<html> <head> <meta http-equiv="Content-Type" content="text/html; charset=gb2312"> <meta name="Generator" content="Microsoft Word 14 (filtered medium)"> <style></style> </head> <body bgcolor="white" lang="ZH-CN" link="blue" vlink="purple"> <div class="WordSection1"> Hi,<o:p></o:p> OK. Now I am using JNA to access libvirt. If we add another mutex which used to access “initialized” parameter. This mutex must be pthread_mutex_init firstly and only once.<o:p></o:p> But it seems that there is no way to change libvirt code. I do it as following:<o:p></o:p> <![if !supportLists]>1. <![endif]>Changing libvirt JNA code in Connect.java<o:p></o:p> Old Code:<o:p></o:p> public Connect(String uri) throws LibvirtException {<o:p></o:p> VCP = libvirt.virConnectOpen(uri);<o:p></o:p> // Check for an error<o:p></o:p> processError();<o:p></o:p> ErrorHandler.processError(Libvirt.INSTANCE);<o:p></o:p> }<o:p></o:p> <o:p> </o:p> New Code:<o:p></o:p> public Connect(String uri) throws LibvirtException {<o:p></o:p> synchronized(this.getClass()) {<o:p></o:p> VCP = libvirt.virConnectOpen(uri);<o:p></o:p> }<o:p></o:p> // Check for an error<o:p></o:p> processError();<o:p></o:p> ErrorHandler.processError(Libvirt.INSTANCE);<o:p></o:p> }<o:p></o:p> <o:p> </o:p> This can make sure only that one thread can execute Connect. For a server application, we only need one time. So the performance is OK<o:p></o:p> <o:p> </o:p> <![if !supportLists]>2. <![endif]>Changing libvirt code in libvirt.c<o:p></o:p> Old Code:<o:p></o:p> static int initialized = 0;<o:p></o:p> <o:p> </o:p> New Code:<o:p></o:p> static int volatile initialized = 0;<o:p></o:p> <o:p> </o:p> This can make sure the initialization will be executed once.<o:p></o:p> <o:p> </o:p> Would you give your comments for this solution?<o:p></o:p> <o:p> </o:p> B.R.<o:p></o:p> Benjamin Wang<o:p></o:p> <o:p> </o:p> <div> <div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm"> From: Guannan Ren [mailto:gren@redhat.com] Sent: 2012年9月29日 15:43 To: Benjamin Wang (gendwang) Cc: Daniel Veillard; libvir-list@redhat.com; Yang Zhou (yangzho) Subject: Re: [libvirt] Potential race condition problem<o:p></o:p> </div> </div> <o:p> </o:p> <div> On 09/29/2012 03:07 PM, Benjamin Wang (gendwang) wrote:<o:p></o:p> </div> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> Hi,<o:p></o:p> Currently virInitialize() method defined in libvirt.c has the following code:<o:p></o:p> int<o:p></o:p> virInitialize(void)<o:p></o:p> {<o:p></o:p> if (initialized)<o:p></o:p> return 0;<o:p></o:p> <o:p></o:p> initialized = 1;<o:p></o:p> <o:p></o:p> if (virThreadInitialize() < 0 ||<o:p></o:p> virErrorInitialize() < 0 ||<o:p></o:p> virRandomInitialize(time(NULL) ^ getpid()) ||<o:p></o:p> virNodeSuspendInit() < 0)<o:p></o:p> return -1;<o:p></o:p> <o:p></o:p> ……<o:p></o:p> }<o:p></o:p> <o:p></o:p> When two threads access virInitialize method, there is no lock for the “initialized” parameter. If the first thread enters this method and set “initialized” to 1,<o:p></o:p> the second thread could see that “initialized” is 1(Because initialized is not volatiled, I say could). In some situation, before the first thread finishes all the initialization,<o:p></o:p> the second thread could use some resources which should be initialized in Initialize method. <o:p></o:p> If you have any comments, please let me know. Thanks!<o:p></o:p> <o:p></o:p> B.R.<o:p></o:p> Benjamin Wang<o:p></o:p> <o:p> </o:p> </blockquote> As the comments above the function said, "It's better to call this routine at startup in multithreaded applications to avoid potential race when initializing the library." Guannan <o:p></o:p> </div> </body> </html>