| /****************************************************************************** |
| * This software may be used and distributed according to the terms of |
| * the GNU General Public License (GPL), incorporated herein by reference. |
| * Drivers based on or derived from this code fall under the GPL and must |
| * retain the authorship, copyright and license notice. This file is not |
| * a complete program and may only be used when the entire operating |
| * system is licensed under the GPL. |
| * See the file COPYING in this distribution for more information. |
| * |
| * vxge-main.c: Driver for Exar Corp's X3100 Series 10GbE PCIe I/O |
| * Virtualized Server Adapter. |
| * Copyright(c) 2002-2010 Exar Corp. |
| * |
| * The module loadable parameters that are supported by the driver and a brief |
| * explanation of all the variables: |
| * vlan_tag_strip: |
| * Strip VLAN Tag enable/disable. Instructs the device to remove |
| * the VLAN tag from all received tagged frames that are not |
| * replicated at the internal L2 switch. |
| * 0 - Do not strip the VLAN tag. |
| * 1 - Strip the VLAN tag. |
| * |
| * addr_learn_en: |
| * Enable learning the mac address of the guest OS interface in |
| * a virtualization environment. |
| * 0 - DISABLE |
| * 1 - ENABLE |
| * |
| * max_config_port: |
| * Maximum number of port to be supported. |
| * MIN -1 and MAX - 2 |
| * |
| * max_config_vpath: |
| * This configures the maximum no of VPATH configures for each |
| * device function. |
| * MIN - 1 and MAX - 17 |
| * |
| * max_config_dev: |
| * This configures maximum no of Device function to be enabled. |
| * MIN - 1 and MAX - 17 |
| * |
| ******************************************************************************/ |
| |
| #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt |
| |
| #include <linux/bitops.h> |
| #include <linux/if_vlan.h> |
| #include <linux/interrupt.h> |
| #include <linux/pci.h> |
| #include <linux/slab.h> |
| #include <linux/tcp.h> |
| #include <net/ip.h> |
| #include <linux/netdevice.h> |
| #include <linux/etherdevice.h> |
| #include <linux/firmware.h> |
| #include <linux/net_tstamp.h> |
| #include <linux/prefetch.h> |
| #include <linux/module.h> |
| #include "vxge-main.h" |
| #include "vxge-reg.h" |
| |
| MODULE_LICENSE("Dual BSD/GPL"); |
| MODULE_DESCRIPTION("Neterion's X3100 Series 10GbE PCIe I/O" |
| "Virtualized Server Adapter"); |
| |
| static DEFINE_PCI_DEVICE_TABLE(vxge_id_table) = { |
| {PCI_VENDOR_ID_S2IO, PCI_DEVICE_ID_TITAN_WIN, PCI_ANY_ID, |
| PCI_ANY_ID}, |
| {PCI_VENDOR_ID_S2IO, PCI_DEVICE_ID_TITAN_UNI, PCI_ANY_ID, |
| PCI_ANY_ID}, |
| {0} |
| }; |
| |
| MODULE_DEVICE_TABLE(pci, vxge_id_table); |
| |
| VXGE_MODULE_PARAM_INT(vlan_tag_strip, VXGE_HW_VPATH_RPA_STRIP_VLAN_TAG_ENABLE); |
| VXGE_MODULE_PARAM_INT(addr_learn_en, VXGE_HW_MAC_ADDR_LEARN_DEFAULT); |
| VXGE_MODULE_PARAM_INT(max_config_port, VXGE_MAX_CONFIG_PORT); |
| VXGE_MODULE_PARAM_INT(max_config_vpath, VXGE_USE_DEFAULT); |
| VXGE_MODULE_PARAM_INT(max_mac_vpath, VXGE_MAX_MAC_ADDR_COUNT); |
| VXGE_MODULE_PARAM_INT(max_config_dev, VXGE_MAX_CONFIG_DEV); |
| |
| static u16 vpath_selector[VXGE_HW_MAX_VIRTUAL_PATHS] = |
| {0, 1, 3, 3, 7, 7, 7, 7, 15, 15, 15, 15, 15, 15, 15, 15, 31}; |
| static unsigned int bw_percentage[VXGE_HW_MAX_VIRTUAL_PATHS] = |
| {[0 ...(VXGE_HW_MAX_VIRTUAL_PATHS - 1)] = 0xFF}; |
| module_param_array(bw_percentage, uint, NULL, 0); |
| |
| static struct vxge_drv_config *driver_config; |
| static enum vxge_hw_status vxge_reset_all_vpaths(struct vxgedev *vdev); |
| |
| static inline int is_vxge_card_up(struct vxgedev *vdev) |
| { |
| return test_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| } |
| |
| static inline void VXGE_COMPLETE_VPATH_TX(struct vxge_fifo *fifo) |
| { |
| struct sk_buff **skb_ptr = NULL; |
| struct sk_buff **temp; |
| #define NR_SKB_COMPLETED 128 |
| struct sk_buff *completed[NR_SKB_COMPLETED]; |
| int more; |
| |
| do { |
| more = 0; |
| skb_ptr = completed; |
| |
| if (__netif_tx_trylock(fifo->txq)) { |
| vxge_hw_vpath_poll_tx(fifo->handle, &skb_ptr, |
| NR_SKB_COMPLETED, &more); |
| __netif_tx_unlock(fifo->txq); |
| } |
| |
| /* free SKBs */ |
| for (temp = completed; temp != skb_ptr; temp++) |
| dev_kfree_skb_irq(*temp); |
| } while (more); |
| } |
| |
| static inline void VXGE_COMPLETE_ALL_TX(struct vxgedev *vdev) |
| { |
| int i; |
| |
| /* Complete all transmits */ |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| VXGE_COMPLETE_VPATH_TX(&vdev->vpaths[i].fifo); |
| } |
| |
| static inline void VXGE_COMPLETE_ALL_RX(struct vxgedev *vdev) |
| { |
| int i; |
| struct vxge_ring *ring; |
| |
| /* Complete all receives*/ |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| ring = &vdev->vpaths[i].ring; |
| vxge_hw_vpath_poll_rx(ring->handle); |
| } |
| } |
| |
| /* |
| * vxge_callback_link_up |
| * |
| * This function is called during interrupt context to notify link up state |
| * change. |
| */ |
| static void vxge_callback_link_up(struct __vxge_hw_device *hldev) |
| { |
| struct net_device *dev = hldev->ndev; |
| struct vxgedev *vdev = netdev_priv(dev); |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| vdev->ndev->name, __func__, __LINE__); |
| netdev_notice(vdev->ndev, "Link Up\n"); |
| vdev->stats.link_up++; |
| |
| netif_carrier_on(vdev->ndev); |
| netif_tx_wake_all_queues(vdev->ndev); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", vdev->ndev->name, __func__, __LINE__); |
| } |
| |
| /* |
| * vxge_callback_link_down |
| * |
| * This function is called during interrupt context to notify link down state |
| * change. |
| */ |
| static void vxge_callback_link_down(struct __vxge_hw_device *hldev) |
| { |
| struct net_device *dev = hldev->ndev; |
| struct vxgedev *vdev = netdev_priv(dev); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d", vdev->ndev->name, __func__, __LINE__); |
| netdev_notice(vdev->ndev, "Link Down\n"); |
| |
| vdev->stats.link_down++; |
| netif_carrier_off(vdev->ndev); |
| netif_tx_stop_all_queues(vdev->ndev); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", vdev->ndev->name, __func__, __LINE__); |
| } |
| |
| /* |
| * vxge_rx_alloc |
| * |
| * Allocate SKB. |
| */ |
| static struct sk_buff * |
| vxge_rx_alloc(void *dtrh, struct vxge_ring *ring, const int skb_size) |
| { |
| struct net_device *dev; |
| struct sk_buff *skb; |
| struct vxge_rx_priv *rx_priv; |
| |
| dev = ring->ndev; |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| ring->ndev->name, __func__, __LINE__); |
| |
| rx_priv = vxge_hw_ring_rxd_private_get(dtrh); |
| |
| /* try to allocate skb first. this one may fail */ |
| skb = netdev_alloc_skb(dev, skb_size + |
| VXGE_HW_HEADER_ETHERNET_II_802_3_ALIGN); |
| if (skb == NULL) { |
| vxge_debug_mem(VXGE_ERR, |
| "%s: out of memory to allocate SKB", dev->name); |
| ring->stats.skb_alloc_fail++; |
| return NULL; |
| } |
| |
| vxge_debug_mem(VXGE_TRACE, |
| "%s: %s:%d Skb : 0x%p", ring->ndev->name, |
| __func__, __LINE__, skb); |
| |
| skb_reserve(skb, VXGE_HW_HEADER_ETHERNET_II_802_3_ALIGN); |
| |
| rx_priv->skb = skb; |
| rx_priv->skb_data = NULL; |
| rx_priv->data_size = skb_size; |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", ring->ndev->name, __func__, __LINE__); |
| |
| return skb; |
| } |
| |
| /* |
| * vxge_rx_map |
| */ |
| static int vxge_rx_map(void *dtrh, struct vxge_ring *ring) |
| { |
| struct vxge_rx_priv *rx_priv; |
| dma_addr_t dma_addr; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| ring->ndev->name, __func__, __LINE__); |
| rx_priv = vxge_hw_ring_rxd_private_get(dtrh); |
| |
| rx_priv->skb_data = rx_priv->skb->data; |
| dma_addr = pci_map_single(ring->pdev, rx_priv->skb_data, |
| rx_priv->data_size, PCI_DMA_FROMDEVICE); |
| |
| if (unlikely(pci_dma_mapping_error(ring->pdev, dma_addr))) { |
| ring->stats.pci_map_fail++; |
| return -EIO; |
| } |
| vxge_debug_mem(VXGE_TRACE, |
| "%s: %s:%d 1 buffer mode dma_addr = 0x%llx", |
| ring->ndev->name, __func__, __LINE__, |
| (unsigned long long)dma_addr); |
| vxge_hw_ring_rxd_1b_set(dtrh, dma_addr, rx_priv->data_size); |
| |
| rx_priv->data_dma = dma_addr; |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", ring->ndev->name, __func__, __LINE__); |
| |
| return 0; |
| } |
| |
| /* |
| * vxge_rx_initial_replenish |
| * Allocation of RxD as an initial replenish procedure. |
| */ |
| static enum vxge_hw_status |
| vxge_rx_initial_replenish(void *dtrh, void *userdata) |
| { |
| struct vxge_ring *ring = (struct vxge_ring *)userdata; |
| struct vxge_rx_priv *rx_priv; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| ring->ndev->name, __func__, __LINE__); |
| if (vxge_rx_alloc(dtrh, ring, |
| VXGE_LL_MAX_FRAME_SIZE(ring->ndev)) == NULL) |
| return VXGE_HW_FAIL; |
| |
| if (vxge_rx_map(dtrh, ring)) { |
| rx_priv = vxge_hw_ring_rxd_private_get(dtrh); |
| dev_kfree_skb(rx_priv->skb); |
| |
| return VXGE_HW_FAIL; |
| } |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", ring->ndev->name, __func__, __LINE__); |
| |
| return VXGE_HW_OK; |
| } |
| |
| static inline void |
| vxge_rx_complete(struct vxge_ring *ring, struct sk_buff *skb, u16 vlan, |
| int pkt_length, struct vxge_hw_ring_rxd_info *ext_info) |
| { |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| ring->ndev->name, __func__, __LINE__); |
| skb_record_rx_queue(skb, ring->driver_id); |
| skb->protocol = eth_type_trans(skb, ring->ndev); |
| |
| u64_stats_update_begin(&ring->stats.syncp); |
| ring->stats.rx_frms++; |
| ring->stats.rx_bytes += pkt_length; |
| |
| if (skb->pkt_type == PACKET_MULTICAST) |
| ring->stats.rx_mcast++; |
| u64_stats_update_end(&ring->stats.syncp); |
| |
| vxge_debug_rx(VXGE_TRACE, |
| "%s: %s:%d skb protocol = %d", |
| ring->ndev->name, __func__, __LINE__, skb->protocol); |
| |
| if (ext_info->vlan && |
| ring->vlan_tag_strip == VXGE_HW_VPATH_RPA_STRIP_VLAN_TAG_ENABLE) |
| __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), ext_info->vlan); |
| napi_gro_receive(ring->napi_p, skb); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", ring->ndev->name, __func__, __LINE__); |
| } |
| |
| static inline void vxge_re_pre_post(void *dtr, struct vxge_ring *ring, |
| struct vxge_rx_priv *rx_priv) |
| { |
| pci_dma_sync_single_for_device(ring->pdev, |
| rx_priv->data_dma, rx_priv->data_size, PCI_DMA_FROMDEVICE); |
| |
| vxge_hw_ring_rxd_1b_set(dtr, rx_priv->data_dma, rx_priv->data_size); |
| vxge_hw_ring_rxd_pre_post(ring->handle, dtr); |
| } |
| |
| static inline void vxge_post(int *dtr_cnt, void **first_dtr, |
| void *post_dtr, struct __vxge_hw_ring *ringh) |
| { |
| int dtr_count = *dtr_cnt; |
| if ((*dtr_cnt % VXGE_HW_RXSYNC_FREQ_CNT) == 0) { |
| if (*first_dtr) |
| vxge_hw_ring_rxd_post_post_wmb(ringh, *first_dtr); |
| *first_dtr = post_dtr; |
| } else |
| vxge_hw_ring_rxd_post_post(ringh, post_dtr); |
| dtr_count++; |
| *dtr_cnt = dtr_count; |
| } |
| |
| /* |
| * vxge_rx_1b_compl |
| * |
| * If the interrupt is because of a received frame or if the receive ring |
| * contains fresh as yet un-processed frames, this function is called. |
| */ |
| static enum vxge_hw_status |
| vxge_rx_1b_compl(struct __vxge_hw_ring *ringh, void *dtr, |
| u8 t_code, void *userdata) |
| { |
| struct vxge_ring *ring = (struct vxge_ring *)userdata; |
| struct net_device *dev = ring->ndev; |
| unsigned int dma_sizes; |
| void *first_dtr = NULL; |
| int dtr_cnt = 0; |
| int data_size; |
| dma_addr_t data_dma; |
| int pkt_length; |
| struct sk_buff *skb; |
| struct vxge_rx_priv *rx_priv; |
| struct vxge_hw_ring_rxd_info ext_info; |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| ring->ndev->name, __func__, __LINE__); |
| |
| do { |
| prefetch((char *)dtr + L1_CACHE_BYTES); |
| rx_priv = vxge_hw_ring_rxd_private_get(dtr); |
| skb = rx_priv->skb; |
| data_size = rx_priv->data_size; |
| data_dma = rx_priv->data_dma; |
| prefetch(rx_priv->skb_data); |
| |
| vxge_debug_rx(VXGE_TRACE, |
| "%s: %s:%d skb = 0x%p", |
| ring->ndev->name, __func__, __LINE__, skb); |
| |
| vxge_hw_ring_rxd_1b_get(ringh, dtr, &dma_sizes); |
| pkt_length = dma_sizes; |
| |
| pkt_length -= ETH_FCS_LEN; |
| |
| vxge_debug_rx(VXGE_TRACE, |
| "%s: %s:%d Packet Length = %d", |
| ring->ndev->name, __func__, __LINE__, pkt_length); |
| |
| vxge_hw_ring_rxd_1b_info_get(ringh, dtr, &ext_info); |
| |
| /* check skb validity */ |
| vxge_assert(skb); |
| |
| prefetch((char *)skb + L1_CACHE_BYTES); |
| if (unlikely(t_code)) { |
| if (vxge_hw_ring_handle_tcode(ringh, dtr, t_code) != |
| VXGE_HW_OK) { |
| |
| ring->stats.rx_errors++; |
| vxge_debug_rx(VXGE_TRACE, |
| "%s: %s :%d Rx T_code is %d", |
| ring->ndev->name, __func__, |
| __LINE__, t_code); |
| |
| /* If the t_code is not supported and if the |
| * t_code is other than 0x5 (unparseable packet |
| * such as unknown UPV6 header), Drop it !!! |
| */ |
| vxge_re_pre_post(dtr, ring, rx_priv); |
| |
| vxge_post(&dtr_cnt, &first_dtr, dtr, ringh); |
| ring->stats.rx_dropped++; |
| continue; |
| } |
| } |
| |
| if (pkt_length > VXGE_LL_RX_COPY_THRESHOLD) { |
| if (vxge_rx_alloc(dtr, ring, data_size) != NULL) { |
| if (!vxge_rx_map(dtr, ring)) { |
| skb_put(skb, pkt_length); |
| |
| pci_unmap_single(ring->pdev, data_dma, |
| data_size, PCI_DMA_FROMDEVICE); |
| |
| vxge_hw_ring_rxd_pre_post(ringh, dtr); |
| vxge_post(&dtr_cnt, &first_dtr, dtr, |
| ringh); |
| } else { |
| dev_kfree_skb(rx_priv->skb); |
| rx_priv->skb = skb; |
| rx_priv->data_size = data_size; |
| vxge_re_pre_post(dtr, ring, rx_priv); |
| |
| vxge_post(&dtr_cnt, &first_dtr, dtr, |
| ringh); |
| ring->stats.rx_dropped++; |
| break; |
| } |
| } else { |
| vxge_re_pre_post(dtr, ring, rx_priv); |
| |
| vxge_post(&dtr_cnt, &first_dtr, dtr, ringh); |
| ring->stats.rx_dropped++; |
| break; |
| } |
| } else { |
| struct sk_buff *skb_up; |
| |
| skb_up = netdev_alloc_skb(dev, pkt_length + |
| VXGE_HW_HEADER_ETHERNET_II_802_3_ALIGN); |
| if (skb_up != NULL) { |
| skb_reserve(skb_up, |
| VXGE_HW_HEADER_ETHERNET_II_802_3_ALIGN); |
| |
| pci_dma_sync_single_for_cpu(ring->pdev, |
| data_dma, data_size, |
| PCI_DMA_FROMDEVICE); |
| |
| vxge_debug_mem(VXGE_TRACE, |
| "%s: %s:%d skb_up = %p", |
| ring->ndev->name, __func__, |
| __LINE__, skb); |
| memcpy(skb_up->data, skb->data, pkt_length); |
| |
| vxge_re_pre_post(dtr, ring, rx_priv); |
| |
| vxge_post(&dtr_cnt, &first_dtr, dtr, |
| ringh); |
| /* will netif_rx small SKB instead */ |
| skb = skb_up; |
| skb_put(skb, pkt_length); |
| } else { |
| vxge_re_pre_post(dtr, ring, rx_priv); |
| |
| vxge_post(&dtr_cnt, &first_dtr, dtr, ringh); |
| vxge_debug_rx(VXGE_ERR, |
| "%s: vxge_rx_1b_compl: out of " |
| "memory", dev->name); |
| ring->stats.skb_alloc_fail++; |
| break; |
| } |
| } |
| |
| if ((ext_info.proto & VXGE_HW_FRAME_PROTO_TCP_OR_UDP) && |
| !(ext_info.proto & VXGE_HW_FRAME_PROTO_IP_FRAG) && |
| (dev->features & NETIF_F_RXCSUM) && /* Offload Rx side CSUM */ |
| ext_info.l3_cksum == VXGE_HW_L3_CKSUM_OK && |
| ext_info.l4_cksum == VXGE_HW_L4_CKSUM_OK) |
| skb->ip_summed = CHECKSUM_UNNECESSARY; |
| else |
| skb_checksum_none_assert(skb); |
| |
| |
| if (ring->rx_hwts) { |
| struct skb_shared_hwtstamps *skb_hwts; |
| u32 ns = *(u32 *)(skb->head + pkt_length); |
| |
| skb_hwts = skb_hwtstamps(skb); |
| skb_hwts->hwtstamp = ns_to_ktime(ns); |
| skb_hwts->syststamp.tv64 = 0; |
| } |
| |
| /* rth_hash_type and rth_it_hit are non-zero regardless of |
| * whether rss is enabled. Only the rth_value is zero/non-zero |
| * if rss is disabled/enabled, so key off of that. |
| */ |
| if (ext_info.rth_value) |
| skb_set_hash(skb, ext_info.rth_value, |
| PKT_HASH_TYPE_L3); |
| |
| vxge_rx_complete(ring, skb, ext_info.vlan, |
| pkt_length, &ext_info); |
| |
| ring->budget--; |
| ring->pkts_processed++; |
| if (!ring->budget) |
| break; |
| |
| } while (vxge_hw_ring_rxd_next_completed(ringh, &dtr, |
| &t_code) == VXGE_HW_OK); |
| |
| if (first_dtr) |
| vxge_hw_ring_rxd_post_post_wmb(ringh, first_dtr); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", |
| __func__, __LINE__); |
| return VXGE_HW_OK; |
| } |
| |
| /* |
| * vxge_xmit_compl |
| * |
| * If an interrupt was raised to indicate DMA complete of the Tx packet, |
| * this function is called. It identifies the last TxD whose buffer was |
| * freed and frees all skbs whose data have already DMA'ed into the NICs |
| * internal memory. |
| */ |
| static enum vxge_hw_status |
| vxge_xmit_compl(struct __vxge_hw_fifo *fifo_hw, void *dtr, |
| enum vxge_hw_fifo_tcode t_code, void *userdata, |
| struct sk_buff ***skb_ptr, int nr_skb, int *more) |
| { |
| struct vxge_fifo *fifo = (struct vxge_fifo *)userdata; |
| struct sk_buff *skb, **done_skb = *skb_ptr; |
| int pkt_cnt = 0; |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Entered....", __func__, __LINE__); |
| |
| do { |
| int frg_cnt; |
| skb_frag_t *frag; |
| int i = 0, j; |
| struct vxge_tx_priv *txd_priv = |
| vxge_hw_fifo_txdl_private_get(dtr); |
| |
| skb = txd_priv->skb; |
| frg_cnt = skb_shinfo(skb)->nr_frags; |
| frag = &skb_shinfo(skb)->frags[0]; |
| |
| vxge_debug_tx(VXGE_TRACE, |
| "%s: %s:%d fifo_hw = %p dtr = %p " |
| "tcode = 0x%x", fifo->ndev->name, __func__, |
| __LINE__, fifo_hw, dtr, t_code); |
| /* check skb validity */ |
| vxge_assert(skb); |
| vxge_debug_tx(VXGE_TRACE, |
| "%s: %s:%d skb = %p itxd_priv = %p frg_cnt = %d", |
| fifo->ndev->name, __func__, __LINE__, |
| skb, txd_priv, frg_cnt); |
| if (unlikely(t_code)) { |
| fifo->stats.tx_errors++; |
| vxge_debug_tx(VXGE_ERR, |
| "%s: tx: dtr %p completed due to " |
| "error t_code %01x", fifo->ndev->name, |
| dtr, t_code); |
| vxge_hw_fifo_handle_tcode(fifo_hw, dtr, t_code); |
| } |
| |
| /* for unfragmented skb */ |
| pci_unmap_single(fifo->pdev, txd_priv->dma_buffers[i++], |
| skb_headlen(skb), PCI_DMA_TODEVICE); |
| |
| for (j = 0; j < frg_cnt; j++) { |
| pci_unmap_page(fifo->pdev, |
| txd_priv->dma_buffers[i++], |
| skb_frag_size(frag), PCI_DMA_TODEVICE); |
| frag += 1; |
| } |
| |
| vxge_hw_fifo_txdl_free(fifo_hw, dtr); |
| |
| /* Updating the statistics block */ |
| u64_stats_update_begin(&fifo->stats.syncp); |
| fifo->stats.tx_frms++; |
| fifo->stats.tx_bytes += skb->len; |
| u64_stats_update_end(&fifo->stats.syncp); |
| |
| *done_skb++ = skb; |
| |
| if (--nr_skb <= 0) { |
| *more = 1; |
| break; |
| } |
| |
| pkt_cnt++; |
| if (pkt_cnt > fifo->indicate_max_pkts) |
| break; |
| |
| } while (vxge_hw_fifo_txdl_next_completed(fifo_hw, |
| &dtr, &t_code) == VXGE_HW_OK); |
| |
| *skb_ptr = done_skb; |
| if (netif_tx_queue_stopped(fifo->txq)) |
| netif_tx_wake_queue(fifo->txq); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", |
| fifo->ndev->name, __func__, __LINE__); |
| return VXGE_HW_OK; |
| } |
| |
| /* select a vpath to transmit the packet */ |
| static u32 vxge_get_vpath_no(struct vxgedev *vdev, struct sk_buff *skb) |
| { |
| u16 queue_len, counter = 0; |
| if (skb->protocol == htons(ETH_P_IP)) { |
| struct iphdr *ip; |
| struct tcphdr *th; |
| |
| ip = ip_hdr(skb); |
| |
| if (!ip_is_fragment(ip)) { |
| th = (struct tcphdr *)(((unsigned char *)ip) + |
| ip->ihl*4); |
| |
| queue_len = vdev->no_of_vpath; |
| counter = (ntohs(th->source) + |
| ntohs(th->dest)) & |
| vdev->vpath_selector[queue_len - 1]; |
| if (counter >= queue_len) |
| counter = queue_len - 1; |
| } |
| } |
| return counter; |
| } |
| |
| static enum vxge_hw_status vxge_search_mac_addr_in_list( |
| struct vxge_vpath *vpath, u64 del_mac) |
| { |
| struct list_head *entry, *next; |
| list_for_each_safe(entry, next, &vpath->mac_addr_list) { |
| if (((struct vxge_mac_addrs *)entry)->macaddr == del_mac) |
| return TRUE; |
| } |
| return FALSE; |
| } |
| |
| static int vxge_mac_list_add(struct vxge_vpath *vpath, struct macInfo *mac) |
| { |
| struct vxge_mac_addrs *new_mac_entry; |
| u8 *mac_address = NULL; |
| |
| if (vpath->mac_addr_cnt >= VXGE_MAX_LEARN_MAC_ADDR_CNT) |
| return TRUE; |
| |
| new_mac_entry = kzalloc(sizeof(struct vxge_mac_addrs), GFP_ATOMIC); |
| if (!new_mac_entry) { |
| vxge_debug_mem(VXGE_ERR, |
| "%s: memory allocation failed", |
| VXGE_DRIVER_NAME); |
| return FALSE; |
| } |
| |
| list_add(&new_mac_entry->item, &vpath->mac_addr_list); |
| |
| /* Copy the new mac address to the list */ |
| mac_address = (u8 *)&new_mac_entry->macaddr; |
| memcpy(mac_address, mac->macaddr, ETH_ALEN); |
| |
| new_mac_entry->state = mac->state; |
| vpath->mac_addr_cnt++; |
| |
| if (is_multicast_ether_addr(mac->macaddr)) |
| vpath->mcast_addr_cnt++; |
| |
| return TRUE; |
| } |
| |
| /* Add a mac address to DA table */ |
| static enum vxge_hw_status |
| vxge_add_mac_addr(struct vxgedev *vdev, struct macInfo *mac) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_vpath *vpath; |
| enum vxge_hw_vpath_mac_addr_add_mode duplicate_mode; |
| |
| if (is_multicast_ether_addr(mac->macaddr)) |
| duplicate_mode = VXGE_HW_VPATH_MAC_ADDR_ADD_DUPLICATE; |
| else |
| duplicate_mode = VXGE_HW_VPATH_MAC_ADDR_REPLACE_DUPLICATE; |
| |
| vpath = &vdev->vpaths[mac->vpath_no]; |
| status = vxge_hw_vpath_mac_addr_add(vpath->handle, mac->macaddr, |
| mac->macmask, duplicate_mode); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "DA config add entry failed for vpath:%d", |
| vpath->device_id); |
| } else |
| if (FALSE == vxge_mac_list_add(vpath, mac)) |
| status = -EPERM; |
| |
| return status; |
| } |
| |
| static int vxge_learn_mac(struct vxgedev *vdev, u8 *mac_header) |
| { |
| struct macInfo mac_info; |
| u8 *mac_address = NULL; |
| u64 mac_addr = 0, vpath_vector = 0; |
| int vpath_idx = 0; |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_vpath *vpath = NULL; |
| struct __vxge_hw_device *hldev; |
| |
| hldev = pci_get_drvdata(vdev->pdev); |
| |
| mac_address = (u8 *)&mac_addr; |
| memcpy(mac_address, mac_header, ETH_ALEN); |
| |
| /* Is this mac address already in the list? */ |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; vpath_idx++) { |
| vpath = &vdev->vpaths[vpath_idx]; |
| if (vxge_search_mac_addr_in_list(vpath, mac_addr)) |
| return vpath_idx; |
| } |
| |
| memset(&mac_info, 0, sizeof(struct macInfo)); |
| memcpy(mac_info.macaddr, mac_header, ETH_ALEN); |
| |
| /* Any vpath has room to add mac address to its da table? */ |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; vpath_idx++) { |
| vpath = &vdev->vpaths[vpath_idx]; |
| if (vpath->mac_addr_cnt < vpath->max_mac_addr_cnt) { |
| /* Add this mac address to this vpath */ |
| mac_info.vpath_no = vpath_idx; |
| mac_info.state = VXGE_LL_MAC_ADDR_IN_DA_TABLE; |
| status = vxge_add_mac_addr(vdev, &mac_info); |
| if (status != VXGE_HW_OK) |
| return -EPERM; |
| return vpath_idx; |
| } |
| } |
| |
| mac_info.state = VXGE_LL_MAC_ADDR_IN_LIST; |
| vpath_idx = 0; |
| mac_info.vpath_no = vpath_idx; |
| /* Is the first vpath already selected as catch-basin ? */ |
| vpath = &vdev->vpaths[vpath_idx]; |
| if (vpath->mac_addr_cnt > vpath->max_mac_addr_cnt) { |
| /* Add this mac address to this vpath */ |
| if (FALSE == vxge_mac_list_add(vpath, &mac_info)) |
| return -EPERM; |
| return vpath_idx; |
| } |
| |
| /* Select first vpath as catch-basin */ |
| vpath_vector = vxge_mBIT(vpath->device_id); |
| status = vxge_hw_mgmt_reg_write(vpath->vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof( |
| struct vxge_hw_mrpcim_reg, |
| rts_mgr_cbasin_cfg), |
| vpath_vector); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_tx(VXGE_ERR, |
| "%s: Unable to set the vpath-%d in catch-basin mode", |
| VXGE_DRIVER_NAME, vpath->device_id); |
| return -EPERM; |
| } |
| |
| if (FALSE == vxge_mac_list_add(vpath, &mac_info)) |
| return -EPERM; |
| |
| return vpath_idx; |
| } |
| |
| /** |
| * vxge_xmit |
| * @skb : the socket buffer containing the Tx data. |
| * @dev : device pointer. |
| * |
| * This function is the Tx entry point of the driver. Neterion NIC supports |
| * certain protocol assist features on Tx side, namely CSO, S/G, LSO. |
| */ |
| static netdev_tx_t |
| vxge_xmit(struct sk_buff *skb, struct net_device *dev) |
| { |
| struct vxge_fifo *fifo = NULL; |
| void *dtr_priv; |
| void *dtr = NULL; |
| struct vxgedev *vdev = NULL; |
| enum vxge_hw_status status; |
| int frg_cnt, first_frg_len; |
| skb_frag_t *frag; |
| int i = 0, j = 0, avail; |
| u64 dma_pointer; |
| struct vxge_tx_priv *txdl_priv = NULL; |
| struct __vxge_hw_fifo *fifo_hw; |
| int offload_type; |
| int vpath_no = 0; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| dev->name, __func__, __LINE__); |
| |
| /* A buffer with no data will be dropped */ |
| if (unlikely(skb->len <= 0)) { |
| vxge_debug_tx(VXGE_ERR, |
| "%s: Buffer has no data..", dev->name); |
| dev_kfree_skb(skb); |
| return NETDEV_TX_OK; |
| } |
| |
| vdev = netdev_priv(dev); |
| |
| if (unlikely(!is_vxge_card_up(vdev))) { |
| vxge_debug_tx(VXGE_ERR, |
| "%s: vdev not initialized", dev->name); |
| dev_kfree_skb(skb); |
| return NETDEV_TX_OK; |
| } |
| |
| if (vdev->config.addr_learn_en) { |
| vpath_no = vxge_learn_mac(vdev, skb->data + ETH_ALEN); |
| if (vpath_no == -EPERM) { |
| vxge_debug_tx(VXGE_ERR, |
| "%s: Failed to store the mac address", |
| dev->name); |
| dev_kfree_skb(skb); |
| return NETDEV_TX_OK; |
| } |
| } |
| |
| if (vdev->config.tx_steering_type == TX_MULTIQ_STEERING) |
| vpath_no = skb_get_queue_mapping(skb); |
| else if (vdev->config.tx_steering_type == TX_PORT_STEERING) |
| vpath_no = vxge_get_vpath_no(vdev, skb); |
| |
| vxge_debug_tx(VXGE_TRACE, "%s: vpath_no= %d", dev->name, vpath_no); |
| |
| if (vpath_no >= vdev->no_of_vpath) |
| vpath_no = 0; |
| |
| fifo = &vdev->vpaths[vpath_no].fifo; |
| fifo_hw = fifo->handle; |
| |
| if (netif_tx_queue_stopped(fifo->txq)) |
| return NETDEV_TX_BUSY; |
| |
| avail = vxge_hw_fifo_free_txdl_count_get(fifo_hw); |
| if (avail == 0) { |
| vxge_debug_tx(VXGE_ERR, |
| "%s: No free TXDs available", dev->name); |
| fifo->stats.txd_not_free++; |
| goto _exit0; |
| } |
| |
| /* Last TXD? Stop tx queue to avoid dropping packets. TX |
| * completion will resume the queue. |
| */ |
| if (avail == 1) |
| netif_tx_stop_queue(fifo->txq); |
| |
| status = vxge_hw_fifo_txdl_reserve(fifo_hw, &dtr, &dtr_priv); |
| if (unlikely(status != VXGE_HW_OK)) { |
| vxge_debug_tx(VXGE_ERR, |
| "%s: Out of descriptors .", dev->name); |
| fifo->stats.txd_out_of_desc++; |
| goto _exit0; |
| } |
| |
| vxge_debug_tx(VXGE_TRACE, |
| "%s: %s:%d fifo_hw = %p dtr = %p dtr_priv = %p", |
| dev->name, __func__, __LINE__, |
| fifo_hw, dtr, dtr_priv); |
| |
| if (vlan_tx_tag_present(skb)) { |
| u16 vlan_tag = vlan_tx_tag_get(skb); |
| vxge_hw_fifo_txdl_vlan_set(dtr, vlan_tag); |
| } |
| |
| first_frg_len = skb_headlen(skb); |
| |
| dma_pointer = pci_map_single(fifo->pdev, skb->data, first_frg_len, |
| PCI_DMA_TODEVICE); |
| |
| if (unlikely(pci_dma_mapping_error(fifo->pdev, dma_pointer))) { |
| vxge_hw_fifo_txdl_free(fifo_hw, dtr); |
| fifo->stats.pci_map_fail++; |
| goto _exit0; |
| } |
| |
| txdl_priv = vxge_hw_fifo_txdl_private_get(dtr); |
| txdl_priv->skb = skb; |
| txdl_priv->dma_buffers[j] = dma_pointer; |
| |
| frg_cnt = skb_shinfo(skb)->nr_frags; |
| vxge_debug_tx(VXGE_TRACE, |
| "%s: %s:%d skb = %p txdl_priv = %p " |
| "frag_cnt = %d dma_pointer = 0x%llx", dev->name, |
| __func__, __LINE__, skb, txdl_priv, |
| frg_cnt, (unsigned long long)dma_pointer); |
| |
| vxge_hw_fifo_txdl_buffer_set(fifo_hw, dtr, j++, dma_pointer, |
| first_frg_len); |
| |
| frag = &skb_shinfo(skb)->frags[0]; |
| for (i = 0; i < frg_cnt; i++) { |
| /* ignore 0 length fragment */ |
| if (!skb_frag_size(frag)) |
| continue; |
| |
| dma_pointer = (u64)skb_frag_dma_map(&fifo->pdev->dev, frag, |
| 0, skb_frag_size(frag), |
| DMA_TO_DEVICE); |
| |
| if (unlikely(dma_mapping_error(&fifo->pdev->dev, dma_pointer))) |
| goto _exit2; |
| vxge_debug_tx(VXGE_TRACE, |
| "%s: %s:%d frag = %d dma_pointer = 0x%llx", |
| dev->name, __func__, __LINE__, i, |
| (unsigned long long)dma_pointer); |
| |
| txdl_priv->dma_buffers[j] = dma_pointer; |
| vxge_hw_fifo_txdl_buffer_set(fifo_hw, dtr, j++, dma_pointer, |
| skb_frag_size(frag)); |
| frag += 1; |
| } |
| |
| offload_type = vxge_offload_type(skb); |
| |
| if (offload_type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6)) { |
| int mss = vxge_tcp_mss(skb); |
| if (mss) { |
| vxge_debug_tx(VXGE_TRACE, "%s: %s:%d mss = %d", |
| dev->name, __func__, __LINE__, mss); |
| vxge_hw_fifo_txdl_mss_set(dtr, mss); |
| } else { |
| vxge_assert(skb->len <= |
| dev->mtu + VXGE_HW_MAC_HEADER_MAX_SIZE); |
| vxge_assert(0); |
| goto _exit1; |
| } |
| } |
| |
| if (skb->ip_summed == CHECKSUM_PARTIAL) |
| vxge_hw_fifo_txdl_cksum_set_bits(dtr, |
| VXGE_HW_FIFO_TXD_TX_CKO_IPV4_EN | |
| VXGE_HW_FIFO_TXD_TX_CKO_TCP_EN | |
| VXGE_HW_FIFO_TXD_TX_CKO_UDP_EN); |
| |
| vxge_hw_fifo_txdl_post(fifo_hw, dtr); |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d Exiting...", |
| dev->name, __func__, __LINE__); |
| return NETDEV_TX_OK; |
| |
| _exit2: |
| vxge_debug_tx(VXGE_TRACE, "%s: pci_map_page failed", dev->name); |
| _exit1: |
| j = 0; |
| frag = &skb_shinfo(skb)->frags[0]; |
| |
| pci_unmap_single(fifo->pdev, txdl_priv->dma_buffers[j++], |
| skb_headlen(skb), PCI_DMA_TODEVICE); |
| |
| for (; j < i; j++) { |
| pci_unmap_page(fifo->pdev, txdl_priv->dma_buffers[j], |
| skb_frag_size(frag), PCI_DMA_TODEVICE); |
| frag += 1; |
| } |
| |
| vxge_hw_fifo_txdl_free(fifo_hw, dtr); |
| _exit0: |
| netif_tx_stop_queue(fifo->txq); |
| dev_kfree_skb(skb); |
| |
| return NETDEV_TX_OK; |
| } |
| |
| /* |
| * vxge_rx_term |
| * |
| * Function will be called by hw function to abort all outstanding receive |
| * descriptors. |
| */ |
| static void |
| vxge_rx_term(void *dtrh, enum vxge_hw_rxd_state state, void *userdata) |
| { |
| struct vxge_ring *ring = (struct vxge_ring *)userdata; |
| struct vxge_rx_priv *rx_priv = |
| vxge_hw_ring_rxd_private_get(dtrh); |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| ring->ndev->name, __func__, __LINE__); |
| if (state != VXGE_HW_RXD_STATE_POSTED) |
| return; |
| |
| pci_unmap_single(ring->pdev, rx_priv->data_dma, |
| rx_priv->data_size, PCI_DMA_FROMDEVICE); |
| |
| dev_kfree_skb(rx_priv->skb); |
| rx_priv->skb_data = NULL; |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", |
| ring->ndev->name, __func__, __LINE__); |
| } |
| |
| /* |
| * vxge_tx_term |
| * |
| * Function will be called to abort all outstanding tx descriptors |
| */ |
| static void |
| vxge_tx_term(void *dtrh, enum vxge_hw_txdl_state state, void *userdata) |
| { |
| struct vxge_fifo *fifo = (struct vxge_fifo *)userdata; |
| skb_frag_t *frag; |
| int i = 0, j, frg_cnt; |
| struct vxge_tx_priv *txd_priv = vxge_hw_fifo_txdl_private_get(dtrh); |
| struct sk_buff *skb = txd_priv->skb; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| if (state != VXGE_HW_TXDL_STATE_POSTED) |
| return; |
| |
| /* check skb validity */ |
| vxge_assert(skb); |
| frg_cnt = skb_shinfo(skb)->nr_frags; |
| frag = &skb_shinfo(skb)->frags[0]; |
| |
| /* for unfragmented skb */ |
| pci_unmap_single(fifo->pdev, txd_priv->dma_buffers[i++], |
| skb_headlen(skb), PCI_DMA_TODEVICE); |
| |
| for (j = 0; j < frg_cnt; j++) { |
| pci_unmap_page(fifo->pdev, txd_priv->dma_buffers[i++], |
| skb_frag_size(frag), PCI_DMA_TODEVICE); |
| frag += 1; |
| } |
| |
| dev_kfree_skb(skb); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| } |
| |
| static int vxge_mac_list_del(struct vxge_vpath *vpath, struct macInfo *mac) |
| { |
| struct list_head *entry, *next; |
| u64 del_mac = 0; |
| u8 *mac_address = (u8 *) (&del_mac); |
| |
| /* Copy the mac address to delete from the list */ |
| memcpy(mac_address, mac->macaddr, ETH_ALEN); |
| |
| list_for_each_safe(entry, next, &vpath->mac_addr_list) { |
| if (((struct vxge_mac_addrs *)entry)->macaddr == del_mac) { |
| list_del(entry); |
| kfree((struct vxge_mac_addrs *)entry); |
| vpath->mac_addr_cnt--; |
| |
| if (is_multicast_ether_addr(mac->macaddr)) |
| vpath->mcast_addr_cnt--; |
| return TRUE; |
| } |
| } |
| |
| return FALSE; |
| } |
| |
| /* delete a mac address from DA table */ |
| static enum vxge_hw_status |
| vxge_del_mac_addr(struct vxgedev *vdev, struct macInfo *mac) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_vpath *vpath; |
| |
| vpath = &vdev->vpaths[mac->vpath_no]; |
| status = vxge_hw_vpath_mac_addr_delete(vpath->handle, mac->macaddr, |
| mac->macmask); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "DA config delete entry failed for vpath:%d", |
| vpath->device_id); |
| } else |
| vxge_mac_list_del(vpath, mac); |
| return status; |
| } |
| |
| /** |
| * vxge_set_multicast |
| * @dev: pointer to the device structure |
| * |
| * Entry point for multicast address enable/disable |
| * This function is a driver entry point which gets called by the kernel |
| * whenever multicast addresses must be enabled/disabled. This also gets |
| * called to set/reset promiscuous mode. Depending on the deivce flag, we |
| * determine, if multicast address must be enabled or if promiscuous mode |
| * is to be disabled etc. |
| */ |
| static void vxge_set_multicast(struct net_device *dev) |
| { |
| struct netdev_hw_addr *ha; |
| struct vxgedev *vdev; |
| int i, mcast_cnt = 0; |
| struct __vxge_hw_device *hldev; |
| struct vxge_vpath *vpath; |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct macInfo mac_info; |
| int vpath_idx = 0; |
| struct vxge_mac_addrs *mac_entry; |
| struct list_head *list_head; |
| struct list_head *entry, *next; |
| u8 *mac_address = NULL; |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d", __func__, __LINE__); |
| |
| vdev = netdev_priv(dev); |
| hldev = vdev->devh; |
| |
| if (unlikely(!is_vxge_card_up(vdev))) |
| return; |
| |
| if ((dev->flags & IFF_ALLMULTI) && (!vdev->all_multi_flg)) { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vxge_assert(vpath->is_open); |
| status = vxge_hw_vpath_mcast_enable(vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, "failed to enable " |
| "multicast, status %d", status); |
| vdev->all_multi_flg = 1; |
| } |
| } else if (!(dev->flags & IFF_ALLMULTI) && (vdev->all_multi_flg)) { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vxge_assert(vpath->is_open); |
| status = vxge_hw_vpath_mcast_disable(vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, "failed to disable " |
| "multicast, status %d", status); |
| vdev->all_multi_flg = 0; |
| } |
| } |
| |
| |
| if (!vdev->config.addr_learn_en) { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vxge_assert(vpath->is_open); |
| |
| if (dev->flags & IFF_PROMISC) |
| status = vxge_hw_vpath_promisc_enable( |
| vpath->handle); |
| else |
| status = vxge_hw_vpath_promisc_disable( |
| vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, "failed to %s promisc" |
| ", status %d", dev->flags&IFF_PROMISC ? |
| "enable" : "disable", status); |
| } |
| } |
| |
| memset(&mac_info, 0, sizeof(struct macInfo)); |
| /* Update individual M_CAST address list */ |
| if ((!vdev->all_multi_flg) && netdev_mc_count(dev)) { |
| mcast_cnt = vdev->vpaths[0].mcast_addr_cnt; |
| list_head = &vdev->vpaths[0].mac_addr_list; |
| if ((netdev_mc_count(dev) + |
| (vdev->vpaths[0].mac_addr_cnt - mcast_cnt)) > |
| vdev->vpaths[0].max_mac_addr_cnt) |
| goto _set_all_mcast; |
| |
| /* Delete previous MC's */ |
| for (i = 0; i < mcast_cnt; i++) { |
| list_for_each_safe(entry, next, list_head) { |
| mac_entry = (struct vxge_mac_addrs *)entry; |
| /* Copy the mac address to delete */ |
| mac_address = (u8 *)&mac_entry->macaddr; |
| memcpy(mac_info.macaddr, mac_address, ETH_ALEN); |
| |
| if (is_multicast_ether_addr(mac_info.macaddr)) { |
| for (vpath_idx = 0; vpath_idx < |
| vdev->no_of_vpath; |
| vpath_idx++) { |
| mac_info.vpath_no = vpath_idx; |
| status = vxge_del_mac_addr( |
| vdev, |
| &mac_info); |
| } |
| } |
| } |
| } |
| |
| /* Add new ones */ |
| netdev_for_each_mc_addr(ha, dev) { |
| memcpy(mac_info.macaddr, ha->addr, ETH_ALEN); |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; |
| vpath_idx++) { |
| mac_info.vpath_no = vpath_idx; |
| mac_info.state = VXGE_LL_MAC_ADDR_IN_DA_TABLE; |
| status = vxge_add_mac_addr(vdev, &mac_info); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s:%d Setting individual" |
| "multicast address failed", |
| __func__, __LINE__); |
| goto _set_all_mcast; |
| } |
| } |
| } |
| |
| return; |
| _set_all_mcast: |
| mcast_cnt = vdev->vpaths[0].mcast_addr_cnt; |
| /* Delete previous MC's */ |
| for (i = 0; i < mcast_cnt; i++) { |
| list_for_each_safe(entry, next, list_head) { |
| mac_entry = (struct vxge_mac_addrs *)entry; |
| /* Copy the mac address to delete */ |
| mac_address = (u8 *)&mac_entry->macaddr; |
| memcpy(mac_info.macaddr, mac_address, ETH_ALEN); |
| |
| if (is_multicast_ether_addr(mac_info.macaddr)) |
| break; |
| } |
| |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; |
| vpath_idx++) { |
| mac_info.vpath_no = vpath_idx; |
| status = vxge_del_mac_addr(vdev, &mac_info); |
| } |
| } |
| |
| /* Enable all multicast */ |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vxge_assert(vpath->is_open); |
| |
| status = vxge_hw_vpath_mcast_enable(vpath->handle); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s:%d Enabling all multicasts failed", |
| __func__, __LINE__); |
| } |
| vdev->all_multi_flg = 1; |
| } |
| dev->flags |= IFF_ALLMULTI; |
| } |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| } |
| |
| /** |
| * vxge_set_mac_addr |
| * @dev: pointer to the device structure |
| * |
| * Update entry "0" (default MAC addr) |
| */ |
| static int vxge_set_mac_addr(struct net_device *dev, void *p) |
| { |
| struct sockaddr *addr = p; |
| struct vxgedev *vdev; |
| struct __vxge_hw_device *hldev; |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct macInfo mac_info_new, mac_info_old; |
| int vpath_idx = 0; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| vdev = netdev_priv(dev); |
| hldev = vdev->devh; |
| |
| if (!is_valid_ether_addr(addr->sa_data)) |
| return -EINVAL; |
| |
| memset(&mac_info_new, 0, sizeof(struct macInfo)); |
| memset(&mac_info_old, 0, sizeof(struct macInfo)); |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d Exiting...", |
| __func__, __LINE__); |
| |
| /* Get the old address */ |
| memcpy(mac_info_old.macaddr, dev->dev_addr, dev->addr_len); |
| |
| /* Copy the new address */ |
| memcpy(mac_info_new.macaddr, addr->sa_data, dev->addr_len); |
| |
| /* First delete the old mac address from all the vpaths |
| as we can't specify the index while adding new mac address */ |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; vpath_idx++) { |
| struct vxge_vpath *vpath = &vdev->vpaths[vpath_idx]; |
| if (!vpath->is_open) { |
| /* This can happen when this interface is added/removed |
| to the bonding interface. Delete this station address |
| from the linked list */ |
| vxge_mac_list_del(vpath, &mac_info_old); |
| |
| /* Add this new address to the linked list |
| for later restoring */ |
| vxge_mac_list_add(vpath, &mac_info_new); |
| |
| continue; |
| } |
| /* Delete the station address */ |
| mac_info_old.vpath_no = vpath_idx; |
| status = vxge_del_mac_addr(vdev, &mac_info_old); |
| } |
| |
| if (unlikely(!is_vxge_card_up(vdev))) { |
| memcpy(dev->dev_addr, addr->sa_data, dev->addr_len); |
| return VXGE_HW_OK; |
| } |
| |
| /* Set this mac address to all the vpaths */ |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; vpath_idx++) { |
| mac_info_new.vpath_no = vpath_idx; |
| mac_info_new.state = VXGE_LL_MAC_ADDR_IN_DA_TABLE; |
| status = vxge_add_mac_addr(vdev, &mac_info_new); |
| if (status != VXGE_HW_OK) |
| return -EINVAL; |
| } |
| |
| memcpy(dev->dev_addr, addr->sa_data, dev->addr_len); |
| |
| return status; |
| } |
| |
| /* |
| * vxge_vpath_intr_enable |
| * @vdev: pointer to vdev |
| * @vp_id: vpath for which to enable the interrupts |
| * |
| * Enables the interrupts for the vpath |
| */ |
| static void vxge_vpath_intr_enable(struct vxgedev *vdev, int vp_id) |
| { |
| struct vxge_vpath *vpath = &vdev->vpaths[vp_id]; |
| int msix_id = 0; |
| int tim_msix_id[4] = {0, 1, 0, 0}; |
| int alarm_msix_id = VXGE_ALARM_MSIX_ID; |
| |
| vxge_hw_vpath_intr_enable(vpath->handle); |
| |
| if (vdev->config.intr_type == INTA) |
| vxge_hw_vpath_inta_unmask_tx_rx(vpath->handle); |
| else { |
| vxge_hw_vpath_msix_set(vpath->handle, tim_msix_id, |
| alarm_msix_id); |
| |
| msix_id = vpath->device_id * VXGE_HW_VPATH_MSIX_ACTIVE; |
| vxge_hw_vpath_msix_unmask(vpath->handle, msix_id); |
| vxge_hw_vpath_msix_unmask(vpath->handle, msix_id + 1); |
| |
| /* enable the alarm vector */ |
| msix_id = (vpath->handle->vpath->hldev->first_vp_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE) + alarm_msix_id; |
| vxge_hw_vpath_msix_unmask(vpath->handle, msix_id); |
| } |
| } |
| |
| /* |
| * vxge_vpath_intr_disable |
| * @vdev: pointer to vdev |
| * @vp_id: vpath for which to disable the interrupts |
| * |
| * Disables the interrupts for the vpath |
| */ |
| static void vxge_vpath_intr_disable(struct vxgedev *vdev, int vp_id) |
| { |
| struct vxge_vpath *vpath = &vdev->vpaths[vp_id]; |
| struct __vxge_hw_device *hldev; |
| int msix_id; |
| |
| hldev = pci_get_drvdata(vdev->pdev); |
| |
| vxge_hw_vpath_wait_receive_idle(hldev, vpath->device_id); |
| |
| vxge_hw_vpath_intr_disable(vpath->handle); |
| |
| if (vdev->config.intr_type == INTA) |
| vxge_hw_vpath_inta_mask_tx_rx(vpath->handle); |
| else { |
| msix_id = vpath->device_id * VXGE_HW_VPATH_MSIX_ACTIVE; |
| vxge_hw_vpath_msix_mask(vpath->handle, msix_id); |
| vxge_hw_vpath_msix_mask(vpath->handle, msix_id + 1); |
| |
| /* disable the alarm vector */ |
| msix_id = (vpath->handle->vpath->hldev->first_vp_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE) + VXGE_ALARM_MSIX_ID; |
| vxge_hw_vpath_msix_mask(vpath->handle, msix_id); |
| } |
| } |
| |
| /* list all mac addresses from DA table */ |
| static enum vxge_hw_status |
| vxge_search_mac_addr_in_da_table(struct vxge_vpath *vpath, struct macInfo *mac) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| unsigned char macmask[ETH_ALEN]; |
| unsigned char macaddr[ETH_ALEN]; |
| |
| status = vxge_hw_vpath_mac_addr_get(vpath->handle, |
| macaddr, macmask); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "DA config list entry failed for vpath:%d", |
| vpath->device_id); |
| return status; |
| } |
| |
| while (!ether_addr_equal(mac->macaddr, macaddr)) { |
| status = vxge_hw_vpath_mac_addr_get_next(vpath->handle, |
| macaddr, macmask); |
| if (status != VXGE_HW_OK) |
| break; |
| } |
| |
| return status; |
| } |
| |
| /* Store all mac addresses from the list to the DA table */ |
| static enum vxge_hw_status vxge_restore_vpath_mac_addr(struct vxge_vpath *vpath) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct macInfo mac_info; |
| u8 *mac_address = NULL; |
| struct list_head *entry, *next; |
| |
| memset(&mac_info, 0, sizeof(struct macInfo)); |
| |
| if (vpath->is_open) { |
| list_for_each_safe(entry, next, &vpath->mac_addr_list) { |
| mac_address = |
| (u8 *)& |
| ((struct vxge_mac_addrs *)entry)->macaddr; |
| memcpy(mac_info.macaddr, mac_address, ETH_ALEN); |
| ((struct vxge_mac_addrs *)entry)->state = |
| VXGE_LL_MAC_ADDR_IN_DA_TABLE; |
| /* does this mac address already exist in da table? */ |
| status = vxge_search_mac_addr_in_da_table(vpath, |
| &mac_info); |
| if (status != VXGE_HW_OK) { |
| /* Add this mac address to the DA table */ |
| status = vxge_hw_vpath_mac_addr_add( |
| vpath->handle, mac_info.macaddr, |
| mac_info.macmask, |
| VXGE_HW_VPATH_MAC_ADDR_ADD_DUPLICATE); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "DA add entry failed for vpath:%d", |
| vpath->device_id); |
| ((struct vxge_mac_addrs *)entry)->state |
| = VXGE_LL_MAC_ADDR_IN_LIST; |
| } |
| } |
| } |
| } |
| |
| return status; |
| } |
| |
| /* Store all vlan ids from the list to the vid table */ |
| static enum vxge_hw_status |
| vxge_restore_vpath_vid_table(struct vxge_vpath *vpath) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxgedev *vdev = vpath->vdev; |
| u16 vid; |
| |
| if (!vpath->is_open) |
| return status; |
| |
| for_each_set_bit(vid, vdev->active_vlans, VLAN_N_VID) |
| status = vxge_hw_vpath_vid_add(vpath->handle, vid); |
| |
| return status; |
| } |
| |
| /* |
| * vxge_reset_vpath |
| * @vdev: pointer to vdev |
| * @vp_id: vpath to reset |
| * |
| * Resets the vpath |
| */ |
| static int vxge_reset_vpath(struct vxgedev *vdev, int vp_id) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_vpath *vpath = &vdev->vpaths[vp_id]; |
| int ret = 0; |
| |
| /* check if device is down already */ |
| if (unlikely(!is_vxge_card_up(vdev))) |
| return 0; |
| |
| /* is device reset already scheduled */ |
| if (test_bit(__VXGE_STATE_RESET_CARD, &vdev->state)) |
| return 0; |
| |
| if (vpath->handle) { |
| if (vxge_hw_vpath_reset(vpath->handle) == VXGE_HW_OK) { |
| if (is_vxge_card_up(vdev) && |
| vxge_hw_vpath_recover_from_reset(vpath->handle) |
| != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "vxge_hw_vpath_recover_from_reset" |
| "failed for vpath:%d", vp_id); |
| return status; |
| } |
| } else { |
| vxge_debug_init(VXGE_ERR, |
| "vxge_hw_vpath_reset failed for" |
| "vpath:%d", vp_id); |
| return status; |
| } |
| } else |
| return VXGE_HW_FAIL; |
| |
| vxge_restore_vpath_mac_addr(vpath); |
| vxge_restore_vpath_vid_table(vpath); |
| |
| /* Enable all broadcast */ |
| vxge_hw_vpath_bcast_enable(vpath->handle); |
| |
| /* Enable all multicast */ |
| if (vdev->all_multi_flg) { |
| status = vxge_hw_vpath_mcast_enable(vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, |
| "%s:%d Enabling multicast failed", |
| __func__, __LINE__); |
| } |
| |
| /* Enable the interrupts */ |
| vxge_vpath_intr_enable(vdev, vp_id); |
| |
| smp_wmb(); |
| |
| /* Enable the flow of traffic through the vpath */ |
| vxge_hw_vpath_enable(vpath->handle); |
| |
| smp_wmb(); |
| vxge_hw_vpath_rx_doorbell_init(vpath->handle); |
| vpath->ring.last_status = VXGE_HW_OK; |
| |
| /* Vpath reset done */ |
| clear_bit(vp_id, &vdev->vp_reset); |
| |
| /* Start the vpath queue */ |
| if (netif_tx_queue_stopped(vpath->fifo.txq)) |
| netif_tx_wake_queue(vpath->fifo.txq); |
| |
| return ret; |
| } |
| |
| /* Configure CI */ |
| static void vxge_config_ci_for_tti_rti(struct vxgedev *vdev) |
| { |
| int i = 0; |
| |
| /* Enable CI for RTI */ |
| if (vdev->config.intr_type == MSI_X) { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| struct __vxge_hw_ring *hw_ring; |
| |
| hw_ring = vdev->vpaths[i].ring.handle; |
| vxge_hw_vpath_dynamic_rti_ci_set(hw_ring); |
| } |
| } |
| |
| /* Enable CI for TTI */ |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| struct __vxge_hw_fifo *hw_fifo = vdev->vpaths[i].fifo.handle; |
| vxge_hw_vpath_tti_ci_set(hw_fifo); |
| /* |
| * For Inta (with or without napi), Set CI ON for only one |
| * vpath. (Have only one free running timer). |
| */ |
| if ((vdev->config.intr_type == INTA) && (i == 0)) |
| break; |
| } |
| |
| return; |
| } |
| |
| static int do_vxge_reset(struct vxgedev *vdev, int event) |
| { |
| enum vxge_hw_status status; |
| int ret = 0, vp_id, i; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| if ((event == VXGE_LL_FULL_RESET) || (event == VXGE_LL_START_RESET)) { |
| /* check if device is down already */ |
| if (unlikely(!is_vxge_card_up(vdev))) |
| return 0; |
| |
| /* is reset already scheduled */ |
| if (test_and_set_bit(__VXGE_STATE_RESET_CARD, &vdev->state)) |
| return 0; |
| } |
| |
| if (event == VXGE_LL_FULL_RESET) { |
| netif_carrier_off(vdev->ndev); |
| |
| /* wait for all the vpath reset to complete */ |
| for (vp_id = 0; vp_id < vdev->no_of_vpath; vp_id++) { |
| while (test_bit(vp_id, &vdev->vp_reset)) |
| msleep(50); |
| } |
| |
| netif_carrier_on(vdev->ndev); |
| |
| /* if execution mode is set to debug, don't reset the adapter */ |
| if (unlikely(vdev->exec_mode)) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: execution mode is debug, returning..", |
| vdev->ndev->name); |
| clear_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| netif_tx_stop_all_queues(vdev->ndev); |
| return 0; |
| } |
| } |
| |
| if (event == VXGE_LL_FULL_RESET) { |
| vxge_hw_device_wait_receive_idle(vdev->devh); |
| vxge_hw_device_intr_disable(vdev->devh); |
| |
| switch (vdev->cric_err_event) { |
| case VXGE_HW_EVENT_UNKNOWN: |
| netif_tx_stop_all_queues(vdev->ndev); |
| vxge_debug_init(VXGE_ERR, |
| "fatal: %s: Disabling device due to" |
| "unknown error", |
| vdev->ndev->name); |
| ret = -EPERM; |
| goto out; |
| case VXGE_HW_EVENT_RESET_START: |
| break; |
| case VXGE_HW_EVENT_RESET_COMPLETE: |
| case VXGE_HW_EVENT_LINK_DOWN: |
| case VXGE_HW_EVENT_LINK_UP: |
| case VXGE_HW_EVENT_ALARM_CLEARED: |
| case VXGE_HW_EVENT_ECCERR: |
| case VXGE_HW_EVENT_MRPCIM_ECCERR: |
| ret = -EPERM; |
| goto out; |
| case VXGE_HW_EVENT_FIFO_ERR: |
| case VXGE_HW_EVENT_VPATH_ERR: |
| break; |
| case VXGE_HW_EVENT_CRITICAL_ERR: |
| netif_tx_stop_all_queues(vdev->ndev); |
| vxge_debug_init(VXGE_ERR, |
| "fatal: %s: Disabling device due to" |
| "serious error", |
| vdev->ndev->name); |
| /* SOP or device reset required */ |
| /* This event is not currently used */ |
| ret = -EPERM; |
| goto out; |
| case VXGE_HW_EVENT_SERR: |
| netif_tx_stop_all_queues(vdev->ndev); |
| vxge_debug_init(VXGE_ERR, |
| "fatal: %s: Disabling device due to" |
| "serious error", |
| vdev->ndev->name); |
| ret = -EPERM; |
| goto out; |
| case VXGE_HW_EVENT_SRPCIM_SERR: |
| case VXGE_HW_EVENT_MRPCIM_SERR: |
| ret = -EPERM; |
| goto out; |
| case VXGE_HW_EVENT_SLOT_FREEZE: |
| netif_tx_stop_all_queues(vdev->ndev); |
| vxge_debug_init(VXGE_ERR, |
| "fatal: %s: Disabling device due to" |
| "slot freeze", |
| vdev->ndev->name); |
| ret = -EPERM; |
| goto out; |
| default: |
| break; |
| |
| } |
| } |
| |
| if ((event == VXGE_LL_FULL_RESET) || (event == VXGE_LL_START_RESET)) |
| netif_tx_stop_all_queues(vdev->ndev); |
| |
| if (event == VXGE_LL_FULL_RESET) { |
| status = vxge_reset_all_vpaths(vdev); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "fatal: %s: can not reset vpaths", |
| vdev->ndev->name); |
| ret = -EPERM; |
| goto out; |
| } |
| } |
| |
| if (event == VXGE_LL_COMPL_RESET) { |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| if (vdev->vpaths[i].handle) { |
| if (vxge_hw_vpath_recover_from_reset( |
| vdev->vpaths[i].handle) |
| != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "vxge_hw_vpath_recover_" |
| "from_reset failed for vpath: " |
| "%d", i); |
| ret = -EPERM; |
| goto out; |
| } |
| } else { |
| vxge_debug_init(VXGE_ERR, |
| "vxge_hw_vpath_reset failed for " |
| "vpath:%d", i); |
| ret = -EPERM; |
| goto out; |
| } |
| } |
| |
| if ((event == VXGE_LL_FULL_RESET) || (event == VXGE_LL_COMPL_RESET)) { |
| /* Reprogram the DA table with populated mac addresses */ |
| for (vp_id = 0; vp_id < vdev->no_of_vpath; vp_id++) { |
| vxge_restore_vpath_mac_addr(&vdev->vpaths[vp_id]); |
| vxge_restore_vpath_vid_table(&vdev->vpaths[vp_id]); |
| } |
| |
| /* enable vpath interrupts */ |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| vxge_vpath_intr_enable(vdev, i); |
| |
| vxge_hw_device_intr_enable(vdev->devh); |
| |
| smp_wmb(); |
| |
| /* Indicate card up */ |
| set_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| |
| /* Get the traffic to flow through the vpaths */ |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vxge_hw_vpath_enable(vdev->vpaths[i].handle); |
| smp_wmb(); |
| vxge_hw_vpath_rx_doorbell_init(vdev->vpaths[i].handle); |
| } |
| |
| netif_tx_wake_all_queues(vdev->ndev); |
| } |
| |
| /* configure CI */ |
| vxge_config_ci_for_tti_rti(vdev); |
| |
| out: |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| |
| /* Indicate reset done */ |
| if ((event == VXGE_LL_FULL_RESET) || (event == VXGE_LL_COMPL_RESET)) |
| clear_bit(__VXGE_STATE_RESET_CARD, &vdev->state); |
| return ret; |
| } |
| |
| /* |
| * vxge_reset |
| * @vdev: pointer to ll device |
| * |
| * driver may reset the chip on events of serr, eccerr, etc |
| */ |
| static void vxge_reset(struct work_struct *work) |
| { |
| struct vxgedev *vdev = container_of(work, struct vxgedev, reset_task); |
| |
| if (!netif_running(vdev->ndev)) |
| return; |
| |
| do_vxge_reset(vdev, VXGE_LL_FULL_RESET); |
| } |
| |
| /** |
| * vxge_poll - Receive handler when Receive Polling is used. |
| * @dev: pointer to the device structure. |
| * @budget: Number of packets budgeted to be processed in this iteration. |
| * |
| * This function comes into picture only if Receive side is being handled |
| * through polling (called NAPI in linux). It mostly does what the normal |
| * Rx interrupt handler does in terms of descriptor and packet processing |
| * but not in an interrupt context. Also it will process a specified number |
| * of packets at most in one iteration. This value is passed down by the |
| * kernel as the function argument 'budget'. |
| */ |
| static int vxge_poll_msix(struct napi_struct *napi, int budget) |
| { |
| struct vxge_ring *ring = container_of(napi, struct vxge_ring, napi); |
| int pkts_processed; |
| int budget_org = budget; |
| |
| ring->budget = budget; |
| ring->pkts_processed = 0; |
| vxge_hw_vpath_poll_rx(ring->handle); |
| pkts_processed = ring->pkts_processed; |
| |
| if (ring->pkts_processed < budget_org) { |
| napi_complete(napi); |
| |
| /* Re enable the Rx interrupts for the vpath */ |
| vxge_hw_channel_msix_unmask( |
| (struct __vxge_hw_channel *)ring->handle, |
| ring->rx_vector_no); |
| mmiowb(); |
| } |
| |
| /* We are copying and returning the local variable, in case if after |
| * clearing the msix interrupt above, if the interrupt fires right |
| * away which can preempt this NAPI thread */ |
| return pkts_processed; |
| } |
| |
| static int vxge_poll_inta(struct napi_struct *napi, int budget) |
| { |
| struct vxgedev *vdev = container_of(napi, struct vxgedev, napi); |
| int pkts_processed = 0; |
| int i; |
| int budget_org = budget; |
| struct vxge_ring *ring; |
| |
| struct __vxge_hw_device *hldev = pci_get_drvdata(vdev->pdev); |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| ring = &vdev->vpaths[i].ring; |
| ring->budget = budget; |
| ring->pkts_processed = 0; |
| vxge_hw_vpath_poll_rx(ring->handle); |
| pkts_processed += ring->pkts_processed; |
| budget -= ring->pkts_processed; |
| if (budget <= 0) |
| break; |
| } |
| |
| VXGE_COMPLETE_ALL_TX(vdev); |
| |
| if (pkts_processed < budget_org) { |
| napi_complete(napi); |
| /* Re enable the Rx interrupts for the ring */ |
| vxge_hw_device_unmask_all(hldev); |
| vxge_hw_device_flush_io(hldev); |
| } |
| |
| return pkts_processed; |
| } |
| |
| #ifdef CONFIG_NET_POLL_CONTROLLER |
| /** |
| * vxge_netpoll - netpoll event handler entry point |
| * @dev : pointer to the device structure. |
| * Description: |
| * This function will be called by upper layer to check for events on the |
| * interface in situations where interrupts are disabled. It is used for |
| * specific in-kernel networking tasks, such as remote consoles and kernel |
| * debugging over the network (example netdump in RedHat). |
| */ |
| static void vxge_netpoll(struct net_device *dev) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| struct pci_dev *pdev = vdev->pdev; |
| struct __vxge_hw_device *hldev = pci_get_drvdata(pdev); |
| const int irq = pdev->irq; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| if (pci_channel_offline(pdev)) |
| return; |
| |
| disable_irq(irq); |
| vxge_hw_device_clear_tx_rx(hldev); |
| |
| vxge_hw_device_clear_tx_rx(hldev); |
| VXGE_COMPLETE_ALL_RX(vdev); |
| VXGE_COMPLETE_ALL_TX(vdev); |
| |
| enable_irq(irq); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| } |
| #endif |
| |
| /* RTH configuration */ |
| static enum vxge_hw_status vxge_rth_configure(struct vxgedev *vdev) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_hw_rth_hash_types hash_types; |
| u8 itable[256] = {0}; /* indirection table */ |
| u8 mtable[256] = {0}; /* CPU to vpath mapping */ |
| int index; |
| |
| /* |
| * Filling |
| * - itable with bucket numbers |
| * - mtable with bucket-to-vpath mapping |
| */ |
| for (index = 0; index < (1 << vdev->config.rth_bkt_sz); index++) { |
| itable[index] = index; |
| mtable[index] = index % vdev->no_of_vpath; |
| } |
| |
| /* set indirection table, bucket-to-vpath mapping */ |
| status = vxge_hw_vpath_rts_rth_itable_set(vdev->vp_handles, |
| vdev->no_of_vpath, |
| mtable, itable, |
| vdev->config.rth_bkt_sz); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "RTH indirection table configuration failed " |
| "for vpath:%d", vdev->vpaths[0].device_id); |
| return status; |
| } |
| |
| /* Fill RTH hash types */ |
| hash_types.hash_type_tcpipv4_en = vdev->config.rth_hash_type_tcpipv4; |
| hash_types.hash_type_ipv4_en = vdev->config.rth_hash_type_ipv4; |
| hash_types.hash_type_tcpipv6_en = vdev->config.rth_hash_type_tcpipv6; |
| hash_types.hash_type_ipv6_en = vdev->config.rth_hash_type_ipv6; |
| hash_types.hash_type_tcpipv6ex_en = |
| vdev->config.rth_hash_type_tcpipv6ex; |
| hash_types.hash_type_ipv6ex_en = vdev->config.rth_hash_type_ipv6ex; |
| |
| /* |
| * Because the itable_set() method uses the active_table field |
| * for the target virtual path the RTH config should be updated |
| * for all VPATHs. The h/w only uses the lowest numbered VPATH |
| * when steering frames. |
| */ |
| for (index = 0; index < vdev->no_of_vpath; index++) { |
| status = vxge_hw_vpath_rts_rth_set( |
| vdev->vpaths[index].handle, |
| vdev->config.rth_algorithm, |
| &hash_types, |
| vdev->config.rth_bkt_sz); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "RTH configuration failed for vpath:%d", |
| vdev->vpaths[index].device_id); |
| return status; |
| } |
| } |
| |
| return status; |
| } |
| |
| /* reset vpaths */ |
| static enum vxge_hw_status vxge_reset_all_vpaths(struct vxgedev *vdev) |
| { |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_vpath *vpath; |
| int i; |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| if (vpath->handle) { |
| if (vxge_hw_vpath_reset(vpath->handle) == VXGE_HW_OK) { |
| if (is_vxge_card_up(vdev) && |
| vxge_hw_vpath_recover_from_reset( |
| vpath->handle) != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "vxge_hw_vpath_recover_" |
| "from_reset failed for vpath: " |
| "%d", i); |
| return status; |
| } |
| } else { |
| vxge_debug_init(VXGE_ERR, |
| "vxge_hw_vpath_reset failed for " |
| "vpath:%d", i); |
| return status; |
| } |
| } |
| } |
| |
| return status; |
| } |
| |
| /* close vpaths */ |
| static void vxge_close_vpaths(struct vxgedev *vdev, int index) |
| { |
| struct vxge_vpath *vpath; |
| int i; |
| |
| for (i = index; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| |
| if (vpath->handle && vpath->is_open) { |
| vxge_hw_vpath_close(vpath->handle); |
| vdev->stats.vpaths_open--; |
| } |
| vpath->is_open = 0; |
| vpath->handle = NULL; |
| } |
| } |
| |
| /* open vpaths */ |
| static int vxge_open_vpaths(struct vxgedev *vdev) |
| { |
| struct vxge_hw_vpath_attr attr; |
| enum vxge_hw_status status; |
| struct vxge_vpath *vpath; |
| u32 vp_id = 0; |
| int i; |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vxge_assert(vpath->is_configured); |
| |
| if (!vdev->titan1) { |
| struct vxge_hw_vp_config *vcfg; |
| vcfg = &vdev->devh->config.vp_config[vpath->device_id]; |
| |
| vcfg->rti.urange_a = RTI_T1A_RX_URANGE_A; |
| vcfg->rti.urange_b = RTI_T1A_RX_URANGE_B; |
| vcfg->rti.urange_c = RTI_T1A_RX_URANGE_C; |
| vcfg->tti.uec_a = TTI_T1A_TX_UFC_A; |
| vcfg->tti.uec_b = TTI_T1A_TX_UFC_B; |
| vcfg->tti.uec_c = TTI_T1A_TX_UFC_C(vdev->mtu); |
| vcfg->tti.uec_d = TTI_T1A_TX_UFC_D(vdev->mtu); |
| vcfg->tti.ltimer_val = VXGE_T1A_TTI_LTIMER_VAL; |
| vcfg->tti.rtimer_val = VXGE_T1A_TTI_RTIMER_VAL; |
| } |
| |
| attr.vp_id = vpath->device_id; |
| attr.fifo_attr.callback = vxge_xmit_compl; |
| attr.fifo_attr.txdl_term = vxge_tx_term; |
| attr.fifo_attr.per_txdl_space = sizeof(struct vxge_tx_priv); |
| attr.fifo_attr.userdata = &vpath->fifo; |
| |
| attr.ring_attr.callback = vxge_rx_1b_compl; |
| attr.ring_attr.rxd_init = vxge_rx_initial_replenish; |
| attr.ring_attr.rxd_term = vxge_rx_term; |
| attr.ring_attr.per_rxd_space = sizeof(struct vxge_rx_priv); |
| attr.ring_attr.userdata = &vpath->ring; |
| |
| vpath->ring.ndev = vdev->ndev; |
| vpath->ring.pdev = vdev->pdev; |
| |
| status = vxge_hw_vpath_open(vdev->devh, &attr, &vpath->handle); |
| if (status == VXGE_HW_OK) { |
| vpath->fifo.handle = |
| (struct __vxge_hw_fifo *)attr.fifo_attr.userdata; |
| vpath->ring.handle = |
| (struct __vxge_hw_ring *)attr.ring_attr.userdata; |
| vpath->fifo.tx_steering_type = |
| vdev->config.tx_steering_type; |
| vpath->fifo.ndev = vdev->ndev; |
| vpath->fifo.pdev = vdev->pdev; |
| |
| u64_stats_init(&vpath->fifo.stats.syncp); |
| u64_stats_init(&vpath->ring.stats.syncp); |
| |
| if (vdev->config.tx_steering_type) |
| vpath->fifo.txq = |
| netdev_get_tx_queue(vdev->ndev, i); |
| else |
| vpath->fifo.txq = |
| netdev_get_tx_queue(vdev->ndev, 0); |
| vpath->fifo.indicate_max_pkts = |
| vdev->config.fifo_indicate_max_pkts; |
| vpath->fifo.tx_vector_no = 0; |
| vpath->ring.rx_vector_no = 0; |
| vpath->ring.rx_hwts = vdev->rx_hwts; |
| vpath->is_open = 1; |
| vdev->vp_handles[i] = vpath->handle; |
| vpath->ring.vlan_tag_strip = vdev->vlan_tag_strip; |
| vdev->stats.vpaths_open++; |
| } else { |
| vdev->stats.vpath_open_fail++; |
| vxge_debug_init(VXGE_ERR, "%s: vpath: %d failed to " |
| "open with status: %d", |
| vdev->ndev->name, vpath->device_id, |
| status); |
| vxge_close_vpaths(vdev, 0); |
| return -EPERM; |
| } |
| |
| vp_id = vpath->handle->vpath->vp_id; |
| vdev->vpaths_deployed |= vxge_mBIT(vp_id); |
| } |
| |
| return VXGE_HW_OK; |
| } |
| |
| /** |
| * adaptive_coalesce_tx_interrupts - Changes the interrupt coalescing |
| * if the interrupts are not within a range |
| * @fifo: pointer to transmit fifo structure |
| * Description: The function changes boundary timer and restriction timer |
| * value depends on the traffic |
| * Return Value: None |
| */ |
| static void adaptive_coalesce_tx_interrupts(struct vxge_fifo *fifo) |
| { |
| fifo->interrupt_count++; |
| if (jiffies > fifo->jiffies + HZ / 100) { |
| struct __vxge_hw_fifo *hw_fifo = fifo->handle; |
| |
| fifo->jiffies = jiffies; |
| if (fifo->interrupt_count > VXGE_T1A_MAX_TX_INTERRUPT_COUNT && |
| hw_fifo->rtimer != VXGE_TTI_RTIMER_ADAPT_VAL) { |
| hw_fifo->rtimer = VXGE_TTI_RTIMER_ADAPT_VAL; |
| vxge_hw_vpath_dynamic_tti_rtimer_set(hw_fifo); |
| } else if (hw_fifo->rtimer != 0) { |
| hw_fifo->rtimer = 0; |
| vxge_hw_vpath_dynamic_tti_rtimer_set(hw_fifo); |
| } |
| fifo->interrupt_count = 0; |
| } |
| } |
| |
| /** |
| * adaptive_coalesce_rx_interrupts - Changes the interrupt coalescing |
| * if the interrupts are not within a range |
| * @ring: pointer to receive ring structure |
| * Description: The function increases of decreases the packet counts within |
| * the ranges of traffic utilization, if the interrupts due to this ring are |
| * not within a fixed range. |
| * Return Value: Nothing |
| */ |
| static void adaptive_coalesce_rx_interrupts(struct vxge_ring *ring) |
| { |
| ring->interrupt_count++; |
| if (jiffies > ring->jiffies + HZ / 100) { |
| struct __vxge_hw_ring *hw_ring = ring->handle; |
| |
| ring->jiffies = jiffies; |
| if (ring->interrupt_count > VXGE_T1A_MAX_INTERRUPT_COUNT && |
| hw_ring->rtimer != VXGE_RTI_RTIMER_ADAPT_VAL) { |
| hw_ring->rtimer = VXGE_RTI_RTIMER_ADAPT_VAL; |
| vxge_hw_vpath_dynamic_rti_rtimer_set(hw_ring); |
| } else if (hw_ring->rtimer != 0) { |
| hw_ring->rtimer = 0; |
| vxge_hw_vpath_dynamic_rti_rtimer_set(hw_ring); |
| } |
| ring->interrupt_count = 0; |
| } |
| } |
| |
| /* |
| * vxge_isr_napi |
| * @irq: the irq of the device. |
| * @dev_id: a void pointer to the hldev structure of the Titan device |
| * @ptregs: pointer to the registers pushed on the stack. |
| * |
| * This function is the ISR handler of the device when napi is enabled. It |
| * identifies the reason for the interrupt and calls the relevant service |
| * routines. |
| */ |
| static irqreturn_t vxge_isr_napi(int irq, void *dev_id) |
| { |
| struct net_device *dev; |
| struct __vxge_hw_device *hldev; |
| u64 reason; |
| enum vxge_hw_status status; |
| struct vxgedev *vdev = (struct vxgedev *)dev_id; |
| |
| vxge_debug_intr(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| dev = vdev->ndev; |
| hldev = pci_get_drvdata(vdev->pdev); |
| |
| if (pci_channel_offline(vdev->pdev)) |
| return IRQ_NONE; |
| |
| if (unlikely(!is_vxge_card_up(vdev))) |
| return IRQ_HANDLED; |
| |
| status = vxge_hw_device_begin_irq(hldev, vdev->exec_mode, &reason); |
| if (status == VXGE_HW_OK) { |
| vxge_hw_device_mask_all(hldev); |
| |
| if (reason & |
| VXGE_HW_TITAN_GENERAL_INT_STATUS_VPATH_TRAFFIC_INT( |
| vdev->vpaths_deployed >> |
| (64 - VXGE_HW_MAX_VIRTUAL_PATHS))) { |
| |
| vxge_hw_device_clear_tx_rx(hldev); |
| napi_schedule(&vdev->napi); |
| vxge_debug_intr(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| return IRQ_HANDLED; |
| } else |
| vxge_hw_device_unmask_all(hldev); |
| } else if (unlikely((status == VXGE_HW_ERR_VPATH) || |
| (status == VXGE_HW_ERR_CRITICAL) || |
| (status == VXGE_HW_ERR_FIFO))) { |
| vxge_hw_device_mask_all(hldev); |
| vxge_hw_device_flush_io(hldev); |
| return IRQ_HANDLED; |
| } else if (unlikely(status == VXGE_HW_ERR_SLOT_FREEZE)) |
| return IRQ_HANDLED; |
| |
| vxge_debug_intr(VXGE_TRACE, "%s:%d Exiting...", __func__, __LINE__); |
| return IRQ_NONE; |
| } |
| |
| #ifdef CONFIG_PCI_MSI |
| |
| static irqreturn_t vxge_tx_msix_handle(int irq, void *dev_id) |
| { |
| struct vxge_fifo *fifo = (struct vxge_fifo *)dev_id; |
| |
| adaptive_coalesce_tx_interrupts(fifo); |
| |
| vxge_hw_channel_msix_mask((struct __vxge_hw_channel *)fifo->handle, |
| fifo->tx_vector_no); |
| |
| vxge_hw_channel_msix_clear((struct __vxge_hw_channel *)fifo->handle, |
| fifo->tx_vector_no); |
| |
| VXGE_COMPLETE_VPATH_TX(fifo); |
| |
| vxge_hw_channel_msix_unmask((struct __vxge_hw_channel *)fifo->handle, |
| fifo->tx_vector_no); |
| |
| mmiowb(); |
| |
| return IRQ_HANDLED; |
| } |
| |
| static irqreturn_t vxge_rx_msix_napi_handle(int irq, void *dev_id) |
| { |
| struct vxge_ring *ring = (struct vxge_ring *)dev_id; |
| |
| adaptive_coalesce_rx_interrupts(ring); |
| |
| vxge_hw_channel_msix_mask((struct __vxge_hw_channel *)ring->handle, |
| ring->rx_vector_no); |
| |
| vxge_hw_channel_msix_clear((struct __vxge_hw_channel *)ring->handle, |
| ring->rx_vector_no); |
| |
| napi_schedule(&ring->napi); |
| return IRQ_HANDLED; |
| } |
| |
| static irqreturn_t |
| vxge_alarm_msix_handle(int irq, void *dev_id) |
| { |
| int i; |
| enum vxge_hw_status status; |
| struct vxge_vpath *vpath = (struct vxge_vpath *)dev_id; |
| struct vxgedev *vdev = vpath->vdev; |
| int msix_id = (vpath->handle->vpath->vp_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE) + VXGE_ALARM_MSIX_ID; |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| /* Reduce the chance of losing alarm interrupts by masking |
| * the vector. A pending bit will be set if an alarm is |
| * generated and on unmask the interrupt will be fired. |
| */ |
| vxge_hw_vpath_msix_mask(vdev->vpaths[i].handle, msix_id); |
| vxge_hw_vpath_msix_clear(vdev->vpaths[i].handle, msix_id); |
| mmiowb(); |
| |
| status = vxge_hw_vpath_alarm_process(vdev->vpaths[i].handle, |
| vdev->exec_mode); |
| if (status == VXGE_HW_OK) { |
| vxge_hw_vpath_msix_unmask(vdev->vpaths[i].handle, |
| msix_id); |
| mmiowb(); |
| continue; |
| } |
| vxge_debug_intr(VXGE_ERR, |
| "%s: vxge_hw_vpath_alarm_process failed %x ", |
| VXGE_DRIVER_NAME, status); |
| } |
| return IRQ_HANDLED; |
| } |
| |
| static int vxge_alloc_msix(struct vxgedev *vdev) |
| { |
| int j, i, ret = 0; |
| int msix_intr_vect = 0, temp; |
| vdev->intr_cnt = 0; |
| |
| start: |
| /* Tx/Rx MSIX Vectors count */ |
| vdev->intr_cnt = vdev->no_of_vpath * 2; |
| |
| /* Alarm MSIX Vectors count */ |
| vdev->intr_cnt++; |
| |
| vdev->entries = kcalloc(vdev->intr_cnt, sizeof(struct msix_entry), |
| GFP_KERNEL); |
| if (!vdev->entries) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: memory allocation failed", |
| VXGE_DRIVER_NAME); |
| ret = -ENOMEM; |
| goto alloc_entries_failed; |
| } |
| |
| vdev->vxge_entries = kcalloc(vdev->intr_cnt, |
| sizeof(struct vxge_msix_entry), |
| GFP_KERNEL); |
| if (!vdev->vxge_entries) { |
| vxge_debug_init(VXGE_ERR, "%s: memory allocation failed", |
| VXGE_DRIVER_NAME); |
| ret = -ENOMEM; |
| goto alloc_vxge_entries_failed; |
| } |
| |
| for (i = 0, j = 0; i < vdev->no_of_vpath; i++) { |
| |
| msix_intr_vect = i * VXGE_HW_VPATH_MSIX_ACTIVE; |
| |
| /* Initialize the fifo vector */ |
| vdev->entries[j].entry = msix_intr_vect; |
| vdev->vxge_entries[j].entry = msix_intr_vect; |
| vdev->vxge_entries[j].in_use = 0; |
| j++; |
| |
| /* Initialize the ring vector */ |
| vdev->entries[j].entry = msix_intr_vect + 1; |
| vdev->vxge_entries[j].entry = msix_intr_vect + 1; |
| vdev->vxge_entries[j].in_use = 0; |
| j++; |
| } |
| |
| /* Initialize the alarm vector */ |
| vdev->entries[j].entry = VXGE_ALARM_MSIX_ID; |
| vdev->vxge_entries[j].entry = VXGE_ALARM_MSIX_ID; |
| vdev->vxge_entries[j].in_use = 0; |
| |
| ret = pci_enable_msix(vdev->pdev, vdev->entries, vdev->intr_cnt); |
| if (ret > 0) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: MSI-X enable failed for %d vectors, ret: %d", |
| VXGE_DRIVER_NAME, vdev->intr_cnt, ret); |
| if ((max_config_vpath != VXGE_USE_DEFAULT) || (ret < 3)) { |
| ret = -ENODEV; |
| goto enable_msix_failed; |
| } |
| |
| kfree(vdev->entries); |
| kfree(vdev->vxge_entries); |
| vdev->entries = NULL; |
| vdev->vxge_entries = NULL; |
| /* Try with less no of vector by reducing no of vpaths count */ |
| temp = (ret - 1)/2; |
| vxge_close_vpaths(vdev, temp); |
| vdev->no_of_vpath = temp; |
| goto start; |
| } else if (ret < 0) { |
| ret = -ENODEV; |
| goto enable_msix_failed; |
| } |
| return 0; |
| |
| enable_msix_failed: |
| kfree(vdev->vxge_entries); |
| alloc_vxge_entries_failed: |
| kfree(vdev->entries); |
| alloc_entries_failed: |
| return ret; |
| } |
| |
| static int vxge_enable_msix(struct vxgedev *vdev) |
| { |
| |
| int i, ret = 0; |
| /* 0 - Tx, 1 - Rx */ |
| int tim_msix_id[4] = {0, 1, 0, 0}; |
| |
| vdev->intr_cnt = 0; |
| |
| /* allocate msix vectors */ |
| ret = vxge_alloc_msix(vdev); |
| if (!ret) { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| struct vxge_vpath *vpath = &vdev->vpaths[i]; |
| |
| /* If fifo or ring are not enabled, the MSIX vector for |
| * it should be set to 0. |
| */ |
| vpath->ring.rx_vector_no = (vpath->device_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE) + 1; |
| |
| vpath->fifo.tx_vector_no = (vpath->device_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE); |
| |
| vxge_hw_vpath_msix_set(vpath->handle, tim_msix_id, |
| VXGE_ALARM_MSIX_ID); |
| } |
| } |
| |
| return ret; |
| } |
| |
| static void vxge_rem_msix_isr(struct vxgedev *vdev) |
| { |
| int intr_cnt; |
| |
| for (intr_cnt = 0; intr_cnt < (vdev->no_of_vpath * 2 + 1); |
| intr_cnt++) { |
| if (vdev->vxge_entries[intr_cnt].in_use) { |
| synchronize_irq(vdev->entries[intr_cnt].vector); |
| free_irq(vdev->entries[intr_cnt].vector, |
| vdev->vxge_entries[intr_cnt].arg); |
| vdev->vxge_entries[intr_cnt].in_use = 0; |
| } |
| } |
| |
| kfree(vdev->entries); |
| kfree(vdev->vxge_entries); |
| vdev->entries = NULL; |
| vdev->vxge_entries = NULL; |
| |
| if (vdev->config.intr_type == MSI_X) |
| pci_disable_msix(vdev->pdev); |
| } |
| #endif |
| |
| static void vxge_rem_isr(struct vxgedev *vdev) |
| { |
| struct __vxge_hw_device *hldev; |
| hldev = pci_get_drvdata(vdev->pdev); |
| |
| #ifdef CONFIG_PCI_MSI |
| if (vdev->config.intr_type == MSI_X) { |
| vxge_rem_msix_isr(vdev); |
| } else |
| #endif |
| if (vdev->config.intr_type == INTA) { |
| synchronize_irq(vdev->pdev->irq); |
| free_irq(vdev->pdev->irq, vdev); |
| } |
| } |
| |
| static int vxge_add_isr(struct vxgedev *vdev) |
| { |
| int ret = 0; |
| #ifdef CONFIG_PCI_MSI |
| int vp_idx = 0, intr_idx = 0, intr_cnt = 0, msix_idx = 0, irq_req = 0; |
| int pci_fun = PCI_FUNC(vdev->pdev->devfn); |
| |
| if (vdev->config.intr_type == MSI_X) |
| ret = vxge_enable_msix(vdev); |
| |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: Enabling MSI-X Failed", VXGE_DRIVER_NAME); |
| vxge_debug_init(VXGE_ERR, |
| "%s: Defaulting to INTA", VXGE_DRIVER_NAME); |
| vdev->config.intr_type = INTA; |
| } |
| |
| if (vdev->config.intr_type == MSI_X) { |
| for (intr_idx = 0; |
| intr_idx < (vdev->no_of_vpath * |
| VXGE_HW_VPATH_MSIX_ACTIVE); intr_idx++) { |
| |
| msix_idx = intr_idx % VXGE_HW_VPATH_MSIX_ACTIVE; |
| irq_req = 0; |
| |
| switch (msix_idx) { |
| case 0: |
| snprintf(vdev->desc[intr_cnt], VXGE_INTR_STRLEN, |
| "%s:vxge:MSI-X %d - Tx - fn:%d vpath:%d", |
| vdev->ndev->name, |
| vdev->entries[intr_cnt].entry, |
| pci_fun, vp_idx); |
| ret = request_irq( |
| vdev->entries[intr_cnt].vector, |
| vxge_tx_msix_handle, 0, |
| vdev->desc[intr_cnt], |
| &vdev->vpaths[vp_idx].fifo); |
| vdev->vxge_entries[intr_cnt].arg = |
| &vdev->vpaths[vp_idx].fifo; |
| irq_req = 1; |
| break; |
| case 1: |
| snprintf(vdev->desc[intr_cnt], VXGE_INTR_STRLEN, |
| "%s:vxge:MSI-X %d - Rx - fn:%d vpath:%d", |
| vdev->ndev->name, |
| vdev->entries[intr_cnt].entry, |
| pci_fun, vp_idx); |
| ret = request_irq( |
| vdev->entries[intr_cnt].vector, |
| vxge_rx_msix_napi_handle, |
| 0, |
| vdev->desc[intr_cnt], |
| &vdev->vpaths[vp_idx].ring); |
| vdev->vxge_entries[intr_cnt].arg = |
| &vdev->vpaths[vp_idx].ring; |
| irq_req = 1; |
| break; |
| } |
| |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: MSIX - %d Registration failed", |
| vdev->ndev->name, intr_cnt); |
| vxge_rem_msix_isr(vdev); |
| vdev->config.intr_type = INTA; |
| vxge_debug_init(VXGE_ERR, |
| "%s: Defaulting to INTA" |
| , vdev->ndev->name); |
| goto INTA_MODE; |
| } |
| |
| if (irq_req) { |
| /* We requested for this msix interrupt */ |
| vdev->vxge_entries[intr_cnt].in_use = 1; |
| msix_idx += vdev->vpaths[vp_idx].device_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE; |
| vxge_hw_vpath_msix_unmask( |
| vdev->vpaths[vp_idx].handle, |
| msix_idx); |
| intr_cnt++; |
| } |
| |
| /* Point to next vpath handler */ |
| if (((intr_idx + 1) % VXGE_HW_VPATH_MSIX_ACTIVE == 0) && |
| (vp_idx < (vdev->no_of_vpath - 1))) |
| vp_idx++; |
| } |
| |
| intr_cnt = vdev->no_of_vpath * 2; |
| snprintf(vdev->desc[intr_cnt], VXGE_INTR_STRLEN, |
| "%s:vxge:MSI-X %d - Alarm - fn:%d", |
| vdev->ndev->name, |
| vdev->entries[intr_cnt].entry, |
| pci_fun); |
| /* For Alarm interrupts */ |
| ret = request_irq(vdev->entries[intr_cnt].vector, |
| vxge_alarm_msix_handle, 0, |
| vdev->desc[intr_cnt], |
| &vdev->vpaths[0]); |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: MSIX - %d Registration failed", |
| vdev->ndev->name, intr_cnt); |
| vxge_rem_msix_isr(vdev); |
| vdev->config.intr_type = INTA; |
| vxge_debug_init(VXGE_ERR, |
| "%s: Defaulting to INTA", |
| vdev->ndev->name); |
| goto INTA_MODE; |
| } |
| |
| msix_idx = (vdev->vpaths[0].handle->vpath->vp_id * |
| VXGE_HW_VPATH_MSIX_ACTIVE) + VXGE_ALARM_MSIX_ID; |
| vxge_hw_vpath_msix_unmask(vdev->vpaths[vp_idx].handle, |
| msix_idx); |
| vdev->vxge_entries[intr_cnt].in_use = 1; |
| vdev->vxge_entries[intr_cnt].arg = &vdev->vpaths[0]; |
| } |
| INTA_MODE: |
| #endif |
| |
| if (vdev->config.intr_type == INTA) { |
| snprintf(vdev->desc[0], VXGE_INTR_STRLEN, |
| "%s:vxge:INTA", vdev->ndev->name); |
| vxge_hw_device_set_intr_type(vdev->devh, |
| VXGE_HW_INTR_MODE_IRQLINE); |
| |
| vxge_hw_vpath_tti_ci_set(vdev->vpaths[0].fifo.handle); |
| |
| ret = request_irq((int) vdev->pdev->irq, |
| vxge_isr_napi, |
| IRQF_SHARED, vdev->desc[0], vdev); |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, |
| "%s %s-%d: ISR registration failed", |
| VXGE_DRIVER_NAME, "IRQ", vdev->pdev->irq); |
| return -ENODEV; |
| } |
| vxge_debug_init(VXGE_TRACE, |
| "new %s-%d line allocated", |
| "IRQ", vdev->pdev->irq); |
| } |
| |
| return VXGE_HW_OK; |
| } |
| |
| static void vxge_poll_vp_reset(unsigned long data) |
| { |
| struct vxgedev *vdev = (struct vxgedev *)data; |
| int i, j = 0; |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| if (test_bit(i, &vdev->vp_reset)) { |
| vxge_reset_vpath(vdev, i); |
| j++; |
| } |
| } |
| if (j && (vdev->config.intr_type != MSI_X)) { |
| vxge_hw_device_unmask_all(vdev->devh); |
| vxge_hw_device_flush_io(vdev->devh); |
| } |
| |
| mod_timer(&vdev->vp_reset_timer, jiffies + HZ / 2); |
| } |
| |
| static void vxge_poll_vp_lockup(unsigned long data) |
| { |
| struct vxgedev *vdev = (struct vxgedev *)data; |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxge_vpath *vpath; |
| struct vxge_ring *ring; |
| int i; |
| unsigned long rx_frms; |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| ring = &vdev->vpaths[i].ring; |
| |
| /* Truncated to machine word size number of frames */ |
| rx_frms = ACCESS_ONCE(ring->stats.rx_frms); |
| |
| /* Did this vpath received any packets */ |
| if (ring->stats.prev_rx_frms == rx_frms) { |
| status = vxge_hw_vpath_check_leak(ring->handle); |
| |
| /* Did it received any packets last time */ |
| if ((VXGE_HW_FAIL == status) && |
| (VXGE_HW_FAIL == ring->last_status)) { |
| |
| /* schedule vpath reset */ |
| if (!test_and_set_bit(i, &vdev->vp_reset)) { |
| vpath = &vdev->vpaths[i]; |
| |
| /* disable interrupts for this vpath */ |
| vxge_vpath_intr_disable(vdev, i); |
| |
| /* stop the queue for this vpath */ |
| netif_tx_stop_queue(vpath->fifo.txq); |
| continue; |
| } |
| } |
| } |
| ring->stats.prev_rx_frms = rx_frms; |
| ring->last_status = status; |
| } |
| |
| /* Check every 1 milli second */ |
| mod_timer(&vdev->vp_lockup_timer, jiffies + HZ / 1000); |
| } |
| |
| static netdev_features_t vxge_fix_features(struct net_device *dev, |
| netdev_features_t features) |
| { |
| netdev_features_t changed = dev->features ^ features; |
| |
| /* Enabling RTH requires some of the logic in vxge_device_register and a |
| * vpath reset. Due to these restrictions, only allow modification |
| * while the interface is down. |
| */ |
| if ((changed & NETIF_F_RXHASH) && netif_running(dev)) |
| features ^= NETIF_F_RXHASH; |
| |
| return features; |
| } |
| |
| static int vxge_set_features(struct net_device *dev, netdev_features_t features) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| netdev_features_t changed = dev->features ^ features; |
| |
| if (!(changed & NETIF_F_RXHASH)) |
| return 0; |
| |
| /* !netif_running() ensured by vxge_fix_features() */ |
| |
| vdev->devh->config.rth_en = !!(features & NETIF_F_RXHASH); |
| if (vxge_reset_all_vpaths(vdev) != VXGE_HW_OK) { |
| dev->features = features ^ NETIF_F_RXHASH; |
| vdev->devh->config.rth_en = !!(dev->features & NETIF_F_RXHASH); |
| return -EIO; |
| } |
| |
| return 0; |
| } |
| |
| /** |
| * vxge_open |
| * @dev: pointer to the device structure. |
| * |
| * This function is the open entry point of the driver. It mainly calls a |
| * function to allocate Rx buffers and inserts them into the buffer |
| * descriptors and then enables the Rx part of the NIC. |
| * Return value: '0' on success and an appropriate (-)ve integer as |
| * defined in errno.h file on failure. |
| */ |
| static int vxge_open(struct net_device *dev) |
| { |
| enum vxge_hw_status status; |
| struct vxgedev *vdev; |
| struct __vxge_hw_device *hldev; |
| struct vxge_vpath *vpath; |
| int ret = 0; |
| int i; |
| u64 val64, function_mode; |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d", dev->name, __func__, __LINE__); |
| |
| vdev = netdev_priv(dev); |
| hldev = pci_get_drvdata(vdev->pdev); |
| function_mode = vdev->config.device_hw_info.function_mode; |
| |
| /* make sure you have link off by default every time Nic is |
| * initialized */ |
| netif_carrier_off(dev); |
| |
| /* Open VPATHs */ |
| status = vxge_open_vpaths(vdev); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: fatal: Vpath open failed", vdev->ndev->name); |
| ret = -EPERM; |
| goto out0; |
| } |
| |
| vdev->mtu = dev->mtu; |
| |
| status = vxge_add_isr(vdev); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: fatal: ISR add failed", dev->name); |
| ret = -EPERM; |
| goto out1; |
| } |
| |
| if (vdev->config.intr_type != MSI_X) { |
| netif_napi_add(dev, &vdev->napi, vxge_poll_inta, |
| vdev->config.napi_weight); |
| napi_enable(&vdev->napi); |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vpath->ring.napi_p = &vdev->napi; |
| } |
| } else { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| netif_napi_add(dev, &vpath->ring.napi, |
| vxge_poll_msix, vdev->config.napi_weight); |
| napi_enable(&vpath->ring.napi); |
| vpath->ring.napi_p = &vpath->ring.napi; |
| } |
| } |
| |
| /* configure RTH */ |
| if (vdev->config.rth_steering) { |
| status = vxge_rth_configure(vdev); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: fatal: RTH configuration failed", |
| dev->name); |
| ret = -EPERM; |
| goto out2; |
| } |
| } |
| printk(KERN_INFO "%s: Receive Hashing Offload %s\n", dev->name, |
| hldev->config.rth_en ? "enabled" : "disabled"); |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| |
| /* set initial mtu before enabling the device */ |
| status = vxge_hw_vpath_mtu_set(vpath->handle, vdev->mtu); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: fatal: can not set new MTU", dev->name); |
| ret = -EPERM; |
| goto out2; |
| } |
| } |
| |
| VXGE_DEVICE_DEBUG_LEVEL_SET(VXGE_TRACE, VXGE_COMPONENT_LL, vdev); |
| vxge_debug_init(vdev->level_trace, |
| "%s: MTU is %d", vdev->ndev->name, vdev->mtu); |
| VXGE_DEVICE_DEBUG_LEVEL_SET(VXGE_ERR, VXGE_COMPONENT_LL, vdev); |
| |
| /* Restore the DA, VID table and also multicast and promiscuous mode |
| * states |
| */ |
| if (vdev->all_multi_flg) { |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| vxge_restore_vpath_mac_addr(vpath); |
| vxge_restore_vpath_vid_table(vpath); |
| |
| status = vxge_hw_vpath_mcast_enable(vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, |
| "%s:%d Enabling multicast failed", |
| __func__, __LINE__); |
| } |
| } |
| |
| /* Enable vpath to sniff all unicast/multicast traffic that not |
| * addressed to them. We allow promiscuous mode for PF only |
| */ |
| |
| val64 = 0; |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) |
| val64 |= VXGE_HW_RXMAC_AUTHORIZE_ALL_ADDR_VP(i); |
| |
| vxge_hw_mgmt_reg_write(vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof(struct vxge_hw_mrpcim_reg, |
| rxmac_authorize_all_addr), |
| val64); |
| |
| vxge_hw_mgmt_reg_write(vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof(struct vxge_hw_mrpcim_reg, |
| rxmac_authorize_all_vid), |
| val64); |
| |
| vxge_set_multicast(dev); |
| |
| /* Enabling Bcast and mcast for all vpath */ |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| status = vxge_hw_vpath_bcast_enable(vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, |
| "%s : Can not enable bcast for vpath " |
| "id %d", dev->name, i); |
| if (vdev->config.addr_learn_en) { |
| status = vxge_hw_vpath_mcast_enable(vpath->handle); |
| if (status != VXGE_HW_OK) |
| vxge_debug_init(VXGE_ERR, |
| "%s : Can not enable mcast for vpath " |
| "id %d", dev->name, i); |
| } |
| } |
| |
| vxge_hw_device_setpause_data(vdev->devh, 0, |
| vdev->config.tx_pause_enable, |
| vdev->config.rx_pause_enable); |
| |
| if (vdev->vp_reset_timer.function == NULL) |
| vxge_os_timer(&vdev->vp_reset_timer, vxge_poll_vp_reset, vdev, |
| HZ / 2); |
| |
| /* There is no need to check for RxD leak and RxD lookup on Titan1A */ |
| if (vdev->titan1 && vdev->vp_lockup_timer.function == NULL) |
| vxge_os_timer(&vdev->vp_lockup_timer, vxge_poll_vp_lockup, vdev, |
| HZ / 2); |
| |
| set_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| |
| smp_wmb(); |
| |
| if (vxge_hw_device_link_state_get(vdev->devh) == VXGE_HW_LINK_UP) { |
| netif_carrier_on(vdev->ndev); |
| netdev_notice(vdev->ndev, "Link Up\n"); |
| vdev->stats.link_up++; |
| } |
| |
| vxge_hw_device_intr_enable(vdev->devh); |
| |
| smp_wmb(); |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| vpath = &vdev->vpaths[i]; |
| |
| vxge_hw_vpath_enable(vpath->handle); |
| smp_wmb(); |
| vxge_hw_vpath_rx_doorbell_init(vpath->handle); |
| } |
| |
| netif_tx_start_all_queues(vdev->ndev); |
| |
| /* configure CI */ |
| vxge_config_ci_for_tti_rti(vdev); |
| |
| goto out0; |
| |
| out2: |
| vxge_rem_isr(vdev); |
| |
| /* Disable napi */ |
| if (vdev->config.intr_type != MSI_X) |
| napi_disable(&vdev->napi); |
| else { |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| napi_disable(&vdev->vpaths[i].ring.napi); |
| } |
| |
| out1: |
| vxge_close_vpaths(vdev, 0); |
| out0: |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", |
| dev->name, __func__, __LINE__); |
| return ret; |
| } |
| |
| /* Loop through the mac address list and delete all the entries */ |
| static void vxge_free_mac_add_list(struct vxge_vpath *vpath) |
| { |
| |
| struct list_head *entry, *next; |
| if (list_empty(&vpath->mac_addr_list)) |
| return; |
| |
| list_for_each_safe(entry, next, &vpath->mac_addr_list) { |
| list_del(entry); |
| kfree((struct vxge_mac_addrs *)entry); |
| } |
| } |
| |
| static void vxge_napi_del_all(struct vxgedev *vdev) |
| { |
| int i; |
| if (vdev->config.intr_type != MSI_X) |
| netif_napi_del(&vdev->napi); |
| else { |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| netif_napi_del(&vdev->vpaths[i].ring.napi); |
| } |
| } |
| |
| static int do_vxge_close(struct net_device *dev, int do_io) |
| { |
| enum vxge_hw_status status; |
| struct vxgedev *vdev; |
| struct __vxge_hw_device *hldev; |
| int i; |
| u64 val64, vpath_vector; |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d", |
| dev->name, __func__, __LINE__); |
| |
| vdev = netdev_priv(dev); |
| hldev = pci_get_drvdata(vdev->pdev); |
| |
| if (unlikely(!is_vxge_card_up(vdev))) |
| return 0; |
| |
| /* If vxge_handle_crit_err task is executing, |
| * wait till it completes. */ |
| while (test_and_set_bit(__VXGE_STATE_RESET_CARD, &vdev->state)) |
| msleep(50); |
| |
| if (do_io) { |
| /* Put the vpath back in normal mode */ |
| vpath_vector = vxge_mBIT(vdev->vpaths[0].device_id); |
| status = vxge_hw_mgmt_reg_read(vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof( |
| struct vxge_hw_mrpcim_reg, |
| rts_mgr_cbasin_cfg), |
| &val64); |
| if (status == VXGE_HW_OK) { |
| val64 &= ~vpath_vector; |
| status = vxge_hw_mgmt_reg_write(vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof( |
| struct vxge_hw_mrpcim_reg, |
| rts_mgr_cbasin_cfg), |
| val64); |
| } |
| |
| /* Remove the function 0 from promiscuous mode */ |
| vxge_hw_mgmt_reg_write(vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof(struct vxge_hw_mrpcim_reg, |
| rxmac_authorize_all_addr), |
| 0); |
| |
| vxge_hw_mgmt_reg_write(vdev->devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| (ulong)offsetof(struct vxge_hw_mrpcim_reg, |
| rxmac_authorize_all_vid), |
| 0); |
| |
| smp_wmb(); |
| } |
| |
| if (vdev->titan1) |
| del_timer_sync(&vdev->vp_lockup_timer); |
| |
| del_timer_sync(&vdev->vp_reset_timer); |
| |
| if (do_io) |
| vxge_hw_device_wait_receive_idle(hldev); |
| |
| clear_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| |
| /* Disable napi */ |
| if (vdev->config.intr_type != MSI_X) |
| napi_disable(&vdev->napi); |
| else { |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| napi_disable(&vdev->vpaths[i].ring.napi); |
| } |
| |
| netif_carrier_off(vdev->ndev); |
| netdev_notice(vdev->ndev, "Link Down\n"); |
| netif_tx_stop_all_queues(vdev->ndev); |
| |
| /* Note that at this point xmit() is stopped by upper layer */ |
| if (do_io) |
| vxge_hw_device_intr_disable(vdev->devh); |
| |
| vxge_rem_isr(vdev); |
| |
| vxge_napi_del_all(vdev); |
| |
| if (do_io) |
| vxge_reset_all_vpaths(vdev); |
| |
| vxge_close_vpaths(vdev, 0); |
| |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s: %s:%d Exiting...", dev->name, __func__, __LINE__); |
| |
| clear_bit(__VXGE_STATE_RESET_CARD, &vdev->state); |
| |
| return 0; |
| } |
| |
| /** |
| * vxge_close |
| * @dev: device pointer. |
| * |
| * This is the stop entry point of the driver. It needs to undo exactly |
| * whatever was done by the open entry point, thus it's usually referred to |
| * as the close function.Among other things this function mainly stops the |
| * Rx side of the NIC and frees all the Rx buffers in the Rx rings. |
| * Return value: '0' on success and an appropriate (-)ve integer as |
| * defined in errno.h file on failure. |
| */ |
| static int vxge_close(struct net_device *dev) |
| { |
| do_vxge_close(dev, 1); |
| return 0; |
| } |
| |
| /** |
| * vxge_change_mtu |
| * @dev: net device pointer. |
| * @new_mtu :the new MTU size for the device. |
| * |
| * A driver entry point to change MTU size for the device. Before changing |
| * the MTU the device must be stopped. |
| */ |
| static int vxge_change_mtu(struct net_device *dev, int new_mtu) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| |
| vxge_debug_entryexit(vdev->level_trace, |
| "%s:%d", __func__, __LINE__); |
| if ((new_mtu < VXGE_HW_MIN_MTU) || (new_mtu > VXGE_HW_MAX_MTU)) { |
| vxge_debug_init(vdev->level_err, |
| "%s: mtu size is invalid", dev->name); |
| return -EPERM; |
| } |
| |
| /* check if device is down already */ |
| if (unlikely(!is_vxge_card_up(vdev))) { |
| /* just store new value, will use later on open() */ |
| dev->mtu = new_mtu; |
| vxge_debug_init(vdev->level_err, |
| "%s", "device is down on MTU change"); |
| return 0; |
| } |
| |
| vxge_debug_init(vdev->level_trace, |
| "trying to apply new MTU %d", new_mtu); |
| |
| if (vxge_close(dev)) |
| return -EIO; |
| |
| dev->mtu = new_mtu; |
| vdev->mtu = new_mtu; |
| |
| if (vxge_open(dev)) |
| return -EIO; |
| |
| vxge_debug_init(vdev->level_trace, |
| "%s: MTU changed to %d", vdev->ndev->name, new_mtu); |
| |
| vxge_debug_entryexit(vdev->level_trace, |
| "%s:%d Exiting...", __func__, __LINE__); |
| |
| return 0; |
| } |
| |
| /** |
| * vxge_get_stats64 |
| * @dev: pointer to the device structure |
| * @stats: pointer to struct rtnl_link_stats64 |
| * |
| */ |
| static struct rtnl_link_stats64 * |
| vxge_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *net_stats) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| int k; |
| |
| /* net_stats already zeroed by caller */ |
| for (k = 0; k < vdev->no_of_vpath; k++) { |
| struct vxge_ring_stats *rxstats = &vdev->vpaths[k].ring.stats; |
| struct vxge_fifo_stats *txstats = &vdev->vpaths[k].fifo.stats; |
| unsigned int start; |
| u64 packets, bytes, multicast; |
| |
| do { |
| start = u64_stats_fetch_begin_bh(&rxstats->syncp); |
| |
| packets = rxstats->rx_frms; |
| multicast = rxstats->rx_mcast; |
| bytes = rxstats->rx_bytes; |
| } while (u64_stats_fetch_retry_bh(&rxstats->syncp, start)); |
| |
| net_stats->rx_packets += packets; |
| net_stats->rx_bytes += bytes; |
| net_stats->multicast += multicast; |
| |
| net_stats->rx_errors += rxstats->rx_errors; |
| net_stats->rx_dropped += rxstats->rx_dropped; |
| |
| do { |
| start = u64_stats_fetch_begin_bh(&txstats->syncp); |
| |
| packets = txstats->tx_frms; |
| bytes = txstats->tx_bytes; |
| } while (u64_stats_fetch_retry_bh(&txstats->syncp, start)); |
| |
| net_stats->tx_packets += packets; |
| net_stats->tx_bytes += bytes; |
| net_stats->tx_errors += txstats->tx_errors; |
| } |
| |
| return net_stats; |
| } |
| |
| static enum vxge_hw_status vxge_timestamp_config(struct __vxge_hw_device *devh) |
| { |
| enum vxge_hw_status status; |
| u64 val64; |
| |
| /* Timestamp is passed to the driver via the FCS, therefore we |
| * must disable the FCS stripping by the adapter. Since this is |
| * required for the driver to load (due to a hardware bug), |
| * there is no need to do anything special here. |
| */ |
| val64 = VXGE_HW_XMAC_TIMESTAMP_EN | |
| VXGE_HW_XMAC_TIMESTAMP_USE_LINK_ID(0) | |
| VXGE_HW_XMAC_TIMESTAMP_INTERVAL(0); |
| |
| status = vxge_hw_mgmt_reg_write(devh, |
| vxge_hw_mgmt_reg_type_mrpcim, |
| 0, |
| offsetof(struct vxge_hw_mrpcim_reg, |
| xmac_timestamp), |
| val64); |
| vxge_hw_device_flush_io(devh); |
| devh->config.hwts_en = VXGE_HW_HWTS_ENABLE; |
| return status; |
| } |
| |
| static int vxge_hwtstamp_set(struct vxgedev *vdev, void __user *data) |
| { |
| struct hwtstamp_config config; |
| int i; |
| |
| if (copy_from_user(&config, data, sizeof(config))) |
| return -EFAULT; |
| |
| /* reserved for future extensions */ |
| if (config.flags) |
| return -EINVAL; |
| |
| /* Transmit HW Timestamp not supported */ |
| switch (config.tx_type) { |
| case HWTSTAMP_TX_OFF: |
| break; |
| case HWTSTAMP_TX_ON: |
| default: |
| return -ERANGE; |
| } |
| |
| switch (config.rx_filter) { |
| case HWTSTAMP_FILTER_NONE: |
| vdev->rx_hwts = 0; |
| config.rx_filter = HWTSTAMP_FILTER_NONE; |
| break; |
| |
| case HWTSTAMP_FILTER_ALL: |
| case HWTSTAMP_FILTER_SOME: |
| case HWTSTAMP_FILTER_PTP_V1_L4_EVENT: |
| case HWTSTAMP_FILTER_PTP_V1_L4_SYNC: |
| case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ: |
| case HWTSTAMP_FILTER_PTP_V2_L4_EVENT: |
| case HWTSTAMP_FILTER_PTP_V2_L4_SYNC: |
| case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ: |
| case HWTSTAMP_FILTER_PTP_V2_L2_EVENT: |
| case HWTSTAMP_FILTER_PTP_V2_L2_SYNC: |
| case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ: |
| case HWTSTAMP_FILTER_PTP_V2_EVENT: |
| case HWTSTAMP_FILTER_PTP_V2_SYNC: |
| case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ: |
| if (vdev->devh->config.hwts_en != VXGE_HW_HWTS_ENABLE) |
| return -EFAULT; |
| |
| vdev->rx_hwts = 1; |
| config.rx_filter = HWTSTAMP_FILTER_ALL; |
| break; |
| |
| default: |
| return -ERANGE; |
| } |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| vdev->vpaths[i].ring.rx_hwts = vdev->rx_hwts; |
| |
| if (copy_to_user(data, &config, sizeof(config))) |
| return -EFAULT; |
| |
| return 0; |
| } |
| |
| static int vxge_hwtstamp_get(struct vxgedev *vdev, void __user *data) |
| { |
| struct hwtstamp_config config; |
| |
| config.flags = 0; |
| config.tx_type = HWTSTAMP_TX_OFF; |
| config.rx_filter = (vdev->rx_hwts ? |
| HWTSTAMP_FILTER_ALL : HWTSTAMP_FILTER_NONE); |
| |
| if (copy_to_user(data, &config, sizeof(config))) |
| return -EFAULT; |
| |
| return 0; |
| } |
| |
| /** |
| * vxge_ioctl |
| * @dev: Device pointer. |
| * @ifr: An IOCTL specific structure, that can contain a pointer to |
| * a proprietary structure used to pass information to the driver. |
| * @cmd: This is used to distinguish between the different commands that |
| * can be passed to the IOCTL functions. |
| * |
| * Entry point for the Ioctl. |
| */ |
| static int vxge_ioctl(struct net_device *dev, struct ifreq *rq, int cmd) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| |
| switch (cmd) { |
| case SIOCSHWTSTAMP: |
| return vxge_hwtstamp_set(vdev, rq->ifr_data); |
| case SIOCGHWTSTAMP: |
| return vxge_hwtstamp_get(vdev, rq->ifr_data); |
| default: |
| return -EOPNOTSUPP; |
| } |
| } |
| |
| /** |
| * vxge_tx_watchdog |
| * @dev: pointer to net device structure |
| * |
| * Watchdog for transmit side. |
| * This function is triggered if the Tx Queue is stopped |
| * for a pre-defined amount of time when the Interface is still up. |
| */ |
| static void vxge_tx_watchdog(struct net_device *dev) |
| { |
| struct vxgedev *vdev; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| vdev = netdev_priv(dev); |
| |
| vdev->cric_err_event = VXGE_HW_EVENT_RESET_START; |
| |
| schedule_work(&vdev->reset_task); |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| } |
| |
| /** |
| * vxge_vlan_rx_add_vid |
| * @dev: net device pointer. |
| * @proto: vlan protocol |
| * @vid: vid |
| * |
| * Add the vlan id to the devices vlan id table |
| */ |
| static int |
| vxge_vlan_rx_add_vid(struct net_device *dev, __be16 proto, u16 vid) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| struct vxge_vpath *vpath; |
| int vp_id; |
| |
| /* Add these vlan to the vid table */ |
| for (vp_id = 0; vp_id < vdev->no_of_vpath; vp_id++) { |
| vpath = &vdev->vpaths[vp_id]; |
| if (!vpath->is_open) |
| continue; |
| vxge_hw_vpath_vid_add(vpath->handle, vid); |
| } |
| set_bit(vid, vdev->active_vlans); |
| return 0; |
| } |
| |
| /** |
| * vxge_vlan_rx_kill_vid |
| * @dev: net device pointer. |
| * @proto: vlan protocol |
| * @vid: vid |
| * |
| * Remove the vlan id from the device's vlan id table |
| */ |
| static int |
| vxge_vlan_rx_kill_vid(struct net_device *dev, __be16 proto, u16 vid) |
| { |
| struct vxgedev *vdev = netdev_priv(dev); |
| struct vxge_vpath *vpath; |
| int vp_id; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| |
| /* Delete this vlan from the vid table */ |
| for (vp_id = 0; vp_id < vdev->no_of_vpath; vp_id++) { |
| vpath = &vdev->vpaths[vp_id]; |
| if (!vpath->is_open) |
| continue; |
| vxge_hw_vpath_vid_delete(vpath->handle, vid); |
| } |
| vxge_debug_entryexit(VXGE_TRACE, |
| "%s:%d Exiting...", __func__, __LINE__); |
| clear_bit(vid, vdev->active_vlans); |
| return 0; |
| } |
| |
| static const struct net_device_ops vxge_netdev_ops = { |
| .ndo_open = vxge_open, |
| .ndo_stop = vxge_close, |
| .ndo_get_stats64 = vxge_get_stats64, |
| .ndo_start_xmit = vxge_xmit, |
| .ndo_validate_addr = eth_validate_addr, |
| .ndo_set_rx_mode = vxge_set_multicast, |
| .ndo_do_ioctl = vxge_ioctl, |
| .ndo_set_mac_address = vxge_set_mac_addr, |
| .ndo_change_mtu = vxge_change_mtu, |
| .ndo_fix_features = vxge_fix_features, |
| .ndo_set_features = vxge_set_features, |
| .ndo_vlan_rx_kill_vid = vxge_vlan_rx_kill_vid, |
| .ndo_vlan_rx_add_vid = vxge_vlan_rx_add_vid, |
| .ndo_tx_timeout = vxge_tx_watchdog, |
| #ifdef CONFIG_NET_POLL_CONTROLLER |
| .ndo_poll_controller = vxge_netpoll, |
| #endif |
| }; |
| |
| static int vxge_device_register(struct __vxge_hw_device *hldev, |
| struct vxge_config *config, int high_dma, |
| int no_of_vpath, struct vxgedev **vdev_out) |
| { |
| struct net_device *ndev; |
| enum vxge_hw_status status = VXGE_HW_OK; |
| struct vxgedev *vdev; |
| int ret = 0, no_of_queue = 1; |
| u64 stat; |
| |
| *vdev_out = NULL; |
| if (config->tx_steering_type) |
| no_of_queue = no_of_vpath; |
| |
| ndev = alloc_etherdev_mq(sizeof(struct vxgedev), |
| no_of_queue); |
| if (ndev == NULL) { |
| vxge_debug_init( |
| vxge_hw_device_trace_level_get(hldev), |
| "%s : device allocation failed", __func__); |
| ret = -ENODEV; |
| goto _out0; |
| } |
| |
| vxge_debug_entryexit( |
| vxge_hw_device_trace_level_get(hldev), |
| "%s: %s:%d Entering...", |
| ndev->name, __func__, __LINE__); |
| |
| vdev = netdev_priv(ndev); |
| memset(vdev, 0, sizeof(struct vxgedev)); |
| |
| vdev->ndev = ndev; |
| vdev->devh = hldev; |
| vdev->pdev = hldev->pdev; |
| memcpy(&vdev->config, config, sizeof(struct vxge_config)); |
| vdev->rx_hwts = 0; |
| vdev->titan1 = (vdev->pdev->revision == VXGE_HW_TITAN1_PCI_REVISION); |
| |
| SET_NETDEV_DEV(ndev, &vdev->pdev->dev); |
| |
| ndev->hw_features = NETIF_F_RXCSUM | NETIF_F_SG | |
| NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM | |
| NETIF_F_TSO | NETIF_F_TSO6 | |
| NETIF_F_HW_VLAN_CTAG_TX; |
| if (vdev->config.rth_steering != NO_STEERING) |
| ndev->hw_features |= NETIF_F_RXHASH; |
| |
| ndev->features |= ndev->hw_features | |
| NETIF_F_HW_VLAN_CTAG_RX | NETIF_F_HW_VLAN_CTAG_FILTER; |
| |
| |
| ndev->netdev_ops = &vxge_netdev_ops; |
| |
| ndev->watchdog_timeo = VXGE_LL_WATCH_DOG_TIMEOUT; |
| INIT_WORK(&vdev->reset_task, vxge_reset); |
| |
| vxge_initialize_ethtool_ops(ndev); |
| |
| /* Allocate memory for vpath */ |
| vdev->vpaths = kzalloc((sizeof(struct vxge_vpath)) * |
| no_of_vpath, GFP_KERNEL); |
| if (!vdev->vpaths) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: vpath memory allocation failed", |
| vdev->ndev->name); |
| ret = -ENOMEM; |
| goto _out1; |
| } |
| |
| vxge_debug_init(vxge_hw_device_trace_level_get(hldev), |
| "%s : checksumming enabled", __func__); |
| |
| if (high_dma) { |
| ndev->features |= NETIF_F_HIGHDMA; |
| vxge_debug_init(vxge_hw_device_trace_level_get(hldev), |
| "%s : using High DMA", __func__); |
| } |
| |
| ret = register_netdev(ndev); |
| if (ret) { |
| vxge_debug_init(vxge_hw_device_trace_level_get(hldev), |
| "%s: %s : device registration failed!", |
| ndev->name, __func__); |
| goto _out2; |
| } |
| |
| /* Set the factory defined MAC address initially */ |
| ndev->addr_len = ETH_ALEN; |
| |
| /* Make Link state as off at this point, when the Link change |
| * interrupt comes the state will be automatically changed to |
| * the right state. |
| */ |
| netif_carrier_off(ndev); |
| |
| vxge_debug_init(vxge_hw_device_trace_level_get(hldev), |
| "%s: Ethernet device registered", |
| ndev->name); |
| |
| hldev->ndev = ndev; |
| *vdev_out = vdev; |
| |
| /* Resetting the Device stats */ |
| status = vxge_hw_mrpcim_stats_access( |
| hldev, |
| VXGE_HW_STATS_OP_CLEAR_ALL_STATS, |
| 0, |
| 0, |
| &stat); |
| |
| if (status == VXGE_HW_ERR_PRIVILAGED_OPEARATION) |
| vxge_debug_init( |
| vxge_hw_device_trace_level_get(hldev), |
| "%s: device stats clear returns" |
| "VXGE_HW_ERR_PRIVILAGED_OPEARATION", ndev->name); |
| |
| vxge_debug_entryexit(vxge_hw_device_trace_level_get(hldev), |
| "%s: %s:%d Exiting...", |
| ndev->name, __func__, __LINE__); |
| |
| return ret; |
| _out2: |
| kfree(vdev->vpaths); |
| _out1: |
| free_netdev(ndev); |
| _out0: |
| return ret; |
| } |
| |
| /* |
| * vxge_device_unregister |
| * |
| * This function will unregister and free network device |
| */ |
| static void vxge_device_unregister(struct __vxge_hw_device *hldev) |
| { |
| struct vxgedev *vdev; |
| struct net_device *dev; |
| char buf[IFNAMSIZ]; |
| |
| dev = hldev->ndev; |
| vdev = netdev_priv(dev); |
| |
| vxge_debug_entryexit(vdev->level_trace, "%s: %s:%d", vdev->ndev->name, |
| __func__, __LINE__); |
| |
| strncpy(buf, dev->name, IFNAMSIZ); |
| |
| flush_work(&vdev->reset_task); |
| |
| /* in 2.6 will call stop() if device is up */ |
| unregister_netdev(dev); |
| |
| kfree(vdev->vpaths); |
| |
| /* we are safe to free it now */ |
| free_netdev(dev); |
| |
| vxge_debug_init(vdev->level_trace, "%s: ethernet device unregistered", |
| buf); |
| vxge_debug_entryexit(vdev->level_trace, "%s: %s:%d Exiting...", buf, |
| __func__, __LINE__); |
| } |
| |
| /* |
| * vxge_callback_crit_err |
| * |
| * This function is called by the alarm handler in interrupt context. |
| * Driver must analyze it based on the event type. |
| */ |
| static void |
| vxge_callback_crit_err(struct __vxge_hw_device *hldev, |
| enum vxge_hw_event type, u64 vp_id) |
| { |
| struct net_device *dev = hldev->ndev; |
| struct vxgedev *vdev = netdev_priv(dev); |
| struct vxge_vpath *vpath = NULL; |
| int vpath_idx; |
| |
| vxge_debug_entryexit(vdev->level_trace, |
| "%s: %s:%d", vdev->ndev->name, __func__, __LINE__); |
| |
| /* Note: This event type should be used for device wide |
| * indications only - Serious errors, Slot freeze and critical errors |
| */ |
| vdev->cric_err_event = type; |
| |
| for (vpath_idx = 0; vpath_idx < vdev->no_of_vpath; vpath_idx++) { |
| vpath = &vdev->vpaths[vpath_idx]; |
| if (vpath->device_id == vp_id) |
| break; |
| } |
| |
| if (!test_bit(__VXGE_STATE_RESET_CARD, &vdev->state)) { |
| if (type == VXGE_HW_EVENT_SLOT_FREEZE) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: Slot is frozen", vdev->ndev->name); |
| } else if (type == VXGE_HW_EVENT_SERR) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: Encountered Serious Error", |
| vdev->ndev->name); |
| } else if (type == VXGE_HW_EVENT_CRITICAL_ERR) |
| vxge_debug_init(VXGE_ERR, |
| "%s: Encountered Critical Error", |
| vdev->ndev->name); |
| } |
| |
| if ((type == VXGE_HW_EVENT_SERR) || |
| (type == VXGE_HW_EVENT_SLOT_FREEZE)) { |
| if (unlikely(vdev->exec_mode)) |
| clear_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| } else if (type == VXGE_HW_EVENT_CRITICAL_ERR) { |
| vxge_hw_device_mask_all(hldev); |
| if (unlikely(vdev->exec_mode)) |
| clear_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| } else if ((type == VXGE_HW_EVENT_FIFO_ERR) || |
| (type == VXGE_HW_EVENT_VPATH_ERR)) { |
| |
| if (unlikely(vdev->exec_mode)) |
| clear_bit(__VXGE_STATE_CARD_UP, &vdev->state); |
| else { |
| /* check if this vpath is already set for reset */ |
| if (!test_and_set_bit(vpath_idx, &vdev->vp_reset)) { |
| |
| /* disable interrupts for this vpath */ |
| vxge_vpath_intr_disable(vdev, vpath_idx); |
| |
| /* stop the queue for this vpath */ |
| netif_tx_stop_queue(vpath->fifo.txq); |
| } |
| } |
| } |
| |
| vxge_debug_entryexit(vdev->level_trace, |
| "%s: %s:%d Exiting...", |
| vdev->ndev->name, __func__, __LINE__); |
| } |
| |
| static void verify_bandwidth(void) |
| { |
| int i, band_width, total = 0, equal_priority = 0; |
| |
| /* 1. If user enters 0 for some fifo, give equal priority to all */ |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) { |
| if (bw_percentage[i] == 0) { |
| equal_priority = 1; |
| break; |
| } |
| } |
| |
| if (!equal_priority) { |
| /* 2. If sum exceeds 100, give equal priority to all */ |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) { |
| if (bw_percentage[i] == 0xFF) |
| break; |
| |
| total += bw_percentage[i]; |
| if (total > VXGE_HW_VPATH_BANDWIDTH_MAX) { |
| equal_priority = 1; |
| break; |
| } |
| } |
| } |
| |
| if (!equal_priority) { |
| /* Is all the bandwidth consumed? */ |
| if (total < VXGE_HW_VPATH_BANDWIDTH_MAX) { |
| if (i < VXGE_HW_MAX_VIRTUAL_PATHS) { |
| /* Split rest of bw equally among next VPs*/ |
| band_width = |
| (VXGE_HW_VPATH_BANDWIDTH_MAX - total) / |
| (VXGE_HW_MAX_VIRTUAL_PATHS - i); |
| if (band_width < 2) /* min of 2% */ |
| equal_priority = 1; |
| else { |
| for (; i < VXGE_HW_MAX_VIRTUAL_PATHS; |
| i++) |
| bw_percentage[i] = |
| band_width; |
| } |
| } |
| } else if (i < VXGE_HW_MAX_VIRTUAL_PATHS) |
| equal_priority = 1; |
| } |
| |
| if (equal_priority) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: Assigning equal bandwidth to all the vpaths", |
| VXGE_DRIVER_NAME); |
| bw_percentage[0] = VXGE_HW_VPATH_BANDWIDTH_MAX / |
| VXGE_HW_MAX_VIRTUAL_PATHS; |
| for (i = 1; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) |
| bw_percentage[i] = bw_percentage[0]; |
| } |
| } |
| |
| /* |
| * Vpath configuration |
| */ |
| static int vxge_config_vpaths(struct vxge_hw_device_config *device_config, |
| u64 vpath_mask, struct vxge_config *config_param) |
| { |
| int i, no_of_vpaths = 0, default_no_vpath = 0, temp; |
| u32 txdl_size, txdl_per_memblock; |
| |
| temp = driver_config->vpath_per_dev; |
| if ((driver_config->vpath_per_dev == VXGE_USE_DEFAULT) && |
| (max_config_dev == VXGE_MAX_CONFIG_DEV)) { |
| /* No more CPU. Return vpath number as zero.*/ |
| if (driver_config->g_no_cpus == -1) |
| return 0; |
| |
| if (!driver_config->g_no_cpus) |
| driver_config->g_no_cpus = |
| netif_get_num_default_rss_queues(); |
| |
| driver_config->vpath_per_dev = driver_config->g_no_cpus >> 1; |
| if (!driver_config->vpath_per_dev) |
| driver_config->vpath_per_dev = 1; |
| |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) |
| if (!vxge_bVALn(vpath_mask, i, 1)) |
| continue; |
| else |
| default_no_vpath++; |
| if (default_no_vpath < driver_config->vpath_per_dev) |
| driver_config->vpath_per_dev = default_no_vpath; |
| |
| driver_config->g_no_cpus = driver_config->g_no_cpus - |
| (driver_config->vpath_per_dev * 2); |
| if (driver_config->g_no_cpus <= 0) |
| driver_config->g_no_cpus = -1; |
| } |
| |
| if (driver_config->vpath_per_dev == 1) { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: Disable tx and rx steering, " |
| "as single vpath is configured", VXGE_DRIVER_NAME); |
| config_param->rth_steering = NO_STEERING; |
| config_param->tx_steering_type = NO_STEERING; |
| device_config->rth_en = 0; |
| } |
| |
| /* configure bandwidth */ |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) |
| device_config->vp_config[i].min_bandwidth = bw_percentage[i]; |
| |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) { |
| device_config->vp_config[i].vp_id = i; |
| device_config->vp_config[i].mtu = VXGE_HW_DEFAULT_MTU; |
| if (no_of_vpaths < driver_config->vpath_per_dev) { |
| if (!vxge_bVALn(vpath_mask, i, 1)) { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: vpath: %d is not available", |
| VXGE_DRIVER_NAME, i); |
| continue; |
| } else { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: vpath: %d available", |
| VXGE_DRIVER_NAME, i); |
| no_of_vpaths++; |
| } |
| } else { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: vpath: %d is not configured, " |
| "max_config_vpath exceeded", |
| VXGE_DRIVER_NAME, i); |
| break; |
| } |
| |
| /* Configure Tx fifo's */ |
| device_config->vp_config[i].fifo.enable = |
| VXGE_HW_FIFO_ENABLE; |
| device_config->vp_config[i].fifo.max_frags = |
| MAX_SKB_FRAGS + 1; |
| device_config->vp_config[i].fifo.memblock_size = |
| VXGE_HW_MIN_FIFO_MEMBLOCK_SIZE; |
| |
| txdl_size = device_config->vp_config[i].fifo.max_frags * |
| sizeof(struct vxge_hw_fifo_txd); |
| txdl_per_memblock = VXGE_HW_MIN_FIFO_MEMBLOCK_SIZE / txdl_size; |
| |
| device_config->vp_config[i].fifo.fifo_blocks = |
| ((VXGE_DEF_FIFO_LENGTH - 1) / txdl_per_memblock) + 1; |
| |
| device_config->vp_config[i].fifo.intr = |
| VXGE_HW_FIFO_QUEUE_INTR_DISABLE; |
| |
| /* Configure tti properties */ |
| device_config->vp_config[i].tti.intr_enable = |
| VXGE_HW_TIM_INTR_ENABLE; |
| |
| device_config->vp_config[i].tti.btimer_val = |
| (VXGE_TTI_BTIMER_VAL * 1000) / 272; |
| |
| device_config->vp_config[i].tti.timer_ac_en = |
| VXGE_HW_TIM_TIMER_AC_ENABLE; |
| |
| /* For msi-x with napi (each vector has a handler of its own) - |
| * Set CI to OFF for all vpaths |
| */ |
| device_config->vp_config[i].tti.timer_ci_en = |
| VXGE_HW_TIM_TIMER_CI_DISABLE; |
| |
| device_config->vp_config[i].tti.timer_ri_en = |
| VXGE_HW_TIM_TIMER_RI_DISABLE; |
| |
| device_config->vp_config[i].tti.util_sel = |
| VXGE_HW_TIM_UTIL_SEL_LEGACY_TX_NET_UTIL; |
| |
| device_config->vp_config[i].tti.ltimer_val = |
| (VXGE_TTI_LTIMER_VAL * 1000) / 272; |
| |
| device_config->vp_config[i].tti.rtimer_val = |
| (VXGE_TTI_RTIMER_VAL * 1000) / 272; |
| |
| device_config->vp_config[i].tti.urange_a = TTI_TX_URANGE_A; |
| device_config->vp_config[i].tti.urange_b = TTI_TX_URANGE_B; |
| device_config->vp_config[i].tti.urange_c = TTI_TX_URANGE_C; |
| device_config->vp_config[i].tti.uec_a = TTI_TX_UFC_A; |
| device_config->vp_config[i].tti.uec_b = TTI_TX_UFC_B; |
| device_config->vp_config[i].tti.uec_c = TTI_TX_UFC_C; |
| device_config->vp_config[i].tti.uec_d = TTI_TX_UFC_D; |
| |
| /* Configure Rx rings */ |
| device_config->vp_config[i].ring.enable = |
| VXGE_HW_RING_ENABLE; |
| |
| device_config->vp_config[i].ring.ring_blocks = |
| VXGE_HW_DEF_RING_BLOCKS; |
| |
| device_config->vp_config[i].ring.buffer_mode = |
| VXGE_HW_RING_RXD_BUFFER_MODE_1; |
| |
| device_config->vp_config[i].ring.rxds_limit = |
| VXGE_HW_DEF_RING_RXDS_LIMIT; |
| |
| device_config->vp_config[i].ring.scatter_mode = |
| VXGE_HW_RING_SCATTER_MODE_A; |
| |
| /* Configure rti properties */ |
| device_config->vp_config[i].rti.intr_enable = |
| VXGE_HW_TIM_INTR_ENABLE; |
| |
| device_config->vp_config[i].rti.btimer_val = |
| (VXGE_RTI_BTIMER_VAL * 1000)/272; |
| |
| device_config->vp_config[i].rti.timer_ac_en = |
| VXGE_HW_TIM_TIMER_AC_ENABLE; |
| |
| device_config->vp_config[i].rti.timer_ci_en = |
| VXGE_HW_TIM_TIMER_CI_DISABLE; |
| |
| device_config->vp_config[i].rti.timer_ri_en = |
| VXGE_HW_TIM_TIMER_RI_DISABLE; |
| |
| device_config->vp_config[i].rti.util_sel = |
| VXGE_HW_TIM_UTIL_SEL_LEGACY_RX_NET_UTIL; |
| |
| device_config->vp_config[i].rti.urange_a = |
| RTI_RX_URANGE_A; |
| device_config->vp_config[i].rti.urange_b = |
| RTI_RX_URANGE_B; |
| device_config->vp_config[i].rti.urange_c = |
| RTI_RX_URANGE_C; |
| device_config->vp_config[i].rti.uec_a = RTI_RX_UFC_A; |
| device_config->vp_config[i].rti.uec_b = RTI_RX_UFC_B; |
| device_config->vp_config[i].rti.uec_c = RTI_RX_UFC_C; |
| device_config->vp_config[i].rti.uec_d = RTI_RX_UFC_D; |
| |
| device_config->vp_config[i].rti.rtimer_val = |
| (VXGE_RTI_RTIMER_VAL * 1000) / 272; |
| |
| device_config->vp_config[i].rti.ltimer_val = |
| (VXGE_RTI_LTIMER_VAL * 1000) / 272; |
| |
| device_config->vp_config[i].rpa_strip_vlan_tag = |
| vlan_tag_strip; |
| } |
| |
| driver_config->vpath_per_dev = temp; |
| return no_of_vpaths; |
| } |
| |
| /* initialize device configuratrions */ |
| static void vxge_device_config_init(struct vxge_hw_device_config *device_config, |
| int *intr_type) |
| { |
| /* Used for CQRQ/SRQ. */ |
| device_config->dma_blockpool_initial = |
| VXGE_HW_INITIAL_DMA_BLOCK_POOL_SIZE; |
| |
| device_config->dma_blockpool_max = |
| VXGE_HW_MAX_DMA_BLOCK_POOL_SIZE; |
| |
| if (max_mac_vpath > VXGE_MAX_MAC_ADDR_COUNT) |
| max_mac_vpath = VXGE_MAX_MAC_ADDR_COUNT; |
| |
| #ifndef CONFIG_PCI_MSI |
| vxge_debug_init(VXGE_ERR, |
| "%s: This Kernel does not support " |
| "MSI-X. Defaulting to INTA", VXGE_DRIVER_NAME); |
| *intr_type = INTA; |
| #endif |
| |
| /* Configure whether MSI-X or IRQL. */ |
| switch (*intr_type) { |
| case INTA: |
| device_config->intr_mode = VXGE_HW_INTR_MODE_IRQLINE; |
| break; |
| |
| case MSI_X: |
| device_config->intr_mode = VXGE_HW_INTR_MODE_MSIX_ONE_SHOT; |
| break; |
| } |
| |
| /* Timer period between device poll */ |
| device_config->device_poll_millis = VXGE_TIMER_DELAY; |
| |
| /* Configure mac based steering. */ |
| device_config->rts_mac_en = addr_learn_en; |
| |
| /* Configure Vpaths */ |
| device_config->rth_it_type = VXGE_HW_RTH_IT_TYPE_MULTI_IT; |
| |
| vxge_debug_ll_config(VXGE_TRACE, "%s : Device Config Params ", |
| __func__); |
| vxge_debug_ll_config(VXGE_TRACE, "intr_mode : %d", |
| device_config->intr_mode); |
| vxge_debug_ll_config(VXGE_TRACE, "device_poll_millis : %d", |
| device_config->device_poll_millis); |
| vxge_debug_ll_config(VXGE_TRACE, "rth_en : %d", |
| device_config->rth_en); |
| vxge_debug_ll_config(VXGE_TRACE, "rth_it_type : %d", |
| device_config->rth_it_type); |
| } |
| |
| static void vxge_print_parm(struct vxgedev *vdev, u64 vpath_mask) |
| { |
| int i; |
| |
| vxge_debug_init(VXGE_TRACE, |
| "%s: %d Vpath(s) opened", |
| vdev->ndev->name, vdev->no_of_vpath); |
| |
| switch (vdev->config.intr_type) { |
| case INTA: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Interrupt type INTA", vdev->ndev->name); |
| break; |
| |
| case MSI_X: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Interrupt type MSI-X", vdev->ndev->name); |
| break; |
| } |
| |
| if (vdev->config.rth_steering) { |
| vxge_debug_init(VXGE_TRACE, |
| "%s: RTH steering enabled for TCP_IPV4", |
| vdev->ndev->name); |
| } else { |
| vxge_debug_init(VXGE_TRACE, |
| "%s: RTH steering disabled", vdev->ndev->name); |
| } |
| |
| switch (vdev->config.tx_steering_type) { |
| case NO_STEERING: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Tx steering disabled", vdev->ndev->name); |
| break; |
| case TX_PRIORITY_STEERING: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Unsupported tx steering option", |
| vdev->ndev->name); |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Tx steering disabled", vdev->ndev->name); |
| vdev->config.tx_steering_type = 0; |
| break; |
| case TX_VLAN_STEERING: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Unsupported tx steering option", |
| vdev->ndev->name); |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Tx steering disabled", vdev->ndev->name); |
| vdev->config.tx_steering_type = 0; |
| break; |
| case TX_MULTIQ_STEERING: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Tx multiqueue steering enabled", |
| vdev->ndev->name); |
| break; |
| case TX_PORT_STEERING: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Tx port steering enabled", |
| vdev->ndev->name); |
| break; |
| default: |
| vxge_debug_init(VXGE_ERR, |
| "%s: Unsupported tx steering type", |
| vdev->ndev->name); |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Tx steering disabled", vdev->ndev->name); |
| vdev->config.tx_steering_type = 0; |
| } |
| |
| if (vdev->config.addr_learn_en) |
| vxge_debug_init(VXGE_TRACE, |
| "%s: MAC Address learning enabled", vdev->ndev->name); |
| |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) { |
| if (!vxge_bVALn(vpath_mask, i, 1)) |
| continue; |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: MTU size - %d", vdev->ndev->name, |
| ((vdev->devh))-> |
| config.vp_config[i].mtu); |
| vxge_debug_init(VXGE_TRACE, |
| "%s: VLAN tag stripping %s", vdev->ndev->name, |
| ((vdev->devh))-> |
| config.vp_config[i].rpa_strip_vlan_tag |
| ? "Enabled" : "Disabled"); |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: Max frags : %d", vdev->ndev->name, |
| ((vdev->devh))-> |
| config.vp_config[i].fifo.max_frags); |
| break; |
| } |
| } |
| |
| #ifdef CONFIG_PM |
| /** |
| * vxge_pm_suspend - vxge power management suspend entry point |
| * |
| */ |
| static int vxge_pm_suspend(struct pci_dev *pdev, pm_message_t state) |
| { |
| return -ENOSYS; |
| } |
| /** |
| * vxge_pm_resume - vxge power management resume entry point |
| * |
| */ |
| static int vxge_pm_resume(struct pci_dev *pdev) |
| { |
| return -ENOSYS; |
| } |
| |
| #endif |
| |
| /** |
| * vxge_io_error_detected - called when PCI error is detected |
| * @pdev: Pointer to PCI device |
| * @state: The current pci connection state |
| * |
| * This function is called after a PCI bus error affecting |
| * this device has been detected. |
| */ |
| static pci_ers_result_t vxge_io_error_detected(struct pci_dev *pdev, |
| pci_channel_state_t state) |
| { |
| struct __vxge_hw_device *hldev = pci_get_drvdata(pdev); |
| struct net_device *netdev = hldev->ndev; |
| |
| netif_device_detach(netdev); |
| |
| if (state == pci_channel_io_perm_failure) |
| return PCI_ERS_RESULT_DISCONNECT; |
| |
| if (netif_running(netdev)) { |
| /* Bring down the card, while avoiding PCI I/O */ |
| do_vxge_close(netdev, 0); |
| } |
| |
| pci_disable_device(pdev); |
| |
| return PCI_ERS_RESULT_NEED_RESET; |
| } |
| |
| /** |
| * vxge_io_slot_reset - called after the pci bus has been reset. |
| * @pdev: Pointer to PCI device |
| * |
| * Restart the card from scratch, as if from a cold-boot. |
| * At this point, the card has exprienced a hard reset, |
| * followed by fixups by BIOS, and has its config space |
| * set up identically to what it was at cold boot. |
| */ |
| static pci_ers_result_t vxge_io_slot_reset(struct pci_dev *pdev) |
| { |
| struct __vxge_hw_device *hldev = pci_get_drvdata(pdev); |
| struct net_device *netdev = hldev->ndev; |
| |
| struct vxgedev *vdev = netdev_priv(netdev); |
| |
| if (pci_enable_device(pdev)) { |
| netdev_err(netdev, "Cannot re-enable device after reset\n"); |
| return PCI_ERS_RESULT_DISCONNECT; |
| } |
| |
| pci_set_master(pdev); |
| do_vxge_reset(vdev, VXGE_LL_FULL_RESET); |
| |
| return PCI_ERS_RESULT_RECOVERED; |
| } |
| |
| /** |
| * vxge_io_resume - called when traffic can start flowing again. |
| * @pdev: Pointer to PCI device |
| * |
| * This callback is called when the error recovery driver tells |
| * us that its OK to resume normal operation. |
| */ |
| static void vxge_io_resume(struct pci_dev *pdev) |
| { |
| struct __vxge_hw_device *hldev = pci_get_drvdata(pdev); |
| struct net_device *netdev = hldev->ndev; |
| |
| if (netif_running(netdev)) { |
| if (vxge_open(netdev)) { |
| netdev_err(netdev, |
| "Can't bring device back up after reset\n"); |
| return; |
| } |
| } |
| |
| netif_device_attach(netdev); |
| } |
| |
| static inline u32 vxge_get_num_vfs(u64 function_mode) |
| { |
| u32 num_functions = 0; |
| |
| switch (function_mode) { |
| case VXGE_HW_FUNCTION_MODE_MULTI_FUNCTION: |
| case VXGE_HW_FUNCTION_MODE_SRIOV_8: |
| num_functions = 8; |
| break; |
| case VXGE_HW_FUNCTION_MODE_SINGLE_FUNCTION: |
| num_functions = 1; |
| break; |
| case VXGE_HW_FUNCTION_MODE_SRIOV: |
| case VXGE_HW_FUNCTION_MODE_MRIOV: |
| case VXGE_HW_FUNCTION_MODE_MULTI_FUNCTION_17: |
| num_functions = 17; |
| break; |
| case VXGE_HW_FUNCTION_MODE_SRIOV_4: |
| num_functions = 4; |
| break; |
| case VXGE_HW_FUNCTION_MODE_MULTI_FUNCTION_2: |
| num_functions = 2; |
| break; |
| case VXGE_HW_FUNCTION_MODE_MRIOV_8: |
| num_functions = 8; /* TODO */ |
| break; |
| } |
| return num_functions; |
| } |
| |
| int vxge_fw_upgrade(struct vxgedev *vdev, char *fw_name, int override) |
| { |
| struct __vxge_hw_device *hldev = vdev->devh; |
| u32 maj, min, bld, cmaj, cmin, cbld; |
| enum vxge_hw_status status; |
| const struct firmware *fw; |
| int ret; |
| |
| ret = request_firmware(&fw, fw_name, &vdev->pdev->dev); |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, "%s: Firmware file '%s' not found", |
| VXGE_DRIVER_NAME, fw_name); |
| goto out; |
| } |
| |
| /* Load the new firmware onto the adapter */ |
| status = vxge_update_fw_image(hldev, fw->data, fw->size); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: FW image download to adapter failed '%s'.", |
| VXGE_DRIVER_NAME, fw_name); |
| ret = -EIO; |
| goto out; |
| } |
| |
| /* Read the version of the new firmware */ |
| status = vxge_hw_upgrade_read_version(hldev, &maj, &min, &bld); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: Upgrade read version failed '%s'.", |
| VXGE_DRIVER_NAME, fw_name); |
| ret = -EIO; |
| goto out; |
| } |
| |
| cmaj = vdev->config.device_hw_info.fw_version.major; |
| cmin = vdev->config.device_hw_info.fw_version.minor; |
| cbld = vdev->config.device_hw_info.fw_version.build; |
| /* It's possible the version in /lib/firmware is not the latest version. |
| * If so, we could get into a loop of trying to upgrade to the latest |
| * and flashing the older version. |
| */ |
| if (VXGE_FW_VER(maj, min, bld) == VXGE_FW_VER(cmaj, cmin, cbld) && |
| !override) { |
| ret = -EINVAL; |
| goto out; |
| } |
| |
| printk(KERN_NOTICE "Upgrade to firmware version %d.%d.%d commencing\n", |
| maj, min, bld); |
| |
| /* Flash the adapter with the new firmware */ |
| status = vxge_hw_flash_fw(hldev); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, "%s: Upgrade commit failed '%s'.", |
| VXGE_DRIVER_NAME, fw_name); |
| ret = -EIO; |
| goto out; |
| } |
| |
| printk(KERN_NOTICE "Upgrade of firmware successful! Adapter must be " |
| "hard reset before using, thus requiring a system reboot or a " |
| "hotplug event.\n"); |
| |
| out: |
| release_firmware(fw); |
| return ret; |
| } |
| |
| static int vxge_probe_fw_update(struct vxgedev *vdev) |
| { |
| u32 maj, min, bld; |
| int ret, gpxe = 0; |
| char *fw_name; |
| |
| maj = vdev->config.device_hw_info.fw_version.major; |
| min = vdev->config.device_hw_info.fw_version.minor; |
| bld = vdev->config.device_hw_info.fw_version.build; |
| |
| if (VXGE_FW_VER(maj, min, bld) == VXGE_CERT_FW_VER) |
| return 0; |
| |
| /* Ignore the build number when determining if the current firmware is |
| * "too new" to load the driver |
| */ |
| if (VXGE_FW_VER(maj, min, 0) > VXGE_CERT_FW_VER) { |
| vxge_debug_init(VXGE_ERR, "%s: Firmware newer than last known " |
| "version, unable to load driver\n", |
| VXGE_DRIVER_NAME); |
| return -EINVAL; |
| } |
| |
| /* Firmware 1.4.4 and older cannot be upgraded, and is too ancient to |
| * work with this driver. |
| */ |
| if (VXGE_FW_VER(maj, min, bld) <= VXGE_FW_DEAD_VER) { |
| vxge_debug_init(VXGE_ERR, "%s: Firmware %d.%d.%d cannot be " |
| "upgraded\n", VXGE_DRIVER_NAME, maj, min, bld); |
| return -EINVAL; |
| } |
| |
| /* If file not specified, determine gPXE or not */ |
| if (VXGE_FW_VER(maj, min, bld) >= VXGE_EPROM_FW_VER) { |
| int i; |
| for (i = 0; i < VXGE_HW_MAX_ROM_IMAGES; i++) |
| if (vdev->devh->eprom_versions[i]) { |
| gpxe = 1; |
| break; |
| } |
| } |
| if (gpxe) |
| fw_name = "vxge/X3fw-pxe.ncf"; |
| else |
| fw_name = "vxge/X3fw.ncf"; |
| |
| ret = vxge_fw_upgrade(vdev, fw_name, 0); |
| /* -EINVAL and -ENOENT are not fatal errors for flashing firmware on |
| * probe, so ignore them |
| */ |
| if (ret != -EINVAL && ret != -ENOENT) |
| return -EIO; |
| else |
| ret = 0; |
| |
| if (VXGE_FW_VER(VXGE_CERT_FW_VER_MAJOR, VXGE_CERT_FW_VER_MINOR, 0) > |
| VXGE_FW_VER(maj, min, 0)) { |
| vxge_debug_init(VXGE_ERR, "%s: Firmware %d.%d.%d is too old to" |
| " be used with this driver.", |
| VXGE_DRIVER_NAME, maj, min, bld); |
| return -EINVAL; |
| } |
| |
| return ret; |
| } |
| |
| static int is_sriov_initialized(struct pci_dev *pdev) |
| { |
| int pos; |
| u16 ctrl; |
| |
| pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_SRIOV); |
| if (pos) { |
| pci_read_config_word(pdev, pos + PCI_SRIOV_CTRL, &ctrl); |
| if (ctrl & PCI_SRIOV_CTRL_VFE) |
| return 1; |
| } |
| return 0; |
| } |
| |
| static const struct vxge_hw_uld_cbs vxge_callbacks = { |
| .link_up = vxge_callback_link_up, |
| .link_down = vxge_callback_link_down, |
| .crit_err = vxge_callback_crit_err, |
| }; |
| |
| /** |
| * vxge_probe |
| * @pdev : structure containing the PCI related information of the device. |
| * @pre: List of PCI devices supported by the driver listed in vxge_id_table. |
| * Description: |
| * This function is called when a new PCI device gets detected and initializes |
| * it. |
| * Return value: |
| * returns 0 on success and negative on failure. |
| * |
| */ |
| static int |
| vxge_probe(struct pci_dev *pdev, const struct pci_device_id *pre) |
| { |
| struct __vxge_hw_device *hldev; |
| enum vxge_hw_status status; |
| int ret; |
| int high_dma = 0; |
| u64 vpath_mask = 0; |
| struct vxgedev *vdev; |
| struct vxge_config *ll_config = NULL; |
| struct vxge_hw_device_config *device_config = NULL; |
| struct vxge_hw_device_attr attr; |
| int i, j, no_of_vpath = 0, max_vpath_supported = 0; |
| u8 *macaddr; |
| struct vxge_mac_addrs *entry; |
| static int bus = -1, device = -1; |
| u32 host_type; |
| u8 new_device = 0; |
| enum vxge_hw_status is_privileged; |
| u32 function_mode; |
| u32 num_vfs = 0; |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s:%d", __func__, __LINE__); |
| attr.pdev = pdev; |
| |
| /* In SRIOV-17 mode, functions of the same adapter |
| * can be deployed on different buses |
| */ |
| if (((bus != pdev->bus->number) || (device != PCI_SLOT(pdev->devfn))) && |
| !pdev->is_virtfn) |
| new_device = 1; |
| |
| bus = pdev->bus->number; |
| device = PCI_SLOT(pdev->devfn); |
| |
| if (new_device) { |
| if (driver_config->config_dev_cnt && |
| (driver_config->config_dev_cnt != |
| driver_config->total_dev_cnt)) |
| vxge_debug_init(VXGE_ERR, |
| "%s: Configured %d of %d devices", |
| VXGE_DRIVER_NAME, |
| driver_config->config_dev_cnt, |
| driver_config->total_dev_cnt); |
| driver_config->config_dev_cnt = 0; |
| driver_config->total_dev_cnt = 0; |
| } |
| |
| /* Now making the CPU based no of vpath calculation |
| * applicable for individual functions as well. |
| */ |
| driver_config->g_no_cpus = 0; |
| driver_config->vpath_per_dev = max_config_vpath; |
| |
| driver_config->total_dev_cnt++; |
| if (++driver_config->config_dev_cnt > max_config_dev) { |
| ret = 0; |
| goto _exit0; |
| } |
| |
| device_config = kzalloc(sizeof(struct vxge_hw_device_config), |
| GFP_KERNEL); |
| if (!device_config) { |
| ret = -ENOMEM; |
| vxge_debug_init(VXGE_ERR, |
| "device_config : malloc failed %s %d", |
| __FILE__, __LINE__); |
| goto _exit0; |
| } |
| |
| ll_config = kzalloc(sizeof(struct vxge_config), GFP_KERNEL); |
| if (!ll_config) { |
| ret = -ENOMEM; |
| vxge_debug_init(VXGE_ERR, |
| "device_config : malloc failed %s %d", |
| __FILE__, __LINE__); |
| goto _exit0; |
| } |
| ll_config->tx_steering_type = TX_MULTIQ_STEERING; |
| ll_config->intr_type = MSI_X; |
| ll_config->napi_weight = NEW_NAPI_WEIGHT; |
| ll_config->rth_steering = RTH_STEERING; |
| |
| /* get the default configuration parameters */ |
| vxge_hw_device_config_default_get(device_config); |
| |
| /* initialize configuration parameters */ |
| vxge_device_config_init(device_config, &ll_config->intr_type); |
| |
| ret = pci_enable_device(pdev); |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, |
| "%s : can not enable PCI device", __func__); |
| goto _exit0; |
| } |
| |
| if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s : using 64bit DMA", __func__); |
| |
| high_dma = 1; |
| |
| if (pci_set_consistent_dma_mask(pdev, |
| DMA_BIT_MASK(64))) { |
| vxge_debug_init(VXGE_ERR, |
| "%s : unable to obtain 64bit DMA for " |
| "consistent allocations", __func__); |
| ret = -ENOMEM; |
| goto _exit1; |
| } |
| } else if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(32))) { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s : using 32bit DMA", __func__); |
| } else { |
| ret = -ENOMEM; |
| goto _exit1; |
| } |
| |
| ret = pci_request_region(pdev, 0, VXGE_DRIVER_NAME); |
| if (ret) { |
| vxge_debug_init(VXGE_ERR, |
| "%s : request regions failed", __func__); |
| goto _exit1; |
| } |
| |
| pci_set_master(pdev); |
| |
| attr.bar0 = pci_ioremap_bar(pdev, 0); |
| if (!attr.bar0) { |
| vxge_debug_init(VXGE_ERR, |
| "%s : cannot remap io memory bar0", __func__); |
| ret = -ENODEV; |
| goto _exit2; |
| } |
| vxge_debug_ll_config(VXGE_TRACE, |
| "pci ioremap bar0: %p:0x%llx", |
| attr.bar0, |
| (unsigned long long)pci_resource_start(pdev, 0)); |
| |
| status = vxge_hw_device_hw_info_get(attr.bar0, |
| &ll_config->device_hw_info); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: Reading of hardware info failed." |
| "Please try upgrading the firmware.", VXGE_DRIVER_NAME); |
| ret = -EINVAL; |
| goto _exit3; |
| } |
| |
| vpath_mask = ll_config->device_hw_info.vpath_mask; |
| if (vpath_mask == 0) { |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s: No vpaths available in device", VXGE_DRIVER_NAME); |
| ret = -EINVAL; |
| goto _exit3; |
| } |
| |
| vxge_debug_ll_config(VXGE_TRACE, |
| "%s:%d Vpath mask = %llx", __func__, __LINE__, |
| (unsigned long long)vpath_mask); |
| |
| function_mode = ll_config->device_hw_info.function_mode; |
| host_type = ll_config->device_hw_info.host_type; |
| is_privileged = __vxge_hw_device_is_privilaged(host_type, |
| ll_config->device_hw_info.func_id); |
| |
| /* Check how many vpaths are available */ |
| for (i = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) { |
| if (!((vpath_mask) & vxge_mBIT(i))) |
| continue; |
| max_vpath_supported++; |
| } |
| |
| if (new_device) |
| num_vfs = vxge_get_num_vfs(function_mode) - 1; |
| |
| /* Enable SRIOV mode, if firmware has SRIOV support and if it is a PF */ |
| if (is_sriov(function_mode) && !is_sriov_initialized(pdev) && |
| (ll_config->intr_type != INTA)) { |
| ret = pci_enable_sriov(pdev, num_vfs); |
| if (ret) |
| vxge_debug_ll_config(VXGE_ERR, |
| "Failed in enabling SRIOV mode: %d\n", ret); |
| /* No need to fail out, as an error here is non-fatal */ |
| } |
| |
| /* |
| * Configure vpaths and get driver configured number of vpaths |
| * which is less than or equal to the maximum vpaths per function. |
| */ |
| no_of_vpath = vxge_config_vpaths(device_config, vpath_mask, ll_config); |
| if (!no_of_vpath) { |
| vxge_debug_ll_config(VXGE_ERR, |
| "%s: No more vpaths to configure", VXGE_DRIVER_NAME); |
| ret = 0; |
| goto _exit3; |
| } |
| |
| /* Setting driver callbacks */ |
| attr.uld_callbacks = &vxge_callbacks; |
| |
| status = vxge_hw_device_initialize(&hldev, &attr, device_config); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, |
| "Failed to initialize device (%d)", status); |
| ret = -EINVAL; |
| goto _exit3; |
| } |
| |
| if (VXGE_FW_VER(ll_config->device_hw_info.fw_version.major, |
| ll_config->device_hw_info.fw_version.minor, |
| ll_config->device_hw_info.fw_version.build) >= |
| VXGE_EPROM_FW_VER) { |
| struct eprom_image img[VXGE_HW_MAX_ROM_IMAGES]; |
| |
| status = vxge_hw_vpath_eprom_img_ver_get(hldev, img); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, "%s: Reading of EPROM failed", |
| VXGE_DRIVER_NAME); |
| /* This is a non-fatal error, continue */ |
| } |
| |
| for (i = 0; i < VXGE_HW_MAX_ROM_IMAGES; i++) { |
| hldev->eprom_versions[i] = img[i].version; |
| if (!img[i].is_valid) |
| break; |
| vxge_debug_init(VXGE_TRACE, "%s: EPROM %d, version " |
| "%d.%d.%d.%d", VXGE_DRIVER_NAME, i, |
| VXGE_EPROM_IMG_MAJOR(img[i].version), |
| VXGE_EPROM_IMG_MINOR(img[i].version), |
| VXGE_EPROM_IMG_FIX(img[i].version), |
| VXGE_EPROM_IMG_BUILD(img[i].version)); |
| } |
| } |
| |
| /* if FCS stripping is not disabled in MAC fail driver load */ |
| status = vxge_hw_vpath_strip_fcs_check(hldev, vpath_mask); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, "%s: FCS stripping is enabled in MAC" |
| " failing driver load", VXGE_DRIVER_NAME); |
| ret = -EINVAL; |
| goto _exit4; |
| } |
| |
| /* Always enable HWTS. This will always cause the FCS to be invalid, |
| * due to the fact that HWTS is using the FCS as the location of the |
| * timestamp. The HW FCS checking will still correctly determine if |
| * there is a valid checksum, and the FCS is being removed by the driver |
| * anyway. So no fucntionality is being lost. Since it is always |
| * enabled, we now simply use the ioctl call to set whether or not the |
| * driver should be paying attention to the HWTS. |
| */ |
| if (is_privileged == VXGE_HW_OK) { |
| status = vxge_timestamp_config(hldev); |
| if (status != VXGE_HW_OK) { |
| vxge_debug_init(VXGE_ERR, "%s: HWTS enable failed", |
| VXGE_DRIVER_NAME); |
| ret = -EFAULT; |
| goto _exit4; |
| } |
| } |
| |
| vxge_hw_device_debug_set(hldev, VXGE_ERR, VXGE_COMPONENT_LL); |
| |
| /* set private device info */ |
| pci_set_drvdata(pdev, hldev); |
| |
| ll_config->fifo_indicate_max_pkts = VXGE_FIFO_INDICATE_MAX_PKTS; |
| ll_config->addr_learn_en = addr_learn_en; |
| ll_config->rth_algorithm = RTH_ALG_JENKINS; |
| ll_config->rth_hash_type_tcpipv4 = 1; |
| ll_config->rth_hash_type_ipv4 = 0; |
| ll_config->rth_hash_type_tcpipv6 = 0; |
| ll_config->rth_hash_type_ipv6 = 0; |
| ll_config->rth_hash_type_tcpipv6ex = 0; |
| ll_config->rth_hash_type_ipv6ex = 0; |
| ll_config->rth_bkt_sz = RTH_BUCKET_SIZE; |
| ll_config->tx_pause_enable = VXGE_PAUSE_CTRL_ENABLE; |
| ll_config->rx_pause_enable = VXGE_PAUSE_CTRL_ENABLE; |
| |
| ret = vxge_device_register(hldev, ll_config, high_dma, no_of_vpath, |
| &vdev); |
| if (ret) { |
| ret = -EINVAL; |
| goto _exit4; |
| } |
| |
| ret = vxge_probe_fw_update(vdev); |
| if (ret) |
| goto _exit5; |
| |
| vxge_hw_device_debug_set(hldev, VXGE_TRACE, VXGE_COMPONENT_LL); |
| VXGE_COPY_DEBUG_INFO_TO_LL(vdev, vxge_hw_device_error_level_get(hldev), |
| vxge_hw_device_trace_level_get(hldev)); |
| |
| /* set private HW device info */ |
| vdev->mtu = VXGE_HW_DEFAULT_MTU; |
| vdev->bar0 = attr.bar0; |
| vdev->max_vpath_supported = max_vpath_supported; |
| vdev->no_of_vpath = no_of_vpath; |
| |
| /* Virtual Path count */ |
| for (i = 0, j = 0; i < VXGE_HW_MAX_VIRTUAL_PATHS; i++) { |
| if (!vxge_bVALn(vpath_mask, i, 1)) |
| continue; |
| if (j >= vdev->no_of_vpath) |
| break; |
| |
| vdev->vpaths[j].is_configured = 1; |
| vdev->vpaths[j].device_id = i; |
| vdev->vpaths[j].ring.driver_id = j; |
| vdev->vpaths[j].vdev = vdev; |
| vdev->vpaths[j].max_mac_addr_cnt = max_mac_vpath; |
| memcpy((u8 *)vdev->vpaths[j].macaddr, |
| ll_config->device_hw_info.mac_addrs[i], |
| ETH_ALEN); |
| |
| /* Initialize the mac address list header */ |
| INIT_LIST_HEAD(&vdev->vpaths[j].mac_addr_list); |
| |
| vdev->vpaths[j].mac_addr_cnt = 0; |
| vdev->vpaths[j].mcast_addr_cnt = 0; |
| j++; |
| } |
| vdev->exec_mode = VXGE_EXEC_MODE_DISABLE; |
| vdev->max_config_port = max_config_port; |
| |
| vdev->vlan_tag_strip = vlan_tag_strip; |
| |
| /* map the hashing selector table to the configured vpaths */ |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| vdev->vpath_selector[i] = vpath_selector[i]; |
| |
| macaddr = (u8 *)vdev->vpaths[0].macaddr; |
| |
| ll_config->device_hw_info.serial_number[VXGE_HW_INFO_LEN - 1] = '\0'; |
| ll_config->device_hw_info.product_desc[VXGE_HW_INFO_LEN - 1] = '\0'; |
| ll_config->device_hw_info.part_number[VXGE_HW_INFO_LEN - 1] = '\0'; |
| |
| vxge_debug_init(VXGE_TRACE, "%s: SERIAL NUMBER: %s", |
| vdev->ndev->name, ll_config->device_hw_info.serial_number); |
| |
| vxge_debug_init(VXGE_TRACE, "%s: PART NUMBER: %s", |
| vdev->ndev->name, ll_config->device_hw_info.part_number); |
| |
| vxge_debug_init(VXGE_TRACE, "%s: Neterion %s Server Adapter", |
| vdev->ndev->name, ll_config->device_hw_info.product_desc); |
| |
| vxge_debug_init(VXGE_TRACE, "%s: MAC ADDR: %pM", |
| vdev->ndev->name, macaddr); |
| |
| vxge_debug_init(VXGE_TRACE, "%s: Link Width x%d", |
| vdev->ndev->name, vxge_hw_device_link_width_get(hldev)); |
| |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Firmware version : %s Date : %s", vdev->ndev->name, |
| ll_config->device_hw_info.fw_version.version, |
| ll_config->device_hw_info.fw_date.date); |
| |
| if (new_device) { |
| switch (ll_config->device_hw_info.function_mode) { |
| case VXGE_HW_FUNCTION_MODE_SINGLE_FUNCTION: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Single Function Mode Enabled", vdev->ndev->name); |
| break; |
| case VXGE_HW_FUNCTION_MODE_MULTI_FUNCTION: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Multi Function Mode Enabled", vdev->ndev->name); |
| break; |
| case VXGE_HW_FUNCTION_MODE_SRIOV: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Single Root IOV Mode Enabled", vdev->ndev->name); |
| break; |
| case VXGE_HW_FUNCTION_MODE_MRIOV: |
| vxge_debug_init(VXGE_TRACE, |
| "%s: Multi Root IOV Mode Enabled", vdev->ndev->name); |
| break; |
| } |
| } |
| |
| vxge_print_parm(vdev, vpath_mask); |
| |
| /* Store the fw version for ethttool option */ |
| strcpy(vdev->fw_version, ll_config->device_hw_info.fw_version.version); |
| memcpy(vdev->ndev->dev_addr, (u8 *)vdev->vpaths[0].macaddr, ETH_ALEN); |
| |
| /* Copy the station mac address to the list */ |
| for (i = 0; i < vdev->no_of_vpath; i++) { |
| entry = kzalloc(sizeof(struct vxge_mac_addrs), GFP_KERNEL); |
| if (NULL == entry) { |
| vxge_debug_init(VXGE_ERR, |
| "%s: mac_addr_list : memory allocation failed", |
| vdev->ndev->name); |
| ret = -EPERM; |
| goto _exit6; |
| } |
| macaddr = (u8 *)&entry->macaddr; |
| memcpy(macaddr, vdev->ndev->dev_addr, ETH_ALEN); |
| list_add(&entry->item, &vdev->vpaths[i].mac_addr_list); |
| vdev->vpaths[i].mac_addr_cnt = 1; |
| } |
| |
| kfree(device_config); |
| |
| /* |
| * INTA is shared in multi-function mode. This is unlike the INTA |
| * implementation in MR mode, where each VH has its own INTA message. |
| * - INTA is masked (disabled) as long as at least one function sets |
| * its TITAN_MASK_ALL_INT.ALARM bit. |
| * - INTA is unmasked (enabled) when all enabled functions have cleared |
| * their own TITAN_MASK_ALL_INT.ALARM bit. |
| * The TITAN_MASK_ALL_INT ALARM & TRAFFIC bits are cleared on power up. |
| * Though this driver leaves the top level interrupts unmasked while |
| * leaving the required module interrupt bits masked on exit, there |
| * could be a rougue driver around that does not follow this procedure |
| * resulting in a failure to generate interrupts. The following code is |
| * present to prevent such a failure. |
| */ |
| |
| if (ll_config->device_hw_info.function_mode == |
| VXGE_HW_FUNCTION_MODE_MULTI_FUNCTION) |
| if (vdev->config.intr_type == INTA) |
| vxge_hw_device_unmask_all(hldev); |
| |
| vxge_debug_entryexit(VXGE_TRACE, "%s: %s:%d Exiting...", |
| vdev->ndev->name, __func__, __LINE__); |
| |
| vxge_hw_device_debug_set(hldev, VXGE_ERR, VXGE_COMPONENT_LL); |
| VXGE_COPY_DEBUG_INFO_TO_LL(vdev, vxge_hw_device_error_level_get(hldev), |
| vxge_hw_device_trace_level_get(hldev)); |
| |
| kfree(ll_config); |
| return 0; |
| |
| _exit6: |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| vxge_free_mac_add_list(&vdev->vpaths[i]); |
| _exit5: |
| vxge_device_unregister(hldev); |
| _exit4: |
| vxge_hw_device_terminate(hldev); |
| pci_disable_sriov(pdev); |
| _exit3: |
| iounmap(attr.bar0); |
| _exit2: |
| pci_release_region(pdev, 0); |
| _exit1: |
| pci_disable_device(pdev); |
| _exit0: |
| kfree(ll_config); |
| kfree(device_config); |
| driver_config->config_dev_cnt--; |
| driver_config->total_dev_cnt--; |
| return ret; |
| } |
| |
| /** |
| * vxge_rem_nic - Free the PCI device |
| * @pdev: structure containing the PCI related information of the device. |
| * Description: This function is called by the Pci subsystem to release a |
| * PCI device and free up all resource held up by the device. |
| */ |
| static void vxge_remove(struct pci_dev *pdev) |
| { |
| struct __vxge_hw_device *hldev; |
| struct vxgedev *vdev; |
| int i; |
| |
| hldev = pci_get_drvdata(pdev); |
| if (hldev == NULL) |
| return; |
| |
| vdev = netdev_priv(hldev->ndev); |
| |
| vxge_debug_entryexit(vdev->level_trace, "%s:%d", __func__, __LINE__); |
| vxge_debug_init(vdev->level_trace, "%s : removing PCI device...", |
| __func__); |
| |
| for (i = 0; i < vdev->no_of_vpath; i++) |
| vxge_free_mac_add_list(&vdev->vpaths[i]); |
| |
| vxge_device_unregister(hldev); |
| /* Do not call pci_disable_sriov here, as it will break child devices */ |
| vxge_hw_device_terminate(hldev); |
| iounmap(vdev->bar0); |
| pci_release_region(pdev, 0); |
| pci_disable_device(pdev); |
| driver_config->config_dev_cnt--; |
| driver_config->total_dev_cnt--; |
| |
| vxge_debug_init(vdev->level_trace, "%s:%d Device unregistered", |
| __func__, __LINE__); |
| vxge_debug_entryexit(vdev->level_trace, "%s:%d Exiting...", __func__, |
| __LINE__); |
| } |
| |
| static const struct pci_error_handlers vxge_err_handler = { |
| .error_detected = vxge_io_error_detected, |
| .slot_reset = vxge_io_slot_reset, |
| .resume = vxge_io_resume, |
| }; |
| |
| static struct pci_driver vxge_driver = { |
| .name = VXGE_DRIVER_NAME, |
| .id_table = vxge_id_table, |
| .probe = vxge_probe, |
| .remove = vxge_remove, |
| #ifdef CONFIG_PM |
| .suspend = vxge_pm_suspend, |
| .resume = vxge_pm_resume, |
| #endif |
| .err_handler = &vxge_err_handler, |
| }; |
| |
| static int __init |
| vxge_starter(void) |
| { |
| int ret = 0; |
| |
| pr_info("Copyright(c) 2002-2010 Exar Corp.\n"); |
| pr_info("Driver version: %s\n", DRV_VERSION); |
| |
| verify_bandwidth(); |
| |
| driver_config = kzalloc(sizeof(struct vxge_drv_config), GFP_KERNEL); |
| if (!driver_config) |
| return -ENOMEM; |
| |
| ret = pci_register_driver(&vxge_driver); |
| if (ret) { |
| kfree(driver_config); |
| goto err; |
| } |
| |
| if (driver_config->config_dev_cnt && |
| (driver_config->config_dev_cnt != driver_config->total_dev_cnt)) |
| vxge_debug_init(VXGE_ERR, |
| "%s: Configured %d of %d devices", |
| VXGE_DRIVER_NAME, driver_config->config_dev_cnt, |
| driver_config->total_dev_cnt); |
| err: |
| return ret; |
| } |
| |
| static void __exit |
| vxge_closer(void) |
| { |
| pci_unregister_driver(&vxge_driver); |
| kfree(driver_config); |
| } |
| module_init(vxge_starter); |
| module_exit(vxge_closer); |