Community
cancel
Showing results for 
Search instead for 
Did you mean: 
jasedit
Beginner
30 Views

Crash in construction of UMC::VideoDecoderParams?

I'm using IPP 7.0.1 on an Ubuntu 9.04 (and Suse 10.3) system, and am running into an issue where the system is crashing when the UMC::VideoDecoderParams object is instantiated. In this case, I've got a block of code which looks like this:
[cpp]  boost::shared_ptr<:VIDEODECODERPARAMS> result;
  std::cerr << "Creating VideoDecoderParams . . ." << std::endl;
  result = boost::shared_ptr<:VIDEODECODERPARAMS>(new UMC::VideoDecoderParams());
  std::cerr << "Created VideoDecoderParams." << std::endl;[/cpp]
with the resulting output being:
Creating params . . .
*** glibc detected *** /home/jasedit/bin/TestApp: malloc(): memory corruption: 0x0894fbd0 ***

I've got absolutely no idea what might be causing this - this block of code worked fine undera previous version of IPP (not sure which version exactly, but it's a 6.something release I'm almost certain.)

Any suggestions?
0 Kudos
3 Replies
Chao_Y_Intel
Employee
30 Views

Hello,

Is this 32 bit application, or 64 bit?, and which compiler you are using to build the application?

I just review the code within the function call ( just some construction functions):

new UMC::VideoDecoderParams());

There is no malloc() function call, could you have a debug within the code, and see which line of the code is crashed?

Thanks,
Chao

jasedit
Beginner
30 Views

The two platforms I'm currently testing on are 32-bit - although it will be compiled on a 64-bit version of Ubuntu 10.04 once this is tested and fixed as well. We're using gcc - whichever version is shipped with each distro (I can pull up version numbers if those are useful to help debug the issue.)
I'm not sure what you mean by "have a debug within the code" - I've tried to pull out information from gdb to localize the crash, but it ends up with a corrupt stack trace and provides no information.
Thanks,
Jason
jasedit
Beginner
30 Views

I've been hammering at this, and it appears that something screwy was going on when the IPP system was installed on these machines. Performing a fresh install of IPP has got both machines no longer crashing - one is working fully as expected, and the other appears to have issues in unrelated parts of the code.

Thanks for all the help!
Jason