dpdk PMD
作者：互联网
PMD是Poll Mode Driver的缩写，即基于用户态的轮询机制的驱动
在不考虑vfio的情况下，PMD的结构图如下
虽然PMD是在用户态实现设备驱动，但还是依赖于内核提供的策略。其中uio模块，是内核提供的用户态驱动框架，而igb_uio是DPDK kit中拥有与uio交互，bind指定网卡的内核模块;
当使用DPDK脚本dpdk-devbind来bind网卡时，会通过sysfs与内核交互，让内核使用指定驱动来匹配网卡。具体的行为向/sys/bus/pci/devices/(pci id)/driver_override写入指定驱动名称，或者向/sys/bus/pci/drivers/igb_uio(驱动名称）/new_id写入要绑定网卡的PCI ID。前者是配置设备，让其选择驱动。后者是是配置驱动，让其支持新的PCI设备。按照内核的文档https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-bus-pci，这两个动作都会促使驱动bind新设备。
但是在dpdk-devbind脚本中，还是通过向/sys/bus/pci/drivers/igb_uio(驱动名称）/bind写入pci id来保证bind;
当使用igb_uio bind指定设备后，内核会调用igb_uio注册的struct pci_driver的probe函数，即igbuio_pci_probe;
其会调用uio_register_device，注册uio设备
打开uio设备
应用层DPDK已经可以使用uio设备了。DPDK的应用层代码，会打开uioX设备，在函数pci_uio_alloc_resource中；
当open对应的uio设备时，对应的内核操作为uio_open，其又会调用igb_uio的open函数，流程图如下
igb_uio的默认中断模式为RTE_INTR_MODE_MSIX，在igbuio_pci_enable_interrupts的关键代码如下
static int
igbuio_pci_enable_interrupts(struct rte_uio_pci_dev *udev)
{
int err = 0;
switch (igbuio_intr_mode_preferred) {
case RTE_INTR_MODE_MSIX:
/* Only 1 msi-x vector needed */
#ifndef HAVE_ALLOC_IRQ_VECTORS
msix_entry.entry = 0;
if (pci_enable_msix(udev->pdev, &msix_entry, 1) == 0) {
dev_dbg(&udev->pdev->dev, "using MSI-X");
udev->info.irq_flags = IRQF_NO_THREAD;
udev->info.irq = msix_entry.vector;
udev->mode = RTE_INTR_MODE_MSIX;
break;
}
#else
if (pci_alloc_irq_vectors(udev->pdev, 1, 1, PCI_IRQ_MSIX) == 1) {
dev_dbg(&udev->pdev->dev, "using MSI-X");
udev->info.irq_flags = IRQF_NO_THREAD;
udev->info.irq = pci_irq_vector(udev->pdev, 0);
udev->mode = RTE_INTR_MODE_MSIX;
break;
}
#endif
default:
dev_err(&udev->pdev->dev, "invalid IRQ mode %u",
igbuio_intr_mode_preferred);
udev->info.irq = UIO_IRQ_NONE;
err = -EINVAL;
}
if (udev->info.irq != UIO_IRQ_NONE)
err = request_irq(udev->info.irq, igbuio_pci_irqhandler,
udev->info.irq_flags, udev->info.name,
udev);
dev_info(&udev->pdev->dev, "uio device registered with irq %ld\n",
udev->info.irq);
return err;
}
当打开uio设备时，igb_uio注册了一个中断。这时大家应该有个疑问，PMD不是用户态轮询设备吗？为什么还要申请中断，注册中断处理函数呢？这是因为，即使应用层可以通过uio来实现设备驱动，但是设备的某些事件还是需要内核进行响应，然后通知应用层；相当于中断的上半部 和下半部
上半部中断服务函数： 主要通知一个event到user比如： 唤醒在idev->wait等待队列中的task
/**
* This is interrupt handler which will check if the interrupt is for the right device.
* If yes, disable it here and will be enable later.
*/
static irqreturn_t
igbuio_pci_irqhandler(int irq, void *dev_id)
{
struct rte_uio_pci_dev *udev = (struct rte_uio_pci_dev *)dev_id;
struct uio_info *info = &udev->info;
/* Legacy mode need to mask in hardware */
if (udev->mode == RTE_INTR_MODE_LEGACY &&
!pci_check_and_mask_intx(udev->pdev))
return IRQ_NONE;
uio_event_notify(info);
/* Message signal mode, no share IRQ and automasked */
return IRQ_HANDLED;
}
当DPDK的app启动时，会进行EAL初始化，如下图：
以ixgbe驱动为例详细说明：
RTE_PMD_REGISTER_PCI(net_ixgbe, rte_ixgbe_pmd);
#define RTE_PMD_EXPORT_NAME(name, idx) \
static const char RTE_PMD_EXPORT_NAME_ARRAY(this_pmd_name, idx) \
__attribute__((used)) = RTE_STR(name)
/** Helper for PCI device registration from driver (eth, crypto) instance */
#define RTE_PMD_REGISTER_PCI(nm, pci_drv) \
RTE_INIT(pciinitfn_ ##nm) \
{\
(pci_drv).driver.name = RTE_STR(nm);\
rte_pci_register(&pci_drv); \
} \
RTE_PMD_EXPORT_NAME(nm, __COUNTER__)
/* register a driver */
void
rte_pci_register(struct rte_pci_driver *driver)
{
TAILQ_INSERT_TAIL(&rte_pci_bus.driver_list, driver, next);
driver->bus = &rte_pci_bus;
}
RTE_PMD_REGISTER_PCI为dpdk定义的宏，使用了GNU C提供的“__attribute__（constructor）”机制，使得注册设备驱动的过程在main函数执行之前完成。
这样就有了设备驱动类型、设备驱动的初始化函数
使用attribute的constructor属性，在MAIN函数执行前，就执行pmd_driver_register()函数，将pmd_igb_drv驱动挂到全局dev_driver_list链表上。
2、通过读取/sys/bus/pci/devices/目录下的信息，扫描当前系统的PCI设备，并初始化，并按照PCI地址从大到小的顺序挂在到pci_debice_list上。
3、rte_bus_probe------>rte_eal_pci_probe_one_driver();通过比对PCI_ID的vendor_id、device_id、subsystem_vendor_id、subsystem_device_id四个字段判断pci设备和pci驱动是否匹配。
PCI设备和PCI驱动匹配后，调用pci_map_device()函数为该PCI设备创建map resource； pci_uio_map_resource()函数为pci设备在虚拟地址空间映射pci资源，后续直接通过操作内存来操作pci设备
4、rte_bus_probe---------->调用rte_eth_dev_create 以及 eth_ixgbe_pci_probe()初始化PCI设备； 并通过eth_ixgbe_dev_init 函数分配一个网卡设备，并在全局数组rte_eth_dev_data[]中为网卡设备的数据域分配空间。
同时初始化网卡设备。首先设置网卡设备的操作函数集，以及收包、发包函数
/* Launch threads, called at application init(). */
int
rte_eal_init(int argc, char **argv)
{
int i, fctret, ret;
pthread_t thread_id;
static rte_atomic32_t run_once = RTE_ATOMIC32_INIT(0);
const char *p;
static char logid[PATH_MAX];
char cpuset[RTE_CPU_AFFINITY_STR_LEN];
char thread_name[RTE_MAX_THREAD_NAME_LEN];
/* checks if the machine is adequate */
if (!rte_cpu_is_supported()) {
rte_eal_init_alert("unsupported cpu type.");
rte_errno = ENOTSUP;
return -1;
}
//操作静态局部变量run_once确保函数只执行一次
if (!rte_atomic32_test_and_set(&run_once)) {
rte_eal_init_alert("already called initialization.");
rte_errno = EALREADY;
return -1;
}
//dpdk: EAL init args: -c 3 -n 4 --file-prefix vpp --master-lcore 0
p = strrchr(argv[0], '/');
strlcpy(logid, p ? p + 1 : argv[0], sizeof(logid));
thread_id = pthread_self();//获取主线程的线程ID
//初始化结构体struct internal_config
eal_reset_internal_config(&internal_config);
/* set log level as early as possible设置log level */
eal_log_level_parse(argc, argv);
//赋值全局结构struct lcore_config
if (rte_eal_cpu_init() < 0) {//赋值全局结构struct lcore_config
rte_eal_init_alert("Cannot detect lcores.");
rte_errno = ENOTSUP;
return -1;
}
fctret = eal_parse_args(argc, argv);
if (fctret < 0) {
rte_eal_init_alert("Invalid 'command line' arguments.");
rte_errno = EINVAL;
rte_atomic32_clear(&run_once);
return -1;
}
if (eal_plugins_init() < 0) {//EAL的“-d”选项可以指定需要载入的动态链接库)
rte_eal_init_alert("Cannot init plugins\n");
rte_errno = EINVAL;
rte_atomic32_clear(&run_once);
return -1;
}
if (eal_option_device_parse()) {
rte_errno = ENODEV;
rte_atomic32_clear(&run_once);
return -1;
}
rte_config_init();
if (rte_eal_intr_init() < 0) {
rte_eal_init_alert("Cannot init interrupt-handling thread\n");
return -1;
}
/* Put mp channel init before bus scan so that we can init the vdev
* bus through mp channel in the secondary process before the bus scan.
*/
if (rte_mp_channel_init() < 0) {
rte_eal_init_alert("failed to init mp channel\n");
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
rte_errno = EFAULT;
return -1;
}
}
if (rte_bus_scan()) {//scan 扫描设备的device信息
///sys/bus/pci/devices”目录下的所有pci地址，逐个获取对应的pci地址、pci id、sriov使能时的vf个数、亲和的numa、设备地址空间、驱动类型等
rte_eal_init_alert("Cannot scan the buses for devices\n");
rte_errno = ENODEV;
rte_atomic32_clear(&run_once);
return -1;
}
/* autodetect the iova mapping mode (default is iova_pa) */
rte_eal_get_configuration()->iova_mode = rte_bus_get_iommu_class();
/* Workaround for KNI which requires physical address to work */
if (rte_eal_get_configuration()->iova_mode == RTE_IOVA_VA &&
rte_eal_check_module("rte_kni") == 1) {
rte_eal_get_configuration()->iova_mode = RTE_IOVA_PA;
RTE_LOG(WARNING, EAL,
"Some devices want IOVA as VA but PA will be used because.. "
"KNI module inserted\n");
}
if (internal_config.no_hugetlbfs == 0) {
/* rte_config isn't initialized yet */
ret = internal_config.process_type == RTE_PROC_PRIMARY ?
eal_hugepage_info_init() :
eal_hugepage_info_read();
if (ret < 0) {
rte_eal_init_alert("Cannot get hugepage information.");
rte_errno = EACCES;
rte_atomic32_clear(&run_once);
return -1;
}
}
if (internal_config.memory == 0 && internal_config.force_sockets == 0) {
if (internal_config.no_hugetlbfs)
internal_config.memory = MEMSIZE_IF_NO_HUGE_PAGE;
}
if (internal_config.vmware_tsc_map == 1) {
#ifdef RTE_LIBRTE_EAL_VMWARE_TSC_MAP_SUPPORT
rte_cycles_vmware_tsc_map = 1;
RTE_LOG (DEBUG, EAL, "Using VMWARE TSC MAP, "
"you must have monitor_control.pseudo_perfctr = TRUE\n");
#else
RTE_LOG (WARNING, EAL, "Ignoring --vmware-tsc-map because "
"RTE_LIBRTE_EAL_VMWARE_TSC_MAP_SUPPORT is not set\n");
#endif
}
rte_srand(rte_rdtsc());
if (rte_eal_log_init(logid, internal_config.syslog_facility) < 0) {
rte_eal_init_alert("Cannot init logging.");
rte_errno = ENOMEM;
rte_atomic32_clear(&run_once);
return -1;
}
#ifdef VFIO_PRESENT
if (rte_eal_vfio_setup() < 0) {
rte_eal_init_alert("Cannot init VFIO\n");
rte_errno = EAGAIN;
rte_atomic32_clear(&run_once);
return -1;
}
#endif
/* in secondary processes, memory init may allocate additional fbarrays
* not present in primary processes, so to avoid any potential issues,
* initialize memzones first.
*/
if (rte_eal_memzone_init() < 0) {
rte_eal_init_alert("Cannot init memzone\n");
rte_errno = ENODEV;
return -1;
}
if (rte_eal_memory_init() < 0) {
rte_eal_init_alert("Cannot init memory\n");
rte_errno = ENOMEM;
return -1;
}
/* the directories are locked during eal_hugepage_info_init */
eal_hugedirs_unlock();
if (rte_eal_malloc_heap_init() < 0) {
rte_eal_init_alert("Cannot init malloc heap\n");
rte_errno = ENODEV;
return -1;
}
if (rte_eal_tailqs_init() < 0) {
rte_eal_init_alert("Cannot init tail queues for objects\n");
rte_errno = EFAULT;
return -1;
}
if (rte_eal_alarm_init() < 0) {
rte_eal_init_alert("Cannot init interrupt-handling thread\n");
/* rte_eal_alarm_init sets rte_errno on failure. */
return -1;
}
if (rte_eal_timer_init() < 0) {
rte_eal_init_alert("Cannot init HPET or TSC timers\n");
rte_errno = ENOTSUP;
return -1;
}
eal_check_mem_on_local_socket();
eal_thread_init_master(rte_config.master_lcore);
ret = eal_thread_dump_affinity(cpuset, sizeof(cpuset));
RTE_LOG(DEBUG, EAL, "Master lcore %u is ready (tid=%x;cpuset=[%s%s])\n",
rte_config.master_lcore, (int)thread_id, cpuset,
ret == 0 ? "" : "...");
RTE_LCORE_FOREACH_SLAVE(i) {
/*
* create communication pipes between master thread
* and children
*/
if (pipe(lcore_config[i].pipe_master2slave) < 0)
rte_panic("Cannot create pipe\n");
if (pipe(lcore_config[i].pipe_slave2master) < 0)
rte_panic("Cannot create pipe\n");
lcore_config[i].state = WAIT;
/* create a thread for each lcore */
ret = pthread_create(&lcore_config[i].thread_id, NULL,
eal_thread_loop, NULL);
if (ret != 0)
rte_panic("Cannot create thread\n");
/* Set thread_name for aid in debugging. */
snprintf(thread_name, sizeof(thread_name),
"lcore-slave-%d", i);
ret = rte_thread_setname(lcore_config[i].thread_id,
thread_name);
if (ret != 0)
RTE_LOG(DEBUG, EAL,
"Cannot set name for lcore thread\n");
}
/*
* Launch a dummy function on all slave lcores, so that master lcore
* knows they are all ready when this function returns.
*/
rte_eal_mp_remote_launch(sync_func, NULL, SKIP_MASTER);
rte_eal_mp_wait_lcore();
/* initialize services so vdevs register service during bus_probe. */
ret = rte_service_init();
if (ret) {
rte_eal_init_alert("rte_service_init() failed\n");
rte_errno = ENOEXEC;
return -1;
}
/* Probe all the buses and devices/drivers on them
/* 所有的 dirver 都会register a driver
void rte_pci_register(struct rte_pci_driver *driver)
{ TAILQ_INSERT_TAIL(&rte_pci_bus.driver_list, driver, next);
driver->bus = &rte_pci_bus;
*/
／＊遍历pci_device_list和pci_driver_list链表，根据PCI_ID，
将pci_device与pci_driver绑定，并调用pci_driver的 probe函数初始化PCI设备＊／
if (rte_bus_probe()) {//对应 rte_pci_probe 函数 --->pci_probe_all_drivers
//--AILQ_FOREACH(p, &(rte_pci_bus.driver_list), next) ---->rte_pci_probe_one_driver
//----rte_pci_map_device && dr->probe(dr, dev);
//PCI设备和PCI驱动匹配后，调用pci_map_device()函数为该PCI设备创建map resource
rte_eal_init_alert("Cannot probe devices\n");
rte_errno = ENOTSUP;
return -1;
}
#ifdef VFIO_PRESENT
/* Register mp action after probe() so that we got enough info */
if (rte_vfio_is_enabled("vfio") && vfio_mp_sync_setup() < 0)
return -1;
#endif
/* initialize default service/lcore mappings and start running. Ignore
* -ENOTSUP, as it indicates no service coremask passed to EAL.
*/
ret = rte_service_start_with_defaults();
if (ret < 0 && ret != -ENOTSUP) {
rte_errno = ENOEXEC;
return -1;
}
rte_eal_mcfg_complete();
return fctret;
}
/*
读取/sys/bus/pci/devices/0000\:00\:09.0/uio/uio0/maps/map0/目录下的文件，获取UIO设备的map resource。并将其记录在struct pci_map数据结构中。
检查PCI设备和UIO设备在内存总线上的物理地址是否一致。如果一致，对/dev/uioID文件mmap一段内存空间，并将其记录在pci_map->addr和rte_pci_device->mem_resource[].addr中
将所有UIO设备的resource信息都记录在struct mapped_pci_resource数据结构中，并挂到全局链表uio_res_list上
*/
/* Map pci device */
int
rte_pci_map_device(struct rte_pci_device *dev)
{
int ret = -1;
/* try mapping the NIC resources using VFIO if it exists */
switch (dev->kdrv) {
case RTE_KDRV_VFIO:
#ifdef VFIO_PRESENT
if (pci_vfio_is_enabled())
ret = pci_vfio_map_resource(dev);
#endif
break;
case RTE_KDRV_IGB_UIO:
case RTE_KDRV_UIO_GENERIC:
if (rte_eal_using_phys_addrs()) {
/* map resources for devices that use uio */
ret = pci_uio_map_resource(dev);
}
break;
default:
RTE_LOG(DEBUG, EAL,
" Not managed by a supported kernel driver, skipped\n");
ret = 1;
break;
}
return ret;
}
/* map the PCI resource of a PCI device in virtual memory */
int
pci_uio_map_resource(struct rte_pci_device *dev)
{
int i, map_idx = 0, ret;
uint64_t phaddr;
struct mapped_pci_resource *uio_res = NULL;
struct mapped_pci_res_list *uio_res_list =
RTE_TAILQ_CAST(rte_uio_tailq.head, mapped_pci_res_list);
dev->intr_handle.fd = -1;
dev->intr_handle.uio_cfg_fd = -1;
/* secondary processes - use already recorded details */
if (rte_eal_process_type() != RTE_PROC_PRIMARY)
return pci_uio_map_secondary(dev);
/* allocate uio resource open 打开要管理的uio 设备*/
ret = pci_uio_alloc_resource(dev, &uio_res);
if (ret)
return ret;
/* DPDK还需要把PCI设备的BAR映射到应用层
Map all BARs */
for (i = 0; i != PCI_MAX_RESOURCE; i++) {
/* skip empty BAR */
phaddr = dev->mem_resource[i].phys_addr;
if (phaddr == 0)
continue;
ret = pci_uio_map_resource_by_index(dev, i,
uio_res, map_idx);
if (ret)
goto error;
map_idx++;
}
uio_res->nb_maps = map_idx;
TAILQ_INSERT_TAIL(uio_res_list, uio_res, next);
return 0;
error:
for (i = 0; i < map_idx; i++) {
pci_unmap_resource(uio_res->maps[i].addr,
(size_t)uio_res->maps[i].size);
rte_free(uio_res->maps[i].path);
}
pci_uio_free_resource(dev, uio_res);
return -1;
}
int
pci_uio_alloc_resource(struct rte_pci_device *dev,
struct mapped_pci_resource **uio_res)
{
char devname[PATH_MAX]; /* contains the /dev/uioX */
struct rte_pci_addr *loc;
loc = &dev->addr;
snprintf(devname, sizeof(devname), "/dev/uio@pci:%u:%u:%u",
dev->addr.bus, dev->addr.devid, dev->addr.function);
if (access(devname, O_RDWR) < 0) {
RTE_LOG(WARNING, EAL, " "PCI_PRI_FMT" not managed by UIO driver, "
"skipping\n", loc->domain, loc->bus, loc->devid, loc->function);
return 1;
}
/* save fd if in primary process
初始化PCI设备的中断句柄。*/
dev->intr_handle.fd = open(devname, O_RDWR);
if (dev->intr_handle.fd < 0) {
RTE_LOG(ERR, EAL, "Cannot open %s: %s\n",
devname, strerror(errno));
goto error;
}
dev->intr_handle.type = RTE_INTR_HANDLE_UIO;
/* allocate the mapping details for secondary processes*/
*uio_res = rte_zmalloc("UIO_RES", sizeof(**uio_res), 0);
if (*uio_res == NULL) {
RTE_LOG(ERR, EAL,
"%s(): cannot store uio mmap details\n", __func__);
goto error;
}
snprintf((*uio_res)->path, sizeof((*uio_res)->path), "%s", devname);
memcpy(&(*uio_res)->pci_addr, &dev->addr, sizeof((*uio_res)->pci_addr));
return 0;
error:
pci_uio_free_resource(dev, *uio_res);
return -1;
}
4、rte_bus_probe---------->调用rte_eth_dev_create 以及 eth_ixgbe_pci_probe()初始化PCI设备； 并通过eth_ixgbe_dev_init 函数分配一个网卡设备，并在全局数组rte_eth_dev_data[]中为网卡设备的数据域分配空间。
同时初始化网卡设备。首先设置网卡设备的操作函数集，以及收包、发包函数
注册中断处理函数。
rte_intr_callback_register(&(pci_dev->intr_handle),
eth_igb_interrupt_handler, (void *)eth_dev);
int __rte_experimental
rte_eth_dev_create(struct rte_device *device, const char *name,
size_t priv_data_size,
ethdev_bus_specific_init ethdev_bus_specific_init,
void *bus_init_params,
ethdev_init_t ethdev_init, void *init_params)
{
struct rte_eth_dev *ethdev;
int retval;
RTE_FUNC_PTR_OR_ERR_RET(*ethdev_init, -EINVAL);
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
//在全局数组rte_eth_devices[]中分配一个网卡设备。并在全局数组rte_eth_dev_data[]中为网卡设备的数据域分配空间。
ethdev = rte_eth_dev_allocate(name);
if (!ethdev)
return -ENODEV;
if (priv_data_size) {
ethdev->data->dev_private = rte_zmalloc_socket(
name, priv_data_size, RTE_CACHE_LINE_SIZE,
device->numa_node);
if (!ethdev->data->dev_private) {
RTE_LOG(ERR, EAL, "failed to allocate private data");
retval = -ENOMEM;
goto data_alloc_failed;
}
}
} else {
ethdev = rte_eth_dev_attach_secondary(name);
if (!ethdev) {
RTE_LOG(ERR, EAL, "secondary process attach failed, "
"ethdev doesn't exist");
return -ENODEV;
}
}
ethdev->device = device;
if (ethdev_bus_specific_init) {
retval = ethdev_bus_specific_init(ethdev, bus_init_params);
if (retval) {
RTE_LOG(ERR, EAL,
"ethdev bus specific initialisation failed");
goto probe_failed;
}
}
retval = ethdev_init(ethdev, init_params);
if (retval) {
RTE_LOG(ERR, EAL, "ethdev initialisation failed");
goto probe_failed;
}
rte_eth_dev_probing_finish(ethdev);
return retval;
probe_failed:
/* free ports private data if primary process */
if (rte_eal_process_type() == RTE_PROC_PRIMARY)
rte_free(ethdev->data->dev_private);
data_alloc_failed:
rte_eth_dev_release_port(ethdev);
return retval;
}
/*
* This function is based on code in ixgbe_attach() in base/ixgbe.c.
* It returns 0 on success.
*/
static int
eth_ixgbe_dev_init(struct rte_eth_dev *eth_dev, void *init_params __rte_unused)
{
struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev);
struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
struct ixgbe_hw *hw =
IXGBE_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
struct ixgbe_vfta *shadow_vfta =
IXGBE_DEV_PRIVATE_TO_VFTA(eth_dev->data->dev_private);
struct ixgbe_hwstrip *hwstrip =
IXGBE_DEV_PRIVATE_TO_HWSTRIP_BITMAP(eth_dev->data->dev_private);
struct ixgbe_dcb_config *dcb_config =
IXGBE_DEV_PRIVATE_TO_DCB_CFG(eth_dev->data->dev_private);
struct ixgbe_filter_info *filter_info =
IXGBE_DEV_PRIVATE_TO_FILTER_INFO(eth_dev->data->dev_private);
struct ixgbe_bw_conf *bw_conf =
IXGBE_DEV_PRIVATE_TO_BW_CONF(eth_dev->data->dev_private);
uint32_t ctrl_ext;
uint16_t csum;
int diag, i;
PMD_INIT_FUNC_TRACE();
//设置网卡设备的操作函数集，以及收包、发包函数
eth_dev->dev_ops = &ixgbe_eth_dev_ops;
eth_dev->rx_pkt_burst = &ixgbe_recv_pkts;
eth_dev->tx_pkt_burst = &ixgbe_xmit_pkts;
eth_dev->tx_pkt_prepare = &ixgbe_prep_pkts;
/*
* For secondary processes, we don't initialise any further as primary
* has already done this work. Only check we don't need a different
* RX and TX function.
*/
if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
struct ixgbe_tx_queue *txq;
/* TX queue function in primary, set by last queue initialized
* Tx queue may not initialized by primary process
*/
if (eth_dev->data->tx_queues) {
txq = eth_dev->data->tx_queues[eth_dev->data->nb_tx_queues-1];
ixgbe_set_tx_function(eth_dev, txq);
} else {
/* Use default TX function if we get here */
PMD_INIT_LOG(NOTICE, "No TX queues configured yet. "
"Using default TX function.");
}
ixgbe_set_rx_function(eth_dev);
return 0;
}
rte_eth_copy_pci_info(eth_dev, pci_dev);
/* Vendor and Device ID need to be set before init of shared code */
hw->device_id = pci_dev->id.device_id;
hw->vendor_id = pci_dev->id.vendor_id;
hw->hw_addr = (void *)pci_dev->mem_resource[0].addr;
hw->allow_unsupported_sfp = 1;
/* Initialize the shared code (base driver) */
#ifdef RTE_LIBRTE_IXGBE_BYPASS
diag = ixgbe_bypass_init_shared_code(hw);
#else
diag = ixgbe_init_shared_code(hw);
#endif /* RTE_LIBRTE_IXGBE_BYPASS */
if (diag != IXGBE_SUCCESS) {
PMD_INIT_LOG(ERR, "Shared code init failed: %d", diag);
return -EIO;
}
/* pick up the PCI bus settings for reporting later */
ixgbe_get_bus_info(hw);
/* Unlock any pending hardware semaphore */
ixgbe_swfw_lock_reset(hw);
#ifdef RTE_LIBRTE_SECURITY
/* Initialize security_ctx only for primary process*/
if (ixgbe_ipsec_ctx_create(eth_dev))
return -ENOMEM;
#endif
/* Initialize DCB configuration*/
memset(dcb_config, 0, sizeof(struct ixgbe_dcb_config));
ixgbe_dcb_init(hw, dcb_config);
/* Get Hardware Flow Control setting */
hw->fc.requested_mode = ixgbe_fc_full;
hw->fc.current_mode = ixgbe_fc_full;
hw->fc.pause_time = IXGBE_FC_PAUSE;
for (i = 0; i < IXGBE_DCB_MAX_TRAFFIC_CLASS; i++) {
hw->fc.low_water[i] = IXGBE_FC_LO;
hw->fc.high_water[i] = IXGBE_FC_HI;
}
hw->fc.send_xon = 1;
/* Make sure we have a good EEPROM before we read from it */
diag = ixgbe_validate_eeprom_checksum(hw, &csum);
if (diag != IXGBE_SUCCESS) {
PMD_INIT_LOG(ERR, "The EEPROM checksum is not valid: %d", diag);
return -EIO;
}
#ifdef RTE_LIBRTE_IXGBE_BYPASS
diag = ixgbe_bypass_init_hw(hw);
#else
diag = ixgbe_init_hw(hw);
#endif /* RTE_LIBRTE_IXGBE_BYPASS */
/*
* Devices with copper phys will fail to initialise if ixgbe_init_hw()
* is called too soon after the kernel driver unbinding/binding occurs.
* The failure occurs in ixgbe_identify_phy_generic() for all devices,
* but for non-copper devies, ixgbe_identify_sfp_module_generic() is
* also called. See ixgbe_identify_phy_82599(). The reason for the
* failure is not known, and only occuts when virtualisation features
* are disabled in the bios. A delay of 100ms was found to be enough by
* trial-and-error, and is doubled to be safe.
*/
if (diag && (hw->mac.ops.get_media_type(hw) == ixgbe_media_type_copper)) {
rte_delay_ms(200);
diag = ixgbe_init_hw(hw);
}
if (diag == IXGBE_ERR_SFP_NOT_PRESENT)
diag = IXGBE_SUCCESS;
if (diag == IXGBE_ERR_EEPROM_VERSION) {
PMD_INIT_LOG(ERR, "This device is a pre-production adapter/"
"LOM. Please be aware there may be issues associated "
"with your hardware.");
PMD_INIT_LOG(ERR, "If you are experiencing problems "
"please contact your Intel or hardware representative "
"who provided you with this hardware.");
} else if (diag == IXGBE_ERR_SFP_NOT_SUPPORTED)
PMD_INIT_LOG(ERR, "Unsupported SFP+ Module");
if (diag) {
PMD_INIT_LOG(ERR, "Hardware Initialization Failure: %d", diag);
return -EIO;
}
/* Reset the hw statistics */
ixgbe_dev_stats_reset(eth_dev);
/* disable interrupt */
ixgbe_disable_intr(hw);
/* reset mappings for queue statistics hw counters*/
ixgbe_reset_qstat_mappings(hw);
/* Allocate memory for storing MAC addresses */
eth_dev->data->mac_addrs = rte_zmalloc("ixgbe", ETHER_ADDR_LEN *
hw->mac.num_rar_entries, 0);
if (eth_dev->data->mac_addrs == NULL) {
PMD_INIT_LOG(ERR,
"Failed to allocate %u bytes needed to store "
"MAC addresses",
ETHER_ADDR_LEN * hw->mac.num_rar_entries);
return -ENOMEM;
}
/* Copy the permanent MAC address */
ether_addr_copy((struct ether_addr *) hw->mac.perm_addr,
ð_dev->data->mac_addrs[0]);
/* Allocate memory for storing hash filter MAC addresses */
eth_dev->data->hash_mac_addrs = rte_zmalloc("ixgbe", ETHER_ADDR_LEN *
IXGBE_VMDQ_NUM_UC_MAC, 0);
if (eth_dev->data->hash_mac_addrs == NULL) {
PMD_INIT_LOG(ERR,
"Failed to allocate %d bytes needed to store MAC addresses",
ETHER_ADDR_LEN * IXGBE_VMDQ_NUM_UC_MAC);
return -ENOMEM;
}
/* initialize the vfta */
memset(shadow_vfta, 0, sizeof(*shadow_vfta));
/* initialize the hw strip bitmap*/
memset(hwstrip, 0, sizeof(*hwstrip));
/* initialize PF if max_vfs not zero */
ixgbe_pf_host_init(eth_dev);
ctrl_ext = IXGBE_READ_REG(hw, IXGBE_CTRL_EXT);
/* let hardware know driver is loaded */
ctrl_ext |= IXGBE_CTRL_EXT_DRV_LOAD;
/* Set PF Reset Done bit so PF/VF Mail Ops can work */
ctrl_ext |= IXGBE_CTRL_EXT_PFRSTD;
IXGBE_WRITE_REG(hw, IXGBE_CTRL_EXT, ctrl_ext);
IXGBE_WRITE_FLUSH(hw);
if (ixgbe_is_sfp(hw) && hw->phy.sfp_type != ixgbe_sfp_type_not_present)
PMD_INIT_LOG(DEBUG, "MAC: %d, PHY: %d, SFP+: %d",
(int) hw->mac.type, (int) hw->phy.type,
(int) hw->phy.sfp_type);
else
PMD_INIT_LOG(DEBUG, "MAC: %d, PHY: %d",
(int) hw->mac.type, (int) hw->phy.type);
PMD_INIT_LOG(DEBUG, "port %d vendorID=0x%x deviceID=0x%x",
eth_dev->data->port_id, pci_dev->id.vendor_id,
pci_dev->id.device_id);
//注册中断处理函数。
rte_intr_callback_register(intr_handle,
ixgbe_dev_interrupt_handler, eth_dev);
/* enable uio/vfio intr/eventfd mapping */
rte_intr_enable(intr_handle);
/* enable support intr */
ixgbe_enable_intr(eth_dev);
/* initialize filter info */
memset(filter_info, 0,
sizeof(struct ixgbe_filter_info));
/* initialize 5tuple filter list */
TAILQ_INIT(&filter_info->fivetuple_list);
/* initialize flow director filter list & hash */
ixgbe_fdir_filter_init(eth_dev);
/* initialize l2 tunnel filter list & hash */
ixgbe_l2_tn_filter_init(eth_dev);
/* initialize flow filter lists */
ixgbe_filterlist_init();
/* initialize bandwidth configuration info */
memset(bw_conf, 0, sizeof(struct ixgbe_bw_conf));
/* initialize Traffic Manager configuration */
ixgbe_tm_conf_init(eth_dev);
return 0;
}
标签：rte,eal,dev,PMD,init,pci,uio,dpdk 来源： https://www.cnblogs.com/codestack/p/15674071.html