[DeadLock] Зависание в LRESULT MainFrame::onCheckDHTStats

[pavel-pimenov]

From Pavel.Pimenov@gmail.com on April 23, 2013 17:28:02

1. Вешаемся при вызове ConnectivityManager::getInstance()->setup(true);
2. крутимся бесконечно в join() ~Server() { die = true; join(); }

Полный стек:

ntdll.dll!ZwWaitForSingleObject()  + 0xa bytes  
KERNELBASE.dll!WaitForSingleObjectEx()  + 0x9c bytes    
FlylinkDC_x64.exe!Thread::join()  Line 206  C++
FlylinkDC_x64.exe!ConnectionManager::Server::~Server()  Line 213 + 0xa bytes    C++
FlylinkDC_x64.exe!ConnectionManager::Server::`scalar deleting destructor'()  + 0x17 bytes   C++
FlylinkDC_x64.exe!safe_delete<DirectoryListing::Directory>(DirectoryListing::Directory * & p)  Line 110 + 0x33 bytes    C++
FlylinkDC_x64.exe!ConnectionManager::disconnect()  Line 517 C++
FlylinkDC_x64.exe!ConnectivityManager::disconnect()  Line 213   C++
FlylinkDC_x64.exe!ConnectivityManager::startSocket()  Line 47   C++
FlylinkDC_x64.exe!ConnectivityManager::setup(bool settingsChanged)  Line 141 + 0xa bytes    C++

FlylinkDC_x64.exe!MainFrame::onCheckDHTStats(unsigned short formal, unsigned short formal, unsigned short formal, unsigned short formal) Line 3479 + 0x16 bytes C++

Original issue: http://code.google.com/p/flylinkdc/issues/detail?id=1003

[pavel-pimenov]

From Pavel.Pimenov@gmail.com on April 23, 2013 09:55:30

Во вложении картинки не совсем понятно как так получается но по номеру дескриптора (возможно совпадение) идет ожидание VideoPreview

Attachment: issue-1003-part-1.png issue-1003-part-2.png

[pavel-pimenov]

From Pavel.Pimenov@gmail.com on April 24, 2013 03:53:51

This issue was closed by revision r13733 .

Status: Fixed

[pavel-pimenov]

From a.rain...@gmail.com on April 29, 2013 02:37:55

Status: Accepted

[pavel-pimenov]

From a.rain...@gmail.com on April 29, 2013 02:44:37

Дополнительные сведения содержаться в r13750 , r13753 . Исправление r13733 не корректное по причине того, что отпал проброс портов через UPnP.

p.s: скорее всего надо просто провести ревизию компонентов VideoPreview ибо судя по краш-серверу ошибки там есть.

Labels: Stability Component-Logic

[pavel-pimenov]

From a.rain...@gmail.com on May 01, 2013 05:58:00

r13768