NET to GS process for a package

Let's take a look at the process of a package of shared memory first look at how the net process is handled {//packages that were just taken from the lock-free queue are encapsulated in this type of SharedataSharedata SD; sd.channel_id=pkt.channel_id; Sd.data=Pkt.data; Sd.is_data=Pkt.is_data; Sd.size=pkt.size; Auto HR= M_spsharememinter->pusha (SD);//put this package in shared memory}//specifically, how to put the shared memoryBOOLSharememinteroneway::p ushin (sharedata sd) {scoped_lock<interprocess_mutex>Lock(M_mem mirror->mutex);//process lock required, different process accessSharedataex SD2; BOOLhr =m_processmempool.getpkt (sd.size, SD2); if(!hr)return false; Sd2.size=sd.size; sd2.channel_id=sd.channel_id; Sd2.is_data=Sd.is_data; if(sd.is_data) {memcpy (Sd2.data, Sd.data, sd.size); } m_sharememque-push_back (SD2); return true;}BOOLPROCESSMEMPOOLEX::GETPKT (intLen, sharedataex&SD) {    if(Len < -)//determine from which shared memory to take from the length of the        returnm_processmempool0.getpkt (len, SD); Else if(Len < +)        returnm_processmempool1.getpkt (len, SD); returnm_processmempool2.getpkt (len, SD);}BOOLPROCESSMEMPOOL::GETPKT (intLen, sharedataex&SD) {    if(Len >m_memsize) {        //Testwchar_t tips[ -] = {0}; wsprintf (tips, L"warning100o:len:%d, memsize:%d! ", Len, m_memsize); MessageBox (NULL, Tips/*L "100o"*/, L"1ooo", MB_OK); return false; }    if(!m_queue->size ())return false; SD= M_queue->front ();//If you see this, you know why you should initialize that Shared queue.M_queue->Pop_front (); Sd.data=Convertadd (Sd.index); return true;}void* Processmempool::convertadd (intindex) {    return(Char*) M_START_ADDR + (index * m_memsize);//get the corresponding address of the shared memory pool by this index} attached: M_queue initialization code: { for(inti =0; i < M_buffer_num; i++) {Sharedataex sd; Sd.index= i;//This is the index that represents which piece of the memory poolSd.size =0; Sd.total_size=m_memsize; M_queue-push_back (SD); }}BOOLSharememinteroneway::p ushin (sharedata sd) {scoped_lock<interprocess_mutex>Lock(M_mem Mirror,mutex);    Sharedataex SD2; BOOLhr = M_PROCESSMEMPOOL.GETPKT (sd.size, SD2);//A memory address is obtained after calling Getpkt    if(!hr)return false; Sd2.size=sd.size; sd2.channel_id=sd.channel_id; Sd2.is_data=Sd.is_data; if(sd.is_data) {memcpy (Sd2.data, Sd.data, sd.size);//then copy the passed in SD memory to the memory pool in the shared memory, and this memory in SD is that m_recvbuff inside, there is a problem is this m_recvbuff why allocate 100m of memory, need so big? } m_sharememque->push_back (SD2);//then he puts the SD2 in the interactive Shared queue, noting that the memory in this area is still allocated in the shared memory pool, and M_sharememque just saves an address .    return true;}//Such a package is placed in the Shared queue.

NET to GS process for a package