[Intel-wired-lan] [PATCH v3 03/11] igc: Add netdev

Neftin, Sasha sasha.neftin at intel.com
Mon Jul 2 13:05:27 UTC 2018


On 6/27/2018 03:32, Shannon Nelson wrote:
> On 6/24/2018 1:45 AM, Sasha Neftin wrote:
>> Now that we have the ability to configure the basic settings on the 
>> device
>> we can start allocating and configuring a netdev for the interface.
>>
>> Sasha Neftin (v2):
>> added description
>>
>> Sasha Neftin (v3):
>> minor cosmetic changes
>>
>> Signed-off-by: Sasha Neftin <sasha.neftin at intel.com>
>> ---
>>   drivers/net/ethernet/intel/igc/e1000_defines.h |  15 +
>>   drivers/net/ethernet/intel/igc/e1000_hw.h      |   1 +
>>   drivers/net/ethernet/intel/igc/igc.h           |  48 +++
>>   drivers/net/ethernet/intel/igc/igc_main.c      | 490 
>> ++++++++++++++++++++++++-
>>   4 files changed, 551 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/intel/igc/e1000_defines.h 
>> b/drivers/net/ethernet/intel/igc/e1000_defines.h
>> index 6831a0864bb4..66b05898eb00 100644
>> --- a/drivers/net/ethernet/intel/igc/e1000_defines.h
>> +++ b/drivers/net/ethernet/intel/igc/e1000_defines.h
>> @@ -4,6 +4,8 @@
>>   #ifndef _E1000_DEFINES_H_
>>   #define _E1000_DEFINES_H_
>> +#define E1000_CTRL_EXT_DRV_LOAD         0x10000000 /* Drv loaded bit 
>> for FW */
>> +
>>   #define E1000_CTRL_EXT_LINK_MODE_PCIE_SERDES    0x00C00000
>>   /* Priority on PCI. 0=rx,1=fair */
>>   #define E1000_CTRL_PRIOR            0x00000004
>> @@ -28,6 +30,16 @@
>>   #define E1000_PCI_PMCSR            0x44
>>   #define E1000_PCI_PMCSR_D3        0x03
>> +/* Receive Address
>> + * Number of high/low register pairs in the RAR. The RAR (Receive 
>> Address
>> + * Registers) holds the directed and multicast addresses that we 
>> monitor.
>> + * Technically, we have 16 spots.  However, we reserve one of these 
>> spots
>> + * (RAR[15]) for our directed address used by controllers with
>> + * manageability enabled, allowing us room for 15 multicast addresses.
>> + */
>> +#define E1000_RAH_AV        0x80000000 /* Receive descriptor valid */
>> +#define E1000_RAH_POOL_1    0x00040000
>> +
>>   /* Error Codes */
>>   #define E1000_SUCCESS                0
>>   #define E1000_ERR_NVM                1
>> @@ -37,6 +49,9 @@
>>   #define E1000_ERR_MAC_INIT            5
>>   #define E1000_ERR_RESET                9
>> +/* PBA constants */
>> +#define E1000_PBA_34K            0x0022
>> +
>>   /* Device Status */
>>   #define E1000_STATUS_FD        0x00000001      /* Full 
>> duplex.0=half,1=full */
>>   #define E1000_STATUS_LU        0x00000002      /* Link 
>> up.0=no,1=link */
>> diff --git a/drivers/net/ethernet/intel/igc/e1000_hw.h 
>> b/drivers/net/ethernet/intel/igc/e1000_hw.h
>> index b8f82f4ba998..6abe722f0bc4 100644
>> --- a/drivers/net/ethernet/intel/igc/e1000_hw.h
>> +++ b/drivers/net/ethernet/intel/igc/e1000_hw.h
>> @@ -87,6 +87,7 @@ struct e1000_mac_info {
>>       bool autoneg;
>>       bool autoneg_failed;
>> +    bool get_link_status;
>>   };
>>   struct e1000_bus_info {
>> diff --git a/drivers/net/ethernet/intel/igc/igc.h 
>> b/drivers/net/ethernet/intel/igc/igc.h
>> index bd732390bd4b..29df87285b9b 100644
>> --- a/drivers/net/ethernet/intel/igc/igc.h
>> +++ b/drivers/net/ethernet/intel/igc/igc.h
>> @@ -28,15 +28,63 @@
>>   extern char igc_driver_name[];
>>   extern char igc_driver_version[];
>> +/* Transmit and receive queues */
>> +#define IGC_MAX_RX_QUEUES                 4
>> +#define IGC_MAX_TX_QUEUES                 4
>> +
>> +#define MAX_Q_VECTORS                     10
>> +#define MAX_STD_JUMBO_FRAME_SIZE        9216
>> +
>> +enum e1000_state_t {
>> +     __IGC_TESTING,
> 
> This mixing of e1000 and IGC is just wrong...
> Can you tell I don't like it?
> 
Yes, you are right, I agree. Fix will be applied to v4.
>> +    __IGC_RESETTING,
> 
> funky inden >
Fix will be applied to v4.

>> +     __IGC_DOWN,
>> +     __IGC_PTP_TX_IN_PROGRESS,
>> +};
>> +
>> +struct igc_q_vector {
>> +    struct igc_adapter *adapter;    /* backlink */
>> +
>> +    struct napi_struct napi;
>> +};
>> +
>> +struct igc_mac_addr {
>> +    u8 addr[ETH_ALEN];
>> +    u8 queue;
>> +    u8 state; /* bitmask */
>> +};
>> +
>> +#define IGC_MAC_STATE_DEFAULT   0x1
>> +#define IGC_MAC_STATE_MODIFIED  0x2
>> +#define IGC_MAC_STATE_IN_USE    0x4
>> +
>>   /* Board specific private data structure */
>>   struct igc_adapter {
>> +    struct net_device *netdev;
>> +
>> +    unsigned long state;
>> +    unsigned int flags;
> 
> These state and flags fields should probably be specifically defined as 
> u32 or u64.
> 
Let me do not agree with you here. I've a two arguments:
1. general set_bit/clear_bit methods is used this type and do not allow 
pass another type. Actually we use set_bit/clear_bit a lot of.
2. i prefer stay consistently with another driver where is used same value.
> Is there a reason that state is a long and flags an int?  Is there an 
> expectation of different sizes?
> 
>> +    unsigned int num_q_vectors;
>> +    u16 link_speed;
>> +    u16 link_duplex;
>> +
>> +    u8 port_num;
>> +
>>       u8 __iomem *io_addr;
>> +    struct work_struct watchdog_task;
>> +
>> +    int msg_enable;
>> +    u32 max_frame_size;
>>       /* OS defined structs */
>>       struct pci_dev *pdev;
>>       /* structs defined in e1000_hw.h */
>>       struct e1000_hw hw;
>> +
>> +    struct igc_q_vector *q_vector[MAX_Q_VECTORS];
>> +
>> +    struct igc_mac_addr *mac_table;
>>   };
>>   #endif /* _IGC_H_ */
>> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c 
>> b/drivers/net/ethernet/intel/igc/igc_main.c
>> index ec3451dad91e..520f49be8e19 100644
>> --- a/drivers/net/ethernet/intel/igc/igc_main.c
>> +++ b/drivers/net/ethernet/intel/igc/igc_main.c
>> @@ -3,6 +3,8 @@
>>   #include <linux/module.h>
>>   #include <linux/types.h>
>> +#include <linux/if_vlan.h>
>> +#include <linux/aer.h>
>>   #include "igc.h"
>>   #include "e1000_hw.h"
>> @@ -10,6 +12,7 @@
>>   #define DRV_VERSION    "0.0.1-k"
>>   #define DRV_SUMMARY    "Intel(R) 2.5G Ethernet Linux Driver"
>> +static int debug = -1;
>>   char igc_driver_name[] = "igc";
>>   char igc_driver_version[] = DRV_VERSION;
>>   static const char igc_driver_string[] = DRV_SUMMARY;
>> @@ -25,8 +28,371 @@ static const struct pci_device_id igc_pci_tbl[] = {
>>   MODULE_DEVICE_TABLE(pci, igc_pci_tbl);
>> -/* Forward declaration */
>> +/* forward declaration */
>>   static int igc_sw_init(struct igc_adapter *);
>> +static void igc_configure(struct igc_adapter *adapter);
>> +static void igc_power_down_link(struct igc_adapter *adapter);
>> +static void igc_set_default_mac_filter(struct igc_adapter *adapter);
> 
> If you change the order of the functions in the file you shouldn't need 
> the forward decl crutch.
> 
how for example? I am afraid from a mess.
>> +
>> +void igc_reset(struct igc_adapter *adapter)
>> +{
>> +    if (!netif_running(adapter->netdev))
>> +        igc_power_down_link(adapter);
>> +}
>> +
>> +/**
>> + *  igc_power_up_link - Power up the phy/serdes link
>> + *  @adapter: address of board private structure
>> + **/
>> +void igc_power_up_link(struct igc_adapter *adapter)
>> +{
>> +}
>> +
>> +/**
>> + *  igc_power_down_link - Power down the phy/serdes link
>> + *  @adapter: address of board private structure
>> + **/
>> +static void igc_power_down_link(struct igc_adapter *adapter)
>> +{
>> +}
>> +
>> +/**
>> + *  igc_release_hw_control - release control of the h/w to f/w
>> + *  @adapter: address of board private structure
>> + *
>> + *  igc_release_hw_control resets CTRL_EXT:DRV_LOAD bit.
>> + *  For ASF and Pass Through versions of f/w this means that the
>> + *  driver is no longer loaded.
>> + **/
>> +static void igc_release_hw_control(struct igc_adapter *adapter)
>> +{
>> +    struct e1000_hw *hw = &adapter->hw;
>> +    u32 ctrl_ext;
>> +
>> +    /* Let firmware take over control of h/w */
>> +    ctrl_ext = rd32(E1000_CTRL_EXT);
>> +    wr32(E1000_CTRL_EXT,
>> +         ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
>> +}
>> +
>> +/**
>> + *  igc_get_hw_control - get control of the h/w from f/w
>> + *  @adapter: address of board private structure
>> + *
>> + *  igc_get_hw_control sets CTRL_EXT:DRV_LOAD bit.
>> + *  For ASF and Pass Through versions of f/w this means that
>> + *  the driver is loaded.
>> + **/
>> +static void igc_get_hw_control(struct igc_adapter *adapter)
>> +{
>> +    struct e1000_hw *hw = &adapter->hw;
>> +    u32 ctrl_ext;
>> +
>> +    /* Let firmware know the driver has taken over */
>> +    ctrl_ext = rd32(E1000_CTRL_EXT);
>> +    wr32(E1000_CTRL_EXT,
>> +         ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
>> +}
>> +
>> +/**
>> + *  igc_set_mac - Change the Ethernet Address of the NIC
>> + *  @netdev: network interface device structure
>> + *  @p: pointer to an address structure
>> + *
>> + *  Returns 0 on success, negative on failure
>> + **/
>> +static int igc_set_mac(struct net_device *netdev, void *p)
>> +{
>> +    struct igc_adapter *adapter = netdev_priv(netdev);
>> +    struct e1000_hw *hw = &adapter->hw;
>> +    struct sockaddr *addr = p;
>> +
>> +    if (!is_valid_ether_addr(addr->sa_data))
>> +        return -EADDRNOTAVAIL;
>> +
>> +    memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len);
>> +    memcpy(hw->mac.addr, addr->sa_data, netdev->addr_len);
>> +
>> +    /* set the correct pool for the new PF MAC address in entry 0 */
>> +    igc_set_default_mac_filter(adapter);
>> +
>> +    return 0;
>> +}
>> +
>> +static netdev_tx_t igc_xmit_frame(struct sk_buff *skb,
>> +                  struct net_device *netdev)
>> +{
>> +    dev_kfree_skb_any(skb);
>> +    return NETDEV_TX_OK;
>> +}
>> +
>> +/**
>> + *  igc_ioctl - I/O control method
>> + *  @netdev: network interface device structure
>> + *  @ifreq: frequency
>> + *  @cmd: command
>> + **/
>> +static int igc_ioctl(struct net_device *netdev, struct ifreq *ifr, 
>> int cmd)
>> +{
>> +    switch (cmd) {
>> +    default:
>> +        return -EOPNOTSUPP;
>> +    }
>> +}
>> +
>> +/**
>> + *  igc_up - Open the interface and prepare it to handle traffic
>> + *  @adapter: board private structure
>> + **/
>> +int igc_up(struct igc_adapter *adapter)
>> +{
>> +    int i = 0;
>> +
>> +    /* hardware has been reset, we need to reload some things */
>> +    igc_configure(adapter);
>> +
>> +    clear_bit(__IGC_DOWN, &adapter->state);
>> +
>> +    for (i = 0; i < adapter->num_q_vectors; i++)
>> +        napi_enable(&adapter->q_vector[i]->napi);
>> +
>> +    return 0;
>> +}
>> +
>> +/**
>> + *  igc_down - Close the interface
>> + *  @adapter: board private structure
>> + **/
>> +void igc_down(struct igc_adapter *adapter)
>> +{
>> +    struct net_device *netdev = adapter->netdev;
>> +    int i = 0;
>> +
>> +    set_bit(__IGC_DOWN, &adapter->state);
>> +
>> +    /* set trans_start so we don't get spurious watchdogs during 
>> reset */
>> +    netif_trans_update(netdev);
>> +
>> +    netif_carrier_off(netdev);
>> +    netif_tx_stop_all_queues(netdev);
>> +
>> +    for (i = 0; i < adapter->num_q_vectors; i++)
>> +        napi_disable(&adapter->q_vector[i]->napi);
>> +
>> +    adapter->link_speed = 0;
>> +    adapter->link_duplex = 0;
>> +}
>> +
>> +/**
>> + *  igc_change_mtu - Change the Maximum Transfer Unit
>> + *  @netdev: network interface device structure
>> + *  @new_mtu: new value for maximum frame size
>> + *
>> + *  Returns 0 on success, negative on failure
>> + **/
>> +static int igc_change_mtu(struct net_device *netdev, int new_mtu)
>> +{
>> +    struct igc_adapter *adapter = netdev_priv(netdev);
>> +    struct pci_dev *pdev = adapter->pdev;
>> +    int max_frame = new_mtu + ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN;
>> +
>> +    /* adjust max frame to be at least the size of a standard frame */
>> +    if (max_frame < (ETH_FRAME_LEN + ETH_FCS_LEN))
>> +        max_frame = ETH_FRAME_LEN + ETH_FCS_LEN;
>> +
>> +    while (test_and_set_bit(__IGC_RESETTING, &adapter->state))
>> +        usleep_range(1000, 2000);
>> +
>> +    /* igc_down has a dependency on max_frame_size */
>> +    adapter->max_frame_size = max_frame;
>> +
>> +    if (netif_running(netdev))
>> +        igc_down(adapter);
>> +
>> +    dev_info(&pdev->dev, "changing MTU from %d to %d\n",
>> +         netdev->mtu, new_mtu);
>> +    netdev->mtu = new_mtu;
>> +
>> +    if (netif_running(netdev))
>> +        igc_up(adapter);
>> +    else
>> +        igc_reset(adapter);
>> +
>> +    clear_bit(__IGC_RESETTING, &adapter->state);
>> +
>> +    return 0;
>> +}
>> +
>> +/**
>> + *  igc_update_stats - Update the board statistics counters
>> + *  @adapter: board private structure
>> + **/
>> +void igc_update_stats(struct igc_adapter *adapter)
>> +{
>> +}
>> +
>> +/**
>> + *  igc_get_stats - Get System Network Statistics
>> + *  @netdev: network interface device structure
>> + *
>> + *  Returns the address of the device statistics structure.
>> + *  The statistics are updated here and also from the timer callback.
>> + **/
>> +static struct net_device_stats *igc_get_stats(struct net_device *netdev)
>> +{
>> +    struct igc_adapter *adapter = netdev_priv(netdev);
>> +
>> +    if (!test_bit(__IGC_RESETTING, &adapter->state))
>> +        igc_update_stats(adapter);
>> +
>> +    /* only return the current stats */
>> +    return &netdev->stats;
>> +}
>> +
>> +/**
>> + *  igc_configure - configure the hardware for RX and TX
>> + *  @adapter: private board structure
>> + **/
>> +static void igc_configure(struct igc_adapter *adapter)
>> +{
>> +    igc_get_hw_control(adapter);
>> +}
>> +
>> +/**
>> + *  igc_rar_set_index - Sync RAL[index] and RAH[index] registers with 
>> MAC table
>> + *  @adapter: Pointer to adapter structure
>> + *  @index: Index of the RAR entry which need to be synced with MAC 
>> table
>> + **/
>> +static void igc_rar_set_index(struct igc_adapter *adapter, u32 index)
>> +{
>> +    struct e1000_hw *hw = &adapter->hw;
>> +    u32 rar_low, rar_high;
>> +    u8 *addr = adapter->mac_table[index].addr;
>> +
>> +    /* HW expects these to be in network order when they are plugged
>> +     * into the registers which are little endian.  In order to 
>> guarantee
>> +     * that ordering we need to do an leXX_to_cpup here in order to be
>> +     * ready for the byteswap that occurs with writel
>> +     */
>> +    rar_low = le32_to_cpup((__le32 *)(addr));
>> +    rar_high = le16_to_cpup((__le16 *)(addr + 4));
>> +
>> +    /* Indicate to hardware the Address is Valid. */
>> +    if (adapter->mac_table[index].state & IGC_MAC_STATE_IN_USE) {
>> +        if (is_valid_ether_addr(addr))
>> +            rar_high |= E1000_RAH_AV;
>> +
>> +        rar_high |= E1000_RAH_POOL_1 <<
>> +            adapter->mac_table[index].queue;
>> +    }
>> +
>> +    wr32(E1000_RAL(index), rar_low);
>> +    wrfl();
>> +    wr32(E1000_RAH(index), rar_high);
>> +    wrfl();
>> +}
>> +
>> +/* Set default MAC address for the PF in the first RAR entry */
>> +static void igc_set_default_mac_filter(struct igc_adapter *adapter)
>> +{
>> +    struct igc_mac_addr *mac_table = &adapter->mac_table[0];
>> +
>> +    ether_addr_copy(mac_table->addr, adapter->hw.mac.addr);
>> +    mac_table->state = IGC_MAC_STATE_DEFAULT | IGC_MAC_STATE_IN_USE;
>> +
>> +    igc_rar_set_index(adapter, 0);
>> +}
>> +
>> +/**
>> + * igc_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 __igc_open(struct net_device *netdev, bool resuming)
>> +{
>> +    struct igc_adapter *adapter = netdev_priv(netdev);
>> +    struct e1000_hw *hw = &adapter->hw;
>> +    int i = 0;
>> +
>> +    /* disallow open during test */
>> +
>> +    if (test_bit(__IGC_TESTING, &adapter->state)) {
>> +        WARN_ON(resuming);
>> +        return -EBUSY;
>> +    }
>> +
>> +    netif_carrier_off(netdev);
>> +
>> +    igc_power_up_link(adapter);
>> +
>> +    igc_configure(adapter);
>> +
>> +    /* From here on the code is the same as igc_up() */
> 
> Then why not use igc_up()?
> 
Very good point. igc_up is sligthly different, so this comment is wrong. 
I will remove. Fix will be applied in v4.
>> +    clear_bit(__IGC_DOWN, &adapter->state);
>> +
>> +    for (i = 0; i < adapter->num_q_vectors; i++)
>> +        napi_enable(&adapter->q_vector[i]->napi);
>> +
>> +    /* start the watchdog. */
>> +    hw->mac.get_link_status = 1;
>> +    schedule_work(&adapter->watchdog_task);
>> +
>> +    return E1000_SUCCESS;
>> +}
>> +
>> +int igc_open(struct net_device *netdev)
>> +{
>> +    return __igc_open(netdev, false);
>> +}
>> +
>> +/**
>> + * igc_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 driver's 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 __igc_close(struct net_device *netdev, bool suspending)
>> +{
>> +    struct igc_adapter *adapter = netdev_priv(netdev);
>> +
>> +    WARN_ON(test_bit(__IGC_RESETTING, &adapter->state));
>> +
>> +    igc_down(adapter);
>> +
>> +    igc_release_hw_control(adapter);
>> +
>> +    return 0;
>> +}
>> +
>> +int igc_close(struct net_device *netdev)
>> +{
>> +    if (netif_device_present(netdev) || netdev->dismantle)
>> +        return __igc_close(netdev, false);
>> +    return 0;
>> +}
>> +
>> +static const struct net_device_ops igc_netdev_ops = {
>> +    .ndo_open               = igc_open,
>> +    .ndo_stop               = igc_close,
>> +    .ndo_start_xmit         = igc_xmit_frame,
>> +    .ndo_set_mac_address    = igc_set_mac,
>> +    .ndo_change_mtu         = igc_change_mtu,
>> +    .ndo_get_stats          = igc_get_stats,
>> +    .ndo_do_ioctl           = igc_ioctl,
>> +
>> +};
>>   /* PCIe configuration access */
>>   void igc_read_pci_cfg(struct e1000_hw *hw, u32 reg, u16 *value)
>> @@ -73,6 +439,7 @@ s32 igc_write_pcie_cap_reg(struct e1000_hw *hw, u32 
>> reg, u16 *value)
>>   u32 igc_rd32(struct e1000_hw *hw, u32 reg)
>>   {
>> +    struct igc_adapter *igc = container_of(hw, struct igc_adapter, hw);
>>       u8 __iomem *hw_addr = READ_ONCE(hw->hw_addr);
>>       u32 value = 0;
>> @@ -82,8 +449,13 @@ u32 igc_rd32(struct e1000_hw *hw, u32 reg)
>>       value = readl(&hw_addr[reg]);
>>       /* reads should not return all F's */
>> -    if (!(~value) && (!reg || !(~readl(hw_addr))))
>> +    if (!(~value) && (!reg || !(~readl(hw_addr)))) {
>> +        struct net_device *netdev = igc->netdev;
>> +
>>           hw->hw_addr = NULL;
>> +        netif_device_detach(netdev);
>> +        netdev_err(netdev, "PCIe link lost, device now detached\n");
>> +    }
>>       return value;
>>   }
>> @@ -102,6 +474,7 @@ u32 igc_rd32(struct e1000_hw *hw, u32 reg)
>>   static int igc_probe(struct pci_dev *pdev,
>>                const struct pci_device_id *ent)
>>   {
>> +    struct net_device *netdev;
> 
> Reverse xmas tree
> 
Good. Fix will be applied in v4.
>>       struct igc_adapter *adapter;
>>       struct e1000_hw *hw;
>>       int err, pci_using_dac;
>> @@ -136,8 +509,56 @@ static int igc_probe(struct pci_dev *pdev,
>>       if (err)
>>           goto err_pci_reg;
>> +    pci_enable_pcie_error_reporting(pdev);
>> +
>>       pci_set_master(pdev);
>> -    pci_save_state(pdev);
>> +
>> +    err = -ENOMEM;
>> +    netdev = alloc_etherdev_mq(sizeof(struct igc_adapter),
>> +                   IGC_MAX_TX_QUEUES);
>> +
>> +    if (!netdev)
>> +        goto err_alloc_etherdev;
>> +
>> +    SET_NETDEV_DEV(netdev, &pdev->dev);
>> +
>> +    pci_set_drvdata(pdev, netdev);
>> +    adapter = netdev_priv(netdev);
>> +    adapter->netdev = netdev;
>> +    adapter->pdev = pdev;
>> +    hw = &adapter->hw;
>> +    hw->back = adapter;
>> +    adapter->port_num = hw->bus.func;
>> +    adapter->msg_enable = GENMASK(debug - 1, 0);
>> +
>> +    err = pci_save_state(pdev);
>> +    if (err)
>> +        goto err_ioremap;
>> +
>> +    err = -EIO;
>> +    adapter->io_addr = ioremap(pci_resource_start(pdev, 0),
>> +                   pci_resource_len(pdev, 0));
>> +    if (!adapter->io_addr)
>> +        goto err_ioremap;
>> +
>> +    /* hw->hw_addr can be zeroed, so use adapter->io_addr for unmap */
>> +    hw->hw_addr = adapter->io_addr;
>> +
>> +    netdev->netdev_ops = &igc_netdev_ops;
>> +
>> +    netdev->watchdog_timeo = 5 * HZ;
>> +
>> +    strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1);
>> +
>> +    netdev->mem_start = pci_resource_start(pdev, 0);
>> +    netdev->mem_end = pci_resource_end(pdev, 0);
>> +
>> +    /* PCI config space info */
>> +    hw->vendor_id = pdev->vendor;
>> +    hw->device_id = pdev->device;
>> +    hw->revision_id = pdev->revision;
>> +    hw->subsystem_vendor_id = pdev->subsystem_vendor;
>> +    hw->subsystem_device_id = pdev->subsystem_device;
>>       /* setup the private structure */
>>       err = igc_sw_init(adapter);
>> @@ -146,9 +567,49 @@ static int igc_probe(struct pci_dev *pdev,
>>       igc_get_bus_info_pcie(hw);
>> +    /* MTU range: 68 - 9216 */
>> +    netdev->min_mtu = ETH_MIN_MTU;
>> +    netdev->max_mtu = MAX_STD_JUMBO_FRAME_SIZE;
>> +
>> +    /* reset the hardware with the new settings */
>> +    igc_reset(adapter);
>> +
>> +    /* let the f/w know that the h/w is now under the control of the
>> +     * driver.
>> +     */
>> +    igc_get_hw_control(adapter);
>> +
>> +    strncpy(netdev->name, "eth%d", IFNAMSIZ);
> 
> You're stomping on what you just set a few lines earlier.  Of course, 
> some might argue that you shouldn't put anything here at all and let the 
> system name the device.
> 
my apologies I do not understand you.
>> +    err = register_netdev(netdev);
>> +    if (err)
>> +        goto err_register;
>> +
>> +     /* carrier off reporting is important to ethtool even BEFORE 
>> open */
>> +    netif_carrier_off(netdev);
>> +
>> +    dev_info(&pdev->dev, "@SUMMARY@");
>> +    /* print bus type/speed/width info */
>> +    dev_info(&pdev->dev, "%s: (PCIe:%s:%s) ",
>> +         netdev->name,
>> +         ((hw->bus.speed == e1000_bus_speed_2500) ? "2.5GT/s" :
>> +         (hw->bus.speed == e1000_bus_speed_5000) ? "5.0GT/s" :
>> +         "unknown"),
>> +         ((hw->bus.width == e1000_bus_width_pcie_x4) ? "Width x4" :
>> +         (hw->bus.width == e1000_bus_width_pcie_x2) ? "Width x2" :
>> +         (hw->bus.width == e1000_bus_width_pcie_x1) ? "Width x1" :
>> +         "unknown"));
> 
> How about using the new pcie_print_link_status()?
> 
Good idea. I will keep it in mind and replace in future. Currenly in my 
working kernel this method not exist yet.
>> +    netdev_info(netdev, "MAC: %pM\n", netdev->dev_addr);
>> +
>>       return 0;
>> +err_register:
>> +    igc_release_hw_control(adapter);
>>   err_sw_init:
>> +err_ioremap:
>> +    free_netdev(netdev);
>> +err_alloc_etherdev:
>> +    pci_release_selected_regions(pdev,
>> +                     pci_select_bars(pdev, IORESOURCE_MEM));
>>   err_pci_reg:
>>   err_dma:
>>       pci_disable_device(pdev);
>> @@ -166,9 +627,22 @@ static int igc_probe(struct pci_dev *pdev,
>>    **/
>>   static void igc_remove(struct pci_dev *pdev)
>>   {
>> +    struct net_device *netdev = pci_get_drvdata(pdev);
>> +    struct igc_adapter *adapter = netdev_priv(netdev);
>> +
>> +    set_bit(__IGC_DOWN, &adapter->state);
>> +    flush_scheduled_work();
>> +
>> +    /* Release control of h/w to f/w.  If f/w is AMT enabled, this
>> +     * would have already happened in close and is redundant.
>> +     */
>> +    igc_release_hw_control(adapter);
>> +    unregister_netdev(netdev);
>> +
>>       pci_release_selected_regions(pdev,
>>                        pci_select_bars(pdev, IORESOURCE_MEM));
>> +    free_netdev(netdev);
>>       pci_disable_device(pdev);
>>   }
>> @@ -190,6 +664,7 @@ static struct pci_driver igc_driver = {
>>   static int igc_sw_init(struct igc_adapter *adapter)
>>   {
>>       struct e1000_hw *hw = &adapter->hw;
>> +    struct net_device *netdev = adapter->netdev;
>>       struct pci_dev *pdev = adapter->pdev;
>>       /* PCI config space info */
>> @@ -203,6 +678,12 @@ static int igc_sw_init(struct igc_adapter *adapter)
>>       pci_read_config_word(pdev, PCI_COMMAND, &hw->bus.pci_cmd_word);
>> +    /* set default work limits */
>> +    adapter->max_frame_size = netdev->mtu + ETH_HLEN + ETH_FCS_LEN +
>> +                    VLAN_HLEN;
>> +
>> +    set_bit(__IGC_DOWN, &adapter->state);
>> +
>>       return 0;
>>   }
>> @@ -244,4 +725,7 @@ MODULE_AUTHOR("Intel Corporation, 
>> <linux.nics at intel.com>");
>>   MODULE_DESCRIPTION(DRV_SUMMARY);
>>   MODULE_LICENSE("GPL");
>>   MODULE_VERSION(DRV_VERSION);
>> +
>> +module_param(debug, int, 0);
>> +MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)");
>>   /* igc_main.c */
>>
Hello Shannon,
Thanks you very much for a comments.
Sasha


More information about the Intel-wired-lan mailing list