming_status(rx_ring, rx_desc); - I40E_RX_INCREMENT(rx_ring, i); - continue; - } - rx_bi = &rx_ring->rx_bi[i]; - skb = rx_bi->skb; - if (likely(!skb)) { - skb = __netdev_alloc_skb_ip_align(rx_ring->netdev, - rx_ring->rx_hdr_len, - GFP_ATOMIC | - __GFP_NOWARN); - if (!skb) { - rx_ring->rx_stats.alloc_buff_failed++; - failure = true; - break; - } - - /* initialize queue mapping */ - skb_record_rx_queue(skb, rx_ring->queue_index); - /* we are reusing so sync this buffer for CPU use */ - dma_sync_single_range_for_cpu(rx_ring->dev, - rx_ring->rx_bi[0].dma, - i * rx_ring->rx_hdr_len, - rx_ring->rx_hdr_len, - DMA_FROM_DEVICE); - } - rx_packet_len = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >> - I40E_RXD_QW1_LENGTH_PBUF_SHIFT; - rx_header_len = (qword & I40E_RXD_QW1_LENGTH_HBUF_MASK) >> - I40E_RXD_QW1_LENGTH_HBUF_SHIFT; - rx_sph = (qword & I40E_RXD_QW1_LENGTH_SPH_MASK) >> - I40E_RXD_QW1_LENGTH_SPH_SHIFT; - - rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >> - I40E_RXD_QW1_ERROR_SHIFT; - rx_hbo = rx_error & BIT(I40E_RX_DESC_ERROR_HBO_SHIFT); - rx_error &= ~BIT(I40E_RX_DESC_ERROR_HBO_SHIFT); - - rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >> - I40E_RXD_QW1_PTYPE_SHIFT; - /* sync half-page for reading */ - dma_sync_single_range_for_cpu(rx_ring->dev, - rx_bi->page_dma, - rx_bi->page_offset, - PAGE_SIZE / 2, - DMA_FROM_DEVICE); - prefetch(page_address(rx_bi->page) + rx_bi->page_offset); - rx_bi->skb = NULL; - cleaned_count++; - copysize = 0; - if (rx_hbo || rx_sph) { - int len; - - if (rx_hbo) - len = I40E_RX_HDR_SIZE; - else - len = rx_header_len; - memcpy(__skb_put(skb, len), rx_bi->hdr_buf, len); - } else if (skb->len == 0) { - int len; - unsigned char *va = page_address(rx_bi->page) + - rx_bi->page_offset; - - len = min(rx_packet_len, rx_ring->rx_hdr_len); - memcpy(__skb_put(skb, len), va, len); - copysize = len; - rx_packet_len -= len; - } - /* Get the rest of the data if this was a header split */ - if (rx_packet_len) { - skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, - rx_bi->page, - rx_bi->page_offset + copysize, - rx_packet_len, I40E_RXBUFFER_2048); - - /* If the page count is more than 2, then both halves - * of the page are used and we need to free it. Do it - * here instead of in the alloc code. Otherwise one - * of the half-pages might be released between now and - * then, and we wouldn't know which one to use. - * Don't call get_page and free_page since those are - * both expensive atomic operations that just change - * the refcount in opposite directions. Just give the - * page to the stack; he can have our refcount. - */ - if (page_count(rx_bi->page) > 2) { - dma_unmap_page(rx_ring->dev, - rx_bi->page_dma, - PAGE_SIZE, - DMA_FROM_DEVICE); - rx_bi->page = NULL; - rx_bi->page_dma = 0; - rx_ring->rx_stats.realloc_count++; - } else { - get_page(rx_bi->page); - /* switch to the other half-page here; the - * allocation code programs the right addr - * into HW. If we haven't used this half-page, - * the address won't be changed, and HW can - * just use it next time through. - */ - rx_bi->page_offset ^= PAGE_SIZE / 2; - } - - } - I40E_RX_INCREMENT(rx_ring, i); - - if (unlikely( - !(rx_status & BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)))) { - struct i40e_rx_buffer *next_buffer; - - next_buffer = &rx_ring->rx_bi[i]; - next_buffer->skb = skb; - rx_ring->rx_stats.non_eop_descs++; - continue; - } - - /* ERR_MASK will only have valid bits if EOP set */ - if (unlikely(rx_error & BIT(I40E_RX_DESC_ERROR_RXE_SHIFT))) { - dev_kfree_skb_any(skb); - continue; - } - - i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype); - - if (unlikely(rx_status & I40E_RXD_QW1_STATUS_TSYNVALID_MASK)) { - i40e_ptp_rx_hwtstamp(vsi->back, skb, (rx_status & - I40E_RXD_QW1_STATUS_TSYNINDX_MASK) >> - I40E_RXD_QW1_STATUS_TSYNINDX_SHIFT); - rx_ring->last_rx_timestamp = jiffies; - } - - /* probably a little skewed due to removing CRC */ - total_rx_bytes += skb->len; - total_rx_packets++; - - skb->protocol = eth_type_trans(skb, rx_ring->netdev); - - i40e_rx_checksum(vsi, skb, rx_status, rx_error, rx_ptype); - - vlan_tag = rx_status & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT) - ? le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1) - : 0; -#ifdef I40E_FCOE - if (unlikely( - i40e_rx_is_fcoe(rx_ptype) && - !i40e_fcoe_handle_offload(rx_ring, rx_desc, skb))) { - dev_kfree_skb_any(skb); - continue; - } -#endif - i40e_receive_skb(rx_ring, skb, vlan_tag); - - rx_desc->wb.qword1.status_error_len = 0; - - } while (likely(total_rx_packets < budget)); - - u64_stats_update_begin(&rx_ring->syncp); - rx_ring->stats.packets += total_rx_packets; - rx_ring->stats.bytes += total_rx_bytes; - u64_stats_update_end(&rx_ring->syncp); - rx_ring->q_vector->rx.total_packets += total_rx_packets; - rx_ring->q_vector->rx.total_bytes += total_rx_bytes; - - return failure ? budget : total_rx_packets; -} - -/** * i40e_clean_rx_irq_1buf - Reclaim resources after receive; single buffer * @rx_ring: rx ring to clean * @budget: how many cleans we're allowed @@ -2001,10 +1761,7 @@ int i40e_napi_poll(struct napi_struct *napi, int budget) i40e_for_each_ring(ring, q_vector->rx) { int cleaned; - if (ring_is_ps_enabled(ring)) - cleaned = i40e_clean_rx_irq_ps(ring, budget_per_ring); - else - cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring); + cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring); work_done += cleaned; /* if we clean as many as budgeted, we must not be done */ diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h index 6b2b191..5a2d0fe 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h @@ -245,16 +245,9 @@ struct i40e_rx_queue_stats { enum i40e_ring_state_t { __I40E_TX_FDIR_INIT_DONE, __I40E_TX_XPS_INIT_DONE, - __I40E_RX_PS_ENABLED, __I40E_RX_16BYTE_DESC_ENABLED, }; -#define ring_is_ps_enabled(ring) \ - test_bit(__I40E_RX_PS_ENABLED, &(ring)->state) -#define set_ring_ps_enabled(ring) \ - set_bit(__I40E_RX_PS_ENABLED, &(ring)->state) -#define clear_ring_ps_enabled(ring) \ - clear_bit(__I40E_RX_PS_ENABLED, &(ring)->state) #define ring_is_16byte_desc_enabled(ring) \ test_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state) #define set_ring_16byte_desc_enabled(ring) \ -- 2.7.4