X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=drivers%2Fnet%2Fixgb%2Fixgb_main.c;h=a083a9189230d8f0a3dda97f7b3eba485862f1d1;hb=97bf2856c6014879bd04983a3e9dfcdac1e7fe85;hp=26af3785dd56f8e7fd55640c618e7ab1413dfd2b;hpb=c7b5ebbddf7bcd3651947760f423e3783bbe6573;p=linux-2.6.git diff --git a/drivers/net/ixgb/ixgb_main.c b/drivers/net/ixgb/ixgb_main.c index 26af3785d..a083a9189 100644 --- a/drivers/net/ixgb/ixgb_main.c +++ b/drivers/net/ixgb/ixgb_main.c @@ -1,27 +1,27 @@ /******************************************************************************* - - Copyright(c) 1999 - 2004 Intel Corporation. All rights reserved. - - This program is free software; you can redistribute it and/or modify it - under the terms of the GNU General Public License as published by the Free - Software Foundation; either version 2 of the License, or (at your option) - any later version. - - This program is distributed in the hope that it will be useful, but WITHOUT - ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or - FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + Intel PRO/10GbE Linux driver + Copyright(c) 1999 - 2006 Intel Corporation. + + This program is free software; you can redistribute it and/or modify it + under the terms and conditions of the GNU General Public License, + version 2, as published by the Free Software Foundation. + + This program is distributed in the hope it will be useful, but WITHOUT + ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. - + You should have received a copy of the GNU General Public License along with - this program; if not, write to the Free Software Foundation, Inc., 59 - Temple Place - Suite 330, Boston, MA 02111-1307, USA. - - The full GNU General Public License is included in this distribution in the - file called LICENSE. - + this program; if not, write to the Free Software Foundation, Inc., + 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + + The full GNU General Public License is included in this distribution in + the file called "COPYING". + Contact Information: Linux NICS + e1000-devel Mailing List Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497 *******************************************************************************/ @@ -29,9 +29,16 @@ #include "ixgb.h" char ixgb_driver_name[] = "ixgb"; -char ixgb_driver_string[] = "Intel(R) PRO/10GbE Network Driver"; -char ixgb_driver_version[] = "1.0.66"; -char ixgb_copyright[] = "Copyright (c) 2001-2004 Intel Corporation."; +static char ixgb_driver_string[] = "Intel(R) PRO/10GbE Network Driver"; + +#ifndef CONFIG_IXGB_NAPI +#define DRIVERNAPI +#else +#define DRIVERNAPI "-NAPI" +#endif +#define DRV_VERSION "1.0.126-k2"DRIVERNAPI +char ixgb_driver_version[] = DRV_VERSION; +static char ixgb_copyright[] = "Copyright (c) 1999-2006 Intel Corporation."; /* ixgb_pci_tbl - PCI Device ID Table * @@ -44,8 +51,12 @@ char ixgb_copyright[] = "Copyright (c) 2001-2004 Intel Corporation."; static struct pci_device_id ixgb_pci_tbl[] = { {INTEL_VENDOR_ID, IXGB_DEVICE_ID_82597EX, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + {INTEL_VENDOR_ID, IXGB_DEVICE_ID_82597EX_CX4, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, {INTEL_VENDOR_ID, IXGB_DEVICE_ID_82597EX_SR, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + {INTEL_VENDOR_ID, IXGB_DEVICE_ID_82597EX_LR, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, /* required last entry */ {0,} @@ -55,11 +66,14 @@ MODULE_DEVICE_TABLE(pci, ixgb_pci_tbl); /* Local Function Prototypes */ -static inline void ixgb_irq_disable(struct ixgb_adapter *adapter); -static inline void ixgb_irq_enable(struct ixgb_adapter *adapter); int ixgb_up(struct ixgb_adapter *adapter); void ixgb_down(struct ixgb_adapter *adapter, boolean_t kill_watchdog); void ixgb_reset(struct ixgb_adapter *adapter); +int ixgb_setup_tx_resources(struct ixgb_adapter *adapter); +int ixgb_setup_rx_resources(struct ixgb_adapter *adapter); +void ixgb_free_tx_resources(struct ixgb_adapter *adapter); +void ixgb_free_rx_resources(struct ixgb_adapter *adapter); +void ixgb_update_stats(struct ixgb_adapter *adapter); static int ixgb_init_module(void); static void ixgb_exit_module(void); @@ -68,27 +82,20 @@ static void __devexit ixgb_remove(struct pci_dev *pdev); static int ixgb_sw_init(struct ixgb_adapter *adapter); static int ixgb_open(struct net_device *netdev); static int ixgb_close(struct net_device *netdev); -static int ixgb_setup_tx_resources(struct ixgb_adapter *adapter); -static int ixgb_setup_rx_resources(struct ixgb_adapter *adapter); static void ixgb_configure_tx(struct ixgb_adapter *adapter); static void ixgb_configure_rx(struct ixgb_adapter *adapter); static void ixgb_setup_rctl(struct ixgb_adapter *adapter); static void ixgb_clean_tx_ring(struct ixgb_adapter *adapter); static void ixgb_clean_rx_ring(struct ixgb_adapter *adapter); -static void ixgb_free_tx_resources(struct ixgb_adapter *adapter); -static void ixgb_free_rx_resources(struct ixgb_adapter *adapter); static void ixgb_set_multi(struct net_device *netdev); static void ixgb_watchdog(unsigned long data); static int ixgb_xmit_frame(struct sk_buff *skb, struct net_device *netdev); static struct net_device_stats *ixgb_get_stats(struct net_device *netdev); static int ixgb_change_mtu(struct net_device *netdev, int new_mtu); static int ixgb_set_mac(struct net_device *netdev, void *p); -static void ixgb_update_stats(struct ixgb_adapter *adapter); -static irqreturn_t ixgb_intr(int irq, void *data, struct pt_regs *regs); +static irqreturn_t ixgb_intr(int irq, void *data); static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter); -static inline void ixgb_rx_checksum(struct ixgb_adapter *adapter, - struct ixgb_rx_desc *rx_desc, - struct sk_buff *skb); + #ifdef CONFIG_IXGB_NAPI static int ixgb_clean(struct net_device *netdev, int *budget); static boolean_t ixgb_clean_rx_irq(struct ixgb_adapter *adapter, @@ -97,54 +104,59 @@ static boolean_t ixgb_clean_rx_irq(struct ixgb_adapter *adapter, static boolean_t ixgb_clean_rx_irq(struct ixgb_adapter *adapter); #endif static void ixgb_alloc_rx_buffers(struct ixgb_adapter *adapter); -static int ixgb_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd); +void ixgb_set_ethtool_ops(struct net_device *netdev); static void ixgb_tx_timeout(struct net_device *dev); -static void ixgb_tx_timeout_task(struct net_device *dev); +static void ixgb_tx_timeout_task(struct work_struct *work); static void ixgb_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp); static void ixgb_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid); static void ixgb_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid); static void ixgb_restore_vlan(struct ixgb_adapter *adapter); -static int ixgb_notify_reboot(struct notifier_block *, unsigned long event, - void *ptr); -static int ixgb_suspend(struct pci_dev *pdev, uint32_t state); - #ifdef CONFIG_NET_POLL_CONTROLLER /* for netdump / net console */ static void ixgb_netpoll(struct net_device *dev); #endif -struct notifier_block ixgb_notifier_reboot = { - .notifier_call = ixgb_notify_reboot, - .next = NULL, - .priority = 0 -}; +static pci_ers_result_t ixgb_io_error_detected (struct pci_dev *pdev, + enum pci_channel_state state); +static pci_ers_result_t ixgb_io_slot_reset (struct pci_dev *pdev); +static void ixgb_io_resume (struct pci_dev *pdev); /* Exported from other modules */ - extern void ixgb_check_options(struct ixgb_adapter *adapter); -extern int ixgb_ethtool_ioctl(struct net_device *netdev, struct ifreq *ifr); + +static struct pci_error_handlers ixgb_err_handler = { + .error_detected = ixgb_io_error_detected, + .slot_reset = ixgb_io_slot_reset, + .resume = ixgb_io_resume, +}; static struct pci_driver ixgb_driver = { - .name = ixgb_driver_name, + .name = ixgb_driver_name, .id_table = ixgb_pci_tbl, - .probe = ixgb_probe, - .remove = __devexit_p(ixgb_remove), - /* Power Managment Hooks */ - .suspend = NULL, - .resume = NULL + .probe = ixgb_probe, + .remove = __devexit_p(ixgb_remove), + .err_handler = &ixgb_err_handler }; MODULE_AUTHOR("Intel Corporation, "); MODULE_DESCRIPTION("Intel(R) PRO/10GbE Network Driver"); MODULE_LICENSE("GPL"); +MODULE_VERSION(DRV_VERSION); + +#define DEFAULT_DEBUG_LEVEL_SHIFT 3 +static int debug = DEFAULT_DEBUG_LEVEL_SHIFT; +module_param(debug, int, 0); +MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)"); /* some defines for controlling descriptor fetches in h/w */ -#define RXDCTL_PTHRESH_DEFAULT 128 /* chip considers prefech below this */ -#define RXDCTL_HTHRESH_DEFAULT 16 /* chip will only prefetch if tail is - pushed this many descriptors from head */ -#define RXDCTL_WTHRESH_DEFAULT 16 /* chip writes back at this many or RXT0 */ +#define RXDCTL_WTHRESH_DEFAULT 15 /* chip writes back at this many or RXT0 */ +#define RXDCTL_PTHRESH_DEFAULT 0 /* chip considers prefech below + * this */ +#define RXDCTL_HTHRESH_DEFAULT 0 /* chip will only prefetch if tail + * is pushed this many descriptors + * from head */ /** * ixgb_init_module - Driver Registration Routine @@ -153,19 +165,15 @@ MODULE_LICENSE("GPL"); * loaded. All it does is register with the PCI subsystem. **/ -static int __init ixgb_init_module(void) +static int __init +ixgb_init_module(void) { - int ret; printk(KERN_INFO "%s - version %s\n", ixgb_driver_string, ixgb_driver_version); printk(KERN_INFO "%s\n", ixgb_copyright); - ret = pci_module_init(&ixgb_driver); - if (ret >= 0) { - register_reboot_notifier(&ixgb_notifier_reboot); - } - return ret; + return pci_register_driver(&ixgb_driver); } module_init(ixgb_init_module); @@ -177,9 +185,9 @@ module_init(ixgb_init_module); * from memory. **/ -static void __exit ixgb_exit_module(void) +static void __exit +ixgb_exit_module(void) { - unregister_reboot_notifier(&ixgb_notifier_reboot); pci_unregister_driver(&ixgb_driver); } @@ -190,7 +198,8 @@ module_exit(ixgb_exit_module); * @adapter: board private structure **/ -static inline void ixgb_irq_disable(struct ixgb_adapter *adapter) +static void +ixgb_irq_disable(struct ixgb_adapter *adapter) { atomic_inc(&adapter->irq_sem); IXGB_WRITE_REG(&adapter->hw, IMC, ~0); @@ -203,17 +212,19 @@ static inline void ixgb_irq_disable(struct ixgb_adapter *adapter) * @adapter: board private structure **/ -static inline void ixgb_irq_enable(struct ixgb_adapter *adapter) +static void +ixgb_irq_enable(struct ixgb_adapter *adapter) { - if (atomic_dec_and_test(&adapter->irq_sem)) { + if(atomic_dec_and_test(&adapter->irq_sem)) { IXGB_WRITE_REG(&adapter->hw, IMS, IXGB_INT_RXT0 | IXGB_INT_RXDMT0 | IXGB_INT_TXDW | - IXGB_INT_RXO | IXGB_INT_LSC); + IXGB_INT_LSC); IXGB_WRITE_FLUSH(&adapter->hw); } } -int ixgb_up(struct ixgb_adapter *adapter) +int +ixgb_up(struct ixgb_adapter *adapter) { struct net_device *netdev = adapter->netdev; int err; @@ -222,6 +233,7 @@ int ixgb_up(struct ixgb_adapter *adapter) /* hardware has been reset, we need to reload some things */ + ixgb_rar_set(hw, netdev->dev_addr, 0); ixgb_set_multi(netdev); ixgb_restore_vlan(adapter); @@ -231,27 +243,47 @@ int ixgb_up(struct ixgb_adapter *adapter) ixgb_configure_rx(adapter); ixgb_alloc_rx_buffers(adapter); - if ((err = request_irq(adapter->pdev->irq, &ixgb_intr, - SA_SHIRQ | SA_SAMPLE_RANDOM, - netdev->name, netdev))) - return err; - /* disable interrupts and get the hardware into a known state */ IXGB_WRITE_REG(&adapter->hw, IMC, 0xffffffff); - if ((hw->max_frame_size != max_frame) || - (hw->max_frame_size != - (IXGB_READ_REG(hw, MFS) >> IXGB_MFS_SHIFT))) { +#ifdef CONFIG_PCI_MSI + { + boolean_t pcix = (IXGB_READ_REG(&adapter->hw, STATUS) & + IXGB_STATUS_PCIX_MODE) ? TRUE : FALSE; + adapter->have_msi = TRUE; + + if (!pcix) + adapter->have_msi = FALSE; + else if((err = pci_enable_msi(adapter->pdev))) { + DPRINTK(PROBE, ERR, + "Unable to allocate MSI interrupt Error: %d\n", err); + adapter->have_msi = FALSE; + /* proceed to try to request regular interrupt */ + } + } + +#endif + if((err = request_irq(adapter->pdev->irq, &ixgb_intr, + IRQF_SHARED | IRQF_SAMPLE_RANDOM, + netdev->name, netdev))) { + DPRINTK(PROBE, ERR, + "Unable to allocate interrupt Error: %d\n", err); + return err; + } + + if((hw->max_frame_size != max_frame) || + (hw->max_frame_size != + (IXGB_READ_REG(hw, MFS) >> IXGB_MFS_SHIFT))) { hw->max_frame_size = max_frame; IXGB_WRITE_REG(hw, MFS, hw->max_frame_size << IXGB_MFS_SHIFT); - if (hw->max_frame_size > - IXGB_MAX_ENET_FRAME_SIZE_WITHOUT_FCS + ENET_FCS_LENGTH) { + if(hw->max_frame_size > + IXGB_MAX_ENET_FRAME_SIZE_WITHOUT_FCS + ENET_FCS_LENGTH) { uint32_t ctrl0 = IXGB_READ_REG(hw, CTRL0); - if (!(ctrl0 & IXGB_CTRL0_JFE)) { + if(!(ctrl0 & IXGB_CTRL0_JFE)) { ctrl0 |= IXGB_CTRL0_JFE; IXGB_WRITE_REG(hw, CTRL0, ctrl0); } @@ -259,19 +291,32 @@ int ixgb_up(struct ixgb_adapter *adapter) } mod_timer(&adapter->watchdog_timer, jiffies); + +#ifdef CONFIG_IXGB_NAPI + netif_poll_enable(netdev); +#endif ixgb_irq_enable(adapter); return 0; } -void ixgb_down(struct ixgb_adapter *adapter, boolean_t kill_watchdog) +void +ixgb_down(struct ixgb_adapter *adapter, boolean_t kill_watchdog) { struct net_device *netdev = adapter->netdev; ixgb_irq_disable(adapter); free_irq(adapter->pdev->irq, netdev); - if (kill_watchdog) +#ifdef CONFIG_PCI_MSI + if(adapter->have_msi == TRUE) + pci_disable_msi(adapter->pdev); + +#endif + if(kill_watchdog) del_timer_sync(&adapter->watchdog_timer); +#ifdef CONFIG_IXGB_NAPI + netif_poll_disable(netdev); +#endif adapter->link_speed = 0; adapter->link_duplex = 0; netif_carrier_off(netdev); @@ -282,12 +327,13 @@ void ixgb_down(struct ixgb_adapter *adapter, boolean_t kill_watchdog) ixgb_clean_rx_ring(adapter); } -void ixgb_reset(struct ixgb_adapter *adapter) +void +ixgb_reset(struct ixgb_adapter *adapter) { ixgb_adapter_stop(&adapter->hw); - if (!ixgb_init_hw(&adapter->hw)) - IXGB_DBG("ixgb_init_hw failed.\n"); + if(!ixgb_init_hw(&adapter->hw)) + DPRINTK(PROBE, ERR, "ixgb_init_hw failed.\n"); } /** @@ -303,7 +349,8 @@ void ixgb_reset(struct ixgb_adapter *adapter) **/ static int __devinit -ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) +ixgb_probe(struct pci_dev *pdev, + const struct pci_device_id *ent) { struct net_device *netdev = NULL; struct ixgb_adapter *adapter; @@ -314,26 +361,29 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) int i; int err; - if ((err = pci_enable_device(pdev))) + if((err = pci_enable_device(pdev))) return err; - if (!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) { + if(!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK)) && + !(err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK))) { pci_using_dac = 1; } else { - if ((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK))) { - IXGB_ERR("No usable DMA configuration, aborting\n"); - return err; + if((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK)) || + (err = pci_set_consistent_dma_mask(pdev, DMA_32BIT_MASK))) { + printk(KERN_ERR + "ixgb: No usable DMA configuration, aborting\n"); + goto err_dma_mask; } pci_using_dac = 0; } - if ((err = pci_request_regions(pdev, ixgb_driver_name))) - return err; + if((err = pci_request_regions(pdev, ixgb_driver_name))) + goto err_request_regions; pci_set_master(pdev); netdev = alloc_etherdev(sizeof(struct ixgb_adapter)); - if (!netdev) { + if(!netdev) { err = -ENOMEM; goto err_alloc_etherdev; } @@ -342,24 +392,25 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) SET_NETDEV_DEV(netdev, &pdev->dev); pci_set_drvdata(pdev, netdev); - adapter = netdev->priv; + adapter = netdev_priv(netdev); adapter->netdev = netdev; adapter->pdev = pdev; adapter->hw.back = adapter; + adapter->msg_enable = netif_msg_init(debug, DEFAULT_DEBUG_LEVEL_SHIFT); mmio_start = pci_resource_start(pdev, BAR_0); mmio_len = pci_resource_len(pdev, BAR_0); adapter->hw.hw_addr = ioremap(mmio_start, mmio_len); - if (!adapter->hw.hw_addr) { + if(!adapter->hw.hw_addr) { err = -EIO; goto err_ioremap; } - for (i = BAR_1; i <= BAR_5; i++) { - if (pci_resource_len(pdev, i) == 0) + for(i = BAR_1; i <= BAR_5; i++) { + if(pci_resource_len(pdev, i) == 0) continue; - if (pci_resource_flags(pdev, i) & IORESOURCE_IO) { + if(pci_resource_flags(pdev, i) & IORESOURCE_IO) { adapter->hw.io_base = pci_resource_start(pdev, i); break; } @@ -372,9 +423,9 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) netdev->set_multicast_list = &ixgb_set_multi; netdev->set_mac_address = &ixgb_set_mac; netdev->change_mtu = &ixgb_change_mtu; - netdev->do_ioctl = &ixgb_ioctl; + ixgb_set_ethtool_ops(netdev); netdev->tx_timeout = &ixgb_tx_timeout; - netdev->watchdog_timeo = HZ; + netdev->watchdog_timeo = 5 * HZ; #ifdef CONFIG_IXGB_NAPI netdev->poll = &ixgb_clean; netdev->weight = 64; @@ -386,6 +437,7 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) netdev->poll_controller = ixgb_netpoll; #endif + strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1); netdev->mem_start = mmio_start; netdev->mem_end = mmio_start + mmio_len; netdev->base_addr = adapter->hw.io_base; @@ -396,30 +448,37 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) /* setup the private structure */ - if ((err = ixgb_sw_init(adapter))) + if((err = ixgb_sw_init(adapter))) goto err_sw_init; netdev->features = NETIF_F_SG | - NETIF_F_HW_CSUM | - NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_RX | NETIF_F_HW_VLAN_FILTER; + NETIF_F_HW_CSUM | + NETIF_F_HW_VLAN_TX | + NETIF_F_HW_VLAN_RX | + NETIF_F_HW_VLAN_FILTER; #ifdef NETIF_F_TSO netdev->features |= NETIF_F_TSO; #endif +#ifdef NETIF_F_LLTX + netdev->features |= NETIF_F_LLTX; +#endif - if (pci_using_dac) + if(pci_using_dac) netdev->features |= NETIF_F_HIGHDMA; /* make sure the EEPROM is good */ - if (!ixgb_validate_eeprom_checksum(&adapter->hw)) { - printk(KERN_ERR "The EEPROM Checksum Is Not Valid\n"); + if(!ixgb_validate_eeprom_checksum(&adapter->hw)) { + DPRINTK(PROBE, ERR, "The EEPROM Checksum Is Not Valid\n"); err = -EIO; goto err_eeprom; } ixgb_get_ee_mac_addr(&adapter->hw, netdev->dev_addr); + memcpy(netdev->perm_addr, netdev->dev_addr, netdev->addr_len); - if (!is_valid_ether_addr(netdev->dev_addr)) { + if(!is_valid_ether_addr(netdev->perm_addr)) { + DPRINTK(PROBE, ERR, "Invalid MAC Address\n"); err = -EIO; goto err_eeprom; } @@ -430,10 +489,10 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) adapter->watchdog_timer.function = &ixgb_watchdog; adapter->watchdog_timer.data = (unsigned long)adapter; - INIT_WORK(&adapter->tx_timeout_task, - (void (*)(void *))ixgb_tx_timeout_task, netdev); + INIT_WORK(&adapter->tx_timeout_task, ixgb_tx_timeout_task); - if ((err = register_netdev(netdev))) + strcpy(netdev->name, "eth%d"); + if((err = register_netdev(netdev))) goto err_register; /* we're going to reset, so assume we have no link for now */ @@ -441,8 +500,7 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) netif_carrier_off(netdev); netif_stop_queue(netdev); - printk(KERN_INFO "%s: Intel(R) PRO/10GbE Network Connection\n", - netdev->name); + DPRINTK(PROBE, INFO, "Intel(R) PRO/10GbE Network Connection\n"); ixgb_check_options(adapter); /* reset the hardware with the new settings */ @@ -451,14 +509,17 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) cards_found++; return 0; - err_register: - err_sw_init: - err_eeprom: +err_register: +err_sw_init: +err_eeprom: iounmap(adapter->hw.hw_addr); - err_ioremap: +err_ioremap: free_netdev(netdev); - err_alloc_etherdev: +err_alloc_etherdev: pci_release_regions(pdev); +err_request_regions: +err_dma_mask: + pci_disable_device(pdev); return err; } @@ -472,10 +533,11 @@ ixgb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) * memory. **/ -static void __devexit ixgb_remove(struct pci_dev *pdev) +static void __devexit +ixgb_remove(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); unregister_netdev(netdev); @@ -494,7 +556,8 @@ static void __devexit ixgb_remove(struct pci_dev *pdev) * OS network device settings (MTU size). **/ -static int __devinit ixgb_sw_init(struct ixgb_adapter *adapter) +static int __devinit +ixgb_sw_init(struct ixgb_adapter *adapter) { struct ixgb_hw *hw = &adapter->hw; struct net_device *netdev = adapter->netdev; @@ -507,16 +570,17 @@ static int __devinit ixgb_sw_init(struct ixgb_adapter *adapter) hw->subsystem_vendor_id = pdev->subsystem_vendor; hw->subsystem_id = pdev->subsystem_device; - adapter->rx_buffer_len = IXGB_RXBUFFER_2048; - hw->max_frame_size = netdev->mtu + ENET_HEADER_SIZE + ENET_FCS_LENGTH; + adapter->rx_buffer_len = hw->max_frame_size; - if ((hw->device_id == IXGB_DEVICE_ID_82597EX) - || (hw->device_id == IXGB_DEVICE_ID_82597EX_SR)) - hw->mac_type = ixgb_82597; + if((hw->device_id == IXGB_DEVICE_ID_82597EX) + || (hw->device_id == IXGB_DEVICE_ID_82597EX_CX4) + || (hw->device_id == IXGB_DEVICE_ID_82597EX_LR) + || (hw->device_id == IXGB_DEVICE_ID_82597EX_SR)) + hw->mac_type = ixgb_82597; else { /* should never have loaded on this device */ - printk(KERN_ERR "ixgb: unsupported device id\n"); + DPRINTK(PROBE, ERR, "unsupported device id\n"); } /* enable flow control to be programmed */ @@ -541,31 +605,32 @@ static int __devinit ixgb_sw_init(struct ixgb_adapter *adapter) * and the stack is notified that the interface is ready. **/ -static int ixgb_open(struct net_device *netdev) +static int +ixgb_open(struct net_device *netdev) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); int err; /* allocate transmit descriptors */ - if ((err = ixgb_setup_tx_resources(adapter))) + if((err = ixgb_setup_tx_resources(adapter))) goto err_setup_tx; /* allocate receive descriptors */ - if ((err = ixgb_setup_rx_resources(adapter))) + if((err = ixgb_setup_rx_resources(adapter))) goto err_setup_rx; - if ((err = ixgb_up(adapter))) + if((err = ixgb_up(adapter))) goto err_up; return 0; - err_up: +err_up: ixgb_free_rx_resources(adapter); - err_setup_rx: +err_setup_rx: ixgb_free_tx_resources(adapter); - err_setup_tx: +err_setup_tx: ixgb_reset(adapter); return err; @@ -583,9 +648,10 @@ static int ixgb_open(struct net_device *netdev) * hardware, and all transmit and receive resources are freed. **/ -static int ixgb_close(struct net_device *netdev) +static int +ixgb_close(struct net_device *netdev) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); ixgb_down(adapter, TRUE); @@ -602,15 +668,18 @@ static int ixgb_close(struct net_device *netdev) * Return 0 on success, negative on failure **/ -static int ixgb_setup_tx_resources(struct ixgb_adapter *adapter) +int +ixgb_setup_tx_resources(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *txdr = &adapter->tx_ring; struct pci_dev *pdev = adapter->pdev; int size; size = sizeof(struct ixgb_buffer) * txdr->count; - txdr->buffer_info = kmalloc(size, GFP_KERNEL); - if (!txdr->buffer_info) { + txdr->buffer_info = vmalloc(size); + if(!txdr->buffer_info) { + DPRINTK(PROBE, ERR, + "Unable to allocate transmit descriptor ring memory\n"); return -ENOMEM; } memset(txdr->buffer_info, 0, size); @@ -621,8 +690,10 @@ static int ixgb_setup_tx_resources(struct ixgb_adapter *adapter) IXGB_ROUNDUP(txdr->size, 4096); txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma); - if (!txdr->desc) { - kfree(txdr->buffer_info); + if(!txdr->desc) { + vfree(txdr->buffer_info); + DPRINTK(PROBE, ERR, + "Unable to allocate transmit descriptor memory\n"); return -ENOMEM; } memset(txdr->desc, 0, txdr->size); @@ -640,7 +711,8 @@ static int ixgb_setup_tx_resources(struct ixgb_adapter *adapter) * Configure the Tx unit of the MAC after a reset. **/ -static void ixgb_configure_tx(struct ixgb_adapter *adapter) +static void +ixgb_configure_tx(struct ixgb_adapter *adapter) { uint64_t tdba = adapter->tx_ring.dma; uint32_t tdlen = adapter->tx_ring.count * sizeof(struct ixgb_tx_desc); @@ -661,14 +733,8 @@ static void ixgb_configure_tx(struct ixgb_adapter *adapter) IXGB_WRITE_REG(hw, TDH, 0); IXGB_WRITE_REG(hw, TDT, 0); - /* don't set up txdctl, it induces performance problems if - * configured incorrectly - txdctl = TXDCTL_PTHRESH_DEFAULT; // prefetch txds below this threshold - txdctl |= (TXDCTL_HTHRESH_DEFAULT // only prefetch if there are this many ready - << IXGB_TXDCTL_HTHRESH_SHIFT); - IXGB_WRITE_REG (hw, TXDCTL, txdctl); - */ - + /* don't set up txdctl, it induces performance problems if configured + * incorrectly */ /* Set the Tx Interrupt Delay register */ IXGB_WRITE_REG(hw, TIDV, adapter->tx_int_delay); @@ -680,8 +746,8 @@ static void ixgb_configure_tx(struct ixgb_adapter *adapter) /* Setup Transmit Descriptor Settings for this adapter */ adapter->tx_cmd_type = - IXGB_TX_DESC_TYPE - | (adapter->tx_int_delay_enable ? IXGB_TX_DESC_CMD_IDE : 0); + IXGB_TX_DESC_TYPE + | (adapter->tx_int_delay_enable ? IXGB_TX_DESC_CMD_IDE : 0); } /** @@ -691,15 +757,18 @@ static void ixgb_configure_tx(struct ixgb_adapter *adapter) * Returns 0 on success, negative on failure **/ -static int ixgb_setup_rx_resources(struct ixgb_adapter *adapter) +int +ixgb_setup_rx_resources(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *rxdr = &adapter->rx_ring; struct pci_dev *pdev = adapter->pdev; int size; size = sizeof(struct ixgb_buffer) * rxdr->count; - rxdr->buffer_info = kmalloc(size, GFP_KERNEL); - if (!rxdr->buffer_info) { + rxdr->buffer_info = vmalloc(size); + if(!rxdr->buffer_info) { + DPRINTK(PROBE, ERR, + "Unable to allocate receive descriptor ring\n"); return -ENOMEM; } memset(rxdr->buffer_info, 0, size); @@ -711,8 +780,10 @@ static int ixgb_setup_rx_resources(struct ixgb_adapter *adapter) rxdr->desc = pci_alloc_consistent(pdev, rxdr->size, &rxdr->dma); - if (!rxdr->desc) { - kfree(rxdr->buffer_info); + if(!rxdr->desc) { + vfree(rxdr->buffer_info); + DPRINTK(PROBE, ERR, + "Unable to allocate receive descriptors\n"); return -ENOMEM; } memset(rxdr->desc, 0, rxdr->size); @@ -728,7 +799,8 @@ static int ixgb_setup_rx_resources(struct ixgb_adapter *adapter) * @adapter: Board private structure **/ -static void ixgb_setup_rctl(struct ixgb_adapter *adapter) +static void +ixgb_setup_rctl(struct ixgb_adapter *adapter) { uint32_t rctl; @@ -737,27 +809,20 @@ static void ixgb_setup_rctl(struct ixgb_adapter *adapter) rctl &= ~(3 << IXGB_RCTL_MO_SHIFT); rctl |= - IXGB_RCTL_BAM | IXGB_RCTL_RDMTS_1_2 | - IXGB_RCTL_RXEN | IXGB_RCTL_CFF | - (adapter->hw.mc_filter_type << IXGB_RCTL_MO_SHIFT); + IXGB_RCTL_BAM | IXGB_RCTL_RDMTS_1_2 | + IXGB_RCTL_RXEN | IXGB_RCTL_CFF | + (adapter->hw.mc_filter_type << IXGB_RCTL_MO_SHIFT); rctl |= IXGB_RCTL_SECRC; - switch (adapter->rx_buffer_len) { - case IXGB_RXBUFFER_2048: - default: + if (adapter->rx_buffer_len <= IXGB_RXBUFFER_2048) rctl |= IXGB_RCTL_BSIZE_2048; - break; - case IXGB_RXBUFFER_4096: + else if (adapter->rx_buffer_len <= IXGB_RXBUFFER_4096) rctl |= IXGB_RCTL_BSIZE_4096; - break; - case IXGB_RXBUFFER_8192: + else if (adapter->rx_buffer_len <= IXGB_RXBUFFER_8192) rctl |= IXGB_RCTL_BSIZE_8192; - break; - case IXGB_RXBUFFER_16384: + else if (adapter->rx_buffer_len <= IXGB_RXBUFFER_16384) rctl |= IXGB_RCTL_BSIZE_16384; - break; - } IXGB_WRITE_REG(&adapter->hw, RCTL, rctl); } @@ -769,7 +834,8 @@ static void ixgb_setup_rctl(struct ixgb_adapter *adapter) * Configure the Rx unit of the MAC after a reset. **/ -static void ixgb_configure_rx(struct ixgb_adapter *adapter) +static void +ixgb_configure_rx(struct ixgb_adapter *adapter) { uint64_t rdba = adapter->rx_ring.dma; uint32_t rdlen = adapter->rx_ring.count * sizeof(struct ixgb_rx_desc); @@ -798,51 +864,21 @@ static void ixgb_configure_rx(struct ixgb_adapter *adapter) IXGB_WRITE_REG(hw, RDH, 0); IXGB_WRITE_REG(hw, RDT, 0); - /* burst 16 or burst when RXT0 */ - rxdctl = RXDCTL_WTHRESH_DEFAULT << IXGB_RXDCTL_WTHRESH_SHIFT - | RXDCTL_HTHRESH_DEFAULT << IXGB_RXDCTL_HTHRESH_SHIFT - | RXDCTL_PTHRESH_DEFAULT << IXGB_RXDCTL_PTHRESH_SHIFT; + /* set up pre-fetching of receive buffers so we get some before we + * run out (default hardware behavior is to run out before fetching + * more). This sets up to fetch if HTHRESH rx descriptors are avail + * and the descriptors in hw cache are below PTHRESH. This avoids + * the hardware behavior of fetching <=512 descriptors in a single + * burst that pre-empts all other activity, usually causing fifo + * overflows. */ + /* use WTHRESH to burst write 16 descriptors or burst when RXT0 */ + rxdctl = RXDCTL_WTHRESH_DEFAULT << IXGB_RXDCTL_WTHRESH_SHIFT | + RXDCTL_HTHRESH_DEFAULT << IXGB_RXDCTL_HTHRESH_SHIFT | + RXDCTL_PTHRESH_DEFAULT << IXGB_RXDCTL_PTHRESH_SHIFT; IXGB_WRITE_REG(hw, RXDCTL, rxdctl); - if (adapter->raidc) { - uint32_t raidc; - uint8_t poll_threshold; - - /* Poll every rx_int_delay period, if RBD exists - * Receive Backlog Detection is set to - * Rx Descriptors - * max is 0x3F == set to poll when 504 RxDesc left - * min is 0 */ - - /* polling times are 1 == 0.8192us - 2 == 1.6384us - 3 == 3.2768us etc - ... - 511 == 418 us - */ -#define IXGB_RAIDC_POLL_DEFAULT 122 /* set to poll every ~100 us under load - also known as 10000 interrupts / sec */ - - /* divide this by 2^3 (8) to get a register size count */ - poll_threshold = ((adapter->rx_ring.count - 1) >> 3); - /* poll at half of that size */ - poll_threshold >>= 1; - /* make sure its not bigger than our max */ - poll_threshold &= 0x3F; - - raidc = IXGB_RAIDC_EN | /* turn on raidc style moderation */ - IXGB_RAIDC_RXT_GATE | /* don't interrupt with rxt0 while - in RBD mode (polling) */ - (IXGB_RAIDC_POLL_DEFAULT << IXGB_RAIDC_POLL_SHIFT) | - /* this sets the regular "min interrupt delay" */ - (adapter->rx_int_delay << IXGB_RAIDC_DELAY_SHIFT) | - poll_threshold; - - IXGB_WRITE_REG(hw, RAIDC, raidc); - } - /* Enable Receive Checksum Offload for TCP and UDP */ - if (adapter->rx_csum == TRUE) { + if(adapter->rx_csum == TRUE) { rxcsum = IXGB_READ_REG(hw, RXCSUM); rxcsum |= IXGB_RXCSUM_TUOFL; IXGB_WRITE_REG(hw, RXCSUM, rxcsum); @@ -860,13 +896,14 @@ static void ixgb_configure_rx(struct ixgb_adapter *adapter) * Free all transmit software resources **/ -static void ixgb_free_tx_resources(struct ixgb_adapter *adapter) +void +ixgb_free_tx_resources(struct ixgb_adapter *adapter) { struct pci_dev *pdev = adapter->pdev; ixgb_clean_tx_ring(adapter); - kfree(adapter->tx_ring.buffer_info); + vfree(adapter->tx_ring.buffer_info); adapter->tx_ring.buffer_info = NULL; pci_free_consistent(pdev, adapter->tx_ring.size, @@ -875,33 +912,45 @@ static void ixgb_free_tx_resources(struct ixgb_adapter *adapter) adapter->tx_ring.desc = NULL; } +static void +ixgb_unmap_and_free_tx_resource(struct ixgb_adapter *adapter, + struct ixgb_buffer *buffer_info) +{ + struct pci_dev *pdev = adapter->pdev; + + if (buffer_info->dma) + pci_unmap_page(pdev, buffer_info->dma, buffer_info->length, + PCI_DMA_TODEVICE); + + if (buffer_info->skb) + dev_kfree_skb_any(buffer_info->skb); + + buffer_info->skb = NULL; + buffer_info->dma = 0; + buffer_info->time_stamp = 0; + /* these fields must always be initialized in tx + * buffer_info->length = 0; + * buffer_info->next_to_watch = 0; */ +} + /** * ixgb_clean_tx_ring - Free Tx Buffers * @adapter: board private structure **/ -static void ixgb_clean_tx_ring(struct ixgb_adapter *adapter) +static void +ixgb_clean_tx_ring(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *tx_ring = &adapter->tx_ring; struct ixgb_buffer *buffer_info; - struct pci_dev *pdev = adapter->pdev; unsigned long size; unsigned int i; /* Free all the Tx ring sk_buffs */ - for (i = 0; i < tx_ring->count; i++) { + for(i = 0; i < tx_ring->count; i++) { buffer_info = &tx_ring->buffer_info[i]; - if (buffer_info->skb) { - - pci_unmap_page(pdev, - buffer_info->dma, - buffer_info->length, PCI_DMA_TODEVICE); - - dev_kfree_skb(buffer_info->skb); - - buffer_info->skb = NULL; - } + ixgb_unmap_and_free_tx_resource(adapter, buffer_info); } size = sizeof(struct ixgb_buffer) * tx_ring->count; @@ -925,14 +974,15 @@ static void ixgb_clean_tx_ring(struct ixgb_adapter *adapter) * Free all receive software resources **/ -static void ixgb_free_rx_resources(struct ixgb_adapter *adapter) +void +ixgb_free_rx_resources(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *rx_ring = &adapter->rx_ring; struct pci_dev *pdev = adapter->pdev; ixgb_clean_rx_ring(adapter); - kfree(rx_ring->buffer_info); + vfree(rx_ring->buffer_info); rx_ring->buffer_info = NULL; pci_free_consistent(pdev, rx_ring->size, rx_ring->desc, rx_ring->dma); @@ -945,7 +995,8 @@ static void ixgb_free_rx_resources(struct ixgb_adapter *adapter) * @adapter: board private structure **/ -static void ixgb_clean_rx_ring(struct ixgb_adapter *adapter) +static void +ixgb_clean_rx_ring(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *rx_ring = &adapter->rx_ring; struct ixgb_buffer *buffer_info; @@ -955,9 +1006,9 @@ static void ixgb_clean_rx_ring(struct ixgb_adapter *adapter) /* Free all the Rx ring sk_buffs */ - for (i = 0; i < rx_ring->count; i++) { + for(i = 0; i < rx_ring->count; i++) { buffer_info = &rx_ring->buffer_info[i]; - if (buffer_info->skb) { + if(buffer_info->skb) { pci_unmap_single(pdev, buffer_info->dma, @@ -992,12 +1043,13 @@ static void ixgb_clean_rx_ring(struct ixgb_adapter *adapter) * Returns 0 on success, negative on failure **/ -static int ixgb_set_mac(struct net_device *netdev, void *p) +static int +ixgb_set_mac(struct net_device *netdev, void *p) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); struct sockaddr *addr = p; - if (!is_valid_ether_addr(addr->sa_data)) + if(!is_valid_ether_addr(addr->sa_data)) return -EADDRNOTAVAIL; memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len); @@ -1017,9 +1069,10 @@ static int ixgb_set_mac(struct net_device *netdev, void *p) * promiscuous mode, and all-multi behavior. **/ -static void ixgb_set_multi(struct net_device *netdev) +static void +ixgb_set_multi(struct net_device *netdev) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); struct ixgb_hw *hw = &adapter->hw; struct dev_mc_list *mc_ptr; uint32_t rctl; @@ -1029,16 +1082,16 @@ static void ixgb_set_multi(struct net_device *netdev) rctl = IXGB_READ_REG(hw, RCTL); - if (netdev->flags & IFF_PROMISC) { + if(netdev->flags & IFF_PROMISC) { rctl |= (IXGB_RCTL_UPE | IXGB_RCTL_MPE); - } else if (netdev->flags & IFF_ALLMULTI) { + } else if(netdev->flags & IFF_ALLMULTI) { rctl |= IXGB_RCTL_MPE; rctl &= ~IXGB_RCTL_UPE; } else { rctl &= ~(IXGB_RCTL_UPE | IXGB_RCTL_MPE); } - if (netdev->mc_count > IXGB_MAX_NUM_MULTICAST_ADDRESSES) { + if(netdev->mc_count > IXGB_MAX_NUM_MULTICAST_ADDRESSES) { rctl |= IXGB_RCTL_MPE; IXGB_WRITE_REG(hw, RCTL, rctl); } else { @@ -1046,10 +1099,10 @@ static void ixgb_set_multi(struct net_device *netdev) IXGB_WRITE_REG(hw, RCTL, rctl); - for (i = 0, mc_ptr = netdev->mc_list; mc_ptr; - i++, mc_ptr = mc_ptr->next) + for(i = 0, mc_ptr = netdev->mc_list; mc_ptr; + i++, mc_ptr = mc_ptr->next) memcpy(&mta[i * IXGB_ETH_LENGTH_OF_ADDRESS], - mc_ptr->dmi_addr, IXGB_ETH_LENGTH_OF_ADDRESS); + mc_ptr->dmi_addr, IXGB_ETH_LENGTH_OF_ADDRESS); ixgb_mc_addr_list_update(hw, mta, netdev->mc_count, 0); } @@ -1060,12 +1113,12 @@ static void ixgb_set_multi(struct net_device *netdev) * @data: pointer to netdev cast into an unsigned long **/ -static void ixgb_watchdog(unsigned long data) +static void +ixgb_watchdog(unsigned long data) { struct ixgb_adapter *adapter = (struct ixgb_adapter *)data; struct net_device *netdev = adapter->netdev; struct ixgb_desc_ring *txdr = &adapter->tx_ring; - unsigned int i; ixgb_check_for_link(&adapter->hw); @@ -1074,21 +1127,20 @@ static void ixgb_watchdog(unsigned long data) netif_stop_queue(netdev); } - if (adapter->hw.link_up) { - if (!netif_carrier_ok(netdev)) { - printk(KERN_INFO "ixgb: %s NIC Link is Up %d Mbps %s\n", - netdev->name, 10000, "Full Duplex"); + if(adapter->hw.link_up) { + if(!netif_carrier_ok(netdev)) { + DPRINTK(LINK, INFO, + "NIC Link is Up 10000 Mbps Full Duplex\n"); adapter->link_speed = 10000; adapter->link_duplex = FULL_DUPLEX; netif_carrier_on(netdev); netif_wake_queue(netdev); } } else { - if (netif_carrier_ok(netdev)) { + if(netif_carrier_ok(netdev)) { adapter->link_speed = 0; adapter->link_duplex = 0; - printk(KERN_INFO - "ixgb: %s NIC Link is Down\n", netdev->name); + DPRINTK(LINK, INFO, "NIC Link is Down\n"); netif_carrier_off(netdev); netif_stop_queue(netdev); @@ -1097,8 +1149,8 @@ static void ixgb_watchdog(unsigned long data) ixgb_update_stats(adapter); - if (!netif_carrier_ok(netdev)) { - if (IXGB_DESC_UNUSED(txdr) + 1 < txdr->count) { + if(!netif_carrier_ok(netdev)) { + if(IXGB_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 * to get done, so reset controller to flush Tx. @@ -1107,12 +1159,8 @@ static void ixgb_watchdog(unsigned long data) } } - /* Early detection of hung controller */ - i = txdr->next_to_clean; - if (txdr->buffer_info[i].dma && - time_after(jiffies, txdr->buffer_info[i].time_stamp + HZ) && - !(IXGB_READ_REG(&adapter->hw, STATUS) & IXGB_STATUS_TXOFF)) - netif_stop_queue(netdev); + /* Force detection of hung controller every watchdog period */ + adapter->detect_tx_hung = TRUE; /* generate an interrupt to force clean up of any stragglers */ IXGB_WRITE_REG(&adapter->hw, ICS, IXGB_INT_TXDW); @@ -1125,7 +1173,7 @@ static void ixgb_watchdog(unsigned long data) #define IXGB_TX_FLAGS_VLAN 0x00000002 #define IXGB_TX_FLAGS_TSO 0x00000004 -static inline boolean_t +static int ixgb_tso(struct ixgb_adapter *adapter, struct sk_buff *skb) { #ifdef NETIF_F_TSO @@ -1133,10 +1181,18 @@ ixgb_tso(struct ixgb_adapter *adapter, struct sk_buff *skb) unsigned int i; uint8_t ipcss, ipcso, tucss, tucso, hdr_len; uint16_t ipcse, tucse, mss; + int err; + + if (likely(skb_is_gso(skb))) { + struct ixgb_buffer *buffer_info; + if (skb_header_cloned(skb)) { + err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC); + if (err) + return err; + } - if (likely(skb_shinfo(skb)->tso_size)) { hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2)); - mss = skb_shinfo(skb)->tso_size; + mss = skb_shinfo(skb)->gso_size; skb->nh.iph->tot_len = 0; skb->nh.iph->check = 0; skb->h.th->check = ~csum_tcpudp_magic(skb->nh.iph->saddr, @@ -1151,6 +1207,8 @@ ixgb_tso(struct ixgb_adapter *adapter, struct sk_buff *skb) i = adapter->tx_ring.next_to_use; context_desc = IXGB_CONTEXT_DESC(adapter->tx_ring, i); + buffer_info = &adapter->tx_ring.buffer_info[i]; + WARN_ON(buffer_info->dma != 0); context_desc->ipcss = ipcss; context_desc->ipcso = ipcso; @@ -1161,59 +1219,55 @@ ixgb_tso(struct ixgb_adapter *adapter, struct sk_buff *skb) context_desc->mss = cpu_to_le16(mss); context_desc->hdr_len = hdr_len; context_desc->status = 0; - context_desc->cmd_type_len = cpu_to_le32(IXGB_CONTEXT_DESC_TYPE - | - IXGB_CONTEXT_DESC_CMD_TSE - | - IXGB_CONTEXT_DESC_CMD_IP - | - IXGB_CONTEXT_DESC_CMD_TCP - | - IXGB_CONTEXT_DESC_CMD_RS - | - IXGB_CONTEXT_DESC_CMD_IDE - | (skb->len - - (hdr_len))); - - if (++i == adapter->tx_ring.count) - i = 0; + context_desc->cmd_type_len = cpu_to_le32( + IXGB_CONTEXT_DESC_TYPE + | IXGB_CONTEXT_DESC_CMD_TSE + | IXGB_CONTEXT_DESC_CMD_IP + | IXGB_CONTEXT_DESC_CMD_TCP + | IXGB_CONTEXT_DESC_CMD_IDE + | (skb->len - (hdr_len))); + + + if(++i == adapter->tx_ring.count) i = 0; adapter->tx_ring.next_to_use = i; - return TRUE; + return 1; } #endif - return FALSE; + return 0; } -static inline boolean_t +static boolean_t ixgb_tx_csum(struct ixgb_adapter *adapter, struct sk_buff *skb) { struct ixgb_context_desc *context_desc; unsigned int i; uint8_t css, cso; - if (likely(skb->ip_summed == CHECKSUM_HW)) { + if(likely(skb->ip_summed == CHECKSUM_PARTIAL)) { + struct ixgb_buffer *buffer_info; css = skb->h.raw - skb->data; - cso = (skb->h.raw + skb->csum) - skb->data; + cso = css + skb->csum_offset; i = adapter->tx_ring.next_to_use; context_desc = IXGB_CONTEXT_DESC(adapter->tx_ring, i); + buffer_info = &adapter->tx_ring.buffer_info[i]; + WARN_ON(buffer_info->dma != 0); context_desc->tucss = css; context_desc->tucso = cso; context_desc->tucse = 0; /* zero out any previously existing data in one instruction */ - *(uint32_t *) & (context_desc->ipcss) = 0; + *(uint32_t *)&(context_desc->ipcss) = 0; context_desc->status = 0; context_desc->hdr_len = 0; context_desc->mss = 0; context_desc->cmd_type_len = - cpu_to_le32(IXGB_CONTEXT_DESC_TYPE - | IXGB_TX_DESC_CMD_RS | IXGB_TX_DESC_CMD_IDE); + cpu_to_le32(IXGB_CONTEXT_DESC_TYPE + | IXGB_TX_DESC_CMD_IDE); - if (++i == adapter->tx_ring.count) - i = 0; + if(++i == adapter->tx_ring.count) i = 0; adapter->tx_ring.next_to_use = i; return TRUE; @@ -1225,7 +1279,7 @@ ixgb_tx_csum(struct ixgb_adapter *adapter, struct sk_buff *skb) #define IXGB_MAX_TXD_PWR 14 #define IXGB_MAX_DATA_PER_TXD (1<len; unsigned int offset = 0, size, count = 0, i; + unsigned int mss = skb_shinfo(skb)->gso_size; unsigned int nr_frags = skb_shinfo(skb)->nr_frags; unsigned int f; + len -= skb->data_len; i = tx_ring->next_to_use; - while (len) { + while(len) { buffer_info = &tx_ring->buffer_info[i]; - size = min(len, IXGB_MAX_JUMBO_FRAME_SIZE); + size = min(len, IXGB_MAX_DATA_PER_TXD); + /* Workaround for premature desc write-backs + * in TSO mode. Append 4-byte sentinel desc */ + if (unlikely(mss && !nr_frags && size == len && size > 8)) + size -= 4; + buffer_info->length = size; + WARN_ON(buffer_info->dma != 0); buffer_info->dma = - pci_map_single(adapter->pdev, - skb->data + offset, size, PCI_DMA_TODEVICE); + pci_map_single(adapter->pdev, + skb->data + offset, + size, + PCI_DMA_TODEVICE); buffer_info->time_stamp = jiffies; + buffer_info->next_to_watch = 0; len -= size; offset += size; count++; - if (++i == tx_ring->count) - i = 0; + if(++i == tx_ring->count) i = 0; } - for (f = 0; f < nr_frags; f++) { + for(f = 0; f < nr_frags; f++) { struct skb_frag_struct *frag; frag = &skb_shinfo(skb)->frags[f]; len = frag->size; offset = 0; - while (len) { + while(len) { buffer_info = &tx_ring->buffer_info[i]; - size = min(len, IXGB_MAX_JUMBO_FRAME_SIZE); + size = min(len, IXGB_MAX_DATA_PER_TXD); + + /* Workaround for premature desc write-backs + * in TSO mode. Append 4-byte sentinel desc */ + if (unlikely(mss && !nr_frags && size == len + && size > 8)) + size -= 4; + buffer_info->length = size; buffer_info->dma = - pci_map_page(adapter->pdev, - frag->page, - frag->page_offset + offset, - size, PCI_DMA_TODEVICE); + pci_map_page(adapter->pdev, + frag->page, + frag->page_offset + offset, + size, + PCI_DMA_TODEVICE); buffer_info->time_stamp = jiffies; + buffer_info->next_to_watch = 0; len -= size; offset += size; count++; - if (++i == tx_ring->count) - i = 0; + if(++i == tx_ring->count) i = 0; } } i = (i == 0) ? tx_ring->count - 1 : i - 1; @@ -1288,9 +1360,8 @@ ixgb_tx_map(struct ixgb_adapter *adapter, struct sk_buff *skb, return count; } -static inline void -ixgb_tx_queue(struct ixgb_adapter *adapter, int count, int vlan_id, - int tx_flags) +static void +ixgb_tx_queue(struct ixgb_adapter *adapter, int count, int vlan_id,int tx_flags) { struct ixgb_desc_ring *tx_ring = &adapter->tx_ring; struct ixgb_tx_desc *tx_desc = NULL; @@ -1300,36 +1371,35 @@ ixgb_tx_queue(struct ixgb_adapter *adapter, int count, int vlan_id, uint8_t popts = 0; unsigned int i; - if (tx_flags & IXGB_TX_FLAGS_TSO) { + if(tx_flags & IXGB_TX_FLAGS_TSO) { cmd_type_len |= IXGB_TX_DESC_CMD_TSE; popts |= (IXGB_TX_DESC_POPTS_IXSM | IXGB_TX_DESC_POPTS_TXSM); } - if (tx_flags & IXGB_TX_FLAGS_CSUM) + if(tx_flags & IXGB_TX_FLAGS_CSUM) popts |= IXGB_TX_DESC_POPTS_TXSM; - if (tx_flags & IXGB_TX_FLAGS_VLAN) { + if(tx_flags & IXGB_TX_FLAGS_VLAN) { cmd_type_len |= IXGB_TX_DESC_CMD_VLE; } i = tx_ring->next_to_use; - while (count--) { + while(count--) { buffer_info = &tx_ring->buffer_info[i]; tx_desc = IXGB_TX_DESC(*tx_ring, i); tx_desc->buff_addr = cpu_to_le64(buffer_info->dma); tx_desc->cmd_type_len = - cpu_to_le32(cmd_type_len | buffer_info->length); + cpu_to_le32(cmd_type_len | buffer_info->length); tx_desc->status = status; tx_desc->popts = popts; tx_desc->vlan = cpu_to_le16(vlan_id); - if (++i == tx_ring->count) - i = 0; + if(++i == tx_ring->count) i = 0; } - tx_desc->cmd_type_len |= cpu_to_le32(IXGB_TX_DESC_CMD_EOP - | IXGB_TX_DESC_CMD_RS); + tx_desc->cmd_type_len |= cpu_to_le32(IXGB_TX_DESC_CMD_EOP + | IXGB_TX_DESC_CMD_RS ); /* Force memory writes to complete before letting h/w * know there are new descriptors to fetch. (Only @@ -1341,51 +1411,115 @@ ixgb_tx_queue(struct ixgb_adapter *adapter, int count, int vlan_id, IXGB_WRITE_REG(&adapter->hw, TDT, i); } +static int __ixgb_maybe_stop_tx(struct net_device *netdev, int size) +{ + struct ixgb_adapter *adapter = netdev_priv(netdev); + struct ixgb_desc_ring *tx_ring = &adapter->tx_ring; + + netif_stop_queue(netdev); + /* Herbert's original patch had: + * smp_mb__after_netif_stop_queue(); + * but since that doesn't exist yet, just open code it. */ + smp_mb(); + + /* We need to check again in a case another CPU has just + * made room available. */ + if (likely(IXGB_DESC_UNUSED(tx_ring) < size)) + return -EBUSY; + + /* A reprieve! */ + netif_start_queue(netdev); + ++adapter->restart_queue; + return 0; +} + +static int ixgb_maybe_stop_tx(struct net_device *netdev, + struct ixgb_desc_ring *tx_ring, int size) +{ + if (likely(IXGB_DESC_UNUSED(tx_ring) >= size)) + return 0; + return __ixgb_maybe_stop_tx(netdev, size); +} + + /* Tx Descriptors needed, worst case */ #define TXD_USE_COUNT(S) (((S) >> IXGB_MAX_TXD_PWR) + \ (((S) & (IXGB_MAX_DATA_PER_TXD - 1)) ? 1 : 0)) -#define DESC_NEEDED TXD_USE_COUNT(IXGB_MAX_DATA_PER_TXD) + \ - MAX_SKB_FRAGS * TXD_USE_COUNT(PAGE_SIZE) + 1 +#define DESC_NEEDED TXD_USE_COUNT(IXGB_MAX_DATA_PER_TXD) /* skb->date */ + \ + MAX_SKB_FRAGS * TXD_USE_COUNT(PAGE_SIZE) + 1 /* for context */ \ + + 1 /* one more needed for sentinel TSO workaround */ -static int ixgb_xmit_frame(struct sk_buff *skb, struct net_device *netdev) +static int +ixgb_xmit_frame(struct sk_buff *skb, struct net_device *netdev) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); unsigned int first; unsigned int tx_flags = 0; unsigned long flags; int vlan_id = 0; + int tso; - if (skb->len <= 0) { + if(skb->len <= 0) { dev_kfree_skb_any(skb); return 0; } +#ifdef NETIF_F_LLTX + local_irq_save(flags); + if (!spin_trylock(&adapter->tx_lock)) { + /* Collision - tell upper layer to requeue */ + local_irq_restore(flags); + return NETDEV_TX_LOCKED; + } +#else spin_lock_irqsave(&adapter->tx_lock, flags); - if (unlikely(IXGB_DESC_UNUSED(&adapter->tx_ring) < DESC_NEEDED)) { +#endif + + if (unlikely(ixgb_maybe_stop_tx(netdev, &adapter->tx_ring, + DESC_NEEDED))) { netif_stop_queue(netdev); spin_unlock_irqrestore(&adapter->tx_lock, flags); - return 1; + return NETDEV_TX_BUSY; } + +#ifndef NETIF_F_LLTX spin_unlock_irqrestore(&adapter->tx_lock, flags); +#endif - if (adapter->vlgrp && vlan_tx_tag_present(skb)) { + if(adapter->vlgrp && vlan_tx_tag_present(skb)) { tx_flags |= IXGB_TX_FLAGS_VLAN; vlan_id = vlan_tx_tag_get(skb); } first = adapter->tx_ring.next_to_use; + + tso = ixgb_tso(adapter, skb); + if (tso < 0) { + dev_kfree_skb_any(skb); +#ifdef NETIF_F_LLTX + spin_unlock_irqrestore(&adapter->tx_lock, flags); +#endif + return NETDEV_TX_OK; + } - if (ixgb_tso(adapter, skb)) + if (likely(tso)) tx_flags |= IXGB_TX_FLAGS_TSO; - else if (ixgb_tx_csum(adapter, skb)) + else if(ixgb_tx_csum(adapter, skb)) tx_flags |= IXGB_TX_FLAGS_CSUM; ixgb_tx_queue(adapter, ixgb_tx_map(adapter, skb, first), vlan_id, - tx_flags); + tx_flags); netdev->trans_start = jiffies; - return 0; +#ifdef NETIF_F_LLTX + /* Make sure there is space in the ring for the next send. */ + ixgb_maybe_stop_tx(netdev, &adapter->tx_ring, DESC_NEEDED); + + spin_unlock_irqrestore(&adapter->tx_lock, flags); + +#endif + return NETDEV_TX_OK; } /** @@ -1393,22 +1527,24 @@ static int ixgb_xmit_frame(struct sk_buff *skb, struct net_device *netdev) * @netdev: network interface device structure **/ -static void ixgb_tx_timeout(struct net_device *netdev) +static void +ixgb_tx_timeout(struct net_device *netdev) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); /* Do the reset outside of interrupt context */ schedule_work(&adapter->tx_timeout_task); } -static void ixgb_tx_timeout_task(struct net_device *netdev) +static void +ixgb_tx_timeout_task(struct work_struct *work) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = + container_of(work, struct ixgb_adapter, tx_timeout_task); - netif_device_detach(netdev); + adapter->tx_timeout_count++; ixgb_down(adapter, TRUE); ixgb_up(adapter); - netif_device_attach(netdev); } /** @@ -1419,9 +1555,10 @@ static void ixgb_tx_timeout_task(struct net_device *netdev) * The statistics are actually updated from the timer callback. **/ -static struct net_device_stats *ixgb_get_stats(struct net_device *netdev) +static struct net_device_stats * +ixgb_get_stats(struct net_device *netdev) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); return &adapter->net_stats; } @@ -1434,37 +1571,25 @@ static struct net_device_stats *ixgb_get_stats(struct net_device *netdev) * Returns 0 on success, negative on failure **/ -static int ixgb_change_mtu(struct net_device *netdev, int new_mtu) +static int +ixgb_change_mtu(struct net_device *netdev, int new_mtu) { - struct ixgb_adapter *adapter = netdev->priv; - uint32_t old_mtu = adapter->rx_buffer_len; + struct ixgb_adapter *adapter = netdev_priv(netdev); int max_frame = new_mtu + ENET_HEADER_SIZE + ENET_FCS_LENGTH; + int old_max_frame = netdev->mtu + ENET_HEADER_SIZE + ENET_FCS_LENGTH; - if ((max_frame < IXGB_MIN_ENET_FRAME_SIZE_WITHOUT_FCS + ENET_FCS_LENGTH) - || (max_frame > IXGB_MAX_JUMBO_FRAME_SIZE + ENET_FCS_LENGTH)) { - IXGB_ERR("Invalid MTU setting\n"); + + if((max_frame < IXGB_MIN_ENET_FRAME_SIZE_WITHOUT_FCS + ENET_FCS_LENGTH) + || (max_frame > IXGB_MAX_JUMBO_FRAME_SIZE + ENET_FCS_LENGTH)) { + DPRINTK(PROBE, ERR, "Invalid MTU setting %d\n", new_mtu); return -EINVAL; } - if ((max_frame <= - IXGB_MAX_ENET_FRAME_SIZE_WITHOUT_FCS + ENET_FCS_LENGTH) - || (max_frame <= IXGB_RXBUFFER_2048)) { - adapter->rx_buffer_len = IXGB_RXBUFFER_2048; - - } else if (max_frame <= IXGB_RXBUFFER_4096) { - adapter->rx_buffer_len = IXGB_RXBUFFER_4096; - - } else if (max_frame <= IXGB_RXBUFFER_8192) { - adapter->rx_buffer_len = IXGB_RXBUFFER_8192; - - } else { - adapter->rx_buffer_len = IXGB_RXBUFFER_16384; - } + adapter->rx_buffer_len = max_frame; netdev->mtu = new_mtu; - if (old_mtu != adapter->rx_buffer_len && netif_running(netdev)) { - + if ((old_max_frame != max_frame) && netif_running(netdev)) { ixgb_down(adapter, TRUE); ixgb_up(adapter); } @@ -1477,16 +1602,42 @@ static int ixgb_change_mtu(struct net_device *netdev, int new_mtu) * @adapter: board private structure **/ -static void ixgb_update_stats(struct ixgb_adapter *adapter) +void +ixgb_update_stats(struct ixgb_adapter *adapter) { + struct net_device *netdev = adapter->netdev; + struct pci_dev *pdev = adapter->pdev; + + /* Prevent stats update while adapter is being reset */ + if (pdev->error_state && pdev->error_state != pci_channel_io_normal) + return; + + if((netdev->flags & IFF_PROMISC) || (netdev->flags & IFF_ALLMULTI) || + (netdev->mc_count > IXGB_MAX_NUM_MULTICAST_ADDRESSES)) { + u64 multi = IXGB_READ_REG(&adapter->hw, MPRCL); + u32 bcast_l = IXGB_READ_REG(&adapter->hw, BPRCL); + u32 bcast_h = IXGB_READ_REG(&adapter->hw, BPRCH); + u64 bcast = ((u64)bcast_h << 32) | bcast_l; + + multi |= ((u64)IXGB_READ_REG(&adapter->hw, MPRCH) << 32); + /* fix up multicast stats by removing broadcasts */ + if(multi >= bcast) + multi -= bcast; + + adapter->stats.mprcl += (multi & 0xFFFFFFFF); + adapter->stats.mprch += (multi >> 32); + adapter->stats.bprcl += bcast_l; + adapter->stats.bprch += bcast_h; + } else { + adapter->stats.mprcl += IXGB_READ_REG(&adapter->hw, MPRCL); + adapter->stats.mprch += IXGB_READ_REG(&adapter->hw, MPRCH); + adapter->stats.bprcl += IXGB_READ_REG(&adapter->hw, BPRCL); + adapter->stats.bprch += IXGB_READ_REG(&adapter->hw, BPRCH); + } adapter->stats.tprl += IXGB_READ_REG(&adapter->hw, TPRL); adapter->stats.tprh += IXGB_READ_REG(&adapter->hw, TPRH); adapter->stats.gprcl += IXGB_READ_REG(&adapter->hw, GPRCL); adapter->stats.gprch += IXGB_READ_REG(&adapter->hw, GPRCH); - adapter->stats.bprcl += IXGB_READ_REG(&adapter->hw, BPRCL); - adapter->stats.bprch += IXGB_READ_REG(&adapter->hw, BPRCH); - adapter->stats.mprcl += IXGB_READ_REG(&adapter->hw, MPRCL); - adapter->stats.mprch += IXGB_READ_REG(&adapter->hw, MPRCH); adapter->stats.uprcl += IXGB_READ_REG(&adapter->hw, UPRCL); adapter->stats.uprch += IXGB_READ_REG(&adapter->hw, UPRCH); adapter->stats.vprcl += IXGB_READ_REG(&adapter->hw, VPRCL); @@ -1558,8 +1709,6 @@ static void ixgb_update_stats(struct ixgb_adapter *adapter) adapter->stats.icbc + adapter->stats.ecbc + adapter->stats.mpc; - adapter->net_stats.rx_dropped = adapter->stats.mpc; - /* see above * adapter->net_stats.rx_length_errors = adapter->stats.rlec; */ @@ -1583,50 +1732,47 @@ static void ixgb_update_stats(struct ixgb_adapter *adapter) * ixgb_intr - Interrupt Handler * @irq: interrupt number * @data: pointer to a network interface device structure - * @pt_regs: CPU registers structure **/ -static irqreturn_t ixgb_intr(int irq, void *data, struct pt_regs *regs) +static irqreturn_t +ixgb_intr(int irq, void *data) { struct net_device *netdev = data; - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); struct ixgb_hw *hw = &adapter->hw; - uint32_t icr = IXGB_READ_REG(&adapter->hw, ICR); + uint32_t icr = IXGB_READ_REG(hw, ICR); #ifndef CONFIG_IXGB_NAPI unsigned int i; #endif - if (unlikely(!icr)) - return IRQ_NONE; /* Not our interrupt */ + if(unlikely(!icr)) + return IRQ_NONE; /* Not our interrupt */ - if (unlikely(icr & (IXGB_INT_RXSEQ | IXGB_INT_LSC))) { + if(unlikely(icr & (IXGB_INT_RXSEQ | IXGB_INT_LSC))) { mod_timer(&adapter->watchdog_timer, jiffies); } + #ifdef CONFIG_IXGB_NAPI - if (netif_rx_schedule_prep(netdev)) { + if(netif_rx_schedule_prep(netdev)) { /* Disable interrupts and register for poll. The flush - of the posted write is intentionally left out. - */ + of the posted write is intentionally left out. + */ atomic_inc(&adapter->irq_sem); - IXGB_WRITE_REG(hw, IMC, ~0); + IXGB_WRITE_REG(&adapter->hw, IMC, ~0); __netif_rx_schedule(netdev); } #else - for (i = 0; i < IXGB_MAX_INTR; i++) - if (!ixgb_clean_rx_irq(adapter) & !ixgb_clean_tx_irq(adapter)) + /* yes, that is actually a & and it is meant to make sure that + * every pass through this for loop checks both receive and + * transmit queues for completed descriptors, intended to + * avoid starvation issues and assist tx/rx fairness. */ + for(i = 0; i < IXGB_MAX_INTR; i++) + if(!ixgb_clean_rx_irq(adapter) & + !ixgb_clean_tx_irq(adapter)) break; - /* if RAIDC:EN == 1 and ICR:RXDMT0 == 1, we need to - * set IMS:RXDMT0 to 1 to restart the RBD timer (POLL) - */ - if ((icr & IXGB_INT_RXDMT0) && adapter->raidc) { - /* ready the timer by writing the clear reg */ - IXGB_WRITE_REG(hw, IMC, IXGB_INT_RXDMT0); - /* now restart it, h/w will decide if its necessary */ - IXGB_WRITE_REG(hw, IMS, IXGB_INT_RXDMT0); - } -#endif +#endif return IRQ_HANDLED; } @@ -1636,25 +1782,28 @@ static irqreturn_t ixgb_intr(int irq, void *data, struct pt_regs *regs) * @adapter: board private structure **/ -static int ixgb_clean(struct net_device *netdev, int *budget) +static int +ixgb_clean(struct net_device *netdev, int *budget) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); int work_to_do = min(*budget, netdev->quota); + int tx_cleaned; int work_done = 0; - ixgb_clean_tx_irq(adapter); + tx_cleaned = ixgb_clean_tx_irq(adapter); ixgb_clean_rx_irq(adapter, &work_done, work_to_do); *budget -= work_done; netdev->quota -= work_done; - if (work_done < work_to_do || !netif_running(netdev)) { + /* if no Tx and not enough Rx work done, exit the polling mode */ + if((!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) { netif_rx_complete(netdev); - /* RAIDC will be automatically restarted by irq_enable */ ixgb_irq_enable(adapter); + return 0; } - return (work_done >= work_to_do); + return 1; } #endif @@ -1663,11 +1812,11 @@ static int ixgb_clean(struct net_device *netdev, int *budget) * @adapter: board private structure **/ -static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter) +static boolean_t +ixgb_clean_tx_irq(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *tx_ring = &adapter->tx_ring; struct net_device *netdev = adapter->netdev; - struct pci_dev *pdev = adapter->pdev; struct ixgb_tx_desc *tx_desc, *eop_desc; struct ixgb_buffer *buffer_info; unsigned int i, eop; @@ -1677,9 +1826,9 @@ static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter) eop = tx_ring->buffer_info[i].next_to_watch; eop_desc = IXGB_TX_DESC(*tx_ring, eop); - while (eop_desc->status & cpu_to_le32(IXGB_TX_DESC_STATUS_DD)) { + while(eop_desc->status & IXGB_TX_DESC_STATUS_DD) { - for (cleaned = FALSE; !cleaned;) { + for(cleaned = FALSE; !cleaned; ) { tx_desc = IXGB_TX_DESC(*tx_ring, i); buffer_info = &tx_ring->buffer_info[i]; @@ -1688,28 +1837,12 @@ static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter) IXGB_TX_DESC_POPTS_IXSM)) adapter->hw_csum_tx_good++; - if (buffer_info->dma) { - - pci_unmap_page(pdev, - buffer_info->dma, - buffer_info->length, - PCI_DMA_TODEVICE); - - buffer_info->dma = 0; - } - - if (buffer_info->skb) { - - dev_kfree_skb_any(buffer_info->skb); + ixgb_unmap_and_free_tx_resource(adapter, buffer_info); - buffer_info->skb = NULL; - } - - *(uint32_t *) & (tx_desc->status) = 0; + *(uint32_t *)&(tx_desc->status) = 0; cleaned = (i == eop); - if (++i == tx_ring->count) - i = 0; + if(++i == tx_ring->count) i = 0; } eop = tx_ring->buffer_info[i].next_to_watch; @@ -1718,13 +1851,44 @@ static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter) tx_ring->next_to_clean = i; - spin_lock(&adapter->tx_lock); - if (cleaned && netif_queue_stopped(netdev) && netif_carrier_ok(netdev) - && (IXGB_DESC_UNUSED(tx_ring) > IXGB_TX_QUEUE_WAKE)) { + if (unlikely(netif_queue_stopped(netdev))) { + spin_lock(&adapter->tx_lock); + if (netif_queue_stopped(netdev) && netif_carrier_ok(netdev) && + (IXGB_DESC_UNUSED(tx_ring) >= DESC_NEEDED)) + netif_wake_queue(netdev); + spin_unlock(&adapter->tx_lock); + } - netif_wake_queue(netdev); + if(adapter->detect_tx_hung) { + /* 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[eop].dma && + time_after(jiffies, tx_ring->buffer_info[eop].time_stamp + HZ) + && !(IXGB_READ_REG(&adapter->hw, STATUS) & + IXGB_STATUS_TXOFF)) { + /* detected Tx unit hang */ + DPRINTK(DRV, ERR, "Detected Tx Unit Hang\n" + " TDH <%x>\n" + " TDT <%x>\n" + " next_to_use <%x>\n" + " next_to_clean <%x>\n" + "buffer_info[next_to_clean]\n" + " time_stamp <%lx>\n" + " next_to_watch <%x>\n" + " jiffies <%lx>\n" + " next_to_watch.status <%x>\n", + IXGB_READ_REG(&adapter->hw, TDH), + IXGB_READ_REG(&adapter->hw, TDT), + tx_ring->next_to_use, + tx_ring->next_to_clean, + tx_ring->buffer_info[eop].time_stamp, + eop, + jiffies, + eop_desc->status); + netif_stop_queue(netdev); + } } - spin_unlock(&adapter->tx_lock); return cleaned; } @@ -1736,22 +1900,23 @@ static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter) * @sk_buff: socket buffer with received data **/ -static inline void +static void ixgb_rx_checksum(struct ixgb_adapter *adapter, - struct ixgb_rx_desc *rx_desc, struct sk_buff *skb) + struct ixgb_rx_desc *rx_desc, + struct sk_buff *skb) { /* Ignore Checksum bit is set OR * TCP Checksum has not been calculated */ - if ((rx_desc->status & IXGB_RX_DESC_STATUS_IXSM) || - (!(rx_desc->status & IXGB_RX_DESC_STATUS_TCPCS))) { + if((rx_desc->status & IXGB_RX_DESC_STATUS_IXSM) || + (!(rx_desc->status & IXGB_RX_DESC_STATUS_TCPCS))) { skb->ip_summed = CHECKSUM_NONE; return; } /* At this point we know the hardware did the TCP checksum */ /* now look at the TCP checksum error bit */ - if (rx_desc->errors & IXGB_RX_DESC_ERRORS_TCPE) { + if(rx_desc->errors & IXGB_RX_DESC_ERRORS_TCPE) { /* let the stack verify checksum errors */ skb->ip_summed = CHECKSUM_NONE; adapter->hw_csum_rx_error++; @@ -1779,7 +1944,6 @@ ixgb_clean_rx_irq(struct ixgb_adapter *adapter) struct pci_dev *pdev = adapter->pdev; struct ixgb_rx_desc *rx_desc, *next_rxd; struct ixgb_buffer *buffer_info, *next_buffer, *next2_buffer; - struct sk_buff *skb, *next_skb; uint32_t length; unsigned int i, j; boolean_t cleaned = FALSE; @@ -1788,18 +1952,27 @@ ixgb_clean_rx_irq(struct ixgb_adapter *adapter) rx_desc = IXGB_RX_DESC(*rx_ring, i); buffer_info = &rx_ring->buffer_info[i]; - while (rx_desc->status & IXGB_RX_DESC_STATUS_DD) { + while(rx_desc->status & IXGB_RX_DESC_STATUS_DD) { + struct sk_buff *skb, *next_skb; + u8 status; +#ifdef CONFIG_IXGB_NAPI + if(*work_done >= work_to_do) + break; + + (*work_done)++; +#endif + status = rx_desc->status; skb = buffer_info->skb; + buffer_info->skb = NULL; + prefetch(skb->data); - if (++i == rx_ring->count) - i = 0; + if(++i == rx_ring->count) i = 0; next_rxd = IXGB_RX_DESC(*rx_ring, i); prefetch(next_rxd); - if ((j = i + 1) == rx_ring->count) - j = 0; + if((j = i + 1) == rx_ring->count) j = 0; next2_buffer = &rx_ring->buffer_info[j]; prefetch(next2_buffer); @@ -1807,35 +1980,24 @@ ixgb_clean_rx_irq(struct ixgb_adapter *adapter) next_skb = next_buffer->skb; prefetch(next_skb); -#ifdef CONFIG_IXGB_NAPI - if (*work_done >= work_to_do) - break; - - (*work_done)++; -#endif - cleaned = TRUE; pci_unmap_single(pdev, buffer_info->dma, - buffer_info->length, PCI_DMA_FROMDEVICE); + buffer_info->length, + PCI_DMA_FROMDEVICE); length = le16_to_cpu(rx_desc->length); - if (unlikely(!(rx_desc->status & IXGB_RX_DESC_STATUS_EOP))) { + if(unlikely(!(status & IXGB_RX_DESC_STATUS_EOP))) { /* All receives must fit into a single buffer */ IXGB_DBG("Receive packet consumed multiple buffers " - "length<%x>\n", length); + "length<%x>\n", length); dev_kfree_skb_irq(skb); - rx_desc->status = 0; - buffer_info->skb = NULL; - - rx_desc = next_rxd; - buffer_info = next_buffer; - continue; + goto rxdesc_done; } if (unlikely(rx_desc->errors @@ -1844,13 +2006,27 @@ ixgb_clean_rx_irq(struct ixgb_adapter *adapter) IXGB_RX_DESC_ERRORS_RXE))) { dev_kfree_skb_irq(skb); - rx_desc->status = 0; - buffer_info->skb = NULL; + goto rxdesc_done; + } - rx_desc = next_rxd; - buffer_info = next_buffer; - continue; + /* code added for copybreak, this should improve + * performance for small packets with large amounts + * of reassembly being done in the stack */ +#define IXGB_CB_LENGTH 256 + if (length < IXGB_CB_LENGTH) { + struct sk_buff *new_skb = + netdev_alloc_skb(netdev, length + NET_IP_ALIGN); + if (new_skb) { + skb_reserve(new_skb, NET_IP_ALIGN); + 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; + } } + /* end copybreak code */ /* Good Receive */ skb_put(skb, length); @@ -1860,31 +2036,29 @@ ixgb_clean_rx_irq(struct ixgb_adapter *adapter) skb->protocol = eth_type_trans(skb, netdev); #ifdef CONFIG_IXGB_NAPI - if (adapter->vlgrp - && (rx_desc->status & IXGB_RX_DESC_STATUS_VP)) { + if(adapter->vlgrp && (status & IXGB_RX_DESC_STATUS_VP)) { vlan_hwaccel_receive_skb(skb, adapter->vlgrp, - le16_to_cpu(rx_desc-> - special & - IXGB_RX_DESC_SPECIAL_VLAN_MASK)); + le16_to_cpu(rx_desc->special) & + IXGB_RX_DESC_SPECIAL_VLAN_MASK); } else { netif_receive_skb(skb); } -#else /* CONFIG_IXGB_NAPI */ - if (adapter->vlgrp - && (rx_desc->status & IXGB_RX_DESC_STATUS_VP)) { +#else /* CONFIG_IXGB_NAPI */ + if(adapter->vlgrp && (status & IXGB_RX_DESC_STATUS_VP)) { vlan_hwaccel_rx(skb, adapter->vlgrp, - le16_to_cpu(rx_desc-> - special & - IXGB_RX_DESC_SPECIAL_VLAN_MASK)); + le16_to_cpu(rx_desc->special) & + IXGB_RX_DESC_SPECIAL_VLAN_MASK); } else { netif_rx(skb); } -#endif /* CONFIG_IXGB_NAPI */ +#endif /* CONFIG_IXGB_NAPI */ netdev->last_rx = jiffies; +rxdesc_done: + /* clean up descriptor, might be written over by hw */ rx_desc->status = 0; - buffer_info->skb = NULL; + /* use prefetched values */ rx_desc = next_rxd; buffer_info = next_buffer; } @@ -1901,7 +2075,8 @@ ixgb_clean_rx_irq(struct ixgb_adapter *adapter) * @adapter: address of board private structure **/ -static void ixgb_alloc_rx_buffers(struct ixgb_adapter *adapter) +static void +ixgb_alloc_rx_buffers(struct ixgb_adapter *adapter) { struct ixgb_desc_ring *rx_ring = &adapter->rx_ring; struct net_device *netdev = adapter->netdev; @@ -1917,20 +2092,22 @@ static void ixgb_alloc_rx_buffers(struct ixgb_adapter *adapter) buffer_info = &rx_ring->buffer_info[i]; cleancount = IXGB_DESC_UNUSED(rx_ring); - /* lessen this to 4 if we're - * in the midst of raidc and rbd is occuring - * because we don't want to delay returning buffers when low - */ - num_group_tail_writes = adapter->raidc ? 4 : IXGB_RX_BUFFER_WRITE; - - /* leave one descriptor unused */ - while (--cleancount > 0) { - rx_desc = IXGB_RX_DESC(*rx_ring, i); + num_group_tail_writes = IXGB_RX_BUFFER_WRITE; - skb = dev_alloc_skb(adapter->rx_buffer_len + NET_IP_ALIGN); + /* leave three descriptors unused */ + while(--cleancount > 2) { + /* recycle! its good for you */ + skb = buffer_info->skb; + if (skb) { + skb_trim(skb, 0); + goto map_skb; + } + skb = netdev_alloc_skb(netdev, adapter->rx_buffer_len + + NET_IP_ALIGN); if (unlikely(!skb)) { /* Better luck next round */ + adapter->alloc_rx_buff_failed++; break; } @@ -1940,52 +2117,38 @@ static void ixgb_alloc_rx_buffers(struct ixgb_adapter *adapter) */ skb_reserve(skb, NET_IP_ALIGN); - skb->dev = netdev; - buffer_info->skb = skb; buffer_info->length = adapter->rx_buffer_len; - buffer_info->dma = - pci_map_single(pdev, - skb->data, - adapter->rx_buffer_len, PCI_DMA_FROMDEVICE); +map_skb: + buffer_info->dma = pci_map_single(pdev, + skb->data, + adapter->rx_buffer_len, + PCI_DMA_FROMDEVICE); + rx_desc = IXGB_RX_DESC(*rx_ring, i); rx_desc->buff_addr = cpu_to_le64(buffer_info->dma); + /* guarantee DD bit not set now before h/w gets descriptor + * this is the rest of the workaround for h/w double + * writeback. */ + rx_desc->status = 0; - if ((i & ~(num_group_tail_writes - 1)) == i) { - /* Force memory writes to complete before letting h/w - * know there are new descriptors to fetch. (Only - * applicable for weak-ordered memory model archs, - * such as IA-64). */ - wmb(); - - IXGB_WRITE_REG(&adapter->hw, RDT, i); - } - if (++i == rx_ring->count) - i = 0; + if(++i == rx_ring->count) i = 0; buffer_info = &rx_ring->buffer_info[i]; } - rx_ring->next_to_use = i; -} - -/** - * ixgb_ioctl - perform a command - e.g: ethtool:get_driver_info. - * @param netdev network interface device structure - * @param ifr data to be used/filled in by the ioctl command - * @param cmd ioctl command to execute - **/ - -static int ixgb_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) -{ - switch (cmd) { - case SIOCETHTOOL: - return ixgb_ethtool_ioctl(netdev, ifr); - default: - return -EOPNOTSUPP; + if (likely(rx_ring->next_to_use != i)) { + rx_ring->next_to_use = i; + if (unlikely(i-- == 0)) + i = (rx_ring->count - 1); + + /* Force memory writes to complete before letting h/w + * know there are new descriptors to fetch. (Only + * applicable for weak-ordered memory model archs, such + * as IA-64). */ + wmb(); + IXGB_WRITE_REG(&adapter->hw, RDT, i); } - - return 0; } /** @@ -1997,13 +2160,13 @@ static int ixgb_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) static void ixgb_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); uint32_t ctrl, rctl; ixgb_irq_disable(adapter); adapter->vlgrp = grp; - if (grp) { + if(grp) { /* enable VLAN tag insert/strip */ ctrl = IXGB_READ_REG(&adapter->hw, CTRL0); ctrl |= IXGB_CTRL0_VME; @@ -2032,9 +2195,10 @@ ixgb_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp) ixgb_irq_enable(adapter); } -static void ixgb_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid) +static void +ixgb_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); uint32_t vfta, index; /* add VID to filter table */ @@ -2045,19 +2209,20 @@ static void ixgb_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid) ixgb_write_vfta(&adapter->hw, index, vfta); } -static void ixgb_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid) +static void +ixgb_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid) { - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); uint32_t vfta, index; ixgb_irq_disable(adapter); - if (adapter->vlgrp) + if(adapter->vlgrp) adapter->vlgrp->vlan_devices[vid] = NULL; ixgb_irq_enable(adapter); - /* remove VID from filter table */ + /* remove VID from filter table*/ index = (vid >> 5) & 0x7F; vfta = IXGB_READ_REG_ARRAY(&adapter->hw, VFTA, index); @@ -2065,81 +2230,130 @@ static void ixgb_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid) ixgb_write_vfta(&adapter->hw, index, vfta); } -static void ixgb_restore_vlan(struct ixgb_adapter *adapter) +static void +ixgb_restore_vlan(struct ixgb_adapter *adapter) { ixgb_vlan_rx_register(adapter->netdev, adapter->vlgrp); - if (adapter->vlgrp) { + if(adapter->vlgrp) { uint16_t vid; - for (vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) { - if (!adapter->vlgrp->vlan_devices[vid]) + for(vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) { + if(!adapter->vlgrp->vlan_devices[vid]) continue; ixgb_vlan_rx_add_vid(adapter->netdev, vid); } } } +#ifdef CONFIG_NET_POLL_CONTROLLER +/* + * Polling 'interrupt' - used by things like netconsole to send skbs + * without having to re-enable interrupts. It's not called while + * the interrupt routine is executing. + */ + +static void ixgb_netpoll(struct net_device *dev) +{ + struct ixgb_adapter *adapter = netdev_priv(dev); + + disable_irq(adapter->pdev->irq); + ixgb_intr(adapter->pdev->irq, dev); + enable_irq(adapter->pdev->irq); +} +#endif + /** - * ixgb_notify_reboot - handles OS notification of reboot event. - * @param nb notifier block, unused - * @param event Event being passed to driver to act upon - * @param p A pointer to our net device - **/ -static int -ixgb_notify_reboot(struct notifier_block *nb, unsigned long event, void *p) + * ixgb_io_error_detected() - called when PCI error is detected + * @pdev pointer to pci device with error + * @state pci channel state after error + * + * This callback is called by the PCI subsystem whenever + * a PCI bus error is detected. + */ +static pci_ers_result_t ixgb_io_error_detected (struct pci_dev *pdev, + enum pci_channel_state state) { - struct pci_dev *pdev = NULL; - - switch (event) { - case SYS_DOWN: - case SYS_HALT: - case SYS_POWER_OFF: - while ((pdev = pci_find_device(PCI_ANY_ID, PCI_ANY_ID, pdev))) { - if (pci_dev_driver(pdev) == &ixgb_driver) - ixgb_suspend(pdev, 3); - } - } - return NOTIFY_DONE; + struct net_device *netdev = pci_get_drvdata(pdev); + struct ixgb_adapter *adapter = netdev_priv(netdev); + + if(netif_running(netdev)) + ixgb_down(adapter, TRUE); + + pci_disable_device(pdev); + + /* Request a slot reset. */ + return PCI_ERS_RESULT_NEED_RESET; } /** - * ixgb_suspend - driver suspend function called from notify. - * @param pdev pci driver structure used for passing to - * @param state power state to enter - **/ -static int ixgb_suspend(struct pci_dev *pdev, uint32_t state) + * ixgb_io_slot_reset - called after the pci bus has been reset. + * @pdev pointer to pci device with error + * + * This callback is called after the PCI buss has been reset. + * Basically, this tries to restart the card from scratch. + * This is a shortened version of the device probe/discovery code, + * it resembles the first-half of the ixgb_probe() routine. + */ +static pci_ers_result_t ixgb_io_slot_reset (struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); - struct ixgb_adapter *adapter = netdev->priv; + struct ixgb_adapter *adapter = netdev_priv(netdev); - netif_device_detach(netdev); + if(pci_enable_device(pdev)) { + DPRINTK(PROBE, ERR, "Cannot re-enable PCI device after reset.\n"); + return PCI_ERS_RESULT_DISCONNECT; + } - if (netif_running(netdev)) - ixgb_down(adapter, TRUE); + /* Perform card reset only on one instance of the card */ + if (0 != PCI_FUNC (pdev->devfn)) + return PCI_ERS_RESULT_RECOVERED; - pci_save_state(pdev, adapter->pci_state); + pci_set_master(pdev); - state = (state > 0) ? 3 : 0; - pci_set_power_state(pdev, state); - msec_delay(200); + netif_carrier_off(netdev); + netif_stop_queue(netdev); + ixgb_reset(adapter); - return 0; + /* Make sure the EEPROM is good */ + if(!ixgb_validate_eeprom_checksum(&adapter->hw)) { + DPRINTK(PROBE, ERR, "After reset, the EEPROM checksum is not valid.\n"); + return PCI_ERS_RESULT_DISCONNECT; + } + ixgb_get_ee_mac_addr(&adapter->hw, netdev->dev_addr); + memcpy(netdev->perm_addr, netdev->dev_addr, netdev->addr_len); + + if(!is_valid_ether_addr(netdev->perm_addr)) { + DPRINTK(PROBE, ERR, "After reset, invalid MAC address.\n"); + return PCI_ERS_RESULT_DISCONNECT; + } + + return PCI_ERS_RESULT_RECOVERED; } -#ifdef CONFIG_NET_POLL_CONTROLLER -/* - * Polling 'interrupt' - used by things like netconsole to send skbs - * without having to re-enable interrupts. It's not called while - * the interrupt routine is executing. +/** + * ixgb_io_resume - called when its OK to resume normal operations + * @pdev pointer to pci device with error + * + * The error recovery driver tells us that its OK to resume + * normal operation. Implementation resembles the second-half + * of the ixgb_probe() routine. */ - -static void ixgb_netpoll(struct net_device *dev) +static void ixgb_io_resume (struct pci_dev *pdev) { - struct ixgb_adapter *adapter = dev->priv; - disable_irq(adapter->pdev->irq); - ixgb_intr(adapter->pdev->irq, dev, NULL); - enable_irq(adapter->pdev->irq); + struct net_device *netdev = pci_get_drvdata(pdev); + struct ixgb_adapter *adapter = netdev_priv(netdev); + + pci_set_master(pdev); + + if(netif_running(netdev)) { + if(ixgb_up(adapter)) { + printk ("ixgb: can't bring device back up after reset\n"); + return; + } + } + + netif_device_attach(netdev); + mod_timer(&adapter->watchdog_timer, jiffies); } -#endif /* ixgb_main.c */