 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Java Forum / General / August 2006JNI EXCEPTION_ACCESS_VIOLATION
I have a java application which calls into a VC++ DLL via JNI. Within the DLL, I call some code via a 3rd party API which ultimately does some disk IO. When the "File Writing" code is hit, my application crashes. I get the following error (I'll spare you the entire dump) An unexpected exception has been detected in native code outside the VM. Unexpected Signal : EXCEPTION_ACCESS_VIOLATION (0xc0000005) occurred at PC=0x7C9111DE Function=wcsncpy+0x74F Library=C:\WINDOWS\system32\ntdll.dll Here are a couple of interesting points... The code will crash when running under XP Service Pack 2 The code will NOT crash when running under Windows 2000 Service Pack 4 When running on XP, the code will crash if the java progam makes either of the following calls before making the JNI call: UIManager.setLookAndFeel("com.sun.java.swing.plaf.windows.WindowsLookAndFeel"); - or - UIManager.setLookAndFeel(UIManager.getSystemLookAndFeelClassName()); When running on XP, the code will NOT crash if I omit the setLookAndFeel() call all together, or use an alternate L&F (Metal for example) I have spent the past couple of days trying to find clues to this problem on the newsgroups, and have had absolutely no luck. If anyone has any experience with this problem, I would appreciate any leads. Also - if anyone could recommend any tools and/or techniques that would be useful in debugging a VM crash such as I described above, that would be appreciated as well. Thanks! Ed Sandfort > I have a java application which calls into a VC++ DLL via JNI. Within > the DLL, I call some code via a 3rd party API which ultimately does [quoted text clipped - 9 lines] > > Here are a couple of interesting points... You have an error in your native code that leaves the JVM in an unstable state. The fact that it appears to work correctly in some configurations is not an indication that the code is ok. Your error is likely due to slopping pointer handling, failure to manage memory properly, uninitialized variables, failure to check return values, and things like that. More information can be found here: http://c-faq.com/strangeprob/segv.html http://c-faq.com/strangeprob/funnybugs.html http://c-faq.com/malloc/crash.html /gordon  Signature[ don't email me support questions or followups ] g o r d o n + n e w s @ b a l d e r 1 3 . s e >> I have a java application which calls into a VC++ DLL via JNI. Within >> the DLL, I call some code via a 3rd party API which ultimately does [quoted text clipped - 12 lines] > You have an error in your native code that leaves the JVM in an > unstable state. Of course. Microsoft NEVER make a mistake. Sun NEVER make a mistake. The third party NEVER make a mistake. It can't possibly be any one else that is wrong; it must be the OP. > The fact that it appears to work correctly in some configurations is > not an indication that the code is ok. [quoted text clipped - 9 lines] > > /gordon Based on my experience there is a 1% chance that the problem is not the OP. I personally would expect that the problem is caused by the L&F resulting in something becoming null where the 3rd party library doesn't handle that possibility. Honestly I would suggest the "if it hurts, don't do it" solution. SignatureBill Medland > Of course. Microsoft NEVER make a mistake. Sun NEVER make a mistake. > The third party NEVER make a mistake. It can't possibly be any one > else that is wrong; it must be the OP. At first I thought you were being ironic, but then you wrote: > Based on my experience there is a 1% chance that the problem is not > the OP. I agree, even if I think that 1% is somewhat high. /gordon Signature[ don't email me support questions or followups ] g o r d o n + n e w s @ b a l d e r 1 3 . s e Thanks Gordon, Your point is well taken, but I'm fairly confident this is not the problem. I experienced the error in a more complex system, and was able to duplicate it with a VERY simple test program. It's unlikely the test code contains are any of the errors you describe in your response. I should also mention that when the crash does happen, the last line of code executed in my program seems to be somewhat randon according to the print logs I include. It is as if the severe error occurs off in another thread, and any correlation between the timing of the system crash, and the line of my code that is currently being executed can't be established with any regularity. Thanks again! Ed Sandfort Hi there. I am not experienced (I never did) in coding a dll and access it via Java, but, I would say to you that, if you do not mind, try to use the same DLL via C++ or other windows native language, and, use the same thing on the test app that your Java app uses, and, see if the crash still occurs. This way, you will be able to isolate if the error is internal of the dll or of the way you are using it in Java - even if the error only happens sometimes under certain so quasai conditions - but, as said before, it does not mean that the code is entirely ok. Hope I Helped, - Inácio Ferrarini > I have a java application which calls into a VC++ DLL via JNI. Within > the DLL, I call some code via a 3rd party API which ultimately does [quoted text clipped - 37 lines] > > Ed Sandfort Hello Inácio, Thanks for the suggestion. If I understand you correctly, you are suggesting that I access the C++ DLL's functions via a C++ program, to take the JVM out of the equation. I have done this already, and this scenario works fine without any crash. I suspect there is some memory conflict condition that I am causing by accessing the Windows XP disk read/write code (presumably in ntdll.dll ??????) running from within a JVM, and by calling the WindowLookAndFeel code into play. I have been researching this issue pretty rigorously for the past few days, and have found other posting reporting similar problems. Unfortunately, it seems that nobody has been able to offer a satisfying explanation, and in my particular case, I have to use the Windows L&F, so simply omitting it is not an option. Short of a complete explanation on how to fix this problem, I could really use advise on how to get more information from the crash that I can use to track the problem down. Thanks again for your post. Ed Sandfort Free MagazinesGet these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...
|
Reply to this Message
|
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2008 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.