Skip to content

Commit

Permalink
Merge pull request #3 from zhangxianyu777/zmx
Browse files Browse the repository at this point in the history 
收包过程中的数据链路层+xdp流程解析
  • Loading branch information
@zhangxianyu777
zhangxianyu777 authored Mar 2, 2025
2 parents 7bdfd4c + 215d2ba commit 5b5b6fc
Show file tree
Hide file tree
Showing 36 changed files with 597 additions and 53 deletions.
1 change: 1 addition & 0 deletions linux-6.2/fs/anon_inodes.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -140,6 +140,7 @@ static struct file *__anon_inode_getfile(const char *name,
* hence saving memory and avoiding code duplication for the file/inode/dentry
* setup. Returns the newly created file* or an error pointer.
*/
//通过将新文件实例连接到一个匿名元节点和一个描述文件“类”的字段，创建一个新文件实实例
struct file *anon_inode_getfile(const char *name,
const struct file_operations *fops,
void *priv, int flags)
Expand Down
57 changes: 55 additions & 2 deletions linux-6.2/fs/eventpoll.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -174,6 +174,7 @@ struct epitem {
* structure and represents the main data structure for the eventpoll
* interface.
*/
//epoll内核对象
struct eventpoll {
/*
* This mutex is used to ensure that files are not removed
Expand Down Expand Up @@ -812,6 +813,7 @@ static __poll_t __ep_eventpoll_poll(struct file *file, poll_table *wait, int dep
init_poll_funcptr(&pt, NULL);

/* Insert inside our poll wait queue */
//加入等待队列
poll_wait(file, &ep->poll_wait, wait);

/*
Expand Down Expand Up @@ -844,13 +846,16 @@ static __poll_t __ep_eventpoll_poll(struct file *file, poll_table *wait, int dep
* the ep->mtx so we need to start from depth=1, such that mutex_lock_nested()
* is correctly annotated.
*/
//关联 fd 和事件触发回调函数 ep_poll_callback
static __poll_t ep_item_poll(const struct epitem *epi, poll_table *pt,
int depth)
{
//获取文件描述符
struct file *file = epi->ffd.file;
__poll_t res;

pt->_key = epi->event.events;
//判断文件是否是 epoll 类型的文件
if (!is_file_epoll(file))
res = vfs_poll(file, pt);
else
Expand Down Expand Up @@ -935,6 +940,7 @@ void eventpoll_release_file(struct file *file)
mutex_unlock(&epmutex);
}

//申请eventpoll结构并初始化
static int ep_alloc(struct eventpoll **pep)
{
int error;
Expand All @@ -943,14 +949,18 @@ static int ep_alloc(struct eventpoll **pep)

user = get_current_user();
error = -ENOMEM;
//申请内存
ep = kzalloc(sizeof(*ep), GFP_KERNEL);
if (unlikely(!ep))
goto free_uid;

mutex_init(&ep->mtx);
rwlock_init(&ep->lock);
//初始化等待队列头
init_waitqueue_head(&ep->wq);
//初始化就绪列表
init_waitqueue_head(&ep->poll_wait);
//初始化红黑树指针
INIT_LIST_HEAD(&ep->rdllist);
ep->rbr = RB_ROOT_CACHED;
ep->ovflist = EP_UNACTIVE_PTR;
Expand Down Expand Up @@ -1133,10 +1143,13 @@ static inline bool chain_epi_lockless(struct epitem *epi)
* queues are used should be detected accordingly. This is detected using
* cmpxchg() operation.
*/
//唤醒相应进程
static int ep_poll_callback(wait_queue_entry_t *wait, unsigned mode, int sync, void *key)
{
int pwake = 0;
//获取wait中对应的epitem结构
struct epitem *epi = ep_item_from_wait(wait);
//获取epitem中eventpoll结构
struct eventpoll *ep = epi->ep;
__poll_t pollflags = key_to_poll(key);
unsigned long flags;
Expand Down Expand Up @@ -1175,6 +1188,7 @@ static int ep_poll_callback(wait_queue_entry_t *wait, unsigned mode, int sync, v
ep_pm_stay_awake_rcu(epi);
} else if (!ep_is_linked(epi)) {
/* In the usual case, add event to ready list. */
//将当前epitem添加到eventpoll的就绪队列中
if (list_add_tail_lockless(&epi->rdllink, &ep->rdllist))
ep_pm_stay_awake_rcu(epi);
}
Expand All @@ -1183,6 +1197,7 @@ static int ep_poll_callback(wait_queue_entry_t *wait, unsigned mode, int sync, v
* Wake up ( if active ) both the eventpoll wait list and the ->poll()
* wait list.
*/
//查看eventpoll的等待队列上是否有等待
if (waitqueue_active(&ep->wq)) {
if ((epi->event.events & EPOLLEXCLUSIVE) &&
!(pollflags & POLLFREE)) {
Expand Down Expand Up @@ -1238,6 +1253,7 @@ static int ep_poll_callback(wait_queue_entry_t *wait, unsigned mode, int sync, v
* This is the callback that is used to add our wait queue to the
* target file wakeup lists.
*/
//实际调用函数 whead为ep->poll_wait
static void ep_ptable_queue_proc(struct file *file, wait_queue_head_t *whead,
poll_table *pt)
{
Expand All @@ -1253,13 +1269,14 @@ static void ep_ptable_queue_proc(struct file *file, wait_queue_head_t *whead,
epq->epi = NULL;
return;
}

//初始化回调方法，为ep_poll_callback
init_waitqueue_func_entry(&pwq->wait, ep_poll_callback);
pwq->whead = whead;
pwq->base = epi;
if (epi->event.events & EPOLLEXCLUSIVE)
add_wait_queue_exclusive(whead, &pwq->wait);
else
//放入socket的等待队列whead
add_wait_queue(whead, &pwq->wait);
pwq->next = epi->pwqlist;
epi->pwqlist = pwq;
Expand Down Expand Up @@ -1443,12 +1460,15 @@ static int attach_epitem(struct file *file, struct epitem *epi)
/*
* Must be called with "mtx" held.
*/
//插入socket到eventpoll
static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
struct file *tfile, int fd, int full_check)
{
int error, pwake = 0;
__poll_t revents;
//声明epitem结构
struct epitem *epi;
//socket等待队列
struct ep_pqueue epq;
struct eventpoll *tep = NULL;

Expand All @@ -1462,14 +1482,18 @@ static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
return -ENOSPC;
percpu_counter_inc(&ep->user->epoll_watches);

//申请epi空间
if (!(epi = kmem_cache_zalloc(epi_cache, GFP_KERNEL))) {
percpu_counter_dec(&ep->user->epoll_watches);
return -ENOMEM;
}

/* Item initialization follow here ... */
//初始化epi
INIT_LIST_HEAD(&epi->rdllink);
//设置epi中ep为实际eventpoll对象
epi->ep = ep;
//设置epi的ffd为实际file及fd
ep_set_ffd(&epi->ffd, tfile, fd);
epi->event = *event;
epi->next = EP_UNACTIVE_PTR;
Expand All @@ -1492,6 +1516,7 @@ static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
* Add the current item to the RB tree. All RB tree operations are
* protected by "mtx", and ep_insert() is called with "mtx" held.
*/
//将epi插入红黑树中
ep_rbtree_insert(ep, epi);
if (tep)
mutex_unlock(&tep->mtx);
Expand All @@ -1511,7 +1536,9 @@ static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
}

/* Initialize the poll table using the queue callback */
//初始化epq等待队列
epq.epi = epi;
//注册回调函数为ep_ptable_queue_proc
init_poll_funcptr(&epq.pt, ep_ptable_queue_proc);

/*
Expand All @@ -1521,6 +1548,7 @@ static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
* this operation completes, the poll callback can start hitting
* the new item.
*/
//调用刚刚注册的函数
revents = ep_item_poll(epi, &epq.pt, 1);

/*
Expand Down Expand Up @@ -1755,6 +1783,7 @@ static struct timespec64 *ep_timeout_to_timespec(struct timespec64 *to, long ms)
* woken, and in that case the ep_poll loop will remove the entry anyways, not
* try to reuse it.
*/
//epoll唤醒的回调函数
static int ep_autoremove_wake_function(struct wait_queue_entry *wq_entry,
unsigned int mode, int sync, void *key)
{
Expand All @@ -1781,6 +1810,7 @@ static int ep_autoremove_wake_function(struct wait_queue_entry *wq_entry,
* Return: the number of ready events which have been fetched, or an
* error code, in case of error.
*/
//阻塞等待
static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
int maxevents, struct timespec64 *timeout)
{
Expand Down Expand Up @@ -1811,6 +1841,7 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
* timeout, the user by definition should not care and will have to
* recheck again.
*/

eavail = ep_events_available(ep);

while (1) {
Expand All @@ -1820,6 +1851,7 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
* 0 events and there's still timeout left over, we go
* trying again in search of more luck.
*/
//发送到用户态
res = ep_send_events(ep, events, maxevents);
if (res)
return res;
Expand Down Expand Up @@ -1852,6 +1884,7 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
* performance issue if a process is killed, causing all of its
* threads to wake up without being removed normally.
*/
//定义等待队列，设置回调函数
init_wait(&wait);
wait.func = ep_autoremove_wake_function;

Expand All @@ -1870,12 +1903,15 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
* period of time although events are pending, so lock is
* important.
*/
//判断就绪队列上有没有事件就绪
eavail = ep_events_available(ep);
if (!eavail)
//将队列添加到eventpoll的wq队列
__add_wait_queue_exclusive(&ep->wq, &wait);

write_unlock_irq(&ep->lock);

//主动进入睡眠态 阻塞自己
if (!eavail)
timed_out = !schedule_hrtimeout_range(to, slack,
HRTIMER_MODE_ABS);
Expand Down Expand Up @@ -1986,10 +2022,13 @@ static void clear_tfile_check_list(void)
/*
* Open an eventpoll file descriptor.
*/
//epoll对象创建
static int do_epoll_create(int flags)
{
int error, fd;
//eventpoll对象创建
struct eventpoll *ep = NULL;
//关联file结构 (eventpoll对象储存在file的private中,而file 被关联在当前进程的已打开文件中)
struct file *file;

/* Check the EPOLL_* constant for consistency. */
Expand All @@ -2000,25 +2039,31 @@ static int do_epoll_create(int flags)
/*
* Create the internal data structure ("struct eventpoll").
*/
//创建对象
error = ep_alloc(&ep);
if (error < 0)
return error;
/*
* Creates all the items needed to setup an eventpoll file. That is,
* a file structure and a free file descriptor.
*/
//获取未使用的文件描述符
fd = get_unused_fd_flags(O_RDWR | (flags & O_CLOEXEC));
if (fd < 0) {
error = fd;
goto out_free_ep;
}
//创建 eventpoll 文件对象，设置其文件操作结构体为 eventpoll_fops
//通过将新文件实例连接到一个匿名元节点和一个描述文件“类”的字段，创建一个新文件实实例
file = anon_inode_getfile("[eventpoll]", &eventpoll_fops, ep,
O_RDWR | (flags & O_CLOEXEC));
if (IS_ERR(file)) {
error = PTR_ERR(file);
goto out_free_fd;
}
//关联文件和 eventpoll 对象 （eventpoll中的file 是存储结构的file）
ep->file = file;
//将文件描述符 fd 安装到当前进程的文件描述符表（fdt->fd[fd]）
fd_install(fd, file);
return fd;

Expand All @@ -2034,6 +2079,7 @@ SYSCALL_DEFINE1(epoll_create1, int, flags)
return do_epoll_create(flags);
}

//创建epoll内核对象
SYSCALL_DEFINE1(epoll_create, int, size)
{
if (size <= 0)
Expand All @@ -2053,7 +2099,7 @@ static inline int epoll_mutex_lock(struct mutex *mutex, int depth,
return 0;
return -EAGAIN;
}

//实际为epoll添加socket函数
int do_epoll_ctl(int epfd, int op, int fd, struct epoll_event *epds,
bool nonblock)
{
Expand All @@ -2070,6 +2116,7 @@ int do_epoll_ctl(int epfd, int op, int fd, struct epoll_event *epds,
goto error_return;

/* Get the "struct file *" for the target file */
//获得file内核对象
tf = fdget(fd);
if (!tf.file)
goto error_fput;
Expand Down Expand Up @@ -2109,6 +2156,7 @@ int do_epoll_ctl(int epfd, int op, int fd, struct epoll_event *epds,
* At this point it is safe to assume that the "private_data" contains
* our own data structure.
*/
//获取eventpoll对象
ep = f.file->private_data;

/*
Expand Down Expand Up @@ -2162,6 +2210,7 @@ int do_epoll_ctl(int epfd, int op, int fd, struct epoll_event *epds,
case EPOLL_CTL_ADD:
if (!epi) {
epds->events |= EPOLLERR | EPOLLHUP;
//插入对应socket
error = ep_insert(ep, epds, tf.file, fd, full_check);
} else
error = -EEXIST;
Expand Down Expand Up @@ -2204,6 +2253,7 @@ int do_epoll_ctl(int epfd, int op, int fd, struct epoll_event *epds,
* the eventpoll file that enables the insertion/removal/change of
* file descriptors inside the interest set.
*/
//为epoll添加socket
SYSCALL_DEFINE4(epoll_ctl, int, epfd, int, op, int, fd,
struct epoll_event __user *, event)
{
Expand All @@ -2220,6 +2270,7 @@ SYSCALL_DEFINE4(epoll_ctl, int, epfd, int, op, int, fd,
* Implement the event wait interface for the eventpoll file. It is the kernel
* part of the user space epoll_wait(2).
*/
//阻塞等待
static int do_epoll_wait(int epfd, struct epoll_event __user *events,
int maxevents, struct timespec64 *to)
{
Expand Down Expand Up @@ -2255,13 +2306,15 @@ static int do_epoll_wait(int epfd, struct epoll_event __user *events,
ep = f.file->private_data;

/* Time to fish for events ... */
//实际函数
error = ep_poll(ep, events, maxevents, to);

error_fput:
fdput(f);
return error;
}

//epoll_wait 阻塞等待
SYSCALL_DEFINE4(epoll_wait, int, epfd, struct epoll_event __user *, events,
int, maxevents, int, timeout)
{
Expand Down
Loading

0 comments on commit 5b5b6fc

Please sign in to comment.