spsoft / spserver Goto Github PK

你好，我下载了spserver-0.9.5.src.tar.gz，运行testiocpecho这个示例�
��序，然后使用客户端程序和示例程序建立连接后客户端发出
消息，结果服务端（testiocpecho）程序未执行到
virtual int handle( SP_Request * request, SP_Response * response ) 
{//…………}
请问是哪出了问题？

Hi there,

I was trying to compile "sptunnel" using VS2010 I get an error when something 
is dependent on "spopenssl" (wasn't possible to compile due to the lack of 
spopenssl.lib )

But there's no MSVC project for "spopenssl" only linux g++ makefile.

Can you provide it? Or where can I find it?

Thanks in advance
Ricardo
Portugal

                memcpy(buff+len,fcr,sizeof(FileConnect_rsp));
        SP_Message * msg = new SP_Message();
        msg->getMsg()->append(buff,buffsize);
        /*SP_Buffer* spbuf=new SP_Buffer();
        spbuf->reset();
        spbuf->append(buff,buffsize);*/
        //response->getFromSid()->
        //response->getReply()->getMsg()->append(_message);
        response->addMessage(msg);

What is the expected output? What do you see instead?
在addMessage()或getMsg()->append()时出现
in yd_server.exe: 0xC0000005: Access violation reading location 0x00000208.
跟踪时,sputils中的append中mCount和nMaxCount均未分配
if( mCount >= mMaxCount ) {
        mMaxCount = ( mMaxCount * 3 ) / 2 + 1;
        mFirst = (void**)realloc( mFirst, sizeof( void * ) * mMaxCount );
        assert( NULL != mFirst );
        memset( mFirst + mCount, 0, ( mMaxCount - mCount ) * sizeof( void * ) );
    }
What version of the product are you using? On what operating system?


Please provide any additional information below.

Hello.

I'm writing simple http-server which processes some signals such as 
SIGUSRx, and using SP_LFServer for it (like in testhttp.cpp). But I've 
found that my server shuts down on the next request after it received 
SIGUSRx signal. After some investigation I've found that SP_LFServer uses 
pause() in SP_LFServer::runForever(), which "causes the calling process (or 
thread) to sleep until a signal is delivered that either terminates the 
process or causes the invocation of a signal-catching function", which is 
my case. Isn't is more convenient to use sigwait(), specifying SIGKILL and 
SIGTERM for it, than simple pause()? Below is the code I'm using and it 
seems it works for me.

void SP_LFServer :: runForever()
{
    run();
    sigset_t ss;
    sigemptyset(&ss);
    sigaddset(&ss, SIGKILL);
    sigaddset(&ss, SIGTERM);
    int signum;
    sigwait(&ss, &signum);
    struct sigaction sa;
    sigaction(signum, 0, &sa);
    sa.sa_handler(signum);
    //pause();
}

testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0
testhttp[20478]: session(1024.14) read error, errno 11, status 0

What steps will reproduce the problem?
1. create sp_httpserver using the new method in main thread
2. Start the sp_httpserver in an another thread user method run
3. close the sp_hrrpserver in main thread

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?
windows XP ver spserver-0.9.2

Please provide any additional information below.

I have tested the sample program but there is no memory leak!
thank you!

spserver-read-only> make
make[1]: Entering directory
`/home/taf2/project/google-mongrel-esi/trunk/pesi/spserver-read-only/spserver'
gcc -Wall -D_REENTRANT -D_GNU_SOURCE -g -fPIC -I/home/taf2/libevent/ -c
sputils.cpp -o sputils.o
gcc -Wall -D_REENTRANT -D_GNU_SOURCE -g -fPIC -I/home/taf2/libevent/ -c
spioutils.cpp -o spioutils.o
gcc -Wall -D_REENTRANT -D_GNU_SOURCE -g -fPIC -I/home/taf2/libevent/ -c
spiochannel.cpp -o spiochannel.o
spiochannel.cpp:19:20: error: config.h: No such file or directory
make[1]: *** [spiochannel.o] Error 1
make[1]: Leaving directory
`/home/taf2/project/google-mongrel-esi/trunk/pesi/spserver-read-only/spserver'
make: *** [all] Error 2

What steps will reproduce the problem?
1.用ab进行压力测试. ab -c 100 -n 1000 http://192.168.13.38:8080/
2.调用 shutdown()关机
3.产生泄露
{91756} normal block at 0x00AC9428, 16 bytes long.
 Data: <0 B  RB         > 30 FF 42 00 90 52 42 00 00 00 00 00 01 00 00 00 
{642} normal block at 0x0155CDD8, 24 bytes long.
 Data: <             RB > E8 13 00 00 20 06 00 00 10 06 00 00 90 52 42 00 

What is the expected output? What do you see instead?
泄露只是在多用户情况下出现,单线程时没有问题.另外有时候
,服务不能退出.阻塞在
SP_IocpServer :: ~SP_IocpServer()
{
...
    for( ; mIsRunning; ) sleep( 1 );
}
What version of the product are you using? On what operating system?

ver 0.9.4

Please provide any additional information below.

前面已经有人提到过主动发送的问题，有人给出了一个连接��
�析，但是现在打不开该连接，希望能贴出来，还希望能给出�
��方的接口，好方便以后的跟进更新。

libevent 在 iPhone 上编译不通过，是不是 spserver 
完全没可能使用？

iPhone 是基于 darwin， cpu 是 arm， gcc 是 4.2.1

调用了 SP_IocpLFServer.shutdown() 
之后，客户端依然可以连接成功，但是不会再接受数据，再��
�
run时会出错，因为端口没有被释放。

是我调用方法不对还是BUG？

俺在开发预期的是 SP_IocpLFServer.shutdown() 
之后，接受未完成的数据接收，然后断开所有连
接，停止accept新连接，关闭socket，直到下次执行start。

1.Is it possible that we want to manage other events(timer or fd event) by 
the spserver architecture at the same time?
The spserver just accept a socket from far-end now.
Could we added a tunnel or method to push some events into eventbase of 
spserver?

2.The SP_MsgDecoder just simply decode the message. I think some scenarios 
that maybe decode the message according to the status of SP_Session.
The SP_MsgDecoder::decode(SP_Buffer *inBuffer) is limited. I think that 
something need to changes like as SP_MsgDecoder::decode( SP_Session 
*session ) and to give programmer a factory pattern to generate the custom 
SP_Session(to replace SP_Session * session = new SP_Session( sid )
 at SP_EventCallback :: onAccept()).

Best Regards and Thank you.
Person

请教spiocpserver如何"主动"发送消息给已连接的client? 

谢谢帮忙

What steps will reproduce the problem?
1. 在主线程中创建Sp_httpServer
2. 在子线程中启动
3. 在主线程中启动，等待子线程关闭退出，删除Sp_httpServer
出现内存泄露，应该是 httpResponse 未清楚

code：
DWORD  WINAPI StartServer(LPVOID para)
{
    SP_IocpServer* pServer = (SP_IocpServer*)para;
    pServer->run();
    while (pServer->isRunning())
    {
        Sleep(1000);
    }
    return 0;
}

void CHostListView::OnStartlisten() 
{
    int port = 80, maxThreads = 50;
    const char * serverType = "hahs";

    if( 0 != sp_initsock() ) ASSERT( 0 );
    m_pIocpServer = 
        new SP_IocpServer( "", port, new SP_HttpHandlerAdapterFactory( new
CHttpServerHandlerFactory(this->m_hWnd) ) );
    m_pIocpServer->setTimeout( 60 );
    m_pIocpServer->setMaxThreads( maxThreads );
    m_pIocpServer->setReqQueueSize( 100, "HTTP/1.1 500 Sorry, server is busy
now!\r\n" );
    DWORD dwThreadId1;
    m_hStartServerThread = CreateThread( NULL, 0, StartServer,
(LPVOID)(m_pIocpServer), 0, &dwThreadId1 );
    sp_closelog();
}
关闭退出
    if (m_pIocpServer)
    {
        if (m_pIocpServer->isRunning())
        {
            m_pIocpServer->shutdown();
        }
        WaitForSingleObject(m_hStartServerThread,INFINITE);
        delete m_pIocpServer;
    }
What is the expected output? What do you see instead?

内存泄露
What version of the product are you using? On what operating system?

spserver-0.9.2
Please provide any additional information below.

谢谢

What steps will reproduce the problem?
1. 客户端通过国外的代理服务器链接spserver
2. 在服务端执行SP_IocpEventCallback :: eventLoop函数中的
GetQueuedCompletionStatus时返回错误，提示121，信号量超时
3.

What is the expected output? What do you see instead?
不知道是什么原因？请指教，谢谢！

What version of the product are you using? On what operating system?


Please provide any additional information below.

下载了你的SPSERVER测试,发现使用很方便,但有一个问题我比较�
��惑,在testiocphttp
里,怎么断开一个客户端连接呢.在例子里好象是keep-alive,这样�
��些恶意的连接不断开
怎么办?我还不是很清楚这个过程,望不吝赐教.3X

[deleted issue]

Hi,

XySSL is no longer maintained.

PolarSSL (located at http://www.polarssl.org) is the official fork of the
project. For best compliance it is advised to upgrade accordingly.

If you need any help, don't hesitate to ask me.

Best regards,
Paul

I'm using spmemvfs on windows too, but i had to change some code to make it
build and work in multi thread environment. Add this on top of spmemvfs.c:

[code]
#if ! defined(_WIN32)
 #include <pthread.h>
 #define spmem_critical_section_t         pthread_mutex_t
 #define spmem_init_critical_section      pthread_mutex_init
 #define spmem_destroy_critical_section   pthread_mutex_destroy
 #define spmem_enter_critical_section     pthread_mutex_lock
 #define spmem_exit_critical_section      pthread_mutex_unlock
 #define spmem_init_default_arg           NULL
#else
 #include <windows.h>
 #define __func__ __FUNCTION__
 #define spmem_critical_section           CRITICAL_SECTION
 #define spmem_init_critical_section      InitializeCriticalSectionAndSpinCount
 #define spmem_destroy_critical_section   DeleteCriticalSection
 #define spmem_enter_critical_section     EnterCriticalSection
 #define spmem_exit_critical_section      LeaveCriticalSection
 #define spmem_init_default_arg           0x80000400
#endif
[/code]

and replace every instance of pthread_mutex_* with spmem sersions.

What steps will reproduce the problem?
Client timout ,server Memory leak
fixed:
void SP_IocpEventCallback :: onResponse( void * queueData, void * arg )

Line:419
Add :
//Modify By wang 2010-04-15
sidList->take( i );
msg->getFailure()->add( sid );

If you are not sure whether this is a bug, please go to
http://groups.google.com/group/spserver to request help.

What steps will reproduce the problem?
1.client ./teststress -h 192.168.21.188 -c 30000 -m 5
2.server ./testecho
3.

What is the expected output? What do you see instead?
正常处理后，客户端会统计传递消息个数和网络读写的平均��
�度。
服务端异常输出如下：
session(24638.0) error, r 0(0), w 52(0), i 0, o 0, s 640(28032)

 error, r 0(0), w 52(0), i 0, o 0, s 0(28672)testecho[31802]: session(28993.0) invalid, unknown FROM


What version of the product are you using? On what operating system?


Please provide any additional information below.

What steps will reproduce the problem?
1. upgrade vc6 workspace to vs2008
2. build testiocpecho.exe
3. telnet to localhost 3333, type some text and quit, again and again

What is the expected output? What do you see instead?
VS2008 said:
HEAP[testiocpecho.exe]: HEAP: Free Heap block 3f6b68 modified at 3f6d14
after it was freed

What version of the product are you using? On what operating system?
latest svn source, on WinXP SP3 with VS2008

Please provide any additional information below.

What steps will reproduce the problem?
1. Compiling 0.6
2.
3.

What is the expected output? What do you see instead?

It should compile, but complains of a missing config.h instead. I just 
created it empty and it seems to work

What version of the product are you using? On what operating system?


Please provide any additional information below.

我在阅读代码时，发现多个线程都会调用eventArg->getEventHeap()��
�获取SP_IocpEventHeap并进行操作，这样做不会影响数据的完整性
吗？

在测试testecho时，发现SP_EchoHandler::start()中添加了response reply 
message，
经过多次调用后，该message被放入了session的outlist中。但是在��
�码中没有发现write事件
触发过程。但是最后message还是被发送到客户。
请问这是怎么回事？

SPServer have handle leak

用下载的代码反复测试,在windows下只能连接3962次,下一次就提�
��失败.echo server
设定的最大连接数是2万,还差的远呢.

2520#2011-07-28 14:57:08 session(5120.8329) invalid, unknown FROM
2520#2011-07-28 14:57:09 System busy, session.count 4097 [4096], queue.length 0 
[10]
2520#2011-07-28 14:57:09 session(5120.8330) write error, errno 10038, status 2, 
count 1
2520#2011-07-28 14:57:09 session(5120.8330) error, r 0(0), w 0(0), i 0, o 1, s 
4097(1023), t 4096

使用spserver做了一个服务器,原来运行几天后客户说报错,后来�
��现是这样错误,以为是spserver自带的连接数比较小,就调大了��
�了,想不到只是延长了错误出现的时间而已

很高兴看到国人也在做一些开源的东西，而且在不断的改进��
�中，但是，我想提一点：
1、整个源代码中没有注释，令人看起来不是很方便，我想这�
��是为什么**的软件不能做大做强
的原因所在吧。
2、整个项目没有阐述基本原理的一些东西，比如网页的、chm�
��，我看其他一些开源软件的时
候，基本上都会有这样的内容，函数说明呀，什么的。

就提这么2点，以之鼓励。

What steps will reproduce the problem?
1. 初始化AF_UNIX流fds[0], fds[1];
2. 
使用libevent接口，为fds[0]，注册EV_READ和EV_WRITE事件callback函数e
vread，
evwrite，其中evread将数据从fd中读取出来，evwrite不做任何与fd��
�关的事情。
3. 向fds[1]中write字符串"54321".
4. 连续调用libevent接口event_base_loop( mEventBase, EVLOOP_ONCE )三次。

问题：
为什么 write 事件的callback函数evwrite被调用了三次。

When you call SP_IocpServer::shutdown the clients don't receive FIN or RST 
packets. Therefore the connection remains open at the clients' end eventhough 
the server does not accept incoming packets anymore. You should see a FIN or 
RST packet sent to clients so that they can close the connections properly.

To reproduce:
1. Call SP_IocpServer::shutdown
2. Look at network log (from Wireshark, for example) and note that no FIN or 
RST packets were returned to clients.

I'm using version 0.95 of spserver. My operating system is Windows 7. My build 
environment is MinGW's gcc toolchain. I had to make a custom makefile for the 
project for MinGW. I can see from the debugger that closesocket (sp_close 
macro) is called when shutdown is called, but it doesn't result in FIN or RST 
packet. So this may actually be a problem in Windows itself (?).

Does this occur with other build environments?

stephen，您好！

我是一名学生，最近想用spserver iocphttp做个web server，
virtual void handle( SP_HttpRequest * request, SP_HttpResponse * response )
这个怎么改才能实现异步读文件，在发送数据呢？

还是需要另外写？代码包装的很深，不知道从哪做起？请指��
�下！谢谢

libevent-1.4.7-stable-iocp-1.zip和libevent-main-iocp-1.zip有什么区别啊

例子里的是字符串数据,如果发送文件如何做?

If you are not sure whether this is a bug, please go to
http://groups.google.com/group/spserver to request help.

What steps will reproduce the problem?
1.用ab   -c 5  -n 5000000 
"http://192.168.7.133:8080/?op=test&key=cshb252&value=benchmark_test"测试工��
�测试使用spserver写的icop的web服务器，内存不断的增长，增长�
��去后就一直不能降下来，如果一直运行下去，内存被耗尽。

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?
spsserver 0.9.5， xp sp3

Please provide any additional information below.

What steps will reproduce the problem?
1. Download spserver-0.8.5.src.tar.gz
2. Follow installation/compilation instructions in spserver/README
3. View compilation error after typing 'make'

What is the expected output? What do you see instead?
I expected the project to compile, but arrived at the following compilation
error:
avacado:~/software/srelos/spserver-0.8.5/spserver inajam$ make
gcc -Wall -D_REENTRANT -D_GNU_SOURCE -g -fPIC
-I/Users/inajam/software/srelos/libevent-1.3e/build/include -c
spioutils.cpp -o spioutils.o
spioutils.cpp: In static member function 'static void
SP_IOUtils::inetNtoa(in_addr*, char*, int)':
spioutils.cpp:25: error: 'struct in_addr' has no member named 's_net'
spioutils.cpp:25: error: 'struct in_addr' has no member named 's_host'
spioutils.cpp:25: error: 'struct in_addr' has no member named 's_lh'
spioutils.cpp:25: error: 'struct in_addr' has no member named 's_impno'
make: *** [spioutils.o] Error 1

What version of the product are you using? On what operating system?
I'm using spserver v0.8.5, and Mac OS X v10.4.10

Please provide any additional information below.
I looked @ spioutils.cpp and tried to locate what definition of struct
in_addr it was using.  It appears (as best I can find) that it's accessing
the in_addr defined on OS X in /usr/include/netinet/in.h, which looks like
this:

/*
 * Internet address (a structure for historical reasons)
 */
struct in_addr {
        in_addr_t s_addr;
};

This would explain why compiling spserver on OS X fails as mentioned above.

spserver is actually exactly what I'm looking for to embed an HTTP server
in an existing c++ component.  I'll try to get access to a Linux machine to
see if compilation behaves differently there.

如题,非SP_IocpDispatcher模式

最近在开发一个embedded http server
硬件cpu 4*4一共16核, 16g内存，rhel 5.1 64位
测试工具loadrunner
以下测试结果为每秒点击数

libevent 11500 (cpu 90%)
spserver hahs 20 (cpu 1%)
spserver lf   20 (cpu 1%)
asio example http 8000 (cpu 320%)
asio example http2 15000 (cpu 250%)
asio example http3 15000 (cpu 90%)

各个程序的功能都一样
spserver低得有点离谱，不知道是怎么回事
    virtual void handle( SP_HttpRequest * request, SP_HttpResponse * response ) {

        string uri = request->getURI();



        if (uri.find("crossdomain.xml") != string::npos)
        {
            response->setStatusCode( 200 );

            string res = "<cross-domain-policy><allow-access-from domain=\"*\"
to-ports=\"*\" /></cross-domain-policy>";

            response->addHeader("Content-Type", "application/xml");
            response->addHeader("Content-Length",
(boost::lexical_cast<std::string>(res.size())).c_str());


            response->appendContent(res.c_str());

        }



另外一段代码
    int port = 9001, maxThreads = 10, maxConnections = 20000;
    int timeout = 60, reqQueueSize = 20000;

//      sp_openlog( "testhttp", LOG_CONS | LOG_PID | LOG_PERROR, LOG_USER );
        assert( 0 == sp_initsock() );

        SP_LFServer server( "0.0.0.0", port, new SP_HttpHandlerAdapterFactory(
new SP_HttpEchoHandlerFactory() ) );
        //SP_Server server( "0.0.0.0", port, new SP_HttpHandlerAdapterFactory(
new SP_HttpEchoHandlerFactory() ) );

        server.setTimeout( timeout );
        server.setMaxThreads( maxThreads );
        server.setReqQueueSize( reqQueueSize, "HTTP/1.1 500 Sorry, server is busy
now!\r\n" );
server.setMaxConnections( maxConnections );

        server.runForever();

多一些文档spserver就无敌了

数据的发送最终是通过
int spwin32_writev(SOCKET sock, const struct iovec *vector, DWORD count)

{
  DWORD sent = -1;
  WSASend(sock, (LPWSABUF)vector, count, &sent, 0, NULL, NULL);
    return sent;

}
来完成的，WSASend的后面两个参数为NULL，所以发送投递是不会
被GQCS出列的.不被
GQCS出列那么怎么知道实际完成的字节数呢？
那么是不是存在这样一种情况：WSASend产生的发送投递本来是�
��发送100个字节，但是
实际却只发送了80字节？
据我个人的理解，sent中只是返回WSASend往内核的完成队列中投
递的字节数，并不表示
真正发送完成的字节数。
这样不是存在完成字节数与实际投递数不符的情况吗

0.9.5的版本。

SP_IocpEventHelper :: timeoutNext 有以下代码：

        DWORD ret = ( first->tv_sec - curr.tv_sec ) * 1000
        + ( first->tv_usec - curr.tv_usec ) / 1000;

    if( ret < 0 ) ret = 0;

    return ret;

ret 永远都会是>=0 的。

gcc -Wall -D_REENTRANT -D_GNU_SOURCE -g -fPIC -I/usr/local/include/ -c
testthreadpool.cpp -o testthreadpool.o
testthreadpool.cpp: In function ¡®void threadFunc(void*)¡¯:
testthreadpool.cpp:21: 错误从'void *' 到 ' int' 的转换损失精度
make: *** [testthreadpool.o] ´íÎó 1

编译什么的都通过了，生成了可执行文件，比如说，测试使��
�的testecho，我使用ldd查看需要的动态库也都OK 但是启动时候 
提示undefined symbol: _ZTI17SP_HandlerFactory
  一般怎么回事？ 不知道如何去查？

What steps will reproduce the problem?
1. gcc -Wall -D_REENTRANT -D_GNU_SOURCE -g -fPIC -I/Users/eugene/libevent/ -c 
spiochannel.cpp -o spiochannel.o
spiochannel.cpp: In member function 'virtual int 
SP_DefaultIOChannel::write_vec(iovec*, int)':
spiochannel.cpp:158: error: 'writev' was not declared in this scope

What is the expected output? What do you see instead?
self explanatory

What version of the product are you using? On what operating system?
spserver 0.9.2
Mac OS X 10.4.11

Please provide any additional information below.

Had the same issue as in "3", applied the patch.

Thanks

旧版本中speventcb.cpp的onAccept函数中，
event_set( session->getReadEvent(), clientFD, EV_READ, onRead, session );
调用后，紧接着又调用了AddEvent(),而AddEvent()中又调用了event_se
t,这里是特意这样调
用的吗，还是我对libevent的函数使用没有理解对？

而在新版本中speventcb.cpp的onAccept函数中，
event_set( session->getReadEvent(), clientFD, EV_READ, onRead, session );

这里注册了回调函数，但没看到event_add的调用，同旧版本完��
�不同了，去掉了AddEvent的调
用，是有特别的用意吗

望不吝赐教,谢谢

test

If you are not sure whether this is a bug, please go to
http://groups.google.com/group/spserver to request help.

What steps will reproduce the problem?
1. run testecho(change sp_session uint16_t to unsigned int).
2. use a multi-thread client in other client to do connect, send,recv (don't 
close fd until program exit.
3. about 100,000 connections, ubuntu looks very slowly, and killed testecho.

I use top. there is 700M memory free(Total 2G), and CPU only takes 30%. 

What is the expected output? What do you see instead?
   work normally.
   testecho was killed by system.

What version of the product are you using? On what operating system?
0.9.5 under ubuntu 10.10

Please provide any additional information below.
Jul 18 21:00:24 opensips kernel: [10762.781159] Mem-Info:
Jul 18 21:00:24 opensips kernel: [10762.781160] DMA per-cpu:
Jul 18 21:00:24 opensips kernel: [10762.781162] CPU    0: hi:    0, btch:   1 
usd:   0
Jul 18 21:00:24 opensips kernel: [10762.781164] CPU    1: hi:    0, btch:   1 
usd:   0
Jul 18 21:00:24 opensips kernel: [10762.781166] Normal per-cpu:
Jul 18 21:00:24 opensips kernel: [10762.781168] CPU    0: hi:  186, btch:  31 
usd:  40
Jul 18 21:00:24 opensips kernel: [10762.781170] CPU    1: hi:  186, btch:  31 
usd:  30
Jul 18 21:00:24 opensips kernel: [10762.781171] HighMem per-cpu:
Jul 18 21:00:24 opensips kernel: [10762.781173] CPU    0: hi:  186, btch:  31 
usd: 155
Jul 18 21:00:24 opensips kernel: [10762.781175] CPU    1: hi:  186, btch:  31 
usd:  21
Jul 18 21:00:24 opensips kernel: [10762.781178] active_anon:46948 
inactive_anon:16673 isolated_anon:0
Jul 18 21:00:24 opensips kernel: [10762.781180]  active_file:11359 
inactive_file:20878 isolated_file:0
Jul 18 21:00:24 opensips kernel: [10762.781181]  unevictable:0 dirty:14 
writeback:0 unstable:0
Jul 18 21:00:24 opensips kernel: [10762.781182]  free:190739 
slab_reclaimable:19024 slab_unreclaimable:57150
Jul 18 21:00:24 opensips kernel: [10762.781183]  mapped:13250 shmem:9981 
pagetables:1152 bounce:0
Jul 18 21:00:24 opensips kernel: [10762.781188] DMA free:3524kB min:64kB 
low:80kB high:96kB active_anon:0kB inactive_anon:0kB active_file:0kB 
inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB 
present:15808kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB 
slab_reclaimable:356kB slab_unreclaimable:1084kB kernel_stack:0kB 
pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 
all_unreclaimable? yes
Jul 18 21:00:24 opensips kernel: [10762.781194] lowmem_reserve[]: 0 865 1980 
1980
Jul 18 21:00:24 opensips kernel: [10762.781201] Normal free:3664kB min:3728kB 
low:4660kB high:5592kB active_anon:0kB inactive_anon:0kB active_file:220kB 
inactive_file:40kB unevictable:0kB isolated(anon):0kB isolated(file):0kB 
present:885944kB mlocked:0kB dirty:0kB writeback:0kB mapped:4kB shmem:0kB 
slab_reclaimable:75740kB slab_unreclaimable:227516kB kernel_stack:2456kB 
pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:450 
all_unreclaimable? yes
Jul 18 21:00:24 opensips kernel: [10762.781207] lowmem_reserve[]: 0 0 8922 8922
Jul 18 21:00:24 opensips kernel: [10762.781214] HighMem free:755768kB min:512kB 
low:1712kB high:2912kB active_anon:187792kB inactive_anon:66692kB 
active_file:45216kB inactive_file:83472kB unevictable:0kB isolated(anon):0kB 
isolated(file):0kB present:1142020kB mlocked:0kB dirty:56kB writeback:0kB 
mapped:52996kB shmem:39924kB slab_reclaimable:0kB slab_unreclaimable:0kB 
kernel_stack:0kB pagetables:4608kB unstable:0kB bounce:0kB writeback_tmp:0kB 
pages_scanned:0 all_unreclaimable? no
Jul 18 21:00:24 opensips kernel: [10762.781220] lowmem_reserve[]: 0 0 0 0
Jul 18 21:00:24 opensips kernel: [10762.781223] DMA: 5*4kB 4*8kB 5*16kB 10*32kB 
14*64kB 17*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 3524kB
Jul 18 21:00:24 opensips kernel: [10762.781232] Normal: 228*4kB 66*8kB 5*16kB 
1*32kB 1*64kB 0*128kB 2*256kB 1*512kB 1*1024kB 0*2048kB 0*4096kB = 3664kB
Jul 18 21:00:24 opensips kernel: [10762.781239] HighMem: 564*4kB 637*8kB 
784*16kB 488*32kB 364*64kB 101*128kB 40*256kB 14*512kB 5*1024kB 3*2048kB 
160*4096kB = 755768kB
Jul 18 21:00:24 opensips kernel: [10762.781248] 42223 total pagecache pages
Jul 18 21:00:24 opensips kernel: [10762.781249] 0 pages in swap cache
Jul 18 21:00:24 opensips kernel: [10762.781251] Swap cache stats: add 0, delete 
0, find 0/0
Jul 18 21:00:24 opensips kernel: [10762.781253] Free swap  = 6141948kB
Jul 18 21:00:24 opensips kernel: [10762.781254] Total swap = 6141948kB
Jul 18 21:00:24 opensips kernel: [10762.781254] Total swap = 6141948kB
Jul 18 21:00:24 opensips kernel: [10762.784591] 515079 pages RAM
Jul 18 21:00:24 opensips kernel: [10762.784593] 287754 pages HighMem
Jul 18 21:00:24 opensips kernel: [10762.784595] 9031 pages reserved
Jul 18 21:00:24 opensips kernel: [10762.784596] 82449 pages shared
Jul 18 21:00:24 opensips kernel: [10762.784597] 294015 pages non-shared
Jul 18 21:00:24 opensips kernel: [10762.784600] Out of memory: kill process 
1369 (gnome-session) score 65119 or a child
Jul 18 21:00:24 opensips kernel: [10762.784604] Killed process 1521 
(gnome-panel) vsz:86088kB, anon-rss:3644kB, file-rss:13840kB
Jul 18 21:00:24 opensips kernel: [10762.855130] Xorg invoked oom-killer: 
gfp_mask=0xd0, order=0, oom_adj=0
Jul 18 21:00:24 opensips kernel: [10762.855134] Xorg cpuset=/ mems_allowed=0
Jul 18 21:00:24 opensips kernel: [10762.855137] Pid: 1049, comm: Xorg Not 
tainted 2.6.35-30-generic #54-Ubuntu
Jul 18 21:00:24 opensips kernel: [10762.855139] Call Trace:
Jul 18 21:00:24 opensips kernel: [10762.855146]  [<c01dd51a>] 
dump_header+0x7a/0xb0
Jul 18 21:00:24 opensips kernel: [10762.855149]  [<c01dd5ac>] 
oom_kill_process+0x5c/0x160
Jul 18 21:00:24 opensips kernel: [10762.855151]  [<c01ddb19>] ? 
select_bad_process+0xa9/0xe0
Jul 18 21:00:24 opensips kernel: [10762.855154]  [<c01ddba1>] 
__out_of_memory+0x51/0xb0
Jul 18 21:00:24 opensips kernel: [10762.855156]  [<c01ddc58>] 
out_of_memory+0x58/0xd0
Jul 18 21:00:24 opensips kernel: [10762.855159]  [<c01e0b86>] 
__alloc_pages_slowpath+0x496/0x4b0
Jul 18 21:00:24 opensips kernel: [10762.855162]  [<c01e0d0f>] 
__alloc_pages_nodemask+0x16f/0x1c0
Jul 18 21:00:24 opensips kernel: [10762.855164]  [<c01e0d7c>] 
__get_free_pages+0x1c/0x30
Jul 18 21:00:24 opensips kernel: [10762.855167]  [<c02299c1>] 
__pollwait+0xa1/0xe0
Jul 18 21:00:24 opensips kernel: [10762.855169]  [<c0229985>] ? 
__pollwait+0x65/0xe0
Jul 18 21:00:24 opensips kernel: [10762.855173]  [<c0571b1c>] 
unix_poll+0x1c/0xa0
Jul 18 21:00:24 opensips kernel: [10762.855176]  [<c04e72f4>] 
sock_poll+0x14/0x20
Jul 18 21:00:24 opensips kernel: [10762.855178]  [<c022962b>] 
do_select+0x37b/0x670
Jul 18 21:00:24 opensips kernel: [10762.855181]  [<c03535dd>] ? 
kobject_put+0x1d/0x50
Jul 18 21:00:24 opensips kernel: [10762.855183]  [<c0229920>] ? 
__pollwait+0x0/0xe0
Jul 18 21:00:24 opensips kernel: [10762.855185]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855188]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855190]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855192]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855194]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855196]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855198]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855200]  [<c0229a00>] ? 
pollwake+0x0/0x60
Jul 18 21:00:24 opensips kernel: [10762.855202]  [<c0229a00>] ? 
pollwake+0x0/0x60



Jul 18 21:00:24 opensips kernel: [10762.855223]  [<c0229fb0>] 
core_sys_select+0x140/0x240
Jul 18 21:00:24 opensips kernel: [10762.855226]  [<c016b2da>] ? 
hrtimer_try_to_cancel+0x3a/0xc0
Jul 18 21:00:24 opensips kernel: [10762.855229]  [<c0227bb2>] ? 
vfs_ioctl+0x32/0xb0
Jul 18 21:00:24 opensips kernel: [10762.855231]  [<c022a2b1>] 
sys_select+0x31/0xc0
Jul 18 21:00:24 opensips kernel: [10762.855233]  [<c0165660>] ? 
sys_clock_gettime+0x50/0xa0
Jul 18 21:00:24 opensips kernel: [10762.855236]  [<c05cc254>] 
syscall_call+0x7/0xb
Jul 18 21:00:24 opensips kernel: [10762.855239]  [<c05c0000>] ? 
pcibios_setup+0x93/0x3ac
Jul 18 21:00:24 opensips kernel: [10762.855241] Mem-Info:
Jul 18 21:00:24 opensips kernel: [10762.855242] DMA per-cpu:
Jul 18 21:00:24 opensips kernel: [10762.855244] CPU    0: hi:    0, btch:   1 
usd:   0
Jul 18 21:00:24 opensips kernel: [10762.855245] CPU    1: hi:    0, btch:   1 
usd:   0
Jul 18 21:00:24 opensips kernel: [10762.855247] Normal per-cpu:
Jul 18 21:00:24 opensips kernel: [10762.855248] CPU    0: hi:  186, btch:  31 
usd:  31
Jul 18 21:00:24 opensips kernel: [10762.855250] CPU    1: hi:  186, btch:  31 
usd:  30
Jul 18 21:00:24 opensips kernel: [10762.855251] HighMem per-cpu:
Jul 18 21:00:24 opensips kernel: [10762.855252] CPU    0: hi:  186, btch:  31 
usd: 175
Jul 18 21:00:24 opensips kernel: [10762.855254] CPU    1: hi:  186, btch:  31 
usd:   0
Jul 18 21:00:24 opensips kernel: [10762.855257] active_anon:42267 
inactive_anon:16487 isolated_anon:0
Jul 18 21:00:24 opensips kernel: [10762.855258]  active_file:12010 
inactive_file:18540 isolated_file:33
Jul 18 21:00:24 opensips kernel: [10762.855259]  unevictable:0 dirty:14 
writeback:0 unstable:0
Jul 18 21:00:24 opensips kernel: [10762.855259]  free:197553 
slab_reclaimable:19004 slab_unreclaimable:57150
Jul 18 21:00:24 opensips kernel: [10762.855260]  mapped:12134 shmem:9764 
pagetables:935 bounce:0
Jul 18 21:00:24 opensips kernel: [10762.855265] DMA free:3524kB min:64kB 
low:80kB high:96kB active_anon:0kB inactive_anon:0kB active_file:0kB 
inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB 
present:15808kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB 
slab_reclaimable:356kB slab_unreclaimable:1084kB kernel_stack:0kB 
pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 
all_unreclaimable? no
Jul 18 21:00:24 opensips kernel: [10762.855268] lowmem_reserve[]: 0 865 1980 
1980
Jul 18 21:00:24 opensips kernel: [10762.855274] Normal free:3684kB min:3728kB 
low:4660kB high:5592kB active_anon:0kB inactive_anon:0kB active_file:96kB 
inactive_file:236kB unevictable:0kB isolated(anon):0kB isolated(file):132kB 
present:885944kB mlocked:0kB dirty:0kB writeback:0kB mapped:4kB shmem:0kB 
slab_reclaimable:75660kB slab_unreclaimable:227516kB kernel_stack:2456kB 
pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:559 
all_unreclaimable? no
Jul 18 21:00:24 opensips kernel: [10762.855278] lowmem_reserve[]: 0 0 8922 8922
Jul 18 21:00:24 opensips kernel: [10762.855284] HighMem free:783004kB min:512kB 
low:1712kB high:2912kB active_anon:169068kB inactive_anon:65948kB 
active_file:47944kB inactive_file:73924kB unevictable:0kB isolated(anon):0kB 
isolated(file):0kB present:1142020kB mlocked:0kB dirty:56kB writeback:0kB 
mapped:48532kB shmem:39056kB slab_reclaimable:0kB slab_unreclaimable:0kB 
kernel_stack:0kB pagetables:3740kB unstable:0kB bounce:0kB writeback_tmp:0kB 
pages_scanned:0 all_unreclaimable? no
Jul 18 21:00:24 opensips kernel: [10762.855287] lowmem_reserve[]: 0 0 0 0
Jul 18 21:00:24 opensips kernel: [10762.855290] DMA: 5*4kB 4*8kB 5*16kB 10*32kB 
14*64kB 17*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 3524kB
Jul 18 21:00:24 opensips kernel: [10762.855298] Normal: 207*4kB 67*8kB 11*16kB 
1*32kB 1*64kB 0*128kB 2*256kB 1*512kB 1*1024kB 0*2048kB 0*4096kB = 3684kB

Hi, just a small issue you may not be aware of, VS2010 code analysis outputs 
the following when run against spserver.lib

1>e:\svn\spserver-read-only\spserver\sphttp.cpp(224): warning C6053: Call to 
'strncpy' might not zero-terminate string 'keepAlive': Lines: 155, 156, 158, 
160, 161, 163, 165, 167, 168, 170, 173, 174, 179, 189, 190, 191, 192, 193, 194, 
197, 201, 207, 209, 210, 211, 215, 219, 222, 224
1>e:\svn\spserver-read-only\spserver\sphttpmsg.cpp(66): warning C6011: 
Dereferencing NULL pointer 'line': Lines: 58, 60, 61, 62, 64, 65, 66
1>e:\svn\spserver-read-only\spserver\sphttpmsg.cpp(74): warning C6387: 
'argument 1' might be '0': this does not adhere to the specification for the 
function '_strnicmp': Lines: 58, 60, 61, 62, 64, 65, 66, 68, 70, 71, 72, 74
1>e:\svn\spserver-read-only\spserver\sphttpmsg.cpp(115): warning C6011: 
Dereferencing NULL pointer 'line': Lines: 107, 109, 110, 111, 113, 114, 115
1>e:\svn\spserver-read-only\spserver\sphttpmsg.cpp(147): warning C6031: Return 
value ignored: 'strtok'
1>e:\svn\spserver-read-only\spserver\sphttpmsg.cpp(362): warning C6308: 
'realloc' might return null pointer: assigning null pointer to 'mContent', 
which is passed as an argument to 'realloc', will cause the original memory 
block to be leaked
1>e:\svn\spserver-read-only\spserver\sphttpmsg.cpp(370): warning C6011: 
Dereferencing NULL pointer 'mContent': Lines: 353, 355, 356, 358, 359, 362, 
364, 367, 368, 370
1>e:\svn\spserver-read-only\spserver\spiocpdispatcher.cpp(165): warning C6011: 
Dereferencing NULL pointer 'arg': Lines: 149, 150, 153, 154, 156, 161, 153, 
154, 156, 161, 153, 154, 156, 161, 162, 164, 165
1>e:\svn\spserver-read-only\spserver\spiocpdispatcher.cpp(267): warning C6011: 
Dereferencing NULL pointer 'arg': Lines: 265, 267
1>e:\svn\spserver-read-only\spserver\spiocpdispatcher.cpp(311): warning C6011: 
Dereferencing NULL pointer 'arg': Lines: 309, 311
1>e:\svn\spserver-read-only\spserver\spiocplfserver.cpp(38): warning C6011: 
Dereferencing NULL pointer 'mAcceptArg': Lines: 26, 27, 29, 30, 32, 34, 36, 37, 
38
1>e:\svn\spserver-read-only\spserver\spiocpserver.cpp(268): warning C6011: 
Dereferencing NULL pointer 'arg': Lines: 205, 206, 208, 210, 212, 213, 214, 
216, 217, 219, 223, 224, 226, 227, 228, 229, 230, 231, 233, 234, 236, 242, 243, 
244, 245, 251, 252, 253, 256, 257, 259, 264, 256, 257, 259, 264, 256, 257, 259, 
264, 265, 267, 268
1>e:\svn\spserver-read-only\spserver\spmsgdecoder.cpp(149): warning C6011: 
Dereferencing NULL pointer 'mBuffer': Lines: 133, 138, 140, 144, 145, 147, 148, 
149
1>e:\svn\spserver-read-only\spserver\spmsgdecoder.cpp(271): warning C6011: 
Dereferencing NULL pointer 'ret': Lines: 264, 265, 270, 271
1>e:\svn\spserver-read-only\spserver\spresponse.cpp(45): warning C6011: 
Dereferencing NULL pointer 'p': Lines: 44, 45
1>e:\svn\spserver-read-only\spserver\spsession.cpp(48): warning C6246: Local 
declaration of 'i' hides declaration of the same name in outer scope. For 
additional information, see previous declaration at line '44' of 
'e:\svn\spserver-read-only\spserver\spsession.cpp': Lines: 44
1>e:\svn\spserver-read-only\spserver\spsession.cpp(79): warning C6011: 
Dereferencing NULL pointer 'mArray[avail]': Lines: 63, 65, 66, 67, 68, 67, 68, 
67, 68, 69, 74, 75, 76, 78, 79
1>e:\svn\spserver-read-only\spserver\spsession.cpp(91): warning C6011: 
Dereferencing NULL pointer 'mArray[row]': Lines: 63, 65, 66, 67, 68, 67, 68, 
67, 68, 69, 74, 75, 76, 78, 79, 80, 78, 79, 80, 78, 79, 80, 78, 85, 86, 87, 89, 
91
1>e:\svn\spserver-read-only\spserver\spthreadpool.cpp(112): warning C6246: 
Local declaration of 'thread' hides declaration of the same name in outer 
scope. For additional information, see previous declaration at line '103' of 
'e:\svn\spserver-read-only\spserver\spthreadpool.cpp': Lines: 103
1>e:\svn\spserver-read-only\spserver\spthreadpool.cpp(113): warning C6011: 
Dereferencing NULL pointer 'thread': Lines: 100, 102, 103, 105, 107, 111, 112, 
113
1>e:\svn\spserver-read-only\spserver\sputils.cpp(39): warning C6308: 'realloc' 
might return null pointer: assigning null pointer to 'mFirst', which is passed 
as an argument to 'realloc', will cause the original memory block to be leaked
1>e:\svn\spserver-read-only\spserver\sputils.cpp(44): warning C6011: 
Dereferencing NULL pointer 'mFirst': Lines: 35, 37, 38, 39, 40, 41, 44
1>e:\svn\spserver-read-only\spserver\sputils.cpp(120): warning C6011: 
Dereferencing NULL pointer 'newEntries': Lines: 106, 107, 108, 110, 111, 114, 
115, 118, 119, 120
1>e:\svn\spserver-read-only\spserver\sputils.cpp(130): warning C6011: 
Dereferencing NULL pointer 'mEntries': Lines: 106, 107, 108, 110, 111, 114, 
115, 118, 119, 120, 123, 124, 126, 127, 130
1>e:\svn\spserver-read-only\spserver\sputils.cpp(237): warning C6328: 'char' 
passed as parameter '1' when 'unsigned char' is required in call to 'isspace'
1>e:\svn\spserver-read-only\spserver\sputils.cpp(239): warning C6328: 'const 
char' passed as parameter '1' when 'unsigned char' is required in call to 
'isspace'
1>e:\svn\spserver-read-only\spserver\sputils.cpp(243): warning C6328: 'const 
char' passed as parameter '1' when 'unsigned char' is required in call to 
'isspace'
1>e:\svn\spserver-read-only\spserver\sputils.cpp(251): warning C6328: 'const 
char' passed as parameter '1' when 'unsigned char' is required in call to 
'isspace'
1>e:\svn\spserver-read-only\spserver\sputils.cpp(259): warning C6328: 'const 
char' passed as parameter '1' when 'unsigned char' is required in call to 
'isspace'
1>e:\svn\spserver-read-only\spserver\sputils.cpp(272): warning C6328: 'const 
char' passed as parameter '1' when 'unsigned char' is required in call to 
'isspace'
1>e:\svn\spserver-read-only\spserver\sputils.cpp(283): warning C6328: 'char' 
passed as parameter '1' when 'unsigned char' is required in call to 'isspace'
1>e:\svn\spserver-read-only\spserver\spwin32port.cpp(151): warning C6053: Call 
to '_vsnprintf' might not zero-terminate string 'logTemp': Lines: 128, 129, 
130, 131, 133, 134, 135, 144, 146, 147, 148, 149, 151
1>e:\svn\spserver-read-only\spserver\spwin32port.cpp(284): warning C6031: 
Return value ignored: 'getchar'

I haven't run it on the other projects, but this may be worth checking out. 
Anyway keep up the good work guys :)

spserver在linux和windows的表现差异比较大

作者有没有标准化的测试报告，比如loadrunner之类的，比较有�
��服力

我在自己的代码里测试过WSASend 0字节。
若这个 socket的发送缓冲满了。 那么投递WSASend 
0字节，还是会立即的收到的！！！

这样就存在问题了！ 
实际上他不可写，但是收到的是可写，那么？？？当你send的�
��候，那么他会返回-1，错误为WSAEWOULDBLOCK

请问下发送0字节如何可以作为事件通知啊？我目前想实现此�
��果。

多谢！