Linux中 probe函数何时调用的

所以的驱动教程上都说:只有设备和驱动的名字匹配,BUS就会调用驱动的probe函数,但是有时我们要看看probe函数里面到底做了什么,还有传递给probe函数的参数我们就不知道在哪定义(反正不是我们在驱动里定义的),如果不知道传递进的参数,去看probe函数总是感觉不求甚解的样子(你对系统不求甚解,系统也会对你的要求不求甚解的),心里对自己写出的程序没底,保不齐那天来个bug,就悲剧了。

         这里以static int__devinit sst25l_probe(struct spi_device *spi)为例看看传递进的参数structspi_device *spi到底是什么,在哪定义,什么时候定义,定义了有什么用…?(本着“five W and H”的原则打破沙锅问到底)。首先struct spi_device *spi不是我们定义的驱动里定义的;其次在read,write等函数里都有struct spi_device *spi的影子,不过不是直接传递进去的,而是通过传递进去struct mtd_info *mtd,然后to_sst25l_flash(mtd),即container_of()出包含mtd的struct sst25l_flash *flash,其中flash里的第一个成员就是structspi_device *spi,而此成员的赋值就是将传递给probe中的struct spi_device *spi赋值给struct sst25l_flash *flash的,有代码为证:

static int __devinit sst25l_probe(structspi_device *spi)

{

         structflash_info *flash_info;

         structsst25l_flash *flash;

         ……

         flash->spi = spi;// 将structspi_device *spi赋值给struct sst25l_flash *flash

         mutex_init(&flash->lock);

         dev_set_drvdata(&spi->dev,flash);// &spi->dev ->p->driver_data = flash保持flash

         ……

}

         所以搞清楚structspi_device *spi的来源是搞清楚设备驱动与主控驱动的联系纽带的关键之一,当然要首先搞清楚probe函数什么时候调用才能搞清楚struct spi_device *spi怎么传递的,其重要性不言而喻(虽然言了很多,^-^,有点唐僧了)。我们先从驱动的init开始入手,毕竟这是驱动注册开始的地方,也是一系列后续操作引发的地方:

static int __init sst25l_init(void)

{

         returnspi_register_driver(&sst25l_driver);

}

         里面只有一个函数,最喜欢这样的函数了:

int spi_register_driver(struct spi_driver*sdrv)

{

         sdrv->driver.bus= &spi_bus_type;

         if(sdrv->probe)

                   sdrv->driver.probe= spi_drv_probe;

         if(sdrv->remove)

                   sdrv->driver.remove= spi_drv_remove;

         if(sdrv->shutdown)

                   sdrv->driver.shutdown= spi_drv_shutdown;

         return driver_register(&sdrv->driver);

}

         前面都是赋值,直接最后一个语句:

int driver_register(struct device_driver*drv)

{

         intret;

         structdevice_driver *other;

         ……

         ret = bus_add_driver(drv);

         if(ret)

                   returnret;

         ret= driver_add_groups(drv, drv->groups);

         if(ret)

                   bus_remove_driver(drv);

         returnret;

}

         bus_add_driver(drv)看着就像“好人”:

int bus_add_driver(struct device_driver*drv)

{

         structbus_type *bus;

         structdriver_private *priv;

         interror = 0;

         ……

         if(drv->bus->p->drivers_autoprobe) {

                   error= driver_attach(drv);

                   if(error)

                            goto out_unregister;

         }

         ……

}

         driver_attach看着也很“友善”(函数名中带get,init的一般都不是,如果里面有几个“友善”的,一首歌中已经告诉了我们解决的办法:“xx就像偶尔拨不通的电话号码,多试几次总会回答,……”,如果网上找不到,只好挨个跟踪了,我就这样找的,笨人只好采取笨办法,也是没有办法的办法了):

int driver_attach(struct device_driver*drv)

{

         returnbus_for_each_dev(drv->bus, NULL, drv, __driver_attach);

}

         里面只有一个函数,goon:

int bus_for_each_dev(struct bus_type *bus,struct device *start, void *data, int (*fn)(struct device *, void *))

{

         structklist_iter i;

         structdevice *dev;

         interror = 0;

        

         if(!bus)

                  return -EINVAL;

        

         klist_iter_init_node(&bus->p->klist_devices,&i, (start ? &start->p->knode_bus : NULL));

         while((dev = next_device(&i)) && !error)

                  error = fn(dev,data);

         klist_iter_exit(&i);

         returnerror;

}

         看到这里好像没有我们想要找的attach,只执行了个fn()函数,肿么回事?到回头看看哪里漏了,在bus_for_each_dev中传递了个 __driver_attach,也就是在bus_for_each_dev执行了__driver_attach(dev, data),那么它里面到底执行了什么?

static int __driver_attach(struct device*dev, void *data)

{

         structdevice_driver *drv = data;

        

         if (!driver_match_device(drv, dev))

                  return 0;

        

         if(dev->parent)/* Needed for USB */

                   device_lock(dev->parent);

         device_lock(dev);

         if(!dev->driver)

                   driver_probe_device(drv, dev);

         device_unlock(dev);

         if(dev->parent)

                   device_unlock(dev->parent);

        

         return0;

}

         有个driver_probe_device(drv,dev),继续跟踪:

int driver_probe_device(structdevice_driver *drv, struct device *dev)

{

         intret = 0;

         ……

         ret = really_probe(dev, drv);

         pm_runtime_put_sync(dev);

 

         returnret;

}

         有个really_probe(dev,drv),linux神马的就喜欢这样,经常一个函数传递给另一函数,后一个函数就是在前一个函数前加“do_”、“really_”、“__”,还经常的就是宏定义的或inline的。

static int really_probe(struct device *dev,struct device_driver *drv)

{

         intret = 0;

         ……

         if(dev->bus->probe) {

                   ret = dev->bus->probe(dev);

                   if(ret)

                            gotoprobe_failed;

         }else if (drv->probe) {

                   ret = drv->probe(dev);

                   if(ret)

                            gotoprobe_failed;

         }

         ……

         returnret;

}

         这里如果有总线上的probe函数就调用总线的probe函数,如果没有则调用drv的probe函数。

         在static int__driver_attach(struct device *dev, void *data)中先调用了driver_match_device(drv,dev),用于匹配,成功才继续执行,否则直接返回了。driver_match_device(drv, dev)中:

static inline intdriver_match_device(struct device_driver *drv,

                                           struct device *dev)

{

         returndrv->bus->match ? drv->bus->match(dev, drv) : 1;

}

         即如果match函数的指针不为空,则执行此bus的match函数,也就是为什么资料上老是说总线负责匹配设备和驱动了。这里也传递了参数struct device *dev,到底这个dev来自何方,会在下一篇文章中继续跟踪。

 

注:2019/1/21 在platform的总线模型static int misc_imp706_probe(struct platform_device *pdev),曾遇到打印

imp706_watchdog: probe of imp706_watchdog failed with error -22

追踪代码发现问题出现在really_probe(struct device *dev, struct device_driver *drv)函数中

{
    int ret = 0;
    int local_trigger_count = atomic_read(&deferred_trigger_count);

    atomic_inc(&probe_count);
    pr_debug("bus: '%s': %s: probing driver %s with device %s\n",
         drv->bus->name, __func__, drv->name, dev_name(dev));
    WARN_ON(!list_empty(&dev->devres_head));

    dev->driver = drv;

    /* If using pinctrl, bind pins now before probing */
    ret = pinctrl_bind_pins(dev);
    if (ret)
        goto probe_failed;

    if (driver_sysfs_add(dev)) {
        printk(KERN_ERR "%s: driver_sysfs_add(%s) failed\n",
            __func__, dev_name(dev));
        goto probe_failed;
    }

    if (dev->bus->probe) {
        ret = dev->bus->probe(dev);
        if (ret)
            goto probe_failed;
    } else if (drv->probe) {
        ret = drv->probe(dev);
        if (ret)
            goto probe_failed;

    }

    driver_bound(dev);
    ret = 1;
    pr_debug("bus: '%s': %s: bound device %s to driver %s\n",
         drv->bus->name, __func__, dev_name(dev), drv->name);
    goto done;

probe_failed:
    devres_release_all(dev);
    driver_sysfs_remove(dev);
    dev->driver = NULL;
    dev_set_drvdata(dev, NULL);

    if (ret == -EPROBE_DEFER) {
        /* Driver requested deferred probing */
        dev_info(dev, "Driver %s requests probe deferral\n", drv->name);
        driver_deferred_probe_add(dev);
        /* Did a trigger occur while probing? Need to re-trigger if yes */
        if (local_trigger_count != atomic_read(&deferred_trigger_count))
            driver_deferred_probe_trigger();
    } else if (ret != -ENODEV && ret != -ENXIO) {
        /* driver matched but the probe failed */
        printk(KERN_WARNING
               "%s: probe of %s failed with error %d\n",
               drv->name, dev_name(dev), ret);

    } else {
        pr_debug("%s: probe of %s rejects match %d\n",
               drv->name, dev_name(dev), ret);
    }
    /*
     * Ignore errors returned by ->probe so that the next driver can try
     * its luck.
     */
    ret = 0;
done:
    atomic_dec(&probe_count);
    wake_up(&probe_waitqueue);
    return ret;
}
查看代码发现是misc_imp706_probe函数的返回值ret为非0真值,导致函数进入probe_failed。

你可能感兴趣的:(Linux驱动模型)