1.Linux“线程”
进程与线程之间是有区别的，不过linux内核只提供了轻量进程的支持，未实现线程模型。Linux是一种“多进程单线程”的操作系统。Linux本身只有进程的概念，而其所谓的“线程”本质上在内核里仍然是进程。
大家知道，进程是资源分配的单位，同一进程中的多个线程共享该进程的资源（如作为共享内存的全局变量）。Linux中所谓的“线程”只是在被创建时clone了父进程的资源，因此clone出来的进程表现为“线程”，这一点一定要弄清楚。因此，Linux“线程”这个概念只有在打冒号的情况下才是最准确的。
目前Linux中最流行的线程机制为LinuxThreads，所采用的就是线程－进程“一对一”模型，调度交给核心，而在用户级实现一个包括信号处理在内的线程管理机制。LinuxThreads由Xavier Leroy (Xavier.Leroy@inria.fr)负责开发完成，并已绑定在GLIBC中发行，它实现了一种BiCapitalized面向Linux的Posix 1003.1c “pthread”标准接口。Linuxthread可以支持Intel、Alpha、MIPS等平台上的多处理器系统。
按照POSIX 1003.1c 标准编写的程序与Linuxthread 库相链接即可支持Linux平台上的多线程，在程序中需包含头文件pthread. h，在编译链接时使用命令：
gcc -D -REENTRANT -lpthread xxx. c
其中-REENTRANT宏使得相关库函数(如stdio.h、errno.h中函数) 是可重入的、线程安全的(thread-safe)，-lpthread则意味着链接库目录下的libpthread.a或libpthread.so文件。使用Linuxthread库需要2.0以上版本的Linux内核及相应版本的C库(libc 5.2.18、libc 5.4.12、libc 6)。
2.“线程”控制
线程创建
进程被创建时，系统会为其创建一个主线程，而要在进程中创建新的线程，则可以调用pthread_create：
pthread_create(pthread_t *thread, const pthread_attr_t *attr, void *
(start_routine)(void*), void *arg);
start_routine为新线程的入口函数，arg为传递给start_routine的参数。
每个线程都有自己的线程ID，以便在进程内区分。线程ID在pthread_create调用时回返给创建线程的调用者；一个线程也可以在创建后使用pthread_self()调用获取自己的线程ID：
pthread_self (void) ;
线程退出
线程的退出方式有三：
（1）执行完成后隐式退出；
（2）由线程本身显示调用pthread_exit 函数退出；
pthread_exit (void * retval) ;
（3）被其他线程用pthread_cance函数终止：
pthread_cance (pthread_t thread) ;
在某线程中调用此函数，可以终止由参数thread 指定的线程。
如果一个线程要等待另一个线程的终止，可以使用pthread_join函数，该函数的作用是调用pthread_join的线程将被挂起直到线程ID为参数thread的线程终止：
pthread_join (pthread_t thread, void** threadreturn);
3.线程通信
线程互斥
互斥意味着“排它”，即两个线程不能同时进入被互斥保护的代码。Linux下可以通过pthread_mutex_t 定义互斥体机制完成多线程的互斥操作，该机制的作用是对某个需要互斥的部分，在进入时先得到互斥体，如果没有得到互斥体，表明互斥部分被其它线程拥有，此时欲获取互斥体的线程阻塞，直到拥有该互斥体的线程完成互斥部分的操作为止。
下面的代码实现了对共享全局变量x 用互斥体mutex 进行保护的目的：
int x; // 进程中的全局变量
pthread_mutex_t mutex;
pthread_mutex_init(&mutex, NULL); //按缺省的属性初始化互斥体变量mutex
pthread_mutex_lock(&mutex); // 给互斥体变量加锁
… //对变量x 的操作
phtread_mutex_unlock(&mutex); // 给互斥体变量解除锁
线程同步
同步就是线程等待某个事件的发生。只有当等待的事件发生线程才继续执行，否则线程挂起并放弃处理器。当多个线程协作时，相互作用的任务必须在一定的条件下同步。
Linux下的C语言编程有多种线程同步机制，最典型的是条件变量(condition variable)。pthread_cond_init用来创建一个条件变量，其函数原型为：
pthread_cond_init (pthread_cond_t *cond, const pthread_condattr_t *attr);
pthread_cond_wait和pthread_cond_timedwait用来等待条件变量被设置，值得注意的是这两个等待调用需要一个已经上锁的互斥体mutex，这是为了防止在真正进入等待状态之前别的线程有可能设置该条件变量而产生竞争。pthread_cond_wait的函数原型为：
pthread_cond_wait (pthread_cond_t *cond, pthread_mutex_t *mutex);
pthread_cond_broadcast用于设置条件变量，即使得事件发生，这样等待该事件的线程将不再阻塞：
pthread_cond_broadcast (pthread_cond_t *cond) ;
pthread_cond_signal则用于解除某一个等待线程的阻塞状态：
pthread_cond_signal (pthread_cond_t *cond) ;
pthread_cond_destroy 则用于释放一个条件变量的资源。
在头文件semaphore.h 中定义的信号量则完成了互斥体和条件变量的封装，按照多线程程序设计中访问控制机制，控制对资源的同步访问，提供程序设计人员更方便的调用接口。
sem_init(sem_t *sem, int pshared, unsigned int val);
这个函数初始化一个信号量sem 的值为val，参数pshared 是共享属性控制，表明是否在进程间共享。
sem_wait(sem_t *sem);
调用该函数时，若sem为无状态，调用线程阻塞，等待信号量sem值增加(post )成为有信号状态；若sem为有状态，调用线程顺序执行，但信号量的值减一。
sem_post(sem_t *sem);
调用该函数，信号量sem的值增加，可以从无信号状态变为有信号状态。
4.实例
下面我们还是以名的生产者/消费者问题为例来阐述Linux线程的控制和通信。一组生产者线程与一组消费者线程通过缓冲区发生联系。生产者线程将生产的产品送入缓冲区，消费者线程则从中取出产品。缓冲区有N 个，是一个环形的缓冲池。
#include <stdio.h>
#include <pthread.h>
#define BUFFER_SIZE 16 // 缓冲区数量
struct prodcons
{
// 缓冲区相关数据结构
int buffer[BUFFER_SIZE]; /* 实际数据存放的数组*/
pthread_mutex_t lock; /* 互斥体lock 用于对缓冲区的互斥操作 */
int readpos, writepos; /* 读写指针*/
pthread_cond_t notempty; /* 缓冲区非空的条件变量 */
pthread_cond_t notfull; /* 缓冲区未满的条件变量 */
};
/* 初始化缓冲区结构 */
void init(struct prodcons *b)
{
pthread_mutex_init(&b->lock, NULL);
pthread_cond_init(&b->notempty, NULL);
pthread_cond_init(&b->notfull, NULL);
b->readpos = 0;
b->writepos = 0;
}
/* 将产品放入缓冲区,这里是存入一个整数*/
void put(struct prodcons *b, int data)
{
pthread_mutex_lock(&b->lock);
/* 等待缓冲区未满*/
if ((b->writepos + 1) % BUFFER_SIZE == b->readpos)
{
pthread_cond_wait(&b->notfull, &b->lock);
}
/* 写数据,并移动指针 */
b->buffer[b->writepos] = data;
b->writepos++;
if (b->writepos >= BUFFER_SIZE)
b->writepos = 0;
/* 设置缓冲区非空的条件变量*/
pthread_cond_signal(&b->notempty);
pthread_mutex_unlock(&b->lock);
}
/* 从缓冲区中取出整数*/
int get(struct prodcons *b)
{
int data;
pthread_mutex_lock(&b->lock);
/* 等待缓冲区非空*/
if (b->writepos == b->readpos)
{
pthread_cond_wait(&b->notempty, &b->lock);
}
/* 读数据,移动读指针*/
data = b->buffer[b->readpos];
b->readpos++;
if (b->readpos >= BUFFER_SIZE)
b->readpos = 0;
/* 设置缓冲区未满的条件变量*/
pthread_cond_signal(&b->notfull);
pthread_mutex_unlock(&b->lock);
return data;
}
/* 测试:生产者线程将1 到10000 的整数送入缓冲区,消费者线
程从缓冲区中获取整数,两者都打印信息*/
#define OVER ( - 1)
struct prodcons buffer;
void *producer(void *data)
{
int n;
for (n = 0; n < 10000; n++)
{
printf("%d --->\n", n);
put(&buffer, n);
} put(&buffer, OVER);
return NULL;
}
void *consumer(void *data)
{
int d;
while (1)
{
d = get(&buffer);
if (d == OVER)
break;
printf("--->%d \n", d);
}
return NULL;
}
int main(void)
{
pthread_t th_a, th_b;
void *retval;
init(&buffer);
/* 创建生产者和消费者线程*/
pthread_create(&th_a, NULL, producer, 0);
pthread_create(&th_b, NULL, consumer, 0);
/* 等待两个线程结束*/
pthread_join(th_a, &retval);
pthread_join(th_b, &retval);
return 0;
}
5.WIN32、VxWorks、Linux线程类比
目前为止，笔者已经创作了《基于嵌入式操作系统VxWorks的多任务并发程序设计》（《软件报》2006年5~12期连载）、《深入浅出Win32多线程程序设计》（天极网技术专题）系列，我们来找出这两个系列文章与本文的共通点。
看待技术问题要瞄准其本质，不管是Linux、VxWorks还是WIN32，其涉及到多线程的部分都是那些内容，无非就是线程控制和线程通信，它们的许多函数只是名称不同，其实质含义是等价的，下面我们来列个三大操作系统共同点详细表单：
6.小结
本章讲述了Linux下多线程的控制及线程间通信编程方法，给出了一个生产者/消费者的实例，并将Linux的多线程与WIN32、VxWorks多线程进行了类比，总结了一般规律。鉴于多线程编程已成为开发并发应用程序的主流方法，学好本章的意义也便不言自明。
完
#include <stdio.h>
#include <stdio.h>
#include <pthread.h>
void thread(void)
{
int i;
for(i=0;i<3;i++)
printf("This is a pthread.\n");
}
int main(void)
{
pthread_t id;
int i,ret;
ret=pthread_create(&id,NULL,(void *) thread,NULL);
if(ret!=0){
printf ("Create pthread error!\n");
exit (1);
}
for(i=0;i<3;i++)
printf("This is the main process.\n");
pthread_join(id,NULL);
return (0);
}
编译：
gcc example1.c -lpthread -o example1
#include <pthread.h>
#include <stdio.h>
#include <sys/time.h>
#include <string.h>
#define MAX 10
pthread_t thread[2];
pthread_mutex_t mut;
int number=0, i;
void *thread1()
{
printf ("thread1 : I'm thread 1\n");
for (i = 0; i < MAX; i++)
{
printf("thread1 : number = %d\n",number);
pthread_mutex_lock(&mut);
number++;
pthread_mutex_unlock(&mut);
sleep(2);
}
printf("thread1 :主函数在等我完成任务吗？\n");
pthread_exit(NULL);
}
void *thread2()
{
printf("thread2 : I'm thread 2\n");
for (i = 0; i < MAX; i++)
{
printf("thread2 : number = %d\n",number);
pthread_mutex_lock(&mut);
number++;
pthread_mutex_unlock(&mut);
sleep(3);
}
printf("thread2 :主函数在等我完成任务吗？\n");
pthread_exit(NULL);
}
void thread_create(void)
{
int temp;
memset(&thread, 0, sizeof(thread)); //comment1
//创建线程
if((temp = pthread_create(&thread[0], NULL, thread1, NULL)) != 0) //comment2
printf("线程1创建失败!\n");
else
printf("线程1被创建\n");
if((temp = pthread_create(&thread[1], NULL, thread2, NULL)) != 0) //comment3
printf("线程2创建失败");
else
printf("线程2被创建\n");
}
void thread_wait(void)
{
//等待线程结束
if(thread[0] !=0) { //comment4
pthread_join(thread[0],NULL);
printf("线程1已经结束\n");
}
if(thread[1] !=0) { //comment5
pthread_join(thread[1],NULL);
printf("线程2已经结束\n");
}
}
int main()
{
//用默认属性初始化互斥锁
pthread_mutex_init(&mut,NULL);
printf("我是主函数哦，我正在创建线程，呵呵\n");
thread_create();
printf("我是主函数哦，我正在等待线程完成任务阿，呵呵\n");
thread_wait();
return 0;
}
编译 :
gcc -lpthread -o thread_example lp.c
转自http://www.cnblogs.com/BiffoLee/archive/2011/11/18/2254540.html
另外关于pthred_cond_wait的实现分析为何参数要传入mutex
转载的关于pthread_cond_wait的文章，写的比较详细
pthread_cond_wait()的实现原理
深入理解pthread_cond_wait、pthread_cond_signal
下载了glibc-2.31.tar看下实现
/* Block on condition variable COND. MUTEX should be held by the
calling thread. On success, MUTEX will be held by the calling
thread. */
extern int pthread_cond_wait (pthread_cond_t *__restrict __cond,
pthread_mutex_t *__restrict __mutex)
__nonnull ((1, 2));
versioned_symbol (libpthread, __pthread_cond_wait, pthread_cond_wait,
GLIBC_2_3_2);
versioned_symbol (libpthread, __pthread_cond_timedwait, pthread_cond_timedwait,
GLIBC_2_3_2);
/* See __pthread_cond_wait_common. */
int
__pthread_cond_wait (pthread_cond_t *cond, pthread_mutex_t *mutex)
{
/* clockid is unused when abstime is NULL. */
return __pthread_cond_wait_common (cond, mutex, 0, NULL);
}
/* See __pthread_cond_wait_common. */
int
__pthread_cond_timedwait (pthread_cond_t *cond, pthread_mutex_t *mutex,
const struct timespec *abstime)
{
/* Check parameter validity. This should also tell the compiler that
it can assume that abstime is not NULL. */
if (! valid_nanoseconds (abstime->tv_nsec))
return EINVAL;
/* Relaxed MO is suffice because clock ID bit is only modified
in condition creation. */
unsigned int flags = atomic_load_relaxed (&cond->__data.__wrefs);
clockid_t clockid = (flags & __PTHREAD_COND_CLOCK_MONOTONIC_MASK)
? CLOCK_MONOTONIC : CLOCK_REALTIME;
return __pthread_cond_wait_common (cond, mutex, clockid, abstime);
}
/* This condvar implementation guarantees that all calls to signal and
broadcast and all of the three virtually atomic parts of each call to wait
(i.e., (1) releasing the mutex and blocking, (2) unblocking, and (3) re-
acquiring the mutex) happen in some total order that is consistent with the
happens-before relations in the calling program. However, this order does
not necessarily result in additional happens-before relations being
established (which aligns well with spurious wake-ups being allowed).
All waiters acquire a certain position in a 64b waiter sequence (__wseq).
This sequence determines which waiters are allowed to consume signals.
A broadcast is equal to sending as many signals as are unblocked waiters.
When a signal arrives, it samples the current value of __wseq with a
relaxed-MO load (i.e., the position the next waiter would get). (This is
sufficient because it is consistent with happens-before; the caller can
enforce stronger ordering constraints by calling signal while holding the
mutex.) Only waiters with a position less than the __wseq value observed
by the signal are eligible to consume this signal.
This would be straight-forward to implement if waiters would just spin but
we need to let them block using futexes. Futexes give no guarantee of
waking in FIFO order, so we cannot reliably wake eligible waiters if we
just use a single futex. Also, futex words are 32b in size, but we need
to distinguish more than 1<<32 states because we need to represent the
order of wake-up (and thus which waiters are eligible to consume signals);
blocking in a futex is not atomic with a waiter determining its position in
the waiter sequence, so we need the futex word to reliably notify waiters
that they should not attempt to block anymore because they have been
already signaled in the meantime. While an ABA issue on a 32b value will
be rare, ignoring it when we are aware of it is not the right thing to do
either.
Therefore, we use a 64b counter to represent the waiter sequence (on
architectures which only support 32b atomics, we use a few bits less).
To deal with the blocking using futexes, we maintain two groups of waiters:
* Group G1 consists of waiters that are all eligible to consume signals;
incoming signals will always signal waiters in this group until all
waiters in G1 have been signaled.
* Group G2 consists of waiters that arrive when a G1 is present and still
contains waiters that have not been signaled. When all waiters in G1
are signaled and a new signal arrives, the new signal will convert G2
into the new G1 and create a new G2 for future waiters.
We cannot allocate new memory because of process-shared condvars, so we
have just two slots of groups that change their role between G1 and G2.
Each has a separate futex word, a number of signals available for
consumption, a size (number of waiters in the group that have not been
signaled), and a reference count.
The group reference count is used to maintain the number of waiters that
are using the group's futex. Before a group can change its role, the
reference count must show that no waiters are using the futex anymore; this
prevents ABA issues on the futex word.
To represent which intervals in the waiter sequence the groups cover (and
thus also which group slot contains G1 or G2), we use a 64b counter to
designate the start position of G1 (inclusive), and a single bit in the
waiter sequence counter to represent which group slot currently contains
G2. This allows us to switch group roles atomically wrt. waiters obtaining
a position in the waiter sequence. The G1 start position allows waiters to
figure out whether they are in a group that has already been completely
signaled (i.e., if the current G1 starts at a later position that the
waiter's position). Waiters cannot determine whether they are currently
in G2 or G1 -- but they do not have too because all they are interested in
is whether there are available signals, and they always start in G2 (whose
group slot they know because of the bit in the waiter sequence. Signalers
will simply fill the right group until it is completely signaled and can
be closed (they do not switch group roles until they really have to to
decrease the likelihood of having to wait for waiters still holding a
reference on the now-closed G1).
Signalers maintain the initial size of G1 to be able to determine where
G2 starts (G2 is always open-ended until it becomes G1). They track the
remaining size of a group; when waiters cancel waiting (due to PThreads
cancellation or timeouts), they will decrease this remaining size as well.
To implement condvar destruction requirements (i.e., that
pthread_cond_destroy can be called as soon as all waiters have been
signaled), waiters increment a reference count before starting to wait and
decrement it after they stopped waiting but right before they acquire the
mutex associated with the condvar.
pthread_cond_t thus consists of the following (bits that are used for
flags and are not part of the primary value of each field but necessary
to make some things atomic or because there was no space for them
elsewhere in the data structure):
__wseq: Waiter sequence counter
* LSB is index of current G2.
* Waiters fetch-add while having acquire the mutex associated with the
condvar. Signalers load it and fetch-xor it concurrently.
__g1_start: Starting position of G1 (inclusive)
* LSB is index of current G2.
* Modified by signalers while having acquired the condvar-internal lock
and observed concurrently by waiters.
__g1_orig_size: Initial size of G1
* The two least-significant bits represent the condvar-internal lock.
* Only accessed while having acquired the condvar-internal lock.
__wrefs: Waiter reference counter.
* Bit 2 is true if waiters should run futex_wake when they remove the
last reference. pthread_cond_destroy uses this as futex word.
* Bit 1 is the clock ID (0 == CLOCK_REALTIME, 1 == CLOCK_MONOTONIC).
* Bit 0 is true iff this is a process-shared condvar.
* Simple reference count used by both waiters and pthread_cond_destroy.
(If the format of __wrefs is changed, update nptl_lock_constants.pysym
and the pretty printers.)
For each of the two groups, we have:
__g_refs: Futex waiter reference count.
* LSB is true if waiters should run futex_wake when they remove the
last reference.
* Reference count used by waiters concurrently with signalers that have
acquired the condvar-internal lock.
__g_signals: The number of signals that can still be consumed.
* Used as a futex word by waiters. Used concurrently by waiters and
signalers.
* LSB is true iff this group has been completely signaled (i.e., it is
closed).
__g_size: Waiters remaining in this group (i.e., which have not been
signaled yet.
* Accessed by signalers and waiters that cancel waiting (both do so only
when having acquired the condvar-internal lock.
* The size of G2 is always zero because it cannot be determined until
the group becomes G1.
* Although this is of unsigned type, we rely on using unsigned overflow
rules to make this hold effectively negative values too (in
particular, when waiters in G2 cancel waiting).
A PTHREAD_COND_INITIALIZER condvar has all fields set to zero, which yields
a condvar that has G2 starting at position 0 and a G1 that is closed.
Because waiters do not claim ownership of a group right when obtaining a
position in __wseq but only reference count the group when using futexes
to block, it can happen that a group gets closed before a waiter can
increment the reference count. Therefore, waiters have to check whether
their group is already closed using __g1_start. They also have to perform
this check when spinning when trying to grab a signal from __g_signals.
Note that for these checks, using relaxed MO to load __g1_start is
sufficient because if a waiter can see a sufficiently large value, it could
have also consume a signal in the waiters group.
Waiters try to grab a signal from __g_signals without holding a reference
count, which can lead to stealing a signal from a more recent group after
their own group was already closed. They cannot always detect whether they
in fact did because they do not know when they stole, but they can
conservatively add a signal back to the group they stole from; if they
did so unnecessarily, all that happens is a spurious wake-up. To make this
even less likely, __g1_start contains the index of the current g2 too,
which allows waiters to check if there aliasing on the group slots; if
there wasn't, they didn't steal from the current G1, which means that the
G1 they stole from must have been already closed and they do not need to
fix anything.
It is essential that the last field in pthread_cond_t is __g_signals[1]:
The previous condvar used a pointer-sized field in pthread_cond_t, so a
PTHREAD_COND_INITIALIZER from that condvar implementation might only
initialize 4 bytes to zero instead of the 8 bytes we need (i.e., 44 bytes
in total instead of the 48 we need). __g_signals[1] is not accessed before
the first group switch (G2 starts at index 0), which will set its value to
zero after a harmless fetch-or whose return value is ignored. This
effectively completes initialization.
Limitations:
* This condvar isn't designed to allow for more than
__PTHREAD_COND_MAX_GROUP_SIZE * (1 << 31) calls to __pthread_cond_wait.
* More than __PTHREAD_COND_MAX_GROUP_SIZE concurrent waiters are not
supported.
* Beyond what is allowed as errors by POSIX or documented, we can also
return the following errors:
* EPERM if MUTEX is a recursive mutex and the caller doesn't own it.
* EOWNERDEAD or ENOTRECOVERABLE when using robust mutexes. Unlike
for other errors, this can happen when we re-acquire the mutex; this
isn't allowed by POSIX (which requires all errors to virtually happen
before we release the mutex or change the condvar state), but there's
nothing we can do really.
* When using PTHREAD_MUTEX_PP_* mutexes, we can also return all errors
returned by __pthread_tpp_change_priority. We will already have
released the mutex in such cases, so the caller cannot expect to own
MUTEX.
Other notes:
* Instead of the normal mutex unlock / lock functions, we use
__pthread_mutex_unlock_usercnt(m, 0) / __pthread_mutex_cond_lock(m)
because those will not change the mutex-internal users count, so that it
can be detected when a condvar is still associated with a particular
mutex because there is a waiter blocked on this condvar using this mutex.
*/
static __always_inline int
__pthread_cond_wait_common (pthread_cond_t *cond, pthread_mutex_t *mutex,
clockid_t clockid,
const struct timespec *abstime)
{
const int maxspin = 0;
int err;
int result = 0;
LIBC_PROBE (cond_wait, 2, cond, mutex);
/* clockid will already have been checked by
__pthread_cond_clockwait or pthread_condattr_setclock, or we
don't use it if abstime is NULL, so we don't need to check it
here. */
/* Acquire a position (SEQ) in the waiter sequence (WSEQ). We use an
atomic operation because signals and broadcasts may update the group
switch without acquiring the mutex. We do not need release MO here
because we do not need to establish any happens-before relation with
signalers (see __pthread_cond_signal); modification order alone
establishes a total order of waiters/signals. We do need acquire MO
to synchronize with group reinitialization in
__condvar_quiesce_and_switch_g1. */
uint64_t wseq = __condvar_fetch_add_wseq_acquire (cond, 2);
/* Find our group's index. We always go into what was G2 when we acquired
our position. */
unsigned int g = wseq & 1;
uint64_t seq = wseq >> 1;
/* Increase the waiter reference count. Relaxed MO is sufficient because
we only need to synchronize when decrementing the reference count. */
unsigned int flags = atomic_fetch_add_relaxed (&cond->__data.__wrefs, 8);
int private = __condvar_get_private (flags);
/* Now that we are registered as a waiter, we can release the mutex.
Waiting on the condvar must be atomic with releasing the mutex, so if
the mutex is used to establish a happens-before relation with any
signaler, the waiter must be visible to the latter; thus, we release the
mutex after registering as waiter.
If releasing the mutex fails, we just cancel our registration as a
waiter and confirm that we have woken up. */
err = __pthread_mutex_unlock_usercnt (mutex, 0);
if (__glibc_unlikely (err != 0))
{
__condvar_cancel_waiting (cond, seq, g, private);
__condvar_confirm_wakeup (cond, private);
return err;
}
/* Now wait until a signal is available in our group or it is closed.
Acquire MO so that if we observe a value of zero written after group
switching in __condvar_quiesce_and_switch_g1, we synchronize with that
store and will see the prior update of __g1_start done while switching
groups too. */
unsigned int signals = atomic_load_acquire (cond->__data.__g_signals + g);
do
{
while (1)
{
/* Spin-wait first.
Note that spinning first without checking whether a timeout
passed might lead to what looks like a spurious wake-up even
though we should return ETIMEDOUT (e.g., if the caller provides
an absolute timeout that is clearly in the past). However,
(1) spurious wake-ups are allowed, (2) it seems unlikely that a
user will (ab)use pthread_cond_wait as a check for whether a
point in time is in the past, and (3) spinning first without
having to compare against the current time seems to be the right
choice from a performance perspective for most use cases. */
unsigned int spin = maxspin;
while (signals == 0 && spin > 0)
{
/* Check that we are not spinning on a group that's already
closed. */
if (seq < (__condvar_load_g1_start_relaxed (cond) >> 1))
goto done;
/* TODO Back off. */
/* Reload signals. See above for MO. */
signals = atomic_load_acquire (cond->__data.__g_signals + g);
spin--;
}
/* If our group will be closed as indicated by the flag on signals,
don't bother grabbing a signal. */
if (signals & 1)
goto done;
/* If there is an available signal, don't block. */
if (signals != 0)
break;
/* No signals available after spinning, so prepare to block.
We first acquire a group reference and use acquire MO for that so
that we synchronize with the dummy read-modify-write in
__condvar_quiesce_and_switch_g1 if we read from that. In turn,
in this case this will make us see the closed flag on __g_signals
that designates a concurrent attempt to reuse the group's slot.
We use acquire MO for the __g_signals check to make the
__g1_start check work (see spinning above).
Note that the group reference acquisition will not mask the
release MO when decrementing the reference count because we use
an atomic read-modify-write operation and thus extend the release
sequence. */
atomic_fetch_add_acquire (cond->__data.__g_refs + g, 2);
if (((atomic_load_acquire (cond->__data.__g_signals + g) & 1) != 0)
|| (seq < (__condvar_load_g1_start_relaxed (cond) >> 1)))
{
/* Our group is closed. Wake up any signalers that might be
waiting. */
__condvar_dec_grefs (cond, g, private);
goto done;
}
// Now block.
struct _pthread_cleanup_buffer buffer;
struct _condvar_cleanup_buffer cbuffer;
cbuffer.wseq = wseq;
cbuffer.cond = cond;
cbuffer.mutex = mutex;
cbuffer.private = private;
__pthread_cleanup_push (&buffer, __condvar_cleanup_waiting, &cbuffer);
if (abstime == NULL)
{
/* Block without a timeout. */
err = futex_wait_cancelable (
cond->__data.__g_signals + g, 0, private);
}
else
{
/* Block, but with a timeout.
Work around the fact that the kernel rejects negative timeout
values despite them being valid. */
if (__glibc_unlikely (abstime->tv_sec < 0))
err = ETIMEDOUT;
else
{
err = futex_abstimed_wait_cancelable
(cond->__data.__g_signals + g, 0, clockid, abstime,
private);
}
}
__pthread_cleanup_pop (&buffer, 0);
if (__glibc_unlikely (err == ETIMEDOUT))
{
__condvar_dec_grefs (cond, g, private);
/* If we timed out, we effectively cancel waiting. Note that
we have decremented __g_refs before cancellation, so that a
deadlock between waiting for quiescence of our group in
__condvar_quiesce_and_switch_g1 and us trying to acquire
the lock during cancellation is not possible. */
__condvar_cancel_waiting (cond, seq, g, private);
result = ETIMEDOUT;
goto done;
}
else
__condvar_dec_grefs (cond, g, private);
/* Reload signals. See above for MO. */
signals = atomic_load_acquire (cond->__data.__g_signals + g);
}
}
/* Try to grab a signal. Use acquire MO so that we see an up-to-date value
of __g1_start below (see spinning above for a similar case). In
particular, if we steal from a more recent group, we will also see a
more recent __g1_start below. */
while (!atomic_compare_exchange_weak_acquire (cond->__data.__g_signals + g,
&signals, signals - 2));
/* We consumed a signal but we could have consumed from a more recent group
that aliased with ours due to being in the same group slot. If this
might be the case our group must be closed as visible through
__g1_start. */
uint64_t g1_start = __condvar_load_g1_start_relaxed (cond);
if (seq < (g1_start >> 1))
{
/* We potentially stole a signal from a more recent group but we do not
know which group we really consumed from.
We do not care about groups older than current G1 because they are
closed; we could have stolen from these, but then we just add a
spurious wake-up for the current groups.
We will never steal a signal from current G2 that was really intended
for G2 because G2 never receives signals (until it becomes G1). We
could have stolen a signal from G2 that was conservatively added by a
previous waiter that also thought it stole a signal -- but given that
that signal was added unnecessarily, it's not a problem if we steal
it.
Thus, the remaining case is that we could have stolen from the current
G1, where "current" means the __g1_start value we observed. However,
if the current G1 does not have the same slot index as we do, we did
not steal from it and do not need to undo that. This is the reason
for putting a bit with G2's index into__g1_start as well. */
if (((g1_start & 1) ^ 1) == g)
{
/* We have to conservatively undo our potential mistake of stealing
a signal. We can stop trying to do that when the current G1
changes because other spinning waiters will notice this too and
__condvar_quiesce_and_switch_g1 has checked that there are no
futex waiters anymore before switching G1.
Relaxed MO is fine for the __g1_start load because we need to
merely be able to observe this fact and not have to observe
something else as well.
??? Would it help to spin for a little while to see whether the
current G1 gets closed? This might be worthwhile if the group is
small or close to being closed. */
unsigned int s = atomic_load_relaxed (cond->__data.__g_signals + g);
while (__condvar_load_g1_start_relaxed (cond) == g1_start)
{
/* Try to add a signal. We don't need to acquire the lock
because at worst we can cause a spurious wake-up. If the
group is in the process of being closed (LSB is true), this
has an effect similar to us adding a signal. */
if (((s & 1) != 0)
|| atomic_compare_exchange_weak_relaxed
(cond->__data.__g_signals + g, &s, s + 2))
{
/* If we added a signal, we also need to add a wake-up on
the futex. We also need to do that if we skipped adding
a signal because the group is being closed because
while __condvar_quiesce_and_switch_g1 could have closed
the group, it might stil be waiting for futex waiters to
leave (and one of those waiters might be the one we stole
the signal from, which cause it to block using the
futex). */
futex_wake (cond->__data.__g_signals + g, 1, private);
break;
}
/* TODO Back off. */
}
}
}
done:
/* Confirm that we have been woken. We do that before acquiring the mutex
to allow for execution of pthread_cond_destroy while having acquired the
mutex. */
__condvar_confirm_wakeup (cond, private);
/* Woken up; now re-acquire the mutex. If this doesn't fail, return RESULT,
which is set to ETIMEDOUT if a timeout occured, or zero otherwise. */
err = __pthread_mutex_cond_lock (mutex);
/* XXX Abort on errors that are disallowed by POSIX? */
return (err != 0) ? err : result;
}
源码比较难懂
但可以看见有调用
err = __pthread_mutex_unlock_usercnt (mutex, 0);
...
err = __pthread_mutex_cond_lock (mutex);
应该是调用pthread_cond_wait后先释放锁，然后加入waiter队列等待signal，如果有signal唤醒了该线程则后续执行重新上锁
