一、kobject的定義:
kobject是Linux2.6引入的設(shè)備管理機(jī)制,在內(nèi)核中由struct kobject結(jié)構(gòu)表示,這個(gè)結(jié)構(gòu)使所有設(shè)備在底層都具有統(tǒng)一的接口.kobject提供了基本的對(duì)象管理能力,是構(gòu)成Linux2.6設(shè)備模型的核心結(jié)構(gòu),它與sysfs文件系統(tǒng)緊密聯(lián)系,每個(gè)在內(nèi)核中注冊kobject對(duì)象都對(duì)應(yīng)與sysfs文件系統(tǒng)中的一個(gè)目錄;kobject--->sysfs.dir;其結(jié)構(gòu)定義為:
struct kobject
{
? const char*??????? k_name;????????????? //指向設(shè)備名稱的指針
? char?????????????? name[KOBJ_NAME_LEN]; //設(shè)備名稱
? struct kref??????? kref;??????????????? //內(nèi)核對(duì)象的引用計(jì)數(shù)
? struct list_head?? entry;?????????????? //掛接到當(dāng)前內(nèi)核對(duì)象所在kset中的單元
? struct kobject*??? parent;????????????? //指向父對(duì)象的指針
? struct kset*?????? kset;??????????????? //內(nèi)核對(duì)象所屬kset的指針
? struct kobj_type*? ktype;?????????????? //指向內(nèi)核對(duì)象類型描述符的指針
? struct dentry*???? dentry;????????????? //sysfs文件系統(tǒng)中與該內(nèi)核對(duì)象對(duì)應(yīng)的文件節(jié)點(diǎn)路徑的指針
? wait_queue_head_t? poll;??????????????? //IO等待隊(duì)列;
};
二、kobject相關(guān)函數(shù):
1、void kobject_init(struct kobject* kobj);
?? 該函數(shù)用于初始化kobject對(duì)象,它設(shè)置kobject對(duì)象的引用計(jì)數(shù)為1,entry字段指向自身,其所屬kset對(duì)象的引用計(jì)數(shù)加1;
2、void kobject_cleanup(struct kobject* kobj);
?? void kobject_release(struct kref* ref);
?? 這兩個(gè)函數(shù)用于清除kobject對(duì)象,當(dāng)其引用計(jì)數(shù)為0時(shí),釋放對(duì)象所占用的資源;
3、int kobject_set_name(struct kobject* kobj, const char* format, ...);
?? 該函數(shù)用于設(shè)置指定kobject對(duì)象的名稱;
4、const char* kobject_name(const struct kobject* kobj);
?? 該函數(shù)用于返回指定kobject的名稱;
5、int kobject_rename(struct kobject* kobj, const char* new_name);
?? 該函數(shù)用于為指定kobject對(duì)象重命名;
6、struct kobject* kobject_get(struct kobject* kobj);
?? 該函數(shù)用于將kobject對(duì)象的引用計(jì)數(shù)加1,相當(dāng)于申請(qǐng)了一個(gè)kobject對(duì)象資源,同時(shí)返回該kobject對(duì)象的指針;
7、void kobject_put(struct kobject* kobj);
?? 該函數(shù)用于將kobject對(duì)象的引用計(jì)數(shù)減1,相當(dāng)于釋放了一個(gè)kobject對(duì)象資源;當(dāng)引用計(jì)數(shù)為0時(shí),則調(diào)用kobject_release()釋放該kobject對(duì)象的資源;
8、int kobject_add(struct kobject* kobj);
?? 該函數(shù)用于注冊kobject對(duì)象,即:加入到Linux的設(shè)備層次中,它會(huì)掛接該kobject對(duì)象到kset的list鏈中,增加父目錄各級(jí)kobject對(duì)象的引用計(jì)數(shù),在其parent字段指向的目錄下創(chuàng)建對(duì)應(yīng)的文件節(jié)點(diǎn),并啟動(dòng)該類型kobject對(duì)象的hotplug()函數(shù);
9、void kobject_del(struct kobject* kobj);
?? 該函數(shù)與kobject_add()相反,用于注銷kobject對(duì)象,即:中止該kobject對(duì)象的hotplug()函數(shù),從Linux的設(shè)備層次中刪除該kobject對(duì)象,刪除該kobject對(duì)象在sysfs文件系統(tǒng)中對(duì)應(yīng)的文件節(jié)點(diǎn);
10、int kobject_register(struct kobject* obj);
??? 該函數(shù)用于注冊kobject對(duì)象,它首先會(huì)調(diào)用kobject_init()初始化kobj,然后再調(diào)用kobject_add()完成該內(nèi)核對(duì)象的添加;
11、void kobject_unregister(struct kobject* kobj);
??? 該函數(shù)與kobject_register()相反,用于注銷kobject對(duì)象,它首先調(diào)用kobject_del()從Linux的設(shè)備層次中刪除kobject對(duì)象,再調(diào)用kobject_put()減少該kobject對(duì)象的引用計(jì)數(shù),當(dāng)引用計(jì)數(shù)為0時(shí),則釋放該kobject對(duì)象的資源;
12、struct kobject* kobject_add_dir(struct kobject*, const char* path);
??? 該函數(shù)用于在sysfs文件系統(tǒng)中為該kobject對(duì)象創(chuàng)建對(duì)應(yīng)的目錄;
13、char* kobject_get_path(struct kobject* kobj);
??? 該函數(shù)用于返回該kobject對(duì)象在sysfs文件系統(tǒng)中的對(duì)應(yīng)目錄路徑;
三、kobject的行為:
typedef int __bitwise kobject_action_t;
enum kobject_action
{
? KOBJ_ADD???? = (__force kobject_action_t) 0x01, //exclusive to core
? KOBJ_REMOVE? = (__force kobject_action_t) 0x02, //exclusive to core
? KOBJ_CHANGE? = (__force kobject_action_t) 0x03, //device state change
? KOBJ_MOUNT?? = (__force kobject_action_t) 0x04, //mount event for block devices (broken)
? KOBJ_UMOUNT? = (__force kobject_action_t) 0x05, //umount event for block devices (broken)
? KOBJ_OFFLINE = (__force kobject_action_t) 0x06, //device offline
? KOBJ_ONLINE? = (__force kobject_action_t) 0x07, //device online
};
該枚舉類型用于定義kobject對(duì)象的狀態(tài)更新消息碼,也就是熱插拔事件碼;
備注:struct kobject結(jié)構(gòu)定義于文件include/linux/kobject.h

下面轉(zhuǎn)自：/u1/57901/showart_1803248.html

在LINUX中最讓人不解的大概就是/sys下面的內(nèi)容了

下面首先讓我們來創(chuàng)建一個(gè)簡單的platform設(shè)備,并從這個(gè)設(shè)備的視角進(jìn)行深入,在此篇文章的深入過程中,我們只看kobeject的模型我所使用的內(nèi)核版本號(hào)為2.6.26,操作系統(tǒng)的內(nèi)核版本號(hào)為2.6.27-7,暫未發(fā)現(xiàn)2.6.27-7與2.6.26的重大不同

首先寫一個(gè)簡單的模塊

#include

#include

#include

static int __init test_probe(struct platform_device *pdev)

{

int err = 0;

return err;

}

static int test_remove(struct platform_device *pdev)

{

return 0;

}

static struct platform_device test_device = {

.name = "test_ts",

.id = -1,

};

static struct platform_driver test_driver = {

.probe ? ? ? ? ? ? ? ?= test_probe,

.remove ? ? ? ? ? ? ? ?= test_remove,

.driver ? ? ? ? ? ? ? ?= {

.name ? ? ? ?= "test_ts",

.owner ? ? ? ?= THIS_MODULE,

},

};

static int __devinit test_init(void)

{

platform_device_register(&test_device); ? ? ? ?

return platform_driver_register(&test_driver);

}

static void __exit test_exit(void)

{

platform_device_unregister(&test_device);

platform_driver_unregister(&test_driver);

}

module_init(test_init);

module_exit(test_exit);

MODULE_AUTHOR("zwolf");

MODULE_DESCRIPTION("Module test");

MODULE_LICENSE("GPL");

MODULE_ALIAS("test");

接下來是makefile

#Makefile

obj-m:=test.o

KDIR:=/lib/modules/2.6.27-7-generic/build

PWD:=$(shell pwd)

default:

$(MAKE) -C $(KDIR) M=$(PWD) modules

KDIR中的目錄請(qǐng)改為各位實(shí)際運(yùn)行中的內(nèi)核目錄make之后進(jìn)行模塊的加載 sudo insmod ./test.ko

現(xiàn)在到sys目錄中查看我們的設(shè)備是否已經(jīng)加載上了

首先是/sys/bus/platform/devices/在devices下,每一個(gè)連接文件都代表了一個(gè)設(shè)備ls可看見test_ts,進(jìn)入test_ts,ls可發(fā)現(xiàn)driver這個(gè)鏈接文件,ls-l查看,發(fā)現(xiàn)這個(gè)文件是連到/sys/bus/platform/drivers/test_ts的

這里需要說明的是連接的含義,并不是driver驅(qū)動(dòng)存在于test_ts這個(gè)設(shè)備中,而是test_ts使用的驅(qū)動(dòng)為/sys/bus/platform/drivers/test_ts

現(xiàn)在換到/sys/bus/platform/drivers這個(gè)目錄下

ls查看會(huì)發(fā)現(xiàn)這里的文件都為目錄,而非連接文件,說明這是驅(qū)動(dòng)真正放置的位置

現(xiàn)在進(jìn)入test_ts目錄,然后ls,發(fā)現(xiàn)有一個(gè)test_ts的連接文件,ls –l查看可發(fā)現(xiàn)該文件連接到/sys/devices/platform/test_ts下

回到/sys/bus/platform/devices/下ls –l也會(huì)發(fā)現(xiàn)test_ts連接到/sys/devices/platform/test_ts

為什么test_ts這個(gè)設(shè)備放置于/sys/devices/platform下,而不是/sys/bus/platform/devices下呢

我認(rèn)為和直觀性有關(guān),在sys下有這么幾個(gè)目錄block ?bus ?class ?dev ?devices ?firmware ?kernel ?module ?fs power?

devices很直觀的說明了設(shè)備在這個(gè)目錄下
再來看組成這個(gè)目錄圖的核心,kobject圖,我也叫他層次圖

?

不看大號(hào)綠色箭頭右邊的內(nèi)容的話是不是發(fā)現(xiàn)兩個(gè)架構(gòu)相同?

對(duì)的,kobject的層次決定了目錄的結(jié)構(gòu)

kobeject圖很大,但也不要擔(dān)心,里面的內(nèi)容其實(shí)不多,基礎(chǔ)框架涉及3個(gè)主要結(jié)構(gòu)kset kobject和ktype

在說明test_ts的注冊之前,先讓我們看一下sys下的兩個(gè)基礎(chǔ)目錄bus,devices

首先是bus

bus的注冊在/drivers/base/bus.c里

int __init buses_ini?

t(void)

{

bus_kset = kset_create_and_add("bus", &bus_uevent_ops, NULL);

if (!bus_kset)

return -ENOMEM;

return 0;

}

先看bus_uevent_ops,這是一個(gè)uevent的操作集(我也還沒清楚uevent的用途,所以u(píng)event的內(nèi)容先放著)

然后到kset_create_and_add

struct kset *kset_create_and_add(const char *name,

struct kset_uevent_ops *uevent_ops,

struct kobject *parent_kobj)

//傳遞進(jìn)來的參數(shù)為("bus", &bus_uevent_ops, NULL)

{

struct kset *kset;

int error;

//創(chuàng)建一個(gè)kset容器

kset = kset_create(name, uevent_ops, parent_kobj);

if (!kset)

return NULL;

//注冊創(chuàng)建的kset容器

error = kset_register(kset);

if (error) {

kfree(kset);

return NULL;

}

return kset;

}

首先需要?jiǎng)?chuàng)建一個(gè)kset容器

static struct kset *kset_create(const char *name,

struct kset_uevent_ops *uevent_ops,

struct kobject *parent_kobj)

//傳遞進(jìn)來的參數(shù)為("bus", &bus_uevent_ops, NULL)

{

struct kset *kset;

//為kset分配內(nèi)存

kset = kzalloc(sizeof(*kset), GFP_KERNEL);

if (!kset)

return NULL;

//設(shè)置kset中kobject的名字,這里為bus

kobject_set_name(&kset->kobj, name);

//設(shè)置uevent操作集,這里為bus_uevent_ops

kset->uevent_ops = uevent_ops;

//設(shè)置父對(duì)象,這里為NULL

kset->kobj.parent = parent_kobj;

//設(shè)置容器操作集

kset->kobj.ktype = &kset_ktype;

//設(shè)置父容器

kset->kobj.kset = NULL;

return kset;

}

這里的ktype,也就是kset_ktype是一個(gè)操作集,用于為sys下文件的實(shí)時(shí)反饋?zhàn)龇?wù),例如我們cat name的時(shí)候就要通過ktype提供的show函數(shù),具體什么怎么運(yùn)用,將在后面講解

現(xiàn)在回到kset_create_and_add中的kset_register,將建立好的kset添加進(jìn)sys里

int kset_register(struct kset *k)

{

int err;

if (!k)

return -EINVAL;

//初始化

kset_init(k);

//添加該容器

err = kobject_add_internal(&k->kobj);

if (err)

return err;

kobject_uevent(&k->kobj, KOBJ_ADD);

return 0;

}

kset_init進(jìn)行一些固定的初始化操作,里面沒有我們需要關(guān)心的內(nèi)容

kobject_add_internal為重要的一個(gè)函數(shù),他對(duì)kset里kobj的從屬關(guān)系進(jìn)行解析,搭建正確的架構(gòu)

static int kobject_add_internal(struct kobject *kobj)

{

int error = 0;

struct kobject *parent;

//檢測kobj是否為空

if (!kobj)

return -ENOENT;

//檢測kobj名字是否為空

if (!kobj->name || !kobj->name[0]) {

pr_debug("kobject: (%p): attempted to be registered with empty "

"name!\n", kobj);

WARN_ON(1);

return -EINVAL;

}

//提取父對(duì)象

parent = kobject_get(kobj->parent);

/* join kset if set, use it as parent if we do not already have one */

//父容器存在則設(shè)置父對(duì)象

if (kobj->kset) {//在bus的kset中為空,所以不會(huì)進(jìn)入到下面的代碼

//檢測是否已經(jīng)設(shè)置父對(duì)象

if (!parent)

//無則使用父容器為父對(duì)象

parent = kobject_get(&kobj->kset->kobj);

//添加該kobj到父容器的鏈表中

kobj_kset_join(kobj);

//設(shè)置父對(duì)象

kobj->parent = parent;

}

pr_debug("kobject: '%s' (%p): %s: parent: '%s', set: '%s'\n",

kobject_name(kobj), kobj, __func__,

parent ? kobject_name(parent) : "",

kobj->kset ? kobject_name(&kobj->kset->kobj) : "");

//建立相應(yīng)的目錄

error = create_dir(kobj);

if (error) {

kobj_kset_leave(kobj);

kobject_put(parent);

kobj->parent = NULL;

if (error == -EEXIST)

printk(KERN_ERR "%s failed for %s with "

"-EEXIST, don't try to register things with "

"the same name in the same directory.\n",

__func__, kobject_name(kobj));

else

printk(KERN_ERR "%s failed for %s (%d)\n",

__func__, kobject_name(kobj), error);

dump_stack();

} else

kobj->state_in_sysfs = 1;

return error;

}

至此bus的目錄就建立起來了

模型如下

?

接下來是devices,在/drivers/base/core.c里

int __init devices_init(void)

{

devices_kset = kset_create_and_add("devices", &device_uevent_ops, NULL);

if (!devices_kset)

return -ENOMEM;

return 0;

}

過程和bus的注冊一致,我就不復(fù)述了~

模型如下

然后是platform的注冊

在platform的注冊中,分為兩個(gè)部分,一部分是注冊到devices中,另一部分是注冊到bus中,代碼在/drivers/base/platform.c中

int __init platform_bus_init(void)

{

int error;

//注冊到devices目錄中

error = device_register(&platform_bus);

if (error)

return error;

//注冊到bus目錄中

error = ?bus_register(&platform_bus_type);

if (error)

device_unregister(&platform_bus);

return error;

}

首先是device_register,注冊的參數(shù)為platform_bus,如下所示

struct device platform_bus = {

.bus_id ? ? ? ? ? ? ? ?= "platform",

};

很簡單,只有一個(gè)參數(shù),表明了目錄名

int device_register(struct device *dev)

{

//初始化dev結(jié)構(gòu)

device_initialize(dev);

//添加dev至目錄

return device_add(dev);

}

void device_initialize(struct device *dev)

{

//重要的一步,指明了父容器為devices_kset,而devices_kset的注冊在前面已經(jīng)介紹過了

dev->kobj.kset = devices_kset;

//初始化kobj的ktype為device_ktype

kobject_init(&dev->kobj, &device_ktype);

klist_init(&dev->klist_children, klist_children_get,

klist_children_put);

INIT_LIST_HEAD(&dev->dma_pools);

INIT_LIST_HEAD(&dev->node);

init_MUTEX(&dev->sem);

spin_lock_init(&dev->devres_lock);

INIT_LIST_HEAD(&dev->devres_head);

device_init_wakeup(dev, 0);

set_dev_node(dev, -1);

}

int device_add(struct device *dev)

{

struct device *parent = NULL;

struct class_interface *class_intf;

int error;

dev = get_device(dev);

if (!dev || !strlen(dev->bus_id)) {

error = -EINVAL;

goto Done;

}

pr_debug("device: '%s': %s\n", dev->bus_id, __func__);

parent = get_device(dev->parent);

setup_parent(dev, parent);

if (parent)

set_dev_node(dev, dev_to_node(parent));

//設(shè)置dev->kobj的名字和父對(duì)象,并建立相應(yīng)的目錄

error = kobject_add(&dev->kobj, dev->kobj.parent, "%s", dev->bus_id);

if (error)

goto Error;

if (platform_notify)

platform_notify(dev);

if (dev->bus)

blocking_notifier_call_chain(&dev->bus->p->bus_notifier,

BUS_NOTIFY_ADD_DEVICE, dev);

//建立uevent文件

error = device_create_file(dev, &uevent_attr);

if (error)

goto attrError;

if (MAJOR(dev->devt)) {

error = device_create_file(dev, &devt_attr);

if (error)

goto ueventattrError;

}

//建立subsystem連接文件連接到所屬class,這里沒有設(shè)置class對(duì)象所以不會(huì)建立

error = device_add_class_symlinks(dev);

if (error)

goto SymlinkError;

//建立dev的描述文件,這里沒有設(shè)置描述文件所以不會(huì)建立

error = device_add_attrs(dev);

if (error)

goto AttrsError;

//建立鏈接文件至所屬bus,這里沒有設(shè)置所屬bus所以不會(huì)建立

error = bus_add_device(dev);

if (error)

goto BusError;

//添加power文件,因?yàn)閜latform不屬于設(shè)備,所以不會(huì)建立power文件

error = device_pm_add(dev);

if (error)

goto PMError;

kobject_uevent(&dev->kobj, KOBJ_ADD);

//檢測驅(qū)動(dòng)中有無適合的設(shè)備進(jìn)行匹配,但沒有設(shè)置bus,所以不會(huì)進(jìn)行匹配

bus_attach_device(dev);

if (parent)

klist_add_tail(&dev->knode_parent, &parent->klist_children);

if (dev->class) {

down(&dev->class->sem);

list_add_tail(&dev->node, &dev->class->devices);

list_for_each_entry(class_intf, &dev->class->interfaces, node)

if (class_intf->add_dev)

class_intf->add_dev(dev, class_intf);

up(&dev->class->sem);

}

Done:

put_device(dev);

return error;

PMError:

bus_remove_device(dev);

BusError:

if (dev->bus)

blocking_notifier_call_chain(&dev->bus->p->bus_notifier,

BUS_NOTIFY_DEL_DEVICE, dev);

device_remove_attrs(dev);

AttrsError:

device_remove_class_symlinks(dev);

SymlinkError:

if (MAJOR(dev->devt))

device_remove_file(dev, &devt_attr);

ueventattrError:

device_remove_file(dev, &uevent_attr);

attrError:

kobject_uevent(&dev->kobj, KOBJ_REMOVE);

kobject_del(&dev->kobj);

Error:

cleanup_device_parent(dev);

if (parent)

put_device(parent);

goto Done;

}

在kobject_add-> kobject_add_varg-> kobject_add_internal中

//提取父對(duì)象,因?yàn)闆]有設(shè)置,所以為空

parent = kobject_get(kobj->parent);

//父容器存在則設(shè)置父對(duì)象,在前面的dev->kobj.kset = devices_kset中設(shè)為了devices_kset

if (kobj->kset) {

//檢測是否已經(jīng)設(shè)置父對(duì)象

if (!parent)

//無則使用父容器為父對(duì)象

parent = kobject_get(&kobj->kset->kobj);

//添加該kobj到父容器的鏈表中

kobj_kset_join(kobj);

//設(shè)置父對(duì)象

kobj->parent = parent;

}

現(xiàn)在devices下的platform目錄建立好了,模型如下,其中紅線描繪了目錄關(guān)系

現(xiàn)在到bus_register了

注冊的參數(shù)platform_bus_type如下所示

struct bus_type platform_bus_type = {

.name ? ? ? ? ? ? ? ?= "platform",

.dev_attrs ? ? ? ?= platform_dev_attrs,

.match ? ? ? ? ? ? ? ?= platform_match,

.uevent ? ? ? ? ? ? ? ?= platform_uevent,

.suspend ? ? ? ? ? ? ? ?= platform_suspend,

.suspend_late ? ? ? ?= platform_suspend_late,

.resume_early ? ? ? ?= platform_resume_early,

.resume ? ? ? ? ? ? ? ?= platform_resume,

};

int bus_register(struct bus_type *bus)

{

int retval;

//聲明一個(gè)總線私有數(shù)據(jù)并分配空間

struct bus_type_private *priv;

priv = kzalloc(sizeof(struct bus_type_private), GFP_KERNEL);

if (!priv)

return -ENOMEM;

//互相關(guān)聯(lián)

priv->bus = bus;

bus->p = priv;

BLOCKING_INIT_NOTIFIER_HEAD(&priv->bus_notifier);

//設(shè)置私有數(shù)據(jù)中kobj對(duì)象的名字

retval = kobject_set_name(&priv->subsys.kobj, "%s", bus->name);

if (retval)

goto out;

//設(shè)置父容器為bus_kset,操作集為bus_ktype

priv->subsys.kobj.kset = bus_kset;

priv->subsys.kobj.ktype = &bus_ktype;

priv->drivers_autoprobe = 1;

//注冊bus容器

retval = kset_register(&priv->subsys);

if (retval)

goto out;

//建立uevent屬性文件

retval = bus_create_file(bus, &bus_attr_uevent);

if (retval)

goto bus_uevent_fail;

//建立devices目錄

priv->devices_kset = kset_create_and_add("devices", NULL,

&priv->subsys.kobj);

if (!priv->devices_kset) {

retval = -ENOMEM;

goto bus_devices_fail;

}

//建立drivers目錄

priv->drivers_kset = kset_create_and_add("drivers", NULL,

&priv->subsys.kobj);

if (!priv->drivers_kset) {

retval = -ENOMEM;

goto bus_drivers_fail;

}

//初始化klist_devices和klist_drivers鏈表

klist_init(&priv->klist_devices, klist_devices_get, klist_devices_put);

klist_init(&priv->klist_drivers, NULL, NULL);

//增加probe屬性文件

retval = add_probe_files(bus);

if (retval)

goto bus_probe_files_fail;

//增加總線的屬性文件

retval = bus_add_attrs(bus);

if (retval)

goto bus_attrs_fail;

pr_debug("bus: '%s': registered\n", bus->name);

return 0;

bus_attrs_fail:

remove_probe_files(bus);

bus_probe_files_fail:

kset_unregister(bus->p->drivers_kset);

bus_drivers_fail:

kset_unregister(bus->p->devices_kset);

bus_devices_fail:

bus_remove_file(bus, &bus_attr_uevent);

bus_uevent_fail:

kset_unregister(&bus->p->subsys);

kfree(bus->p);

out:

return retval;

}

在kset_register-> kobject_add_internal中

//提取父對(duì)象,因?yàn)闆]有設(shè)置父對(duì)象,所以為空

parent = kobject_get(kobj->parent);

//父容器存在則設(shè)置父對(duì)象,在上文中設(shè)置了父容器priv->subsys.kobj.kset = bus_kset

if (kobj->kset) {

//檢測是否已經(jīng)設(shè)置父對(duì)象

if (!parent)

//無則使用父容器為父對(duì)象

parent = kobject_get(&kobj->kset->kobj);

//添加該kobj到父容器的鏈表中

kobj_kset_join(kobj);

//設(shè)置父對(duì)象

kobj->parent = parent;

}

在retval = kset_register(&priv->subsys)完成之后platform在bus下的模型如下圖

有印象的話大家還記得在platform下面有兩個(gè)目錄devices和drivers吧~

現(xiàn)在就到這兩個(gè)目錄的注冊了

priv->devices_kset = kset_create_and_add("devices", NULL,&priv->subsys.kobj);

priv->drivers_kset = kset_create_and_add("drivers", NULL, &priv->subsys.kobj);

注意這兩條語句的頭部

priv->devices_kset = kset_create_and_add

priv->drivers_kset = kset_create_and_add

可以清楚的看到bus_type_private下的devices_kset, drivers_kset分別連接到了devices,drivers的kset上

現(xiàn)在來看kset_create_and_add("devices", NULL,&priv->subsys.kobj);

struct kset *kset_create_and_add(const char *name,

struct kset_uevent_ops *uevent_ops,

struct kobject *parent_kobj)

//參數(shù)為"devices", NULL,&priv->subsys.kobj

{

struct kset *kset;

int error;

//創(chuàng)建一個(gè)kset容器

kset = kset_create(name, uevent_ops, parent_kobj);

if (!kset)

return NULL;

//注冊創(chuàng)建的kset容器

error = kset_register(kset);

if (error) {

kfree(kset);

return NULL;

}

return kset;

}

在kset_create 中比較重要的操作為

kset->kobj.ktype = &kset_ktype //設(shè)置了ktype,為kset_ktype

kset->kobj.parent = parent_kobj; //設(shè)置了父對(duì)象,為priv->subsys.kobj,也就是platform_bus_type->p->subsys.kobj

kset->kobj.kset = NULL; ? ?//設(shè)置父容器為空

在kset_register中

//提取父對(duì)象

parent = kobject_get(kobj->parent); //在之前設(shè)置為了

//父容器存在則設(shè)置父對(duì)象,由于父容器為空,不執(zhí)行以下代碼

if (kobj->kset) {

//檢測是否已經(jīng)設(shè)置父對(duì)象

if (!parent)

//無則使用父容器為父對(duì)象

parent = kobject_get(&kobj->kset->kobj);

//添加該kobj到父容器的鏈表中

kobj_kset_join(kobj);

//設(shè)置父對(duì)象

kobj->parent = parent;

}

至此, devices的模型就建立好了,drivers模型的建立和devices是一致的,只是名字不同而已,我就不復(fù)述了,建立好的模型如下

好了~ ?到了這里,bus,devices和platform的基礎(chǔ)模型就就建立好了,就等設(shè)備來注冊了

?

在platform模型設(shè)備的建立中,需要2個(gè)部分的注冊,驅(qū)動(dòng)的注冊和設(shè)備的注冊

platform_device_register(&test_device); ? ? ? ?

platform_driver_register(&test_driver);

首先看platform_device_register

注冊參數(shù)為test_device,結(jié)構(gòu)如下

static struct platform_device test_device = {

.name = "test_ts",

.id = -1,

//. resource

//.dev

};

這個(gè)結(jié)構(gòu)主要描述了設(shè)備的名字,ID和資源和私有數(shù)據(jù),其中資源和私有數(shù)據(jù)我們在這里不使用,將在別的文章中進(jìn)行講解

int platform_device_register(struct platform_device *pdev)

{

//設(shè)備屬性的初始化

device_initialize(&pdev->dev);

//將設(shè)備添加進(jìn)platform里

return platform_device_add(pdev);

}

void device_initialize(struct device *dev)

{

dev->kobj.kset = devices_kset; ? ? ? ? ? ? ? ? ? //設(shè)置kset為devices_kset,則將設(shè)備掛接上了devices目錄

kobject_init(&dev->kobj, &device_ktype); ? ? ? ? ? ? ? ? ? ?//初始化kobeject,置ktype為device_ktype

klist_init(&dev->klist_children, klist_children_get,

klist_children_put);

INIT_LIST_HEAD(&dev->dma_pools);

INIT_LIST_HEAD(&dev->node);

init_MUTEX(&dev->sem);

spin_lock_init(&dev->devres_lock);

INIT_LIST_HEAD(&dev->devres_head);

device_init_wakeup(dev, 0);

set_dev_node(dev, -1);

}

int platform_device_add(struct platform_device *pdev)

{

int i, ret = 0;

if (!pdev)

return -EINVAL;

//檢測是否設(shè)置了dev中的parent,無則賦為platform_bus

if (!pdev->dev.parent)

pdev->dev.parent = &platform_bus;

//設(shè)置dev中的bus為platform_bus_type

pdev->dev.bus = &platform_bus_type;

//檢測id,id為-1表明該設(shè)備只有一個(gè),用設(shè)備名為bus_id

//不為1則表明該設(shè)備有數(shù)個(gè),需要用序號(hào)標(biāo)明bus_id

if (pdev->id != -1)

snprintf(pdev->dev.bus_id, BUS_ID_SIZE, "%s.%d", pdev->name,

pdev->id);

else

strlcpy(pdev->dev.bus_id, pdev->name, BUS_ID_SIZE);

//增加資源到資源樹中

for (i = 0; i < pdev->num_resources; i++) {

struct resource *p, *r = &pdev->resource;

if (r->name == NULL)

r->name = pdev->dev.bus_id;

p = r->parent;

if (!p) {

if (r->flags & IORESOURCE_MEM)

p = &iomem_resource;

else if (r->flags & IORESOURCE_IO)

p = &ioport_resource;

}

if (p && insert_resource(p, r)) {

printk(KERN_ERR "%s: failed to claim resource %d\n",pdev->dev.bus_id, i);

ret = -EBUSY;

goto failed;

}

}

pr_debug("Registering platform device '%s'. Parent at %s\n",pdev->dev.bus_id, pdev->dev.parent->bus_id);

//添加設(shè)備到設(shè)備層次中

ret = device_add(&pdev->dev);

if (ret == 0)

return ret;

failed:

while (--i >= 0)

if (pdev->resource.flags & (IORESOURCE_MEM|IORESOURCE_IO))

release_resource(&pdev->resource);

return ret;

}

int device_add(struct device *dev)

{

struct device *parent = NULL;

struct class_interface *class_intf;

int error;

dev = get_device(dev);

if (!dev || !strlen(dev->bus_id)) {

error = -EINVAL;

goto Done;

}

pr_debug("device: '%s': %s\n", dev->bus_id, __func__);

//取得上層device,而dev->parent的賦值是在platform_device_add中的pdev->dev.parent = &platform_bus完成的

parent = get_device(dev->parent);

//以上層devices為準(zhǔn)重設(shè)dev->kobj.parent

setup_parent(dev, parent);

if (parent)

set_dev_node(dev, dev_to_node(parent));

//設(shè)置dev->kobj的名字和父對(duì)象,并建立相應(yīng)目錄

error = kobject_add(&dev->kobj, dev->kobj.parent, "%s", dev->bus_id);

if (error)

goto Error;

if (platform_notify)

platform_notify(dev);

//一種新型的通知機(jī)制,但是platform中沒有設(shè)置相應(yīng)的結(jié)構(gòu),所以在這里跳過

/* notify clients of device entry (new way) */

if (dev->bus)

blocking_notifier_call_chain(&dev->bus->p->bus_notifier,BUS_NOTIFY_ADD_DEVICE, dev);

//建立uevent文件

error = device_create_file(dev, &uevent_attr);

if (error)

goto attrError;

//設(shè)備有設(shè)備號(hào)則建立dev文件

if (MAJOR(dev->devt)) {

error = device_create_file(dev, &devt_attr);

if (error)

goto ueventattrError;

}

//建立subsystem連接文件連接到所屬class

error = device_add_class_symlinks(dev);

if (error)

goto SymlinkError;

//添加dev的描述文件

error = device_add_attrs(dev);

if (error)

goto AttrsError;

//添加鏈接文件至所屬bus

error = bus_add_device(dev);

if (error)

goto BusError;

//添加power文件

error = device_pm_add(dev);

if (error)

goto PMError;

kobject_uevent(&dev->kobj, KOBJ_ADD);

//檢測驅(qū)動(dòng)中有無適合的設(shè)備進(jìn)行匹配,現(xiàn)在只添加了設(shè)備,還沒有加載驅(qū)動(dòng),所以不會(huì)進(jìn)行匹配

bus_attach_device(dev);

if (parent)

klist_add_tail(&dev->knode_parent, &parent->klist_children);

if (dev->class) {

down(&dev->class->sem);

list_add_tail(&dev->node, &dev->class->devices);

list_for_each_entry(class_intf, &dev->class->interfaces, node)

if (class_intf->add_dev)

class_intf->add_dev(dev, class_intf);

up(&dev->class->sem);

}

Done:

put_device(dev);

return error;

PMError:

bus_remove_device(dev);

BusError:

if (dev->bus)

blocking_notifier_call_chain(&dev->bus->p->bus_notifier,BUS_NOTIFY_DEL_DEVICE, dev);

device_remove_attrs(dev);

AttrsError:

device_remove_class_symlinks(dev);

SymlinkError:

if (MAJOR(dev->devt))

device_remove_file(dev, &devt_attr);

ueventattrError:

device_remove_file(dev, &uevent_attr);

attrError:

kobject_uevent(&dev->kobj, KOBJ_REMOVE);

kobject_del(&dev->kobj);

Error:

cleanup_device_parent(dev);

if (parent)

put_device(parent);

goto Done;

}

static void setup_parent(struct device *dev, struct device *parent)

{

struct kobject *kobj;

//取得上層device的kobj

kobj = get_device_parent(dev, parent);

//kobj不為空則重設(shè)dev->kobj.parent

if (kobj)

dev->kobj.parent = kobj;

}

static struct kobject *get_device_parent(struct device *dev,

struct device *parent)

{

int retval;

//因?yàn)閐ev->class為空,所以跳過這段代碼

if (dev->class) {

struct kobject *kobj = NULL;

struct kobject *parent_kobj;

struct kobject *k;

if (parent == NULL)

parent_kobj = virtual_device_parent(dev);

else if (parent->class)

return &parent->kobj;

else

parent_kobj = &parent->kobj;

spin_lock(&dev->class->class_dirs.list_lock);

list_for_each_entry(k, &dev->class->class_dirs.list, entry)

if (k->parent == parent_kobj) {

kobj = kobject_get(k);

break;

}

spin_unlock(&dev->class->class_dirs.list_lock);

if (kobj)

return kobj;

k = kobject_create();

if (!k)

return NULL;

k->kset = &dev->class->class_dirs;

retval = kobject_add(k, parent_kobj, "%s", dev->class->name);

if (retval < 0) {

kobject_put(k);

return NULL;

}

return k;

}

if (parent)

//返回上層device的kobj

return &parent->kobj;

return NULL;

}

在bus_attach_device中雖然沒有成功進(jìn)行匹配,但是有很重要的一步為之后正確的匹配打下基礎(chǔ)

void bus_attach_device(struct device *dev)

{

struct bus_type *bus = dev->bus;

int ret = 0;

if (bus) {

if (bus->p->drivers_autoprobe)

ret = device_attach(dev);

WARN_ON(ret < 0);

if (ret >= 0)

klist_add_tail(&dev->knode_bus, &bus->p->klist_devices);

}

}

klist_add_tail(&dev->knode_bus, &bus->p->klist_devices)就是這一行

在這一行代碼中將設(shè)備掛載到了bus下的devices鏈表下,這樣,當(dāng)驅(qū)動(dòng)請(qǐng)求匹配的時(shí)候,platform總線就會(huì)歷遍devices鏈表為驅(qū)動(dòng)尋找合適的設(shè)備

現(xiàn)在來看一下test_device的模型

然后platform_driver_unregister,他的參數(shù) test_driver的結(jié)構(gòu)如下

static struct platform_driver test_driver = {

.probe ? ? ? ? ? ? ? ?= test_probe,

.remove ? ? ? ? ? ? ? ?= test_remove,

.driver ? ? ? ? ? ? ? ?= {

.name ? ? ? ?= "test_ts",

.owner ? ? ? ?= THIS_MODULE,

},

};

int platform_driver_register(struct platform_driver *drv)

{

drv->driver.bus = &platform_bus_type;

if (drv->probe)

drv->driver.probe = platform_drv_probe;

if (drv->remove)

drv->driver.remove = platform_drv_remove;

if (drv->shutdown)

drv->driver.shutdown = platform_drv_shutdown;

if (drv->suspend)

drv->driver.suspend = platform_drv_suspend;

if (drv->resume)

drv->driver.resume = platform_drv_resume;

return driver_register(&drv->driver);

}

從上面代碼可以看出,在platform_driver中設(shè)置了probe, remove, shutdown, suspend或resume函數(shù)的話

則drv->driver也會(huì)設(shè)置成platform對(duì)應(yīng)的函數(shù)

int driver_register(struct device_driver *drv)

{

int ret;

struct device_driver *other;

//檢測總線的操作函數(shù)和驅(qū)動(dòng)的操作函數(shù)是否同時(shí)存在,同時(shí)存在則提示使用總線提供的操作函數(shù)

if ((drv->bus->probe && drv->probe) ||

(drv->bus->remove && drv->remove) ||

(drv->bus->shutdown && drv->shutdown))

printk(KERN_WARNING "Driver '%s' needs updating - please use ""bus_type methods\n", drv->name);

//檢測是否已經(jīng)注冊過

other = driver_find(drv->name, drv->bus);

if (other) {

put_driver(other);

printk(KERN_ERR "Error: Driver '%s' is already registered, “"aborting...\n", drv->name);

return -EEXIST;

}

//添加驅(qū)動(dòng)到總線上

ret = bus_add_driver(drv);

if (ret)

return ret;

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

if (ret)

bus_remove_driver(drv);

return ret;

}

int bus_add_driver(struct device_driver *drv)

{

struct bus_type *bus;

struct driver_private *priv;

int error = 0;

//取bus結(jié)構(gòu)

bus = bus_get(drv->bus);

if (!bus)

return -EINVAL;

pr_debug("bus: '%s': add driver %s\n", bus->name, drv->name);

//分配驅(qū)動(dòng)私有數(shù)據(jù)

priv = kzalloc(sizeof(*priv), GFP_KERNEL);

if (!priv) {

error = -ENOMEM;

goto out_put_bus;

}

//初始化klist_devices鏈表

klist_init(&priv->klist_devices, NULL, NULL);

//互相關(guān)聯(lián)

priv->driver = drv;

drv->p = priv;

//設(shè)置私有數(shù)據(jù)的父容器,在這一步中,設(shè)置了kset為platform下的drivers_kset結(jié)構(gòu),也就是drivers呢個(gè)目錄

priv->kobj.kset = bus->p->drivers_kset;

//初始化kobj對(duì)象,設(shè)置容器操作集并建立相應(yīng)的目錄,這里由于沒有提供parent,所以會(huì)使用父容器中的kobj為父對(duì)象

error = kobject_init_and_add(&priv->kobj, &driver_ktype, NULL,

"%s", drv->name);

if (error)

goto out_unregister;

//檢測所屬總線的drivers_autoprobe屬性是否為真

//為真則進(jìn)行與設(shè)備的匹配,到這里,就會(huì)與我們之前注冊的test_device連接上了,至于如何連接,進(jìn)行了什么操作,將在別的文章中詳細(xì)描述

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

error = driver_attach(drv);

if (error)

goto out_unregister;

}

//掛載到所屬總線驅(qū)動(dòng)鏈表上

klist_add_tail(&priv->knode_bus, &bus->p->klist_drivers);

module_add_driver(drv->owner, drv);

//建立uevent屬性文件

error = driver_create_file(drv, &driver_attr_uevent);

if (error) {

printk(KERN_ERR "%s: uevent attr (%s) failed\n",

__func__, drv->name);

}

//建立設(shè)備屬性文件

error = driver_add_attrs(bus, drv);

if (error) {

printk(KERN_ERR "%s: driver_add_attrs(%s) failed\n",__func__, drv->name);

}

error = add_bind_files(drv);

if (error) {

printk(KERN_ERR "%s: add_bind_files(%s) failed\n",__func__, drv->name);

}

kobject_uevent(&priv->kobj, KOBJ_ADD);

return error;

out_unregister:

kobject_put(&priv->kobj);

out_put_bus:

bus_put(bus);

return error;

}

到這里test_driver的模型就建立好了,圖就是最上面的層次圖,我就不再貼了

到這里一個(gè)基本的框架就建立起來了~

下面,我開始對(duì)kobject kset和ktype做分析

先說說關(guān)系,ktype與kobject和kset這兩者之前的關(guān)系較少,讓我畫一個(gè)圖,是這樣的

ktype依賴于kobject,kset也依賴于kobject,而kobject有時(shí)需要kset(所以用了一個(gè)白箭頭),不一定需要ktype(真可憐,連白箭頭都沒有)

首先先說一下這個(gè)可有可無的ktype

到/sys/bus/platform下面可以看見一個(gè)drivers_autoprobe的文件

cat drivers_autoprobe可以查看這個(gè)文件的值

echo 0 > drivers_autoprobe則可以改變這個(gè)文件的值

drivers_autoprobe這個(gè)文件表示的是是否自動(dòng)進(jìn)行初始化

在

void bus_attach_device(struct device *dev)

{

struct bus_type *bus = dev->bus;

int ret = 0;

if (bus) {

if (bus->p->drivers_autoprobe)

ret = device_attach(dev);

WARN_ON(ret < 0);

if (ret >= 0)

klist_add_tail(&dev->knode_bus, &bus->p->klist_devices);

}

}

中可以看見這么一段代碼

if (bus->p->drivers_autoprobe)

ret = device_attach(dev);

bus->p->drivers_autoprobe的值為真則進(jìn)行匹配

而drivers_autoprobe這個(gè)文件則可以動(dòng)態(tài)的修改這個(gè)值選擇是否進(jìn)行匹配

使用外部文件修改內(nèi)核參數(shù),ktype就是提供了這么一種方法

現(xiàn)在讓我們看看ktype是怎么通過kobject進(jìn)行運(yùn)作的

首先是ktype及通過ktype進(jìn)行運(yùn)作的drivers_autoprobe的注冊

ktype的掛載十分簡單,因?yàn)樗呛蚹object是一體的

只有這么下面一句 ? ? ? ?

priv->subsys.kobj.ktype = &bus_ktype;

這樣就將bus_ktype掛載到了platform_bus_type的kobject上

drivers_autoprobe的注冊如下

retval = bus_create_file(bus, &bus_attr_drivers_autoprobe);

bus_attr_drivers_autoprobe這個(gè)結(jié)構(gòu)由一系列的宏進(jìn)行組裝

static BUS_ATTR(drivers_autoprobe, S_IWUSR | S_IRUGO,

show_drivers_autoprobe, store_drivers_autoprobe);

#define BUS_ATTR(_name, _mode, _show, _store) ? ? ? ?\

struct bus_attribute bus_attr_##_name = __ATTR(_name, _mode, _show, _store)

#define __ATTR(_name,_mode,_show,_store) { \

.attr = {.name = __stringify(_name), .mode = _mode }, ? ? ? ?\

.show ? ? ? ?= _show, ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?\

.store ? ? ? ?= _store, ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?\

}

最后bus_attr_drivers_autoprobe的模型如下

struct bus_attribute ?bus_attr_drivers_autoprobe?

{

.attr = {

.name = “drivers_autoprobe”,

.mode = S_IWUSR | S_IRUGO?

}, ? ? ? ?

.show ? ? ? ?= show_drivers_autoprobe, ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

.store ? ? ? ?= store_drivers_autoprobe, ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

}

進(jìn)入到bus_create_file中

int bus_create_file(struct bus_type *bus, struct bus_attribute *attr)

//參數(shù)為(bus, &bus_attr_drivers_autoprobe)

{

int error;

if (bus_get(bus)) {

error = sysfs_create_file(&bus->p->subsys.kobj, &attr->attr);

bus_put(bus);

} else

error = -EINVAL;

return error;

}

int sysfs_create_file(struct kobject * kobj, const struct attribute * attr)

//參數(shù)為(&bus->p->subsys.kobj, &attr->attr)

{

BUG_ON(!kobj || !kobj->sd || !attr);

return sysfs_add_file(kobj->sd, attr, SYSFS_KOBJ_ATTR);

}

int sysfs_add_file(struct sysfs_dirent *dir_sd, const struct attribute *attr,int type)

//參數(shù)為(&bus->p->subsys.kobj ->sd, &attr->attr, SYSFS_KOBJ_ATTR)

{

return sysfs_add_file_mode(dir_sd, attr, type, attr->mode);

}

int sysfs_add_file_mode(struct sysfs_dirent *dir_sd,

const struct attribute *attr, int type, mode_t amode)

//整理一下參數(shù),現(xiàn)在應(yīng)該為

//(&platform_bus_type->p->subsys.kobj ->sd, &bus_attr_drivers_autoprobe->attr, SYSFS_KOBJ_ATTR, &bus_attr_drivers_autoprobe->attr->mode)

{

umode_t mode = (amode & S_IALLUGO) | S_IFREG;

struct sysfs_addrm_cxt acxt;

struct sysfs_dirent *sd;

int rc;

//在這一步中可以看出新建了一個(gè)節(jié)點(diǎn)

sd = sysfs_new_dirent(attr->name, mode, type);

if (!sd)

return -ENOMEM;

//這一步掛載了&bus_attr_drivers_autoprobe->attr到節(jié)點(diǎn)中,為以后提取attr及上層結(jié)構(gòu)做準(zhǔn)備

sd->s_attr.attr = (void *)attr;

// dir_sd也就是上層目錄,在這里為platform_bus_type->p->subsys.kobj ->sd

//也就是/sys/bus/platform這個(gè)目錄

sysfs_addrm_start(&acxt, dir_sd);

rc = sysfs_add_one(&acxt, sd);

sysfs_addrm_finish(&acxt);

if (rc)

sysfs_put(sd);

return rc;

}

struct sysfs_dirent *sysfs_new_dirent(const char *name, umode_t mode, int type)

{

char *dup_name = NULL;

struct sysfs_dirent *sd;

if (type & SYSFS_COPY_NAME) {

name = dup_name = kstrdup(name, GFP_KERNEL);

if (!name)

return NULL;

}

sd = kmem_cache_zalloc(sysfs_dir_cachep, GFP_KERNEL);

if (!sd)

goto err_out1;

if (sysfs_alloc_ino(&sd->s_ino))

goto err_out2;

atomic_set(&sd->s_count, 1);

atomic_set(&sd->s_active, 0);

sd->s_name = name; ? //節(jié)點(diǎn)的名字為&bus_attr_drivers_autoprobe->attr->name ?也就是drivers_autoprobe

sd->s_mode = mode;

sd->s_flags = type; ? //節(jié)點(diǎn)的type為SYSFS_KOBJ_ATTR

return sd;

err_out2:

kmem_cache_free(sysfs_dir_cachep, sd);

err_out1:

kfree(dup_name);

return NULL;

}

現(xiàn)在一切準(zhǔn)備就緒,來看看怎么讀取吧

首先是open,大概流程可以看我的另一篇文章<從文件到設(shè)備>,一直看到ext3_lookup

這里和ext3_lookup不同的是,sys的文件系統(tǒng)是sysfs文件系統(tǒng),所以應(yīng)該使用的lookup函數(shù)為sysfs_lookup(/fs/sysfs/dir.c)

static struct dentry * sysfs_lookup(struct inode *dir, struct dentry *dentry,

struct nameidata *nd)

{

struct dentry *ret = NULL;

struct sysfs_dirent *parent_sd = dentry->d_parent->d_fsdata;

struct sysfs_dirent *sd;

struct inode *inode;

mutex_lock(&sysfs_mutex);

sd = sysfs_find_dirent(parent_sd, dentry->d_name.name);

if (!sd) {

ret = ERR_PTR(-ENOENT);

goto out_unlock;

}

//節(jié)點(diǎn)的初始化在這里

inode = sysfs_get_inode(sd);

if (!inode) {

ret = ERR_PTR(-ENOMEM);

goto out_unlock;

}

dentry->d_op = &sysfs_dentry_ops;

dentry->d_fsdata = sysfs_get(sd);

d_instantiate(dentry, inode);

d_rehash(dentry);

out_unlock:

mutex_unlock(&sysfs_mutex);

return ret;

}

struct inode * sysfs_get_inode(struct sysfs_dirent *sd)

{

struct inode *inode;

inode = iget_locked(sysfs_sb, sd->s_ino);

if (inode && (inode->i_state & I_NEW))

//為節(jié)點(diǎn)賦值

sysfs_init_inode(sd, inode);

return inode;

}

static void sysfs_init_inode(struct sysfs_dirent *sd, struct inode *inode)

{

struct bin_attribute *bin_attr;

inode->i_blocks = 0;

inode->i_mapping->a_ops = &sysfs_aops;

inode->i_mapping->backing_dev_info = &sysfs_backing_dev_info;

inode->i_op = &sysfs_inode_operations;

inode->i_ino = sd->s_ino;

lockdep_set_class(&inode->i_mutex, &sysfs_inode_imutex_key);

if (sd->s_iattr) {

set_inode_attr(inode, sd->s_iattr);

} else

set_default_inode_attr(inode, sd->s_mode);

//判斷類型

switch (sysfs_type(sd)) {

case SYSFS_DIR:

inode->i_op = &sysfs_dir_inode_operations;

inode->i_fop = &sysfs_dir_operations;

inode->i_nlink = sysfs_count_nlink(sd);

break;

//還記得在注冊的時(shí)候有一個(gè)參數(shù)為SYSFS_KOBJ_ATTR賦到了sd->s_flags上面吧

case SYSFS_KOBJ_ATTR:

inode->i_size = PAGE_SIZE;

inode->i_fop = &sysfs_file_operations;

break;

case SYSFS_KOBJ_BIN_ATTR:

bin_attr = sd->s_bin_attr.bin_attr;

inode->i_size = bin_attr->size;

inode->i_fop = &bin_fops;

break;

case SYSFS_KOBJ_LINK:

inode->i_op = &sysfs_symlink_inode_operations;

break;

default:

BUG();

}

unlock_new_inode(inode);

}

sysfs_file_operations的結(jié)構(gòu)如下,之后open和read,write都明了了

const struct file_operations sysfs_file_operations = {

.read ? ? ? ? ? ? ? ?= sysfs_read_file,

.write ? ? ? ? ? ? ? ?= sysfs_write_file,

.llseek ? ? ? ? ? ? ? ?= generic_file_llseek,

.open ? ? ? ? ? ? ? ?= sysfs_open_file,

.release ? ? ? ?= sysfs_release,

.poll ? ? ? ? ? ? ? ?= sysfs_poll,

};

有關(guān)在哪調(diào)用open還是請(qǐng)查閱我的另一篇文章<從文件到設(shè)備>中 nameidata_to_filp之后的操作

好的~ ?現(xiàn)在進(jìn)入到了sysfs_open_file中

static int sysfs_open_file(struct inode *inode, struct file *file)

{

struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;

//要重的取值,在這里取得了drivers_autoprobe的目錄platform的kproject

struct kobject *kobj = attr_sd->s_parent->s_dir.kobj;

struct sysfs_buffer *buffer;

struct sysfs_ops *ops;

int error = -EACCES;

if (!sysfs_get_active_two(attr_sd))

return -ENODEV;

if (kobj->ktype && kobj->ktype->sysfs_ops)

//這里可謂是ktype實(shí)現(xiàn)中的核心,在這里ops設(shè)置成了platform_bus_type中kobject->ktype的sysfs_ops

ops = kobj->ktype->sysfs_ops;

else {

printk(KERN_ERR "missing sysfs attribute operations for ""kobject: %s\n", kobject_name(kobj));

WARN_ON(1);

goto err_out;

}

if (file->f_mode & FMODE_WRITE) {

if (!(inode->i_mode & S_IWUGO) || !ops->store)

goto err_out;

}

if (file->f_mode & FMODE_READ) {

if (!(inode->i_mode & S_IRUGO) || !ops->show)

goto err_out;

}

error = -ENOMEM;

buffer = kzalloc(sizeof(struct sysfs_buffer), GFP_KERNEL);

if (!buffer)

goto err_out;

mutex_init(&buffer->mutex);

buffer->needs_read_fill = 1;

//然后將設(shè)置好的ops掛載到buffer上

buffer->ops = ops;

//再將buffer掛載到file->private_data中

file->private_data = buffer;

error = sysfs_get_open_dirent(attr_sd, buffer);

if (error)

goto err_free;

sysfs_put_active_two(attr_sd);

return 0;

err_free:

kfree(buffer);

err_out:

sysfs_put_active_two(attr_sd);

return error;

}

現(xiàn)在已經(jīng)為read和write操作準(zhǔn)備好了

馬上進(jìn)入到read操作中

整個(gè)流程如上圖所示,如何進(jìn)入到sysfs_read_file在上面open的操作中已經(jīng)說明了

我們就從sysfs_read_file開始分析(該文件在/fs/sysfs/file.c中)

sysfs_read_file(struct file *file, char __user *buf, size_t count, loff_t *ppos)

{

struct sysfs_buffer * buffer = file->private_data;

ssize_t retval = 0;

mutex_lock(&buffer->mutex);

if (buffer->needs_read_fill || *ppos == 0) {

//主要操作在fill_read_buffer中

retval = fill_read_buffer(file->f_path.dentry,buffer);

if (retval)

goto out;

}

pr_debug("%s: count = %zd, ppos = %lld, buf = %s\n",__func__, count, *ppos, buffer->page);

retval = simple_read_from_buffer(buf, count, ppos, buffer->page,

buffer->count);

out:

mutex_unlock(&buffer->mutex);

return retval;

}

static int fill_read_buffer(struct dentry * dentry, struct sysfs_buffer * buffer)

{

struct sysfs_dirent *attr_sd = dentry->d_fsdata;

//取得父目錄的kobject,也就是platform的kobject

struct kobject *kobj = attr_sd->s_parent->s_dir.kobj;

//還記得這個(gè)buffer->ops在什么時(shí)候進(jìn)行賦值的么?

struct sysfs_ops * ops = buffer->ops;

int ret = 0;

ssize_t count;

if (!buffer->page)

buffer->page = (char *) get_zeroed_page(GFP_KERNEL);

if (!buffer->page)

return -ENOMEM;

if (!sysfs_get_active_two(attr_sd))

return -ENODEV;

buffer->event = atomic_read(&attr_sd->s_attr.open->event);

//調(diào)用ops->show ?也就是bus_sysfs_ops->show ? ?具體就是bus_attr_show了

//參數(shù)為父目錄的kobject, bus_attr_drivers_autoprobe->attr,和一段char信息

count = ops->show(kobj, attr_sd->s_attr.attr, buffer->page);

sysfs_put_active_two(attr_sd);

if (count >= (ssize_t)PAGE_SIZE) {

print_symbol("fill_read_buffer: %s returned bad count\n",

(unsigned long)ops->show);

/* Try to struggle along */

count = PAGE_SIZE - 1;

}

if (count >= 0) {

buffer->needs_read_fill = 0;

buffer->count = count;

} else {

ret = count;

}

return ret;

}

現(xiàn)在進(jìn)入bus_attr_show中

static ssize_t bus_attr_show(struct kobject *kobj, struct attribute *attr,char *buf)

{

//提取attr的上層結(jié)構(gòu),也就是bus_attr_drivers_autoprobe

struct bus_attribute *bus_attr = to_bus_attr(attr);

//提取kobj的上上層結(jié)構(gòu),也就是bus_type_private

struct bus_type_private *bus_priv = to_bus(kobj);

ssize_t ret = 0;

if (bus_attr->show)

//終于到了這里,最后的調(diào)用,調(diào)用bus_attr_drivers_autoprobe.show ,也就是show_drivers_autoprobe

//參數(shù)為bus_priv->bus,也就是platform_bus_type , 及一段char信息

ret = bus_attr->show(bus_priv->bus, buf);

return ret;

}

static ssize_t show_drivers_autoprobe(struct bus_type *bus, char *buf)

{

return sprintf(buf, "%d\n", bus->p->drivers_autoprobe);

}

沒什么好介紹了就是打印 buf + bus->p->drivers_autoprobe ? 從結(jié)果來看~ buf是空的

到這里,終于把內(nèi)核的信息給打印出來了,千辛萬苦,層層調(diào)用,就是為了取得上層kobject結(jié)構(gòu),逆運(yùn)算再取得kobject的上層結(jié)構(gòu)

大家是否對(duì)kobject有所了解了呢?~ ?

在對(duì)kobject進(jìn)行介紹之前 ?還是先把write操作講完吧 哈哈~

write操作和read操作重要的步驟基本是一致的,只不過在最后的調(diào)用中

static ssize_t store_drivers_autoprobe(struct bus_type *bus,

const char *buf, size_t count)

{

if (buf[0] == '0')

bus->p->drivers_autoprobe = 0;

else

bus->p->drivers_autoprobe = 1;

return count;

}

不進(jìn)行打印而對(duì)內(nèi)核的參數(shù)進(jìn)行了修改而已

好~ 現(xiàn)在讓我們來看看kobject吧

kobject的結(jié)構(gòu)如下

struct kobject {

const char ? ? ? ? ? ? ? ?*name; ? ? ? ? ?//kobject的名字

struct kref ? ? ? ? ? ? ? ?kref; ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?//kobject的原子操作

struct list_head ? ? ? ?entry;

struct kobject ? ? ? ? ? ? ? ?*parent; ? ? ? ? ? ? ? ? ? ? ? ?//父對(duì)象

struct kset ? ? ? ? ? ? ? ?*kset; ? ? ? ? ? ? ? ? ? ? ? ?//父容器

struct kobj_type ? ? ? ?*ktype; ? ? ? ? ? ? ? ? ? ? ? ?//ktype

struct sysfs_dirent ? ? ? ?*sd; ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?//文件節(jié)點(diǎn)

unsigned int state_initialized:1;

unsigned int state_in_sysfs:1;

unsigned int state_add_uevent_sent:1;

unsigned int state_remove_uevent_sent:1;

};

kobject描述的是較具體的對(duì)象,一個(gè)設(shè)備,一個(gè)驅(qū)動(dòng),一個(gè)總線,一類設(shè)備

在層次圖上可以看出,每個(gè)存在于層次圖中的設(shè)備,驅(qū)動(dòng),總線,類別都有自己的kobject

kobject與kobject之間的層次由kobject中的parent指針決定

而kset指針則表明了kobject的容器

像platform_bus 和test_device的kset都是devices_kset

呢parent和kset有什么不同呢

我認(rèn)為是人工和默認(rèn)的區(qū)別,看下面這張圖 ,藍(lán)框?yàn)閗set,紅框?yàn)閗object

容器提供了一種默認(rèn)的層次~ ?但也可以人工設(shè)置層次

對(duì)于kobject現(xiàn)在我只理解了這么多,歡迎大家指出有疑問的地方

最后是kset,kset比較簡單,看下面的結(jié)構(gòu)

struct kset {

struct list_head list;

spinlock_t list_lock;

struct kobject kobj;

struct kset_uevent_ops *uevent_ops;

};

對(duì)于kset的描述,文檔里也有介紹

/**

* struct kset - a set of kobjects of a specific type, belonging to a specific subsystem.

*

* A kset defines a group of kobjects. ?They can be individually

* different "types" but overall these kobjects all want to be grouped

* together and operated on in the same manner. ?ksets are used to

* define the attribute callbacks and other common events that happen to

* a kobject.

翻譯過來大概就是

結(jié)構(gòu)kset,一個(gè)指定類型的kobject的集合,屬于某一個(gè)指定的子系統(tǒng)

kset定義了一組kobject,它們可以是不同類型組成但卻希望捆在一起有一個(gè)統(tǒng)一的操作

kset通常被定義為回調(diào)屬性和其他通用的事件發(fā)生在kobject上

可能翻譯的不是很好,望大家見諒

從結(jié)構(gòu)中能看出kset比kobject多了3個(gè)屬性

list_head ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?//列表

spinlock_t ? ? ? ? ? ? ? ? ? ? ? ?//共享鎖

kset_uevent_ops ? ? ? ? ? ? ? ?//uevent操作集

list_head ? ? ? ?連接了所有kobject中kset屬性指向自己的kobject

而kset_uevent_ops則用于通知機(jī)制,由于uevent的作用我也沒接觸過,所以暫不解析uevent的機(jī)制了

?