static void
bufferevent_generic_read_timeout_cb(evutil_socket_t fd, short event, void *ctx)
{
struct bufferevent *bev = ctx;
_bufferevent_incref_and_lock(bev);
bufferevent_disable(bev, EV_READ);
_bufferevent_run_eventcb(bev, BEV_EVENT_TIMEOUT|BEV_EVENT_READING);
_bufferevent_decref_and_unlock(bev);
}
static void
write_complete(struct event_overlapped *eo, ev_uintptr_t key,
ev_ssize_t nbytes, int ok)
{
struct bufferevent_async *bev_a = upcast_write(eo);
struct bufferevent *bev = &bev_a->bev.bev;
short what = BEV_EVENT_WRITING;
ev_ssize_t amount_unwritten;
BEV_LOCK(bev);
EVUTIL_ASSERT(bev_a->write_in_progress);
amount_unwritten = bev_a->write_in_progress - nbytes;
evbuffer_commit_write(bev->output, nbytes);
bev_a->write_in_progress = 0;
if (amount_unwritten)
_bufferevent_decrement_write_buckets(&bev_a->bev,
-amount_unwritten);
if (!ok)
bev_async_set_wsa_error(bev, eo);
if (bev_a->ok) {
if (ok && nbytes) {
BEV_RESET_GENERIC_WRITE_TIMEOUT(bev);
if (evbuffer_get_length(bev->output) <=
bev->wm_write.low)
_bufferevent_run_writecb(bev);
bev_async_consider_writing(bev_a);
} else if (!ok) {
what |= BEV_EVENT_ERROR;
bev_a->ok = 0;
_bufferevent_run_eventcb(bev, what);
} else if (!nbytes) {
what |= BEV_EVENT_EOF;
bev_a->ok = 0;
_bufferevent_run_eventcb(bev, what);
}
}
_bufferevent_decref_and_unlock(bev);
}
static void
bev_async_consider_reading(struct bufferevent_async *beva)
{
size_t cur_size;
size_t read_high;
size_t at_most;
int limit;
struct bufferevent *bev = &beva->bev.bev;
/* Don't read if there is a read in progress, or we do not
* want to read. */
if (beva->read_in_progress || beva->bev.connecting)
return;
if (!beva->ok || !(bev->enabled&EV_READ)) {
bev_async_del_read(beva);
return;
}
/* Don't read if we're full */
cur_size = evbuffer_get_length(bev->input);
read_high = bev->wm_read.high;
if (read_high) {
if (cur_size >= read_high) {
bev_async_del_read(beva);
return;
}
at_most = read_high - cur_size;
} else {
at_most = 16384; /* FIXME totally magic. */
}
/* XXXX This over-commits. */
/* XXXX see also not above on cast on _bufferevent_get_write_max() */
limit = (int)_bufferevent_get_read_max(&beva->bev);
if (at_most >= (size_t)limit && limit >= 0)
at_most = limit;
if (beva->bev.read_suspended) {
bev_async_del_read(beva);
return;
}
bufferevent_incref(bev);
if (evbuffer_launch_read(bev->input, at_most, &beva->read_overlapped)) {
beva->ok = 0;
_bufferevent_run_eventcb(bev, BEV_EVENT_ERROR);
bufferevent_decref(bev);
} else {
beva->read_in_progress = at_most;
_bufferevent_decrement_read_buckets(&beva->bev, at_most);
bev_async_add_read(beva);
}
return;
}
/* Callback: Invoked when we are done resolving (or failing to resolve) the
* hostname */
static void
dns_reply_callback(int result, char type, int count, int ttl, void *addresses,
void *arg)
{
struct resolveinfo *info = arg;
struct sockaddr_in sin;
struct sockaddr_in6 sin6;
struct sockaddr *sa = NULL;
int socklen;
EVUTIL_ASSERT(info->bev);
BEV_LOCK(info->bev);
if (result != DNS_ERR_NONE || count == 0) {
_bufferevent_run_eventcb(info->bev, BEV_EVENT_ERROR);
_bufferevent_decref_and_unlock(info->bev);
memset(info, 0, sizeof(*info));
mm_free(info);
return;
}
if (type == DNS_IPv4_A) {
EVUTIL_ASSERT(info->family == AF_INET);
memset(&sin, 0, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_port = info->port;
/* XXX handle multiple addresses better */
sin.sin_addr.s_addr = *(ev_uint32_t*)addresses;
sa = (struct sockaddr*)&sin;
socklen = sizeof(sin);
} else if (type == DNS_IPv6_AAAA) {
EVUTIL_ASSERT(info->family == AF_INET6);
memset(&sin6, 0, sizeof(sin6));
sin6.sin6_family = AF_INET;
sin6.sin6_port = info->port;
/* XXX handle multiple addresses better */
memcpy(sin6.sin6_addr.s6_addr, addresses, 16);
sa = (struct sockaddr*)&sin6;
socklen = sizeof(sin6);
} else {
EVUTIL_ASSERT(info->family == AF_INET ||
info->family == AF_INET6);
return; /* unreachable */
}
bufferevent_socket_connect(info->bev, sa, socklen);
_bufferevent_decref_and_unlock(info->bev);
memset(info, 0, sizeof(*info));
mm_free(info);
}
static void
connect_complete(struct event_overlapped *eo, uintptr_t key,
ev_ssize_t nbytes, int ok)
{
struct bufferevent_async *bev_a = upcast_overlapped(eo);
struct bufferevent *bev = &bev_a->bev.bev; /* XXX locking issue ? */
_bufferevent_incref_and_lock(bev);
EVUTIL_ASSERT(bev_a->bev.connecting);
bev_a->bev.connecting = 0;
_bufferevent_run_eventcb(bev,
ok? BEV_EVENT_CONNECTED : BEV_EVENT_ERROR);
_bufferevent_decref_and_unlock(bev);
}
static void
bev_async_consider_writing(struct bufferevent_async *beva)
{
size_t at_most;
int limit;
struct bufferevent *bev = &beva->bev.bev;
/* Don't write if there's a write in progress, or we do not
* want to write, or when there's nothing left to write. */
if (beva->write_in_progress || beva->bev.connecting)
return;
if (!beva->ok || !(bev->enabled&EV_WRITE) ||
!evbuffer_get_length(bev->output)) {
bev_async_del_write(beva);
return;
}
at_most = evbuffer_get_length(bev->output);
/* This is safe so long as bufferevent_get_write_max never returns
* more than INT_MAX. That's true for now. XXXX */
limit = (int)_bufferevent_get_write_max(&beva->bev);
if (at_most >= (size_t)limit && limit >= 0)
at_most = limit;
if (beva->bev.write_suspended) {
bev_async_del_write(beva);
return;
}
/* XXXX doesn't respect low-water mark very well. */
bufferevent_incref(bev);
if (evbuffer_launch_write(bev->output, at_most,
&beva->write_overlapped)) {
bufferevent_decref(bev);
beva->ok = 0;
_bufferevent_run_eventcb(bev, BEV_EVENT_ERROR);
} else {
beva->write_in_progress = at_most;
_bufferevent_decrement_write_buckets(&beva->bev, at_most);
bev_async_add_write(beva);
}
}
/* bufferevent 非阻塞连接 */
int
bufferevent_socket_connect(struct bufferevent *bev,
struct sockaddr *sa, int socklen)
{
struct bufferevent_private *bufev_p =
EVUTIL_UPCAST(bev, struct bufferevent_private, bev);
evutil_socket_t fd;
int r = 0;
int result=-1;
int ownfd = 0;
_bufferevent_incref_and_lock(bev);
if (!bufev_p)
goto done;
fd = bufferevent_getfd(bev);
//还没设置fd，设置fd。
if (fd < 0) {
if (!sa)
goto done;
fd = socket(sa->sa_family, SOCK_STREAM, 0);
if (fd < 0)
goto done;
if (evutil_make_socket_nonblocking(fd)<0)
goto done;
ownfd = 1;
}
if (sa) {
#ifdef WIN32
if (bufferevent_async_can_connect(bev)) {
bufferevent_setfd(bev, fd);
r = bufferevent_async_connect(bev, fd, sa, socklen);
if (r < 0)
goto freesock;
bufev_p->connecting = 1;
result = 0;
goto done;
} else
#endif
//非阻塞连接
r = evutil_socket_connect(&fd, sa, socklen);
if (r < 0)
goto freesock;
}
#ifdef WIN32
/* ConnectEx() isn't always around, even when IOCP is enabled.
* Here, we borrow the socket object's write handler to fall back
* on a non-blocking connect() when ConnectEx() is unavailable. */
if (BEV_IS_ASYNC(bev)) {
event_assign(&bev->ev_write, bev->ev_base, fd,
EV_WRITE|EV_PERSIST, bufferevent_writecb, bev);
}
#endif
bufferevent_setfd(bev, fd);
if (r == 0) { /* 正在连接，监听读事件。*/
if (! be_socket_enable(bev, EV_WRITE)) {
bufev_p->connecting = 1;
result = 0;
goto done;
}
} else if (r == 1) { /* 连接成功，激活读事件。*/
/* The connect succeeded already. How very BSD of it. */
result = 0;
bufev_p->connecting = 1;
event_active(&bev->ev_write, EV_WRITE, 1);
} else {
/* The connect failed already. How very BSD of it. */
bufev_p->connection_refused = 1;
bufev_p->connecting = 1;
result = 0;
event_active(&bev->ev_write, EV_WRITE, 1);
}
goto done;
freesock:
_bufferevent_run_eventcb(bev, BEV_EVENT_ERROR);
if (ownfd)
evutil_closesocket(fd);
/* do something about the error? */
done:
_bufferevent_decref_and_unlock(bev);
return result;
}
static void
bufferevent_writecb(evutil_socket_t fd, short event, void *arg)
{
struct bufferevent *bufev = arg;
struct bufferevent_private *bufev_p =
EVUTIL_UPCAST(bufev, struct bufferevent_private, bev);
int res = 0;
short what = BEV_EVENT_WRITING;
int connected = 0;
ev_ssize_t atmost = -1;
_bufferevent_incref_and_lock(bufev);
if (event == EV_TIMEOUT) {
/* Note that we only check for event==EV_TIMEOUT. If
* event==EV_TIMEOUT|EV_WRITE, we can safely ignore the
* timeout, since a read has occurred */
what |= BEV_EVENT_TIMEOUT;
goto error;
}
//正在连接服务器...
if (bufev_p->connecting) {
int c = evutil_socket_finished_connecting(fd);
/* we need to fake the error if the connection was refused
* immediately - usually connection to localhost on BSD */
if (bufev_p->connection_refused) { //在bufferevent_socket_connect中被设置
bufev_p->connection_refused = 0;
c = -1;
}
if (c == 0) //正在连接，继续监听可写。
goto done;
bufev_p->connecting = 0;
if (c < 0) { //连接发生错误
event_del(&bufev->ev_write);
event_del(&bufev->ev_read);
_bufferevent_run_eventcb(bufev, BEV_EVENT_ERROR);
goto done;
} else { //连接成功
connected = 1;
#ifdef WIN32
if (BEV_IS_ASYNC(bufev)) {
event_del(&bufev->ev_write);
bufferevent_async_set_connected(bufev);
_bufferevent_run_eventcb(bufev,
BEV_EVENT_CONNECTED);
goto done;
}
#endif
//连接成功，调用用户错误处理函数。比较奇怪。
_bufferevent_run_eventcb(bufev,
BEV_EVENT_CONNECTED);
if (!(bufev->enabled & EV_WRITE) ||
bufev_p->write_suspended) {
event_del(&bufev->ev_write);
goto done;
}
}
}
atmost = _bufferevent_get_write_max(bufev_p);
//写被挂起
if (bufev_p->write_suspended)
goto done;
//存在可写数据
if (evbuffer_get_length(bufev->output)) {
evbuffer_unfreeze(bufev->output, 1);
//将缓冲区数据写入socket。
res = evbuffer_write_atmost(bufev->output, fd, atmost);
evbuffer_freeze(bufev->output, 1);
if (res == -1) { //写发生错误
int err = evutil_socket_geterror(fd);
if (EVUTIL_ERR_RW_RETRIABLE(err)) //socket缓冲区满，一次未写完，继续监听可写事件，等待下次写入。
goto reschedule;
what |= BEV_EVENT_ERROR; //写发生异常错误
} else if (res == 0) {
/* eof case
XXXX Actually, a 0 on write doesn't indicate
an EOF. An ECONNRESET might be more typical.
*/
what |= BEV_EVENT_EOF;
}
if (res <= 0)
goto error;
_bufferevent_decrement_write_buckets(bufev_p, res);
}
//缓冲区数据已写完，删除写事件。
if (evbuffer_get_length(bufev->output) == 0) {
event_del(&bufev->ev_write);
}
/* 低于写低水位，调用用户回调。
*/
if ((res || !connected) &&
evbuffer_get_length(bufev->output) <= bufev->wm_write.low) {
_bufferevent_run_writecb(bufev);
}
goto done;
reschedule:
if (evbuffer_get_length(bufev->output) == 0) {
event_del(&bufev->ev_write);
}
goto done;
error:
bufferevent_disable(bufev, EV_WRITE);
_bufferevent_run_eventcb(bufev, what);
done:
_bufferevent_decref_and_unlock(bufev);
}
static void
bufferevent_readcb(evutil_socket_t fd, short event, void *arg)
{
struct bufferevent *bufev = arg;
struct bufferevent_private *bufev_p =
EVUTIL_UPCAST(bufev, struct bufferevent_private, bev);
struct evbuffer *input;
int res = 0;
short what = BEV_EVENT_READING;
ev_ssize_t howmuch = -1, readmax=-1;
_bufferevent_incref_and_lock(bufev);
if (event == EV_TIMEOUT) {
/* Note that we only check for event==EV_TIMEOUT. If
* event==EV_TIMEOUT|EV_READ, we can safely ignore the
* timeout, since a read has occurred */
what |= BEV_EVENT_TIMEOUT;
goto error;
}
input = bufev->input;
/*
* If we have a high watermark configured then we don't want to
* read more data than would make us reach the watermark.
*/
if (bufev->wm_read.high != 0) {
howmuch = bufev->wm_read.high - evbuffer_get_length(input);
/* 缓冲区超过高水位，挂起读。 */
if (howmuch <= 0) {
bufferevent_wm_suspend_read(bufev);
goto done;
}
}
//因为用户可以限速，所以这么要检测最大的可读大小。
//如果没有限速的话，那么将返回16384字节，即16K
//默认情况下是没有限速的。
readmax = _bufferevent_get_read_max(bufev_p);
if (howmuch < 0 || howmuch > readmax) /* The use of -1 for "unlimited"
* uglifies this code. XXXX */
howmuch = readmax;
if (bufev_p->read_suspended)
goto done;
evbuffer_unfreeze(input, 0);
res = evbuffer_read(input, fd, (int)howmuch); /* XXXX evbuffer_read would do better to take and return ev_ssize_t */
evbuffer_freeze(input, 0);
if (res == -1) {
int err = evutil_socket_geterror(fd);
if (EVUTIL_ERR_RW_RETRIABLE(err)) //EINTER or EAGAIN
goto reschedule;
/* error case */
what |= BEV_EVENT_ERROR;
} else if (res == 0) {
/* eof case */
what |= BEV_EVENT_EOF;
}
if (res <= 0)
goto error;
_bufferevent_decrement_read_buckets(bufev_p, res);
/* 数据大于低水平，调用用户设置的回调。 */
if (evbuffer_get_length(input) >= bufev->wm_read.low)
_bufferevent_run_readcb(bufev);
goto done;
reschedule:
goto done;
error:
bufferevent_disable(bufev, EV_READ);
_bufferevent_run_eventcb(bufev, what);
done:
_bufferevent_decref_and_unlock(bufev);
}
static void
bufferevent_writecb(evutil_socket_t fd, short event, void *arg)
{
struct bufferevent *bufev = arg;
struct bufferevent_private *bufev_p =
EVUTIL_UPCAST(bufev, struct bufferevent_private, bev);
int res = 0;
short what = BEV_EVENT_WRITING;
int connected = 0;
int atmost = -1;
_bufferevent_incref_and_lock(bufev);
if (event == EV_TIMEOUT) {
what |= BEV_EVENT_TIMEOUT;
goto error;
}
if (bufev_p->connecting) {
int c = evutil_socket_finished_connecting(fd);
/* we need to fake the error if the connection was refused
* immediately - usually connection to localhost on BSD */
if (bufev_p->connection_refused) {
bufev_p->connection_refused = 0;
c = -1;
}
if (c == 0)
goto done;
bufev_p->connecting = 0;
if (c < 0) {
event_del(&bufev->ev_write);
event_del(&bufev->ev_read);
_bufferevent_run_eventcb(bufev, BEV_EVENT_ERROR);
goto done;
} else {
connected = 1;
#ifdef WIN32
if (BEV_IS_ASYNC(bufev)) {
event_del(&bufev->ev_write);
bufferevent_async_set_connected(bufev);
_bufferevent_run_eventcb(bufev,
BEV_EVENT_CONNECTED);
goto done;
}
#endif
_bufferevent_run_eventcb(bufev,
BEV_EVENT_CONNECTED);
if (!(bufev->enabled & EV_WRITE) ||
bufev_p->write_suspended) {
event_del(&bufev->ev_write);
goto done;
}
}
}
atmost = _bufferevent_get_write_max(bufev_p);
if (bufev_p->write_suspended)
goto done;
if (evbuffer_get_length(bufev->output)) {
evbuffer_unfreeze(bufev->output, 1);
res = evbuffer_write_atmost(bufev->output, fd, atmost);
evbuffer_freeze(bufev->output, 1);
if (res == -1) {
int err = evutil_socket_geterror(fd);
if (EVUTIL_ERR_RW_RETRIABLE(err))
goto reschedule;
what |= BEV_EVENT_ERROR;
} else if (res == 0) {
/* eof case
XXXX Actually, a 0 on write doesn't indicate
an EOF. An ECONNRESET might be more typical.
*/
what |= BEV_EVENT_EOF;
}
if (res <= 0)
goto error;
_bufferevent_decrement_write_buckets(bufev_p, res);
}
if (evbuffer_get_length(bufev->output) == 0) {
event_del(&bufev->ev_write);
}
/*
* Invoke the user callback if our buffer is drained or below the
* low watermark.
*/
if ((res || !connected) &&
evbuffer_get_length(bufev->output) <= bufev->wm_write.low) {
_bufferevent_run_writecb(bufev);
}
goto done;
reschedule:
if (evbuffer_get_length(bufev->output) == 0) {
event_del(&bufev->ev_write);
}
goto done;
error:
bufferevent_disable(bufev, EV_WRITE);
_bufferevent_run_eventcb(bufev, what);
done:
_bufferevent_decref_and_unlock(bufev);
}
static void
bufferevent_readcb(evutil_socket_t fd, short event, void *arg)
{
struct bufferevent *bufev = arg;
struct bufferevent_private *bufev_p =
EVUTIL_UPCAST(bufev, struct bufferevent_private, bev);
struct evbuffer *input;
int res = 0;
short what = BEV_EVENT_READING;
int howmuch = -1, readmax=-1;
_bufferevent_incref_and_lock(bufev);
if (event == EV_TIMEOUT) {
what |= BEV_EVENT_TIMEOUT;
goto error;
}
input = bufev->input;
/*
* If we have a high watermark configured then we don't want to
* read more data than would make us reach the watermark.
*/
if (bufev->wm_read.high != 0) {
howmuch = bufev->wm_read.high - evbuffer_get_length(input);
/* we somehow lowered the watermark, stop reading */
if (howmuch <= 0) {
bufferevent_wm_suspend_read(bufev);
goto done;
}
}
readmax = _bufferevent_get_read_max(bufev_p);
if (howmuch < 0 || howmuch > readmax) /* The use of -1 for "unlimited"
* uglifies this code. */
howmuch = readmax;
if (bufev_p->read_suspended)
goto done;
evbuffer_unfreeze(input, 0);
res = evbuffer_read(input, fd, howmuch);
evbuffer_freeze(input, 0);
if (res == -1) {
int err = evutil_socket_geterror(fd);
if (EVUTIL_ERR_RW_RETRIABLE(err))
goto reschedule;
/* error case */
what |= BEV_EVENT_ERROR;
} else if (res == 0) {
/* eof case */
what |= BEV_EVENT_EOF;
}
if (res <= 0)
goto error;
_bufferevent_decrement_read_buckets(bufev_p, res);
/* Invoke the user callback - must always be called last */
if (evbuffer_get_length(input) >= bufev->wm_read.low)
_bufferevent_run_readcb(bufev);
goto done;
reschedule:
goto done;
error:
bufferevent_disable(bufev, EV_READ);
_bufferevent_run_eventcb(bufev, what);
done:
_bufferevent_decref_and_unlock(bufev);
}
