diff options
Diffstat (limited to 'drivers/net/e1000')
| -rw-r--r-- | drivers/net/e1000/e1000.h | 51 | ||||
| -rw-r--r-- | drivers/net/e1000/e1000_ethtool.c | 167 | ||||
| -rw-r--r-- | drivers/net/e1000/e1000_hw.c | 53 | ||||
| -rw-r--r-- | drivers/net/e1000/e1000_hw.h | 25 | ||||
| -rw-r--r-- | drivers/net/e1000/e1000_main.c | 814 | ||||
| -rw-r--r-- | drivers/net/e1000/e1000_param.c | 14 |
6 files changed, 768 insertions, 356 deletions
diff --git a/drivers/net/e1000/e1000.h b/drivers/net/e1000/e1000.h index e02e9ba2e18b..27c77306193b 100644 --- a/drivers/net/e1000/e1000.h +++ b/drivers/net/e1000/e1000.h @@ -72,10 +72,6 @@ #include <linux/mii.h> #include <linux/ethtool.h> #include <linux/if_vlan.h> -#ifdef CONFIG_E1000_MQ -#include <linux/cpu.h> -#include <linux/smp.h> -#endif #define BAR_0 0 #define BAR_1 1 @@ -87,6 +83,10 @@ struct e1000_adapter; #include "e1000_hw.h" +#ifdef CONFIG_E1000_MQ +#include <linux/cpu.h> +#include <linux/smp.h> +#endif #ifdef DBG #define E1000_DBG(args...) printk(KERN_DEBUG "e1000: " args) @@ -169,6 +169,13 @@ struct e1000_buffer { uint16_t next_to_watch; }; +#ifdef CONFIG_E1000_MQ +struct e1000_queue_stats { + uint64_t packets; + uint64_t bytes; +}; +#endif + struct e1000_ps_page { struct page *ps_page[PS_PAGE_BUFFERS]; }; struct e1000_ps_page_dma { uint64_t ps_page_dma[PS_PAGE_BUFFERS]; }; @@ -191,10 +198,12 @@ struct e1000_tx_ring { spinlock_t tx_lock; uint16_t tdh; uint16_t tdt; - uint64_t pkt; boolean_t last_tx_tso; +#ifdef CONFIG_E1000_MQ + struct e1000_queue_stats tx_stats; +#endif }; struct e1000_rx_ring { @@ -216,9 +225,17 @@ struct e1000_rx_ring { struct e1000_ps_page *ps_page; struct e1000_ps_page_dma *ps_page_dma; + struct sk_buff *rx_skb_top; + struct sk_buff *rx_skb_prev; + + /* cpu for rx queue */ + int cpu; + uint16_t rdh; uint16_t rdt; - uint64_t pkt; +#ifdef CONFIG_E1000_MQ + struct e1000_queue_stats rx_stats; +#endif }; #define E1000_DESC_UNUSED(R) \ @@ -251,6 +268,9 @@ struct e1000_adapter { uint16_t link_speed; uint16_t link_duplex; spinlock_t stats_lock; +#ifdef CONFIG_E1000_NAPI + spinlock_t tx_queue_lock; +#endif atomic_t irq_sem; struct work_struct tx_timeout_task; struct work_struct watchdog_task; @@ -264,6 +284,7 @@ struct e1000_adapter { #ifdef CONFIG_E1000_MQ struct e1000_tx_ring **cpu_tx_ring; /* per-cpu */ #endif + unsigned long tx_queue_len; uint32_t txd_cmd; uint32_t tx_int_delay; uint32_t tx_abs_int_delay; @@ -271,9 +292,11 @@ struct e1000_adapter { uint64_t gotcl_old; uint64_t tpt_old; uint64_t colc_old; + uint32_t tx_timeout_count; uint32_t tx_fifo_head; uint32_t tx_head_addr; uint32_t tx_fifo_size; + uint8_t tx_timeout_factor; atomic_t tx_fifo_stall; boolean_t pcix_82544; boolean_t detect_tx_hung; @@ -281,14 +304,15 @@ struct e1000_adapter { /* RX */ #ifdef CONFIG_E1000_NAPI boolean_t (*clean_rx) (struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring, - int *work_done, int work_to_do); + struct e1000_rx_ring *rx_ring, + int *work_done, int work_to_do); #else boolean_t (*clean_rx) (struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring); #endif void (*alloc_rx_buf) (struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring, + int cleaned_count); struct e1000_rx_ring *rx_ring; /* One per active queue */ #ifdef CONFIG_E1000_NAPI struct net_device *polling_netdev; /* One per active queue */ @@ -296,13 +320,15 @@ struct e1000_adapter { #ifdef CONFIG_E1000_MQ struct net_device **cpu_netdev; /* per-cpu */ struct call_async_data_struct rx_sched_call_data; - int cpu_for_queue[4]; + cpumask_t cpumask; #endif - int num_queues; + int num_tx_queues; + int num_rx_queues; uint64_t hw_csum_err; uint64_t hw_csum_good; uint64_t rx_hdr_split; + uint32_t alloc_rx_buff_failed; uint32_t rx_int_delay; uint32_t rx_abs_int_delay; boolean_t rx_csum; @@ -330,6 +356,7 @@ struct e1000_adapter { struct e1000_rx_ring test_rx_ring; + u32 *config_space; int msg_enable; #ifdef CONFIG_PCI_MSI boolean_t have_msi; diff --git a/drivers/net/e1000/e1000_ethtool.c b/drivers/net/e1000/e1000_ethtool.c index c88f1a3c1b1d..d252297e4db0 100644 --- a/drivers/net/e1000/e1000_ethtool.c +++ b/drivers/net/e1000/e1000_ethtool.c @@ -80,6 +80,7 @@ static const struct e1000_stats e1000_gstrings_stats[] = { { "tx_deferred_ok", E1000_STAT(stats.dc) }, { "tx_single_coll_ok", E1000_STAT(stats.scc) }, { "tx_multi_coll_ok", E1000_STAT(stats.mcc) }, + { "tx_timeout_count", E1000_STAT(tx_timeout_count) }, { "rx_long_length_errors", E1000_STAT(stats.roc) }, { "rx_short_length_errors", E1000_STAT(stats.ruc) }, { "rx_align_errors", E1000_STAT(stats.algnerrc) }, @@ -93,9 +94,20 @@ static const struct e1000_stats e1000_gstrings_stats[] = { { "rx_csum_offload_good", E1000_STAT(hw_csum_good) }, { "rx_csum_offload_errors", E1000_STAT(hw_csum_err) }, { "rx_header_split", E1000_STAT(rx_hdr_split) }, + { "alloc_rx_buff_failed", E1000_STAT(alloc_rx_buff_failed) }, }; -#define E1000_STATS_LEN \ + +#ifdef CONFIG_E1000_MQ +#define E1000_QUEUE_STATS_LEN \ + (((struct e1000_adapter *)netdev->priv)->num_tx_queues + \ + ((struct e1000_adapter *)netdev->priv)->num_rx_queues) \ + * (sizeof(struct e1000_queue_stats) / sizeof(uint64_t)) +#else +#define E1000_QUEUE_STATS_LEN 0 +#endif +#define E1000_GLOBAL_STATS_LEN \ sizeof(e1000_gstrings_stats) / sizeof(struct e1000_stats) +#define E1000_STATS_LEN (E1000_GLOBAL_STATS_LEN + E1000_QUEUE_STATS_LEN) static const char e1000_gstrings_test[][ETH_GSTRING_LEN] = { "Register test (offline)", "Eeprom test (offline)", "Interrupt test (offline)", "Loopback test (offline)", @@ -183,7 +195,15 @@ e1000_set_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - if(ecmd->autoneg == AUTONEG_ENABLE) { + /* When SoL/IDER sessions are active, autoneg/speed/duplex + * cannot be changed */ + if (e1000_check_phy_reset_block(hw)) { + DPRINTK(DRV, ERR, "Cannot change link characteristics " + "when SoL/IDER is active.\n"); + return -EINVAL; + } + + if (ecmd->autoneg == AUTONEG_ENABLE) { hw->autoneg = 1; if(hw->media_type == e1000_media_type_fiber) hw->autoneg_advertised = ADVERTISED_1000baseT_Full | @@ -567,21 +587,21 @@ e1000_get_drvinfo(struct net_device *netdev, strncpy(drvinfo->driver, e1000_driver_name, 32); strncpy(drvinfo->version, e1000_driver_version, 32); - - /* EEPROM image version # is reported as firware version # for + + /* EEPROM image version # is reported as firmware version # for * 8257{1|2|3} controllers */ e1000_read_eeprom(&adapter->hw, 5, 1, &eeprom_data); switch (adapter->hw.mac_type) { case e1000_82571: case e1000_82572: case e1000_82573: - sprintf(firmware_version, "%d.%d-%d", + sprintf(firmware_version, "%d.%d-%d", (eeprom_data & 0xF000) >> 12, (eeprom_data & 0x0FF0) >> 4, eeprom_data & 0x000F); break; default: - sprintf(firmware_version, "n/a"); + sprintf(firmware_version, "N/A"); } strncpy(drvinfo->fw_version, firmware_version, 32); @@ -623,8 +643,8 @@ e1000_set_ringparam(struct net_device *netdev, struct e1000_rx_ring *rxdr, *rx_old, *rx_new; int i, err, tx_ring_size, rx_ring_size; - tx_ring_size = sizeof(struct e1000_tx_ring) * adapter->num_queues; - rx_ring_size = sizeof(struct e1000_rx_ring) * adapter->num_queues; + tx_ring_size = sizeof(struct e1000_tx_ring) * adapter->num_tx_queues; + rx_ring_size = sizeof(struct e1000_rx_ring) * adapter->num_rx_queues; if (netif_running(adapter->netdev)) e1000_down(adapter); @@ -663,10 +683,10 @@ e1000_set_ringparam(struct net_device *netdev, E1000_MAX_TXD : E1000_MAX_82544_TXD)); E1000_ROUNDUP(txdr->count, REQ_TX_DESCRIPTOR_MULTIPLE); - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_tx_queues; i++) txdr[i].count = txdr->count; + for (i = 0; i < adapter->num_rx_queues; i++) rxdr[i].count = rxdr->count; - } if(netif_running(adapter->netdev)) { /* Try to get new resources before deleting old */ @@ -979,18 +999,17 @@ e1000_free_desc_rings(struct e1000_adapter *adapter) } } - if(txdr->desc) { + if (txdr->desc) { pci_free_consistent(pdev, txdr->size, txdr->desc, txdr->dma); txdr->desc = NULL; } - if(rxdr->desc) { + if (rxdr->desc) { pci_free_consistent(pdev, rxdr->size, rxdr->desc, rxdr->dma); rxdr->desc = NULL; } kfree(txdr->buffer_info); txdr->buffer_info = NULL; - kfree(rxdr->buffer_info); rxdr->buffer_info = NULL; @@ -1327,11 +1346,11 @@ e1000_set_phy_loopback(struct e1000_adapter *adapter) static int e1000_setup_loopback_test(struct e1000_adapter *adapter) { - uint32_t rctl; struct e1000_hw *hw = &adapter->hw; + uint32_t rctl; if (hw->media_type == e1000_media_type_fiber || - hw->media_type == e1000_media_type_internal_serdes) { + hw->media_type == e1000_media_type_internal_serdes) { switch (hw->mac_type) { case e1000_82545: case e1000_82546: @@ -1362,25 +1381,25 @@ e1000_setup_loopback_test(struct e1000_adapter *adapter) static void e1000_loopback_cleanup(struct e1000_adapter *adapter) { + struct e1000_hw *hw = &adapter->hw; uint32_t rctl; uint16_t phy_reg; - struct e1000_hw *hw = &adapter->hw; - rctl = E1000_READ_REG(&adapter->hw, RCTL); + rctl = E1000_READ_REG(hw, RCTL); rctl &= ~(E1000_RCTL_LBM_TCVR | E1000_RCTL_LBM_MAC); - E1000_WRITE_REG(&adapter->hw, RCTL, rctl); + E1000_WRITE_REG(hw, RCTL, rctl); switch (hw->mac_type) { case e1000_82571: case e1000_82572: if (hw->media_type == e1000_media_type_fiber || - hw->media_type == e1000_media_type_internal_serdes){ + hw->media_type == e1000_media_type_internal_serdes) { #define E1000_SERDES_LB_OFF 0x400 E1000_WRITE_REG(hw, SCTL, E1000_SERDES_LB_OFF); msec_delay(10); break; } - /* fall thru for Cu adapters */ + /* Fall Through */ case e1000_82545: case e1000_82546: case e1000_82545_rev_3: @@ -1401,7 +1420,7 @@ static void e1000_create_lbtest_frame(struct sk_buff *skb, unsigned int frame_size) { memset(skb->data, 0xFF, frame_size); - frame_size = (frame_size % 2) ? (frame_size - 1) : frame_size; + frame_size &= ~1; memset(&skb->data[frame_size / 2], 0xAA, frame_size / 2 - 1); memset(&skb->data[frame_size / 2 + 10], 0xBE, 1); memset(&skb->data[frame_size / 2 + 12], 0xAF, 1); @@ -1410,7 +1429,7 @@ e1000_create_lbtest_frame(struct sk_buff *skb, unsigned int frame_size) static int e1000_check_lbtest_frame(struct sk_buff *skb, unsigned int frame_size) { - frame_size = (frame_size % 2) ? (frame_size - 1) : frame_size; + frame_size &= ~1; if(*(skb->data + 3) == 0xFF) { if((*(skb->data + frame_size / 2 + 10) == 0xBE) && (*(skb->data + frame_size / 2 + 12) == 0xAF)) { @@ -1488,14 +1507,25 @@ e1000_run_loopback_test(struct e1000_adapter *adapter) static int e1000_loopback_test(struct e1000_adapter *adapter, uint64_t *data) { - if((*data = e1000_setup_desc_rings(adapter))) goto err_loopback; - if((*data = e1000_setup_loopback_test(adapter))) - goto err_loopback_setup; + /* PHY loopback cannot be performed if SoL/IDER + * sessions are active */ + if (e1000_check_phy_reset_block(&adapter->hw)) { + DPRINTK(DRV, ERR, "Cannot do PHY loopback test " + "when SoL/IDER is active.\n"); + *data = 0; + goto out; + } + + if ((*data = e1000_setup_desc_rings(adapter))) + goto out; + if ((*data = e1000_setup_loopback_test(adapter))) + goto err_loopback; *data = e1000_run_loopback_test(adapter); e1000_loopback_cleanup(adapter); -err_loopback_setup: - e1000_free_desc_rings(adapter); + err_loopback: + e1000_free_desc_rings(adapter); +out: return *data; } @@ -1617,6 +1647,7 @@ e1000_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) case E1000_DEV_ID_82546EB_FIBER: case E1000_DEV_ID_82546GB_FIBER: + case E1000_DEV_ID_82571EB_FIBER: /* Wake events only supported on port A for dual fiber */ if(E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) { wol->supported = 0; @@ -1660,6 +1691,7 @@ e1000_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) case E1000_DEV_ID_82546EB_FIBER: case E1000_DEV_ID_82546GB_FIBER: + case E1000_DEV_ID_82571EB_FIBER: /* Wake events only supported on port A for dual fiber */ if(E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) return wol->wolopts ? -EOPNOTSUPP : 0; @@ -1721,21 +1753,21 @@ e1000_phys_id(struct net_device *netdev, uint32_t data) mod_timer(&adapter->blink_timer, jiffies); msleep_interruptible(data * 1000); del_timer_sync(&adapter->blink_timer); - } - else if(adapter->hw.mac_type < e1000_82573) { - E1000_WRITE_REG(&adapter->hw, LEDCTL, (E1000_LEDCTL_LED2_BLINK_RATE | - E1000_LEDCTL_LED0_BLINK | E1000_LEDCTL_LED2_BLINK | - (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED0_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED1_MODE_SHIFT))); + } else if (adapter->hw.mac_type < e1000_82573) { + E1000_WRITE_REG(&adapter->hw, LEDCTL, + (E1000_LEDCTL_LED2_BLINK_RATE | + E1000_LEDCTL_LED0_BLINK | E1000_LEDCTL_LED2_BLINK | + (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED0_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED1_MODE_SHIFT))); msleep_interruptible(data * 1000); - } - else { - E1000_WRITE_REG(&adapter->hw, LEDCTL, (E1000_LEDCTL_LED2_BLINK_RATE | - E1000_LEDCTL_LED1_BLINK | E1000_LEDCTL_LED2_BLINK | - (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED1_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED0_MODE_SHIFT))); + } else { + E1000_WRITE_REG(&adapter->hw, LEDCTL, + (E1000_LEDCTL_LED2_BLINK_RATE | + E1000_LEDCTL_LED1_BLINK | E1000_LEDCTL_LED2_BLINK | + (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED1_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED0_MODE_SHIFT))); msleep_interruptible(data * 1000); } @@ -1768,19 +1800,43 @@ e1000_get_ethtool_stats(struct net_device *netdev, struct ethtool_stats *stats, uint64_t *data) { struct e1000_adapter *adapter = netdev_priv(netdev); +#ifdef CONFIG_E1000_MQ + uint64_t *queue_stat; + int stat_count = sizeof(struct e1000_queue_stats) / sizeof(uint64_t); + int j, k; +#endif int i; e1000_update_stats(adapter); - for(i = 0; i < E1000_STATS_LEN; i++) { - char *p = (char *)adapter+e1000_gstrings_stats[i].stat_offset; - data[i] = (e1000_gstrings_stats[i].sizeof_stat == + for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) { + char *p = (char *)adapter+e1000_gstrings_stats[i].stat_offset; + data[i] = (e1000_gstrings_stats[i].sizeof_stat == sizeof(uint64_t)) ? *(uint64_t *)p : *(uint32_t *)p; } +#ifdef CONFIG_E1000_MQ + for (j = 0; j < adapter->num_tx_queues; j++) { + queue_stat = (uint64_t *)&adapter->tx_ring[j].tx_stats; + for (k = 0; k < stat_count; k++) + data[i + k] = queue_stat[k]; + i += k; + } + for (j = 0; j < adapter->num_rx_queues; j++) { + queue_stat = (uint64_t *)&adapter->rx_ring[j].rx_stats; + for (k = 0; k < stat_count; k++) + data[i + k] = queue_stat[k]; + i += k; + } +#endif +/* BUG_ON(i != E1000_STATS_LEN); */ } static void e1000_get_strings(struct net_device *netdev, uint32_t stringset, uint8_t *data) { +#ifdef CONFIG_E1000_MQ + struct e1000_adapter *adapter = netdev_priv(netdev); +#endif + uint8_t *p = data; int i; switch(stringset) { @@ -1789,11 +1845,26 @@ e1000_get_strings(struct net_device *netdev, uint32_t stringset, uint8_t *data) E1000_TEST_LEN*ETH_GSTRING_LEN); break; case ETH_SS_STATS: - for (i=0; i < E1000_STATS_LEN; i++) { - memcpy(data + i * ETH_GSTRING_LEN, - e1000_gstrings_stats[i].stat_string, - ETH_GSTRING_LEN); + for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) { + memcpy(p, e1000_gstrings_stats[i].stat_string, + ETH_GSTRING_LEN); + p += ETH_GSTRING_LEN; + } +#ifdef CONFIG_E1000_MQ + for (i = 0; i < adapter->num_tx_queues; i++) { + sprintf(p, "tx_queue_%u_packets", i); + p += ETH_GSTRING_LEN; + sprintf(p, "tx_queue_%u_bytes", i); + p += ETH_GSTRING_LEN; } + for (i = 0; i < adapter->num_rx_queues; i++) { + sprintf(p, "rx_queue_%u_packets", i); + p += ETH_GSTRING_LEN; + sprintf(p, "rx_queue_%u_bytes", i); + p += ETH_GSTRING_LEN; + } +#endif +/* BUG_ON(p - data != E1000_STATS_LEN * ETH_GSTRING_LEN); */ break; } } diff --git a/drivers/net/e1000/e1000_hw.c b/drivers/net/e1000/e1000_hw.c index 136fc031e4ad..2437d362ff63 100644 --- a/drivers/net/e1000/e1000_hw.c +++ b/drivers/net/e1000/e1000_hw.c @@ -318,6 +318,8 @@ e1000_set_mac_type(struct e1000_hw *hw) case E1000_DEV_ID_82546GB_FIBER: case E1000_DEV_ID_82546GB_SERDES: case E1000_DEV_ID_82546GB_PCIE: + case E1000_DEV_ID_82546GB_QUAD_COPPER: + case E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3: hw->mac_type = e1000_82546_rev_3; break; case E1000_DEV_ID_82541EI: @@ -639,6 +641,7 @@ e1000_init_hw(struct e1000_hw *hw) uint16_t cmd_mmrbc; uint16_t stat_mmrbc; uint32_t mta_size; + uint32_t ctrl_ext; DEBUGFUNC("e1000_init_hw"); @@ -735,7 +738,6 @@ e1000_init_hw(struct e1000_hw *hw) break; case e1000_82571: case e1000_82572: - ctrl |= (1 << 22); case e1000_82573: ctrl |= E1000_TXDCTL_COUNT_DESC; break; @@ -775,6 +777,15 @@ e1000_init_hw(struct e1000_hw *hw) */ e1000_clear_hw_cntrs(hw); + if (hw->device_id == E1000_DEV_ID_82546GB_QUAD_COPPER || + hw->device_id == E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3) { + ctrl_ext = E1000_READ_REG(hw, CTRL_EXT); + /* Relaxed ordering must be disabled to avoid a parity + * error crash in a PCI slot. */ + ctrl_ext |= E1000_CTRL_EXT_RO_DIS; + E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext); + } + return ret_val; } @@ -838,6 +849,11 @@ e1000_setup_link(struct e1000_hw *hw) DEBUGFUNC("e1000_setup_link"); + /* In the case of the phy reset being blocked, we already have a link. + * We do not have to set it up again. */ + if (e1000_check_phy_reset_block(hw)) + return E1000_SUCCESS; + /* Read and store word 0x0F of the EEPROM. This word contains bits * that determine the hardware's default PAUSE (flow control) mode, * a bit that determines whether the HW defaults to enabling or @@ -1929,14 +1945,19 @@ e1000_phy_force_speed_duplex(struct e1000_hw *hw) void e1000_config_collision_dist(struct e1000_hw *hw) { - uint32_t tctl; + uint32_t tctl, coll_dist; DEBUGFUNC("e1000_config_collision_dist"); + if (hw->mac_type < e1000_82543) + coll_dist = E1000_COLLISION_DISTANCE_82542; + else + coll_dist = E1000_COLLISION_DISTANCE; + tctl = E1000_READ_REG(hw, TCTL); tctl &= ~E1000_TCTL_COLD; - tctl |= E1000_COLLISION_DISTANCE << E1000_COLD_SHIFT; + tctl |= coll_dist << E1000_COLD_SHIFT; E1000_WRITE_REG(hw, TCTL, tctl); E1000_WRITE_FLUSH(hw); @@ -2982,6 +3003,8 @@ e1000_phy_hw_reset(struct e1000_hw *hw) if (hw->mac_type < e1000_82571) msec_delay(10); + else + udelay(100); E1000_WRITE_REG(hw, CTRL, ctrl); E1000_WRITE_FLUSH(hw); @@ -3881,14 +3904,16 @@ e1000_read_eeprom(struct e1000_hw *hw, return -E1000_ERR_EEPROM; } - /* FLASH reads without acquiring the semaphore are safe in 82573-based - * controllers. - */ - if ((e1000_is_onboard_nvm_eeprom(hw) == TRUE) || - (hw->mac_type != e1000_82573)) { - /* Prepare the EEPROM for reading */ - if(e1000_acquire_eeprom(hw) != E1000_SUCCESS) - return -E1000_ERR_EEPROM; + /* FLASH reads without acquiring the semaphore are safe */ + if (e1000_is_onboard_nvm_eeprom(hw) == TRUE && + hw->eeprom.use_eerd == FALSE) { + switch (hw->mac_type) { + default: + /* Prepare the EEPROM for reading */ + if (e1000_acquire_eeprom(hw) != E1000_SUCCESS) + return -E1000_ERR_EEPROM; + break; + } } if(eeprom->use_eerd == TRUE) { @@ -6720,6 +6745,12 @@ e1000_get_phy_cfg_done(struct e1000_hw *hw) break; } + /* PHY configuration from NVM just starts after EECD_AUTO_RD sets to high. + * Need to wait for PHY configuration completion before accessing NVM + * and PHY. */ + if (hw->mac_type == e1000_82573) + msec_delay(25); + return E1000_SUCCESS; } diff --git a/drivers/net/e1000/e1000_hw.h b/drivers/net/e1000/e1000_hw.h index 7caa35748cea..0b8f6f2b774b 100644 --- a/drivers/net/e1000/e1000_hw.h +++ b/drivers/net/e1000/e1000_hw.h @@ -439,6 +439,7 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw); #define E1000_DEV_ID_82546GB_FIBER 0x107A #define E1000_DEV_ID_82546GB_SERDES 0x107B #define E1000_DEV_ID_82546GB_PCIE 0x108A +#define E1000_DEV_ID_82546GB_QUAD_COPPER 0x1099 #define E1000_DEV_ID_82547EI 0x1019 #define E1000_DEV_ID_82571EB_COPPER 0x105E #define E1000_DEV_ID_82571EB_FIBER 0x105F @@ -449,6 +450,7 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw); #define E1000_DEV_ID_82573E 0x108B #define E1000_DEV_ID_82573E_IAMT 0x108C #define E1000_DEV_ID_82573L 0x109A +#define E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3 0x10B5 #define NODE_ADDRESS_SIZE 6 @@ -1497,6 +1499,7 @@ struct e1000_hw { #define E1000_CTRL_EXT_EE_RST 0x00002000 /* Reinitialize from EEPROM */ #define E1000_CTRL_EXT_IPS 0x00004000 /* Invert Power State */ #define E1000_CTRL_EXT_SPD_BYPS 0x00008000 /* Speed Select Bypass */ +#define E1000_CTRL_EXT_RO_DIS 0x00020000 /* Relaxed Ordering disable */ #define E1000_CTRL_EXT_LINK_MODE_MASK 0x00C00000 #define E1000_CTRL_EXT_LINK_MODE_GMII 0x00000000 #define E1000_CTRL_EXT_LINK_MODE_TBI 0x00C00000 @@ -1954,6 +1957,23 @@ struct e1000_host_command_info { #define E1000_MDALIGN 4096 +/* PCI-Ex registers */ + +/* PCI-Ex Control Register */ +#define E1000_GCR_RXD_NO_SNOOP 0x00000001 +#define E1000_GCR_RXDSCW_NO_SNOOP 0x00000002 +#define E1000_GCR_RXDSCR_NO_SNOOP 0x00000004 +#define E1000_GCR_TXD_NO_SNOOP 0x00000008 +#define E1000_GCR_TXDSCW_NO_SNOOP 0x00000010 +#define E1000_GCR_TXDSCR_NO_SNOOP 0x00000020 + +#define PCI_EX_NO_SNOOP_ALL (E1000_GCR_RXD_NO_SNOOP | \ + E1000_GCR_RXDSCW_NO_SNOOP | \ + E1000_GCR_RXDSCR_NO_SNOOP | \ + E1000_GCR TXD_NO_SNOOP | \ + E1000_GCR_TXDSCW_NO_SNOOP | \ + E1000_GCR_TXDSCR_NO_SNOOP) + #define E1000_GCR_L1_ACT_WITHOUT_L0S_RX 0x08000000 /* Function Active and Power State to MNG */ #define E1000_FACTPS_FUNC0_POWER_STATE_MASK 0x00000003 @@ -2077,7 +2097,10 @@ struct e1000_host_command_info { /* Collision related configuration parameters */ #define E1000_COLLISION_THRESHOLD 15 #define E1000_CT_SHIFT 4 -#define E1000_COLLISION_DISTANCE 64 +/* Collision distance is a 0-based value that applies to + * half-duplex-capable hardware only. */ +#define E1000_COLLISION_DISTANCE 63 +#define E1000_COLLISION_DISTANCE_82542 64 #define E1000_FDX_COLLISION_DISTANCE E1000_COLLISION_DISTANCE #define E1000_HDX_COLLISION_DISTANCE E1000_COLLISION_DISTANCE #define E1000_COLD_SHIFT 12 diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c index 438a931fd55d..d0a5d1656c5f 100644 --- a/drivers/net/e1000/e1000_main.c +++ b/drivers/net/e1000/e1000_main.c @@ -43,7 +43,7 @@ static char e1000_driver_string[] = "Intel(R) PRO/1000 Network Driver"; #else #define DRIVERNAPI "-NAPI" #endif -#define DRV_VERSION "6.1.16-k2"DRIVERNAPI +#define DRV_VERSION "6.3.9-k2"DRIVERNAPI char e1000_driver_version[] = DRV_VERSION; static char e1000_copyright[] = "Copyright (c) 1999-2005 Intel Corporation."; @@ -97,7 +97,9 @@ static struct pci_device_id e1000_pci_tbl[] = { INTEL_E1000_ETHERNET_DEVICE(0x108A), INTEL_E1000_ETHERNET_DEVICE(0x108B), INTEL_E1000_ETHERNET_DEVICE(0x108C), + INTEL_E1000_ETHERNET_DEVICE(0x1099), INTEL_E1000_ETHERNET_DEVICE(0x109A), + INTEL_E1000_ETHERNET_DEVICE(0x10B5), /* required last entry */ {0,} }; @@ -171,9 +173,11 @@ static boolean_t e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, struct e1000_rx_ring *rx_ring); #endif static void e1000_alloc_rx_buffers(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring, + int cleaned_count); static void e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring, + int cleaned_count); static int e1000_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd); static int e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd); @@ -319,7 +323,75 @@ e1000_update_mng_vlan(struct e1000_adapter *adapter) } } } - + +/** + * e1000_release_hw_control - release control of the h/w to f/w + * @adapter: address of board private structure + * + * e1000_release_hw_control resets {CTRL_EXT|FWSM}:DRV_LOAD bit. + * For ASF and Pass Through versions of f/w this means that the + * driver is no longer loaded. For AMT version (only with 82573) i + * of the f/w this means that the netowrk i/f is closed. + * + **/ + +static inline void +e1000_release_hw_control(struct e1000_adapter *adapter) +{ + uint32_t ctrl_ext; + uint32_t swsm; + + /* Let firmware taken over control of h/w */ + switch (adapter->hw.mac_type) { + case e1000_82571: + case e1000_82572: + ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); + E1000_WRITE_REG(&adapter->hw, CTRL_EXT, + ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD); + break; + case e1000_82573: + swsm = E1000_READ_REG(&adapter->hw, SWSM); + E1000_WRITE_REG(&adapter->hw, SWSM, + swsm & ~E1000_SWSM_DRV_LOAD); + default: + break; + } +} + +/** + * e1000_get_hw_control - get control of the h/w from f/w + * @adapter: address of board private structure + * + * e1000_get_hw_control sets {CTRL_EXT|FWSM}:DRV_LOAD bit. + * For ASF and Pass Through versions of f/w this means that + * the driver is loaded. For AMT version (only with 82573) + * of the f/w this means that the netowrk i/f is open. + * + **/ + +static inline void +e1000_get_hw_control(struct e1000_adapter *adapter) +{ + uint32_t ctrl_ext; + uint32_t swsm; + /* Let firmware know the driver has taken over */ + switch (adapter->hw.mac_type) { + case e1000_82571: + case e1000_82572: + ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); + E1000_WRITE_REG(&adapter->hw, CTRL_EXT, + ctrl_ext | E1000_CTRL_EXT_DRV_LOAD); + break; + case e1000_82573: + swsm = E1000_READ_REG(&adapter->hw, SWSM); + E1000_WRITE_REG(&adapter->hw, SWSM, + swsm | E1000_SWSM_DRV_LOAD); + break; + default: + break; + } +} + int e1000_up(struct e1000_adapter *adapter) { @@ -343,8 +415,14 @@ e1000_up(struct e1000_adapter *adapter) e1000_configure_tx(adapter); e1000_setup_rctl(adapter); e1000_configure_rx(adapter); - for (i = 0; i < adapter->num_queues; i++) - adapter->alloc_rx_buf(adapter, &adapter->rx_ring[i]); + /* call E1000_DESC_UNUSED which always leaves + * at least 1 descriptor unused to make sure + * next_to_use != next_to_clean */ + for (i = 0; i < adapter->num_rx_queues; i++) { + struct e1000_rx_ring *ring = &adapter->rx_ring[i]; + adapter->alloc_rx_buf(adapter, ring, + E1000_DESC_UNUSED(ring)); + } #ifdef CONFIG_PCI_MSI if(adapter->hw.mac_type > e1000_82547_rev_2) { @@ -364,6 +442,12 @@ e1000_up(struct e1000_adapter *adapter) return err; } +#ifdef CONFIG_E1000_MQ + e1000_setup_queue_mapping(adapter); +#endif + + adapter->tx_queue_len = netdev->tx_queue_len; + mod_timer(&adapter->watchdog_timer, jiffies); #ifdef CONFIG_E1000_NAPI @@ -378,6 +462,8 @@ void e1000_down(struct e1000_adapter *adapter) { struct net_device *netdev = adapter->netdev; + boolean_t mng_mode_enabled = (adapter->hw.mac_type >= e1000_82571) && + e1000_check_mng_mode(&adapter->hw); e1000_irq_disable(adapter); #ifdef CONFIG_E1000_MQ @@ -396,6 +482,7 @@ e1000_down(struct e1000_adapter *adapter) #ifdef CONFIG_E1000_NAPI netif_poll_disable(netdev); #endif + netdev->tx_queue_len = adapter->tx_queue_len; adapter->link_speed = 0; adapter->link_duplex = 0; netif_carrier_off(netdev); @@ -405,12 +492,16 @@ e1000_down(struct e1000_adapter *adapter) e1000_clean_all_tx_rings(adapter); e1000_clean_all_rx_rings(adapter); - /* If WoL is not enabled and management mode is not IAMT - * Power down the PHY so no link is implied when interface is down */ - if(!adapter->wol && adapter->hw.mac_type >= e1000_82540 && + /* Power down the PHY so no link is implied when interface is down * + * The PHY cannot be powered down if any of the following is TRUE * + * (a) WoL is enabled + * (b) AMT is active + * (c) SoL/IDER session is active */ + if (!adapter->wol && adapter->hw.mac_type >= e1000_82540 && adapter->hw.media_type == e1000_media_type_copper && - !e1000_check_mng_mode(&adapter->hw) && - !(E1000_READ_REG(&adapter->hw, MANC) & E1000_MANC_SMBUS_EN)) { + !(E1000_READ_REG(&adapter->hw, MANC) & E1000_MANC_SMBUS_EN) && + !mng_mode_enabled && + !e1000_check_phy_reset_block(&adapter->hw)) { uint16_t mii_reg; e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &mii_reg); mii_reg |= MII_CR_POWER_DOWN; @@ -422,10 +513,8 @@ e1000_down(struct e1000_adapter *adapter) void e1000_reset(struct e1000_adapter *adapter) { - struct net_device *netdev = adapter->netdev; uint32_t pba, manc; uint16_t fc_high_water_mark = E1000_FC_HIGH_DIFF; - uint16_t fc_low_water_mark = E1000_FC_LOW_DIFF; /* Repartition Pba for greater than 9k mtu * To take effect CTRL.RST is required. @@ -449,15 +538,8 @@ e1000_reset(struct e1000_adapter *adapter) } if((adapter->hw.mac_type != e1000_82573) && - (adapter->rx_buffer_len > E1000_RXBUFFER_8192)) { + (adapter->netdev->mtu > E1000_RXBUFFER_8192)) pba -= 8; /* allocate more FIFO for Tx */ - /* send an XOFF when there is enough space in the - * Rx FIFO to hold one extra full size Rx packet - */ - fc_high_water_mark = netdev->mtu + ENET_HEADER_SIZE + - ETHERNET_FCS_SIZE + 1; - fc_low_water_mark = fc_high_water_mark + 8; - } if(adapter->hw.mac_type == e1000_82547) { @@ -471,10 +553,12 @@ e1000_reset(struct e1000_adapter *adapter) E1000_WRITE_REG(&adapter->hw, PBA, pba); /* flow control settings */ - adapter->hw.fc_high_water = (pba << E1000_PBA_BYTES_SHIFT) - - fc_high_water_mark; - adapter->hw.fc_low_water = (pba << E1000_PBA_BYTES_SHIFT) - - fc_low_water_mark; + /* Set the FC high water mark to 90% of the FIFO size. + * Required to clear last 3 LSB */ + fc_high_water_mark = ((pba * 9216)/10) & 0xFFF8; + + adapter->hw.fc_high_water = fc_high_water_mark; + adapter->hw.fc_low_water = fc_high_water_mark - 8; adapter->hw.fc_pause_time = E1000_FC_PAUSE_TIME; adapter->hw.fc_send_xon = 1; adapter->hw.fc = adapter->hw.original_fc; @@ -517,8 +601,6 @@ e1000_probe(struct pci_dev *pdev, struct net_device *netdev; struct e1000_adapter *adapter; unsigned long mmio_start, mmio_len; - uint32_t ctrl_ext; - uint32_t swsm; static int cards_found = 0; int i, err, pci_using_dac; @@ -712,8 +794,7 @@ e1000_probe(struct pci_dev *pdev, case e1000_82546: case e1000_82546_rev_3: case e1000_82571: - if((E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1) - && (adapter->hw.media_type == e1000_media_type_copper)) { + if(E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1){ e1000_read_eeprom(&adapter->hw, EEPROM_INIT_CONTROL3_PORT_B, 1, &eeprom_data); break; @@ -727,25 +808,36 @@ e1000_probe(struct pci_dev *pdev, if(eeprom_data & eeprom_apme_mask) adapter->wol |= E1000_WUFC_MAG; + /* print bus type/speed/width info */ + { + struct e1000_hw *hw = &adapter->hw; + DPRINTK(PROBE, INFO, "(PCI%s:%s:%s) ", + ((hw->bus_type == e1000_bus_type_pcix) ? "-X" : + (hw->bus_type == e1000_bus_type_pci_express ? " Express":"")), + ((hw->bus_speed == e1000_bus_speed_2500) ? "2.5Gb/s" : + (hw->bus_speed == e1000_bus_speed_133) ? "133MHz" : + (hw->bus_speed == e1000_bus_speed_120) ? "120MHz" : + (hw->bus_speed == e1000_bus_speed_100) ? "100MHz" : + (hw->bus_speed == e1000_bus_speed_66) ? "66MHz" : "33MHz"), + ((hw->bus_width == e1000_bus_width_64) ? "64-bit" : + (hw->bus_width == e1000_bus_width_pciex_4) ? "Width x4" : + (hw->bus_width == e1000_bus_width_pciex_1) ? "Width x1" : + "32-bit")); + } + + for (i = 0; i < 6; i++) + printk("%2.2x%c", netdev->dev_addr[i], i == 5 ? '\n' : ':'); + /* reset the hardware with the new settings */ e1000_reset(adapter); - /* Let firmware know the driver has taken over */ - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext | E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm | E1000_SWSM_DRV_LOAD); - break; - default: - break; - } + /* If the controller is 82573 and f/w is AMT, do not set + * DRV_LOAD until the interface is up. For all other cases, + * let the f/w know that the h/w is now under the control + * of the driver. */ + if (adapter->hw.mac_type != e1000_82573 || + !e1000_check_mng_mode(&adapter->hw)) + e1000_get_hw_control(adapter); strcpy(netdev->name, "eth%d"); if((err = register_netdev(netdev))) @@ -782,8 +874,7 @@ e1000_remove(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); struct e1000_adapter *adapter = netdev_priv(netdev); - uint32_t ctrl_ext; - uint32_t manc, swsm; + uint32_t manc; #ifdef CONFIG_E1000_NAPI int i; #endif @@ -799,26 +890,13 @@ e1000_remove(struct pci_dev *pdev) } } - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm & ~E1000_SWSM_DRV_LOAD); - break; - - default: - break; - } + /* Release control of h/w to f/w. If f/w is AMT enabled, this + * would have already happened in close and is redundant. */ + e1000_release_hw_control(adapter); unregister_netdev(netdev); #ifdef CONFIG_E1000_NAPI - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) __dev_put(&adapter->polling_netdev[i]); #endif @@ -923,15 +1001,34 @@ e1000_sw_init(struct e1000_adapter *adapter) switch (hw->mac_type) { case e1000_82571: case e1000_82572: - adapter->num_queues = 2; + /* These controllers support 2 tx queues, but with a single + * qdisc implementation, multiple tx queues aren't quite as + * interesting. If we can find a logical way of mapping + * flows to a queue, then perhaps we can up the num_tx_queue + * count back to its default. Until then, we run the risk of + * terrible performance due to SACK overload. */ + adapter->num_tx_queues = 1; + adapter->num_rx_queues = 2; break; default: - adapter->num_queues = 1; + adapter->num_tx_queues = 1; + adapter->num_rx_queues = 1; break; } - adapter->num_queues = min(adapter->num_queues, num_online_cpus()); + adapter->num_rx_queues = min(adapter->num_rx_queues, num_online_cpus()); + adapter->num_tx_queues = min(adapter->num_tx_queues, num_online_cpus()); + DPRINTK(DRV, INFO, "Multiqueue Enabled: Rx Queue count = %u %s\n", + adapter->num_rx_queues, + ((adapter->num_rx_queues == 1) + ? ((num_online_cpus() > 1) + ? "(due to unsupported feature in current adapter)" + : "(due to unsupported system configuration)") + : "")); + DPRINTK(DRV, INFO, "Multiqueue Enabled: Tx Queue count = %u\n", + adapter->num_tx_queues); #else - adapter->num_queues = 1; + adapter->num_tx_queues = 1; + adapter->num_rx_queues = 1; #endif if (e1000_alloc_queues(adapter)) { @@ -940,17 +1037,14 @@ e1000_sw_init(struct e1000_adapter *adapter) } #ifdef CONFIG_E1000_NAPI - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_rx_queues; i++) { adapter->polling_netdev[i].priv = adapter; adapter->polling_netdev[i].poll = &e1000_clean; adapter->polling_netdev[i].weight = 64; dev_hold(&adapter->polling_netdev[i]); set_bit(__LINK_STATE_START, &adapter->polling_netdev[i].state); } -#endif - -#ifdef CONFIG_E1000_MQ - e1000_setup_queue_mapping(adapter); + spin_lock_init(&adapter->tx_queue_lock); #endif atomic_set(&adapter->irq_sem, 1); @@ -973,13 +1067,13 @@ e1000_alloc_queues(struct e1000_adapter *adapter) { int size; - size = sizeof(struct e1000_tx_ring) * adapter->num_queues; + size = sizeof(struct e1000_tx_ring) * adapter->num_tx_queues; adapter->tx_ring = kmalloc(size, GFP_KERNEL); if (!adapter->tx_ring) return -ENOMEM; memset(adapter->tx_ring, 0, size); - size = sizeof(struct e1000_rx_ring) * adapter->num_queues; + size = sizeof(struct e1000_rx_ring) * adapter->num_rx_queues; adapter->rx_ring = kmalloc(size, GFP_KERNEL); if (!adapter->rx_ring) { kfree(adapter->tx_ring); @@ -988,7 +1082,7 @@ e1000_alloc_queues(struct e1000_adapter *adapter) memset(adapter->rx_ring, 0, size); #ifdef CONFIG_E1000_NAPI - size = sizeof(struct net_device) * adapter->num_queues; + size = sizeof(struct net_device) * adapter->num_rx_queues; adapter->polling_netdev = kmalloc(size, GFP_KERNEL); if (!adapter->polling_netdev) { kfree(adapter->tx_ring); @@ -998,6 +1092,14 @@ e1000_alloc_queues(struct e1000_adapter *adapter) memset(adapter->polling_netdev, 0, size); #endif +#ifdef CONFIG_E1000_MQ + adapter->rx_sched_call_data.func = e1000_rx_schedule; + adapter->rx_sched_call_data.info = adapter->netdev; + + adapter->cpu_netdev = alloc_percpu(struct net_device *); + adapter->cpu_tx_ring = alloc_percpu(struct e1000_tx_ring *); +#endif + return E1000_SUCCESS; } @@ -1017,14 +1119,15 @@ e1000_setup_queue_mapping(struct e1000_adapter *adapter) lock_cpu_hotplug(); i = 0; for_each_online_cpu(cpu) { - *per_cpu_ptr(adapter->cpu_tx_ring, cpu) = &adapter->tx_ring[i % adapter->num_queues]; + *per_cpu_ptr(adapter->cpu_tx_ring, cpu) = &adapter->tx_ring[i % adapter->num_tx_queues]; /* This is incomplete because we'd like to assign separate * physical cpus to these netdev polling structures and * avoid saturating a subset of cpus. */ - if (i < adapter->num_queues) { + if (i < adapter->num_rx_queues) { *per_cpu_ptr(adapter->cpu_netdev, cpu) = &adapter->polling_netdev[i]; - adapter->cpu_for_queue[i] = cpu; + adapter->rx_ring[i].cpu = cpu; + cpu_set(cpu, adapter->cpumask); } else *per_cpu_ptr(adapter->cpu_netdev, cpu) = NULL; @@ -1071,6 +1174,12 @@ e1000_open(struct net_device *netdev) e1000_update_mng_vlan(adapter); } + /* If AMT is enabled, let the firmware know that the network + * interface is now open */ + if (adapter->hw.mac_type == e1000_82573 && + e1000_check_mng_mode(&adapter->hw)) + e1000_get_hw_control(adapter); + return E1000_SUCCESS; err_up: @@ -1109,6 +1218,13 @@ e1000_close(struct net_device *netdev) E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) { e1000_vlan_rx_kill_vid(netdev, adapter->mng_vlan_id); } + + /* If AMT is enabled, let the firmware know that the network + * interface is now closed */ + if (adapter->hw.mac_type == e1000_82573 && + e1000_check_mng_mode(&adapter->hw)) + e1000_release_hw_control(adapter); + return 0; } @@ -1229,7 +1345,7 @@ e1000_setup_all_tx_resources(struct e1000_adapter *adapter) { int i, err = 0; - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_tx_queues; i++) { err = e1000_setup_tx_resources(adapter, &adapter->tx_ring[i]); if (err) { DPRINTK(PROBE, ERR, @@ -1254,10 +1370,11 @@ e1000_configure_tx(struct e1000_adapter *adapter) uint64_t tdba; struct e1000_hw *hw = &adapter->hw; uint32_t tdlen, tctl, tipg, tarc; + uint32_t ipgr1, ipgr2; /* Setup the HW Tx Head and Tail descriptor pointers */ - switch (adapter->num_queues) { + switch (adapter->num_tx_queues) { case 2: tdba = adapter->tx_ring[1].dma; tdlen = adapter->tx_ring[1].count * @@ -1287,22 +1404,26 @@ e1000_configure_tx(struct e1000_adapter *adapter) /* Set the default values for the Tx Inter Packet Gap timer */ + if (hw->media_type == e1000_media_type_fiber || + hw->media_type == e1000_media_type_internal_serdes) + tipg = DEFAULT_82543_TIPG_IPGT_FIBER; + else + tipg = DEFAULT_82543_TIPG_IPGT_COPPER; + switch (hw->mac_type) { case e1000_82542_rev2_0: case e1000_82542_rev2_1: tipg = DEFAULT_82542_TIPG_IPGT; - tipg |= DEFAULT_82542_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT; - tipg |= DEFAULT_82542_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT; + ipgr1 = DEFAULT_82542_TIPG_IPGR1; + ipgr2 = DEFAULT_82542_TIPG_IPGR2; break; default: - if (hw->media_type == e1000_media_type_fiber || - hw->media_type == e1000_media_type_internal_serdes) - tipg = DEFAULT_82543_TIPG_IPGT_FIBER; - else - tipg = DEFAULT_82543_TIPG_IPGT_COPPER; - tipg |= DEFAULT_82543_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT; - tipg |= DEFAULT_82543_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT; + ipgr1 = DEFAULT_82543_TIPG_IPGR1; + ipgr2 = DEFAULT_82543_TIPG_IPGR2; + break; } + tipg |= ipgr1 << E1000_TIPG_IPGR1_SHIFT; + tipg |= ipgr2 << E1000_TIPG_IPGR2_SHIFT; E1000_WRITE_REG(hw, TIPG, tipg); /* Set the Tx Interrupt Delay register */ @@ -1454,6 +1575,8 @@ setup_rx_desc_die: rxdr->next_to_clean = 0; rxdr->next_to_use = 0; + rxdr->rx_skb_top = NULL; + rxdr->rx_skb_prev = NULL; return 0; } @@ -1475,7 +1598,7 @@ e1000_setup_all_rx_resources(struct e1000_adapter *adapter) { int i, err = 0; - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_rx_queues; i++) { err = e1000_setup_rx_resources(adapter, &adapter->rx_ring[i]); if (err) { DPRINTK(PROBE, ERR, @@ -1510,7 +1633,10 @@ e1000_setup_rctl(struct e1000_adapter *adapter) E1000_RCTL_LBM_NO | E1000_RCTL_RDMTS_HALF | (adapter->hw.mc_filter_type << E1000_RCTL_MO_SHIFT); - if(adapter->hw.tbi_compatibility_on == 1) + if (adapter->hw.mac_type > e1000_82543) + rctl |= E1000_RCTL_SECRC; + + if (adapter->hw.tbi_compatibility_on == 1) rctl |= E1000_RCTL_SBP; else rctl &= ~E1000_RCTL_SBP; @@ -1638,16 +1764,21 @@ e1000_configure_rx(struct e1000_adapter *adapter) } if (hw->mac_type >= e1000_82571) { - /* Reset delay timers after every interrupt */ ctrl_ext = E1000_READ_REG(hw, CTRL_EXT); + /* Reset delay timers after every interrupt */ ctrl_ext |= E1000_CTRL_EXT_CANC; +#ifdef CONFIG_E1000_NAPI + /* Auto-Mask interrupts upon ICR read. */ + ctrl_ext |= E1000_CTRL_EXT_IAME; +#endif E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext); + E1000_WRITE_REG(hw, IAM, ~0); E1000_WRITE_FLUSH(hw); } /* Setup the HW Rx Head and Tail Descriptor Pointers and * the Base and Length of the Rx Descriptor Ring */ - switch (adapter->num_queues) { + switch (adapter->num_rx_queues) { #ifdef CONFIG_E1000_MQ case 2: rdba = adapter->rx_ring[1].dma; @@ -1674,7 +1805,7 @@ e1000_configure_rx(struct e1000_adapter *adapter) } #ifdef CONFIG_E1000_MQ - if (adapter->num_queues > 1) { + if (adapter->num_rx_queues > 1) { uint32_t random[10]; get_random_bytes(&random[0], 40); @@ -1684,7 +1815,7 @@ e1000_configure_rx(struct e1000_adapter *adapter) E1000_WRITE_REG(hw, RSSIM, 0); } - switch (adapter->num_queues) { + switch (adapter->num_rx_queues) { case 2: default: reta = 0x00800080; @@ -1776,7 +1907,7 @@ e1000_free_all_tx_resources(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_tx_queues; i++) e1000_free_tx_resources(adapter, &adapter->tx_ring[i]); } @@ -1789,12 +1920,10 @@ e1000_unmap_and_free_tx_resource(struct e1000_adapter *adapter, buffer_info->dma, buffer_info->length, PCI_DMA_TODEVICE); - buffer_info->dma = 0; } - if(buffer_info->skb) { + if (buffer_info->skb) dev_kfree_skb_any(buffer_info->skb); - buffer_info->skb = NULL; - } + memset(buffer_info, 0, sizeof(struct e1000_buffer)); } /** @@ -1843,7 +1972,7 @@ e1000_clean_all_tx_rings(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_tx_queues; i++) e1000_clean_tx_ring(adapter, &adapter->tx_ring[i]); } @@ -1887,7 +2016,7 @@ e1000_free_all_rx_resources(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) e1000_free_rx_resources(adapter, &adapter->rx_ring[i]); } @@ -1913,8 +2042,6 @@ e1000_clean_rx_ring(struct e1000_adapter *adapter, for(i = 0; i < rx_ring->count; i++) { buffer_info = &rx_ring->buffer_info[i]; if(buffer_info->skb) { - ps_page = &rx_ring->ps_page[i]; - ps_page_dma = &rx_ring->ps_page_dma[i]; pci_unmap_single(pdev, buffer_info->dma, buffer_info->length, @@ -1922,19 +2049,30 @@ e1000_clean_rx_ring(struct e1000_adapter *adapter, dev_kfree_skb(buffer_info->skb); buffer_info->skb = NULL; - - for(j = 0; j < adapter->rx_ps_pages; j++) { - if(!ps_page->ps_page[j]) break; - pci_unmap_single(pdev, - ps_page_dma->ps_page_dma[j], - PAGE_SIZE, PCI_DMA_FROMDEVICE); - ps_page_dma->ps_page_dma[j] = 0; - put_page(ps_page->ps_page[j]); - ps_page->ps_page[j] = NULL; - } + } + ps_page = &rx_ring->ps_page[i]; + ps_page_dma = &rx_ring->ps_page_dma[i]; + for (j = 0; j < adapter->rx_ps_pages; j++) { + if (!ps_page->ps_page[j]) break; + pci_unmap_page(pdev, + ps_page_dma->ps_page_dma[j], + PAGE_SIZE, PCI_DMA_FROMDEVICE); + ps_page_dma->ps_page_dma[j] = 0; + put_page(ps_page->ps_page[j]); + ps_page->ps_page[j] = NULL; } } + /* there also may be some cached data in our adapter */ + if (rx_ring->rx_skb_top) { + dev_kfree_skb(rx_ring->rx_skb_top); + + /* rx_skb_prev will be wiped out by rx_skb_top */ + rx_ring->rx_skb_top = NULL; + rx_ring->rx_skb_prev = NULL; + } + + size = sizeof(struct e1000_buffer) * rx_ring->count; memset(rx_ring->buffer_info, 0, size); size = sizeof(struct e1000_ps_page) * rx_ring->count; @@ -1963,7 +2101,7 @@ e1000_clean_all_rx_rings(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) e1000_clean_rx_ring(adapter, &adapter->rx_ring[i]); } @@ -2005,7 +2143,9 @@ e1000_leave_82542_rst(struct e1000_adapter *adapter) if(netif_running(netdev)) { e1000_configure_rx(adapter); - e1000_alloc_rx_buffers(adapter, &adapter->rx_ring[0]); + /* No need to loop, because 82542 supports only 1 queue */ + struct e1000_rx_ring *ring = &adapter->rx_ring[0]; + adapter->alloc_rx_buf(adapter, ring, E1000_DESC_UNUSED(ring)); } } @@ -2204,7 +2344,7 @@ static void e1000_watchdog_task(struct e1000_adapter *adapter) { struct net_device *netdev = adapter->netdev; - struct e1000_tx_ring *txdr = &adapter->tx_ring[0]; + struct e1000_tx_ring *txdr = adapter->tx_ring; uint32_t link; e1000_check_for_link(&adapter->hw); @@ -2231,6 +2371,21 @@ e1000_watchdog_task(struct e1000_adapter *adapter) adapter->link_duplex == FULL_DUPLEX ? "Full Duplex" : "Half Duplex"); + /* tweak tx_queue_len according to speed/duplex */ + netdev->tx_queue_len = adapter->tx_queue_len; + adapter->tx_timeout_factor = 1; + if (adapter->link_duplex == HALF_DUPLEX) { + switch (adapter->link_speed) { + case SPEED_10: + netdev->tx_queue_len = 10; + adapter->tx_timeout_factor = 8; + break; + case SPEED_100: + netdev->tx_queue_len = 100; + break; + } + } + netif_carrier_on(netdev); netif_wake_queue(netdev); mod_timer(&adapter->phy_info_timer, jiffies + 2 * HZ); @@ -2263,7 +2418,10 @@ e1000_watchdog_task(struct e1000_adapter *adapter) e1000_update_adaptive(&adapter->hw); - if (adapter->num_queues == 1 && !netif_carrier_ok(netdev)) { +#ifdef CONFIG_E1000_MQ + txdr = *per_cpu_ptr(adapter->cpu_tx_ring, smp_processor_id()); +#endif + if (!netif_carrier_ok(netdev)) { if (E1000_DESC_UNUSED(txdr) + 1 < txdr->count) { /* We've lost link, so the controller stops DMA, * but we've got queued Tx work that's never going @@ -2314,6 +2472,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, { #ifdef NETIF_F_TSO struct e1000_context_desc *context_desc; + struct e1000_buffer *buffer_info; unsigned int i; uint32_t cmd_length = 0; uint16_t ipcse = 0, tucse, mss; @@ -2363,6 +2522,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, i = tx_ring->next_to_use; context_desc = E1000_CONTEXT_DESC(*tx_ring, i); + buffer_info = &tx_ring->buffer_info[i]; context_desc->lower_setup.ip_fields.ipcss = ipcss; context_desc->lower_setup.ip_fields.ipcso = ipcso; @@ -2374,14 +2534,16 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, context_desc->tcp_seg_setup.fields.hdr_len = hdr_len; context_desc->cmd_and_length = cpu_to_le32(cmd_length); + buffer_info->time_stamp = jiffies; + if (++i == tx_ring->count) i = 0; tx_ring->next_to_use = i; - return 1; + return TRUE; } #endif - return 0; + return FALSE; } static inline boolean_t @@ -2389,6 +2551,7 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, struct sk_buff *skb) { struct e1000_context_desc *context_desc; + struct e1000_buffer *buffer_info; unsigned int i; uint8_t css; @@ -2396,6 +2559,7 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, css = skb->h.raw - skb->data; i = tx_ring->next_to_use; + buffer_info = &tx_ring->buffer_info[i]; context_desc = E1000_CONTEXT_DESC(*tx_ring, i); context_desc->upper_setup.tcp_fields.tucss = css; @@ -2404,6 +2568,8 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, context_desc->tcp_seg_setup.data = 0; context_desc->cmd_and_length = cpu_to_le32(E1000_TXD_CMD_DEXT); + buffer_info->time_stamp = jiffies; + if (unlikely(++i == tx_ring->count)) i = 0; tx_ring->next_to_use = i; @@ -2688,11 +2854,30 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) * overrun the FIFO, adjust the max buffer len if mss * drops. */ if(mss) { + uint8_t hdr_len; max_per_txd = min(mss << 2, max_per_txd); max_txd_pwr = fls(max_per_txd) - 1; + + /* TSO Workaround for 82571/2 Controllers -- if skb->data + * points to just header, pull a few bytes of payload from + * frags into skb->data */ + hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2)); + if (skb->data_len && (hdr_len == (skb->len - skb->data_len)) && + (adapter->hw.mac_type == e1000_82571 || + adapter->hw.mac_type == e1000_82572)) { + unsigned int pull_size; + pull_size = min((unsigned int)4, skb->data_len); + if (!__pskb_pull_tail(skb, pull_size)) { + printk(KERN_ERR "__pskb_pull_tail failed.\n"); + dev_kfree_skb_any(skb); + return -EFAULT; + } + len = skb->len - skb->data_len; + } } if((mss) || (skb->ip_summed == CHECKSUM_HW)) + /* reserve a descriptor for the offload context */ count++; count++; #else @@ -2726,27 +2911,6 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) if(adapter->pcix_82544) count += nr_frags; -#ifdef NETIF_F_TSO - /* TSO Workaround for 82571/2 Controllers -- if skb->data - * points to just header, pull a few bytes of payload from - * frags into skb->data */ - if (skb_shinfo(skb)->tso_size) { - uint8_t hdr_len; - hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2)); - if (skb->data_len && (hdr_len < (skb->len - skb->data_len)) && - (adapter->hw.mac_type == e1000_82571 || - adapter->hw.mac_type == e1000_82572)) { - unsigned int pull_size; - pull_size = min((unsigned int)4, skb->data_len); - if (!__pskb_pull_tail(skb, pull_size)) { - printk(KERN_ERR "__pskb_pull_tail failed.\n"); - dev_kfree_skb_any(skb); - return -EFAULT; - } - } - } -#endif - if(adapter->hw.tx_pkt_filtering && (adapter->hw.mac_type == e1000_82573) ) e1000_transfer_dhcp_info(adapter, skb); @@ -2833,6 +2997,7 @@ e1000_tx_timeout_task(struct net_device *netdev) { struct e1000_adapter *adapter = netdev_priv(netdev); + adapter->tx_timeout_count++; e1000_down(adapter); e1000_up(adapter); } @@ -2850,7 +3015,7 @@ e1000_get_stats(struct net_device *netdev) { struct e1000_adapter *adapter = netdev_priv(netdev); - e1000_update_stats(adapter); + /* only return the current stats */ return &adapter->net_stats; } @@ -2871,50 +3036,51 @@ e1000_change_mtu(struct net_device *netdev, int new_mtu) if((max_frame < MINIMUM_ETHERNET_FRAME_SIZE) || (max_frame > MAX_JUMBO_FRAME_SIZE)) { DPRINTK(PROBE, ERR, "Invalid MTU setting\n"); - return -EINVAL; - } - -#define MAX_STD_JUMBO_FRAME_SIZE 9234 - /* might want this to be bigger enum check... */ - /* 82571 controllers limit jumbo frame size to 10500 bytes */ - if ((adapter->hw.mac_type == e1000_82571 || - adapter->hw.mac_type == e1000_82572) && - max_frame > MAX_STD_JUMBO_FRAME_SIZE) { - DPRINTK(PROBE, ERR, "MTU > 9216 bytes not supported " - "on 82571 and 82572 controllers.\n"); return -EINVAL; } - if(adapter->hw.mac_type == e1000_82573 && - max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) { - DPRINTK(PROBE, ERR, "Jumbo Frames not supported " - "on 82573\n"); - return -EINVAL; - } - - if(adapter->hw.mac_type > e1000_82547_rev_2) { - adapter->rx_buffer_len = max_frame; - E1000_ROUNDUP(adapter->rx_buffer_len, 1024); - } else { - if(unlikely((adapter->hw.mac_type < e1000_82543) && - (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE))) { - DPRINTK(PROBE, ERR, "Jumbo Frames not supported " - "on 82542\n"); + /* Adapter-specific max frame size limits. */ + switch (adapter->hw.mac_type) { + case e1000_82542_rev2_0: + case e1000_82542_rev2_1: + case e1000_82573: + if (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) { + DPRINTK(PROBE, ERR, "Jumbo Frames not supported.\n"); + return -EINVAL; + } + break; + case e1000_82571: + case e1000_82572: +#define MAX_STD_JUMBO_FRAME_SIZE 9234 + if (max_frame > MAX_STD_JUMBO_FRAME_SIZE) { + DPRINTK(PROBE, ERR, "MTU > 9216 not supported.\n"); return -EINVAL; - - } else { - if(max_frame <= E1000_RXBUFFER_2048) { - adapter->rx_buffer_len = E1000_RXBUFFER_2048; - } else if(max_frame <= E1000_RXBUFFER_4096) { - adapter->rx_buffer_len = E1000_RXBUFFER_4096; - } else if(max_frame <= E1000_RXBUFFER_8192) { - adapter->rx_buffer_len = E1000_RXBUFFER_8192; - } else if(max_frame <= E1000_RXBUFFER_16384) { - adapter->rx_buffer_len = E1000_RXBUFFER_16384; - } } + break; + default: + /* Capable of supporting up to MAX_JUMBO_FRAME_SIZE limit. */ + break; } + /* since the driver code now supports splitting a packet across + * multiple descriptors, most of the fifo related limitations on + * jumbo frame traffic have gone away. + * simply use 2k descriptors for everything. + * + * NOTE: dev_alloc_skb reserves 16 bytes, and typically NET_IP_ALIGN + * means we reserve 2 more, this pushes us to allocate from the next + * larger slab size + * i.e. RXBUFFER_2048 --> size-4096 slab */ + + /* recent hardware supports 1KB granularity */ + if (adapter->hw.mac_type > e1000_82547_rev_2) { + adapter->rx_buffer_len = + ((max_frame < E1000_RXBUFFER_2048) ? + max_frame : E1000_RXBUFFER_2048); + E1000_ROUNDUP(adapter->rx_buffer_len, 1024); + } else + adapter->rx_buffer_len = E1000_RXBUFFER_2048; + netdev->mtu = new_mtu; if(netif_running(netdev)) { @@ -3037,12 +3203,11 @@ e1000_update_stats(struct e1000_adapter *adapter) adapter->net_stats.rx_errors = adapter->stats.rxerrc + adapter->stats.crcerrs + adapter->stats.algnerrc + - adapter->stats.rlec + adapter->stats.mpc + - adapter->stats.cexterr; + adapter->stats.rlec + adapter->stats.cexterr; + adapter->net_stats.rx_dropped = 0; adapter->net_stats.rx_length_errors = adapter->stats.rlec; adapter->net_stats.rx_crc_errors = adapter->stats.crcerrs; adapter->net_stats.rx_frame_errors = adapter->stats.algnerrc; - adapter->net_stats.rx_fifo_errors = adapter->stats.mpc; adapter->net_stats.rx_missed_errors = adapter->stats.mpc; /* Tx Errors */ @@ -3110,12 +3275,24 @@ e1000_intr(int irq, void *data, struct pt_regs *regs) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; uint32_t icr = E1000_READ_REG(hw, ICR); -#if defined(CONFIG_E1000_NAPI) && defined(CONFIG_E1000_MQ) || !defined(CONFIG_E1000_NAPI) +#ifndef CONFIG_E1000_NAPI int i; +#else + /* Interrupt Auto-Mask...upon reading ICR, + * interrupts are masked. No need for the + * IMC write, but it does mean we should + * account for it ASAP. */ + if (likely(hw->mac_type >= e1000_82571)) + atomic_inc(&adapter->irq_sem); #endif - if(unlikely(!icr)) + if (unlikely(!icr)) { +#ifdef CONFIG_E1000_NAPI + if (hw->mac_type >= e1000_82571) + e1000_irq_enable(adapter); +#endif return IRQ_NONE; /* Not our interrupt */ + } if(unlikely(icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC))) { hw->get_link_status = 1; @@ -3123,19 +3300,19 @@ e1000_intr(int irq, void *data, struct pt_regs *regs) } #ifdef CONFIG_E1000_NAPI - atomic_inc(&adapter->irq_sem); - E1000_WRITE_REG(hw, IMC, ~0); - E1000_WRITE_FLUSH(hw); + if (unlikely(hw->mac_type < e1000_82571)) { + atomic_inc(&adapter->irq_sem); + E1000_WRITE_REG(hw, IMC, ~0); + E1000_WRITE_FLUSH(hw); + } #ifdef CONFIG_E1000_MQ if (atomic_read(&adapter->rx_sched_call_data.count) == 0) { - cpu_set(adapter->cpu_for_queue[0], - adapter->rx_sched_call_data.cpumask); - for (i = 1; i < adapter->num_queues; i++) { - cpu_set(adapter->cpu_for_queue[i], - adapter->rx_sched_call_data.cpumask); - atomic_inc(&adapter->irq_sem); - } - atomic_set(&adapter->rx_sched_call_data.count, i); + /* We must setup the cpumask once count == 0 since + * each cpu bit is cleared when the work is done. */ + adapter->rx_sched_call_data.cpumask = adapter->cpumask; + atomic_add(adapter->num_rx_queues - 1, &adapter->irq_sem); + atomic_set(&adapter->rx_sched_call_data.count, + adapter->num_rx_queues); smp_call_async_mask(&adapter->rx_sched_call_data); } else { printk("call_data.count == %u\n", atomic_read(&adapter->rx_sched_call_data.count)); @@ -3187,7 +3364,7 @@ e1000_clean(struct net_device *poll_dev, int *budget) { struct e1000_adapter *adapter; int work_to_do = min(*budget, poll_dev->quota); - int tx_cleaned, i = 0, work_done = 0; + int tx_cleaned = 0, i = 0, work_done = 0; /* Must NOT use netdev_priv macro here. */ adapter = poll_dev->priv; @@ -3198,11 +3375,23 @@ e1000_clean(struct net_device *poll_dev, int *budget) while (poll_dev != &adapter->polling_netdev[i]) { i++; - if (unlikely(i == adapter->num_queues)) + if (unlikely(i == adapter->num_rx_queues)) BUG(); } - tx_cleaned = e1000_clean_tx_irq(adapter, &adapter->tx_ring[i]); + if (likely(adapter->num_tx_queues == 1)) { + /* e1000_clean is called per-cpu. This lock protects + * tx_ring[0] from being cleaned by multiple cpus + * simultaneously. A failure obtaining the lock means + * tx_ring[0] is currently being cleaned anyway. */ + if (spin_trylock(&adapter->tx_queue_lock)) { + tx_cleaned = e1000_clean_tx_irq(adapter, + &adapter->tx_ring[0]); + spin_unlock(&adapter->tx_queue_lock); + } + } else + tx_cleaned = e1000_clean_tx_irq(adapter, &adapter->tx_ring[i]); + adapter->clean_rx(adapter, &adapter->rx_ring[i], &work_done, work_to_do); @@ -3247,17 +3436,19 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter, buffer_info = &tx_ring->buffer_info[i]; cleaned = (i == eop); +#ifdef CONFIG_E1000_MQ + tx_ring->tx_stats.bytes += buffer_info->length; +#endif e1000_unmap_and_free_tx_resource(adapter, buffer_info); - - tx_desc->buffer_addr = 0; - tx_desc->lower.data = 0; - tx_desc->upper.data = 0; + memset(tx_desc, 0, sizeof(struct e1000_tx_desc)); if(unlikely(++i == tx_ring->count)) i = 0; } - tx_ring->pkt++; - +#ifdef CONFIG_E1000_MQ + tx_ring->tx_stats.packets++; +#endif + eop = tx_ring->buffer_info[i].next_to_watch; eop_desc = E1000_TX_DESC(*tx_ring, eop); } @@ -3276,32 +3467,31 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter, /* Detect a transmit hang in hardware, this serializes the * check with the clearing of time_stamp and movement of i */ adapter->detect_tx_hung = FALSE; - if (tx_ring->buffer_info[i].dma && - time_after(jiffies, tx_ring->buffer_info[i].time_stamp + HZ) + if (tx_ring->buffer_info[eop].dma && + time_after(jiffies, tx_ring->buffer_info[eop].time_stamp + + adapter->tx_timeout_factor * HZ) && !(E1000_READ_REG(&adapter->hw, STATUS) & - E1000_STATUS_TXOFF)) { + E1000_STATUS_TXOFF)) { /* detected Tx unit hang */ - i = tx_ring->next_to_clean; - eop = tx_ring->buffer_info[i].next_to_watch; - eop_desc = E1000_TX_DESC(*tx_ring, eop); DPRINTK(DRV, ERR, "Detected Tx Unit Hang\n" + " Tx Queue <%lu>\n" " TDH <%x>\n" " TDT <%x>\n" " next_to_use <%x>\n" " next_to_clean <%x>\n" "buffer_info[next_to_clean]\n" - " dma <%llx>\n" " time_stamp <%lx>\n" " next_to_watch <%x>\n" " jiffies <%lx>\n" " next_to_watch.status <%x>\n", + (unsigned long)((tx_ring - adapter->tx_ring) / + sizeof(struct e1000_tx_ring)), readl(adapter->hw.hw_addr + tx_ring->tdh), readl(adapter->hw.hw_addr + tx_ring->tdt), tx_ring->next_to_use, - i, - (unsigned long long)tx_ring->buffer_info[i].dma, - tx_ring->buffer_info[i].time_stamp, + tx_ring->next_to_clean, + tx_ring->buffer_info[eop].time_stamp, eop, jiffies, eop_desc->upper.fields.status); @@ -3386,20 +3576,23 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, uint32_t length; uint8_t last_byte; unsigned int i; - boolean_t cleaned = FALSE; + int cleaned_count = 0; + boolean_t cleaned = FALSE, multi_descriptor = FALSE; i = rx_ring->next_to_clean; rx_desc = E1000_RX_DESC(*rx_ring, i); while(rx_desc->status & E1000_RXD_STAT_DD) { buffer_info = &rx_ring->buffer_info[i]; + u8 status; #ifdef CONFIG_E1000_NAPI if(*work_done >= work_to_do) break; (*work_done)++; #endif + status = rx_desc->status; cleaned = TRUE; - + cleaned_count++; pci_unmap_single(pdev, buffer_info->dma, buffer_info->length, @@ -3433,18 +3626,40 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, } } - /* Good Receive */ - skb_put(skb, length - ETHERNET_FCS_SIZE); + /* code added for copybreak, this should improve + * performance for small packets with large amounts + * of reassembly being done in the stack */ +#define E1000_CB_LENGTH 256 + if ((length < E1000_CB_LENGTH) && + !rx_ring->rx_skb_top && + /* or maybe (status & E1000_RXD_STAT_EOP) && */ + !multi_descriptor) { + struct sk_buff *new_skb = + dev_alloc_skb(length + NET_IP_ALIGN); + if (new_skb) { + skb_reserve(new_skb, NET_IP_ALIGN); + new_skb->dev = netdev; + memcpy(new_skb->data - NET_IP_ALIGN, + skb->data - NET_IP_ALIGN, + length + NET_IP_ALIGN); + /* save the skb in buffer_info as good */ + buffer_info->skb = skb; + skb = new_skb; + skb_put(skb, length); + } + } + + /* end copybreak code */ /* Receive Checksum Offload */ e1000_rx_checksum(adapter, - (uint32_t)(rx_desc->status) | + (uint32_t)(status) | ((uint32_t)(rx_desc->errors) << 24), rx_desc->csum, skb); skb->protocol = eth_type_trans(skb, netdev); #ifdef CONFIG_E1000_NAPI if(unlikely(adapter->vlgrp && - (rx_desc->status & E1000_RXD_STAT_VP))) { + (status & E1000_RXD_STAT_VP))) { vlan_hwaccel_receive_skb(skb, adapter->vlgrp, le16_to_cpu(rx_desc->special) & E1000_RXD_SPC_VLAN_MASK); @@ -3462,17 +3677,26 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, } #endif /* CONFIG_E1000_NAPI */ netdev->last_rx = jiffies; - rx_ring->pkt++; +#ifdef CONFIG_E1000_MQ + rx_ring->rx_stats.packets++; + rx_ring->rx_stats.bytes += length; +#endif next_desc: rx_desc->status = 0; - buffer_info->skb = NULL; - if(unlikely(++i == rx_ring->count)) i = 0; - rx_desc = E1000_RX_DESC(*rx_ring, i); + /* return some buffers to hardware, one at a time is too slow */ + if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) { + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); + cleaned_count = 0; + } + } rx_ring->next_to_clean = i; - adapter->alloc_rx_buf(adapter, rx_ring); + + cleaned_count = E1000_DESC_UNUSED(rx_ring); + if (cleaned_count) + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); return cleaned; } @@ -3501,6 +3725,7 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, struct sk_buff *skb; unsigned int i, j; uint32_t length, staterr; + int cleaned_count = 0; boolean_t cleaned = FALSE; i = rx_ring->next_to_clean; @@ -3517,6 +3742,7 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, (*work_done)++; #endif cleaned = TRUE; + cleaned_count++; pci_unmap_single(pdev, buffer_info->dma, buffer_info->length, PCI_DMA_FROMDEVICE); @@ -3593,18 +3819,28 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, } #endif /* CONFIG_E1000_NAPI */ netdev->last_rx = jiffies; - rx_ring->pkt++; +#ifdef CONFIG_E1000_MQ + rx_ring->rx_stats.packets++; + rx_ring->rx_stats.bytes += length; +#endif next_desc: rx_desc->wb.middle.status_error &= ~0xFF; buffer_info->skb = NULL; - if(unlikely(++i == rx_ring->count)) i = 0; - rx_desc = E1000_RX_DESC_PS(*rx_ring, i); + /* return some buffers to hardware, one at a time is too slow */ + if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) { + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); + cleaned_count = 0; + } + staterr = le32_to_cpu(rx_desc->wb.middle.status_error); } rx_ring->next_to_clean = i; - adapter->alloc_rx_buf(adapter, rx_ring); + + cleaned_count = E1000_DESC_UNUSED(rx_ring); + if (cleaned_count) + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); return cleaned; } @@ -3616,7 +3852,8 @@ next_desc: static void e1000_alloc_rx_buffers(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring) + struct e1000_rx_ring *rx_ring, + int cleaned_count) { struct net_device *netdev = adapter->netdev; struct pci_dev *pdev = adapter->pdev; @@ -3629,11 +3866,18 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, i = rx_ring->next_to_use; buffer_info = &rx_ring->buffer_info[i]; - while(!buffer_info->skb) { - skb = dev_alloc_skb(bufsz); + while (cleaned_count--) { + if (!(skb = buffer_info->skb)) + skb = dev_alloc_skb(bufsz); + else { + skb_trim(skb, 0); + goto map_skb; + } + if(unlikely(!skb)) { /* Better luck next round */ + adapter->alloc_rx_buff_failed++; break; } @@ -3670,6 +3914,7 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, buffer_info->skb = skb; buffer_info->length = adapter->rx_buffer_len; +map_skb: buffer_info->dma = pci_map_single(pdev, skb->data, adapter->rx_buffer_len, @@ -3718,7 +3963,8 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, static void e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring) + struct e1000_rx_ring *rx_ring, + int cleaned_count) { struct net_device *netdev = adapter->netdev; struct pci_dev *pdev = adapter->pdev; @@ -3734,7 +3980,7 @@ e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, ps_page = &rx_ring->ps_page[i]; ps_page_dma = &rx_ring->ps_page_dma[i]; - while(!buffer_info->skb) { + while (cleaned_count--) { rx_desc = E1000_RX_DESC_PS(*rx_ring, i); for(j = 0; j < PS_PAGE_BUFFERS; j++) { @@ -4106,8 +4352,12 @@ e1000_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid) if((adapter->hw.mng_cookie.status & E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) && - (vid == adapter->mng_vlan_id)) + (vid == adapter->mng_vlan_id)) { + /* release control to f/w */ + e1000_release_hw_control(adapter); return; + } + /* remove VID from filter table */ index = (vid >> 5) & 0x7F; vfta = E1000_READ_REG_ARRAY(&adapter->hw, VFTA, index); @@ -4173,8 +4423,9 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state) { struct net_device *netdev = pci_get_drvdata(pdev); struct e1000_adapter *adapter = netdev_priv(netdev); - uint32_t ctrl, ctrl_ext, rctl, manc, status, swsm; + uint32_t ctrl, ctrl_ext, rctl, manc, status; uint32_t wufc = adapter->wol; + int retval = 0; netif_device_detach(netdev); @@ -4220,13 +4471,21 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state) E1000_WRITE_REG(&adapter->hw, WUC, E1000_WUC_PME_EN); E1000_WRITE_REG(&adapter->hw, WUFC, wufc); - pci_enable_wake(pdev, 3, 1); - pci_enable_wake(pdev, 4, 1); /* 4 == D3 cold */ + retval = pci_enable_wake(pdev, PCI_D3hot, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); } else { E1000_WRITE_REG(&adapter->hw, WUC, 0); E1000_WRITE_REG(&adapter->hw, WUFC, 0); - pci_enable_wake(pdev, 3, 0); - pci_enable_wake(pdev, 4, 0); /* 4 == D3 cold */ + retval = pci_enable_wake(pdev, PCI_D3hot, 0); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 0); /* 4 == D3 cold */ + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); } pci_save_state(pdev); @@ -4237,29 +4496,24 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state) if(manc & E1000_MANC_SMBUS_EN) { manc |= E1000_MANC_ARP_EN; E1000_WRITE_REG(&adapter->hw, MANC, manc); - pci_enable_wake(pdev, 3, 1); - pci_enable_wake(pdev, 4, 1); /* 4 == D3 cold */ + retval = pci_enable_wake(pdev, PCI_D3hot, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); } } - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm & ~E1000_SWSM_DRV_LOAD); - break; - default: - break; - } + /* Release control of h/w to f/w. If f/w is AMT enabled, this + * would have already happened in close and is redundant. */ + e1000_release_hw_control(adapter); pci_disable_device(pdev); - pci_set_power_state(pdev, pci_choose_state(pdev, state)); + + retval = pci_set_power_state(pdev, pci_choose_state(pdev, state)); + if (retval) + DPRINTK(PROBE, ERR, "Error in setting power state\n"); return 0; } @@ -4269,16 +4523,21 @@ e1000_resume(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); struct e1000_adapter *adapter = netdev_priv(netdev); - uint32_t manc, ret_val, swsm; - uint32_t ctrl_ext; + int retval; + uint32_t manc, ret_val; - pci_set_power_state(pdev, PCI_D0); - pci_restore_state(pdev); + retval = pci_set_power_state(pdev, PCI_D0); + if (retval) + DPRINTK(PROBE, ERR, "Error in setting power state\n"); ret_val = pci_enable_device(pdev); pci_set_master(pdev); - pci_enable_wake(pdev, PCI_D3hot, 0); - pci_enable_wake(pdev, PCI_D3cold, 0); + retval = pci_enable_wake(pdev, PCI_D3hot, 0); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 0); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); e1000_reset(adapter); E1000_WRITE_REG(&adapter->hw, WUS, ~0); @@ -4295,21 +4554,13 @@ e1000_resume(struct pci_dev *pdev) E1000_WRITE_REG(&adapter->hw, MANC, manc); } - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext | E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm | E1000_SWSM_DRV_LOAD); - break; - default: - break; - } + /* If the controller is 82573 and f/w is AMT, do not set + * DRV_LOAD until the interface is up. For all other cases, + * let the f/w know that the h/w is now under the control + * of the driver. */ + if (adapter->hw.mac_type != e1000_82573 || + !e1000_check_mng_mode(&adapter->hw)) + e1000_get_hw_control(adapter); return 0; } @@ -4327,6 +4578,9 @@ e1000_netpoll(struct net_device *netdev) disable_irq(adapter->pdev->irq); e1000_intr(adapter->pdev->irq, netdev, NULL); e1000_clean_tx_irq(adapter, adapter->tx_ring); +#ifndef CONFIG_E1000_NAPI + adapter->clean_rx(adapter, adapter->rx_ring); +#endif enable_irq(adapter->pdev->irq); } #endif diff --git a/drivers/net/e1000/e1000_param.c b/drivers/net/e1000/e1000_param.c index ccbbe5ad8e0f..0a7918c62557 100644 --- a/drivers/net/e1000/e1000_param.c +++ b/drivers/net/e1000/e1000_param.c @@ -177,7 +177,7 @@ E1000_PARAM(RxAbsIntDelay, "Receive Absolute Interrupt Delay"); * * Valid Range: 100-100000 (0=off, 1=dynamic) * - * Default Value: 1 + * Default Value: 8000 */ E1000_PARAM(InterruptThrottleRate, "Interrupt Throttling Rate"); @@ -320,7 +320,7 @@ e1000_check_options(struct e1000_adapter *adapter) } else { tx_ring->count = opt.def; } - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_tx_queues; i++) tx_ring[i].count = tx_ring->count; } { /* Receive Descriptor Count */ @@ -346,7 +346,7 @@ e1000_check_options(struct e1000_adapter *adapter) } else { rx_ring->count = opt.def; } - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) rx_ring[i].count = rx_ring->count; } { /* Checksum Offload Enable/Disable */ @@ -388,7 +388,7 @@ e1000_check_options(struct e1000_adapter *adapter) e1000_validate_option(&fc, &opt, adapter); adapter->hw.fc = adapter->hw.original_fc = fc; } else { - adapter->hw.fc = opt.def; + adapter->hw.fc = adapter->hw.original_fc = opt.def; } } { /* Transmit Interrupt Delay */ @@ -584,6 +584,12 @@ e1000_check_copper_options(struct e1000_adapter *adapter) .p = dplx_list }} }; + if (e1000_check_phy_reset_block(&adapter->hw)) { + DPRINTK(PROBE, INFO, + "Link active due to SoL/IDER Session. " + "Speed/Duplex/AutoNeg parameter ignored.\n"); + return; + } if (num_Duplex > bd) { dplx = Duplex[bd]; e1000_validate_option(&dplx, &opt, adapter); |