12 messages in net.java.dev.jna.users[jna-users] JNA problems when working...
FromSent OnAttachments
Stas OskinSep 17, 2008 5:06 am 
Timothy WallSep 17, 2008 6:07 am 
Stas OskinSep 17, 2008 9:55 am 
Timothy WallSep 17, 2008 10:39 am 
Wayne MeissnerSep 17, 2008 10:54 am 
Stas OskinSep 17, 2008 11:07 am 
Timothy WallSep 17, 2008 11:30 am 
Stas OskinSep 18, 2008 2:00 am 
Timothy WallSep 18, 2008 5:50 am 
Stas OskinSep 18, 2008 7:37 am 
Stas OskinSep 18, 2008 11:04 am 
Timothy WallSep 18, 2008 11:29 am 
Actions with this message:
Paste this link in email or IM:
Paste this link in email or IM:
Atom feed for this thread
Paste this URL into your reader:
Subject:[jna-users] JNA problems when working with shared memoryActions...
From:Stas Oskin (stas@gmail.com)
Date:Sep 17, 2008 5:06:02 am
List:net.java.dev.jna.users

Hi.

I have a lot of crashes in Java multi-threaded app, and found out that most of them (memcpy) were related to shared memory section, allocated by a native daemon.

I'm using glibc shmget and shmat to connect and attach it to Pointer, and then using getByteArray to get the information.

After I disabled this part, the Java app became much more stable.

The most recurring issues seem to be pointer corruption (as in my previous emails), while the native daemon application continues to work fine, and more over, I can restart the Java app again and continue to read from the shared memory. It's like the JNA just looses the Ponter somewhere.

Any idea how this could be debugged, and what information I can provide to help catch this possible bug?

Also, what other high-speed IPC methods could be recommended?

Thanks in advance.