e1000_main.c

linux2.6.16版本

	if ((err = register_netdev(netdev)))
		goto err_register;

	DPRINTK(PROBE, INFO, "Intel(R) PRO/1000 Network Connection\n");

	cards_found++;
	return 0;

err_register:
err_sw_init:
err_eeprom:
	iounmap(adapter->hw.hw_addr);
err_ioremap:
	free_netdev(netdev);
err_alloc_etherdev:
	pci_release_regions(pdev);
	return err;
}

/**
 * e1000_remove - Device Removal Routine
 * @pdev: PCI device information struct
 *
 * e1000_remove is called by the PCI subsystem to alert the driver
 * that it should release a PCI device.  The could be caused by a
 * Hot-Plug event, or because the driver is going to be removed from
 * memory.
 **/

static void __devexit
e1000_remove(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct e1000_adapter *adapter = netdev_priv(netdev);
	uint32_t manc;
#ifdef CONFIG_E1000_NAPI
	int i;
#endif

	flush_scheduled_work();

	if (adapter->hw.mac_type >= e1000_82540 &&
	   adapter->hw.media_type == e1000_media_type_copper) {
		manc = E1000_READ_REG(&adapter->hw, MANC);
		if (manc & E1000_MANC_SMBUS_EN) {
			manc |= E1000_MANC_ARP_EN;
			E1000_WRITE_REG(&adapter->hw, MANC, manc);
		}
	}

	/* 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_rx_queues; i++)
		__dev_put(&adapter->polling_netdev[i]);
#endif

	if (!e1000_check_phy_reset_block(&adapter->hw))
		e1000_phy_hw_reset(&adapter->hw);

	kfree(adapter->tx_ring);
	kfree(adapter->rx_ring);
#ifdef CONFIG_E1000_NAPI
	kfree(adapter->polling_netdev);
#endif

	iounmap(adapter->hw.hw_addr);
	pci_release_regions(pdev);

#ifdef CONFIG_E1000_MQ
	free_percpu(adapter->cpu_netdev);
	free_percpu(adapter->cpu_tx_ring);
#endif
	free_netdev(netdev);

	pci_disable_device(pdev);
}

/**
 * e1000_sw_init - Initialize general software structures (struct e1000_adapter)
 * @adapter: board private structure to initialize
 *
 * e1000_sw_init initializes the Adapter private data structure.
 * Fields are initialized based on PCI device information and
 * OS network device settings (MTU size).
 **/

static int __devinit
e1000_sw_init(struct e1000_adapter *adapter)
{
	struct e1000_hw *hw = &adapter->hw;
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
#ifdef CONFIG_E1000_NAPI
	int i;
#endif

	/* PCI config space info */

	hw->vendor_id = pdev->vendor;
	hw->device_id = pdev->device;
	hw->subsystem_vendor_id = pdev->subsystem_vendor;
	hw->subsystem_id = pdev->subsystem_device;

	pci_read_config_byte(pdev, PCI_REVISION_ID, &hw->revision_id);

	pci_read_config_word(pdev, PCI_COMMAND, &hw->pci_cmd_word);

	adapter->rx_buffer_len = E1000_RXBUFFER_2048;
	adapter->rx_ps_bsize0 = E1000_RXBUFFER_256;
	hw->max_frame_size = netdev->mtu +
			     ENET_HEADER_SIZE + ETHERNET_FCS_SIZE;
	hw->min_frame_size = MINIMUM_ETHERNET_FRAME_SIZE;

	/* identify the MAC */

	if (e1000_set_mac_type(hw)) {
		DPRINTK(PROBE, ERR, "Unknown MAC Type\n");
		return -EIO;
	}

	/* initialize eeprom parameters */

	if (e1000_init_eeprom_params(hw)) {
		E1000_ERR("EEPROM initialization failed\n");
		return -EIO;
	}

	switch (hw->mac_type) {
	default:
		break;
	case e1000_82541:
	case e1000_82547:
	case e1000_82541_rev_2:
	case e1000_82547_rev_2:
		hw->phy_init_script = 1;
		break;
	}

	e1000_set_media_type(hw);

	hw->wait_autoneg_complete = FALSE;
	hw->tbi_compatibility_en = TRUE;
	hw->adaptive_ifs = TRUE;

	/* Copper options */

	if (hw->media_type == e1000_media_type_copper) {
		hw->mdix = AUTO_ALL_MODES;
		hw->disable_polarity_correction = FALSE;
		hw->master_slave = E1000_MASTER_SLAVE;
	}

#ifdef CONFIG_E1000_MQ
	/* Number of supported queues */
	switch (hw->mac_type) {
	case e1000_82571:
	case e1000_82572:
		/* 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_tx_queues = 1;
		adapter->num_rx_queues = 1;
		break;
	}
	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_tx_queues = 1;
	adapter->num_rx_queues = 1;
#endif

	if (e1000_alloc_queues(adapter)) {
		DPRINTK(PROBE, ERR, "Unable to allocate memory for queues\n");
		return -ENOMEM;
	}

#ifdef CONFIG_E1000_NAPI
	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);
	}
	spin_lock_init(&adapter->tx_queue_lock);
#endif

	atomic_set(&adapter->irq_sem, 1);
	spin_lock_init(&adapter->stats_lock);

	return 0;
}

/**
 * e1000_alloc_queues - Allocate memory for all rings
 * @adapter: board private structure to initialize
 *
 * We allocate one ring per queue at run-time since we don't know the
 * number of queues at compile-time.  The polling_netdev array is
 * intended for Multiqueue, but should work fine with a single queue.
 **/

static int __devinit
e1000_alloc_queues(struct e1000_adapter *adapter)
{
	int size;

	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_rx_queues;
	adapter->rx_ring = kmalloc(size, GFP_KERNEL);
	if (!adapter->rx_ring) {
		kfree(adapter->tx_ring);
		return -ENOMEM;
	}
	memset(adapter->rx_ring, 0, size);

#ifdef CONFIG_E1000_NAPI
	size = sizeof(struct net_device) * adapter->num_rx_queues;
	adapter->polling_netdev = kmalloc(size, GFP_KERNEL);
	if (!adapter->polling_netdev) {
		kfree(adapter->tx_ring);
		kfree(adapter->rx_ring);
		return -ENOMEM;
	}
	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;
}

#ifdef CONFIG_E1000_MQ
static void __devinit
e1000_setup_queue_mapping(struct e1000_adapter *adapter)
{
	int i, cpu;

	adapter->rx_sched_call_data.func = e1000_rx_schedule;
	adapter->rx_sched_call_data.info = adapter->netdev;
	cpus_clear(adapter->rx_sched_call_data.cpumask);

	adapter->cpu_netdev = alloc_percpu(struct net_device *);
	adapter->cpu_tx_ring = alloc_percpu(struct e1000_tx_ring *);

	lock_cpu_hotplug();
	i = 0;
	for_each_online_cpu(cpu) {
		*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_rx_queues) {
			*per_cpu_ptr(adapter->cpu_netdev, cpu) = &adapter->polling_netdev[i];
			adapter->rx_ring[i].cpu = cpu;
			cpu_set(cpu, adapter->cpumask);
		} else
			*per_cpu_ptr(adapter->cpu_netdev, cpu) = NULL;

		i++;
	}
	unlock_cpu_hotplug();
}
#endif

/**
 * e1000_open - Called when a network interface is made active
 * @netdev: network interface device structure
 *
 * Returns 0 on success, negative value on failure
 *
 * The open entry point is called when a network interface is made
 * active by the system (IFF_UP).  At this point all resources needed
 * for transmit and receive operations are allocated, the interrupt
 * handler is registered with the OS, the watchdog timer is started,
 * and the stack is notified that the interface is ready.
 **/

static int
e1000_open(struct net_device *netdev)
{
	struct e1000_adapter *adapter = netdev_priv(netdev);
	int err;

	/* allocate transmit descriptors */

	if ((err = e1000_setup_all_tx_resources(adapter)))
		goto err_setup_tx;

	/* allocate receive descriptors */

	if ((err = e1000_setup_all_rx_resources(adapter)))
		goto err_setup_rx;

	if ((err = e1000_up(adapter)))
		goto err_up;
	adapter->mng_vlan_id = E1000_MNG_VLAN_NONE;
	if ((adapter->hw.mng_cookie.status &
			  E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) {
		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:
	e1000_free_all_rx_resources(adapter);
err_setup_rx:
	e1000_free_all_tx_resources(adapter);
err_setup_tx:
	e1000_reset(adapter);

	return err;
}

/**
 * e1000_close - Disables a network interface
 * @netdev: network interface device structure
 *
 * Returns 0, this is not allowed to fail
 *
 * The close entry point is called when an interface is de-activated
 * by the OS.  The hardware is still under the drivers control, but
 * needs to be disabled.  A global MAC reset is issued to stop the
 * hardware, and all transmit and receive resources are freed.
 **/

static int
e1000_close(struct net_device *netdev)
{
	struct e1000_adapter *adapter = netdev_priv(netdev);

	e1000_down(adapter);

	e1000_free_all_tx_resources(adapter);
	e1000_free_all_rx_resources(adapter);

	if ((adapter->hw.mng_cookie.status &
			  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;
}

/**
 * e1000_check_64k_bound - check that memory doesn't cross 64kB boundary
 * @adapter: address of board private structure
 * @start: address of beginning of memory
 * @len: length of memory
 **/
static inline boolean_t
e1000_check_64k_bound(struct e1000_adapter *adapter,
		      void *start, unsigned long len)
{
	unsigned long begin = (unsigned long) start;
	unsigned long end = begin + len;

	/* First rev 82545 and 82546 need to not allow any memory
	 * write location to cross 64k boundary due to errata 23 */
	if (adapter->hw.mac_type == e1000_82545 ||
	    adapter->hw.mac_type == e1000_82546) {
		return ((begin ^ (end - 1)) >> 16) != 0 ? FALSE : TRUE;
	}

	return TRUE;
}

/**
 * e1000_setup_tx_resources - allocate Tx resources (Descriptors)
 * @adapter: board private structure
 * @txdr:    tx descriptor ring (for a specific queue) to setup
 *
 * Return 0 on success, negative on failure
 **/

static int
e1000_setup_tx_resources(struct e1000_adapter *adapter,
                         struct e1000_tx_ring *txdr)
{
	struct pci_dev *pdev = adapter->pdev;
	int size;

	size = sizeof(struct e1000_buffer) * txdr->count;

	txdr->buffer_info = vmalloc_node(size, pcibus_to_node(pdev->bus));
	if (!txdr->buffer_info) {
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the transmit descriptor ring\n");
		return -ENOMEM;
	}
	memset(txdr->buffer_info, 0, size);

	/* round up to nearest 4K */

	txdr->size = txdr->count * sizeof(struct e1000_tx_desc);
	E1000_ROUNDUP(txdr->size, 4096);

	txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma);
	if (!txdr->desc) {
setup_tx_desc_die:
		vfree(txdr->buffer_info);
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the transmit descriptor ring\n");
		return -ENOMEM;
	}

	/* Fix for errata 23, can't cross 64kB boundary */